V(D)J Recombination (Advances in Experimental Medicine and Biology Vol 650)

V(D)J Recombination

ADVANCES IN EXPERIMENTAL MEDICINE ANDBIOLOGY

Editorial Board:

NATHAN BACK,StateUniversity ofNew York at Buffalo

IRUNR. COHEN, TheWeizmann Institute ofScience

ABELLAJTHA, N.S. KlineInstitutefor PsychiatricResearch

JOHND. LAMBRIS, University ofPennsylvania

RODOLFO PAOLETfI,UniversityofMilan

RecentVolumes in this Series

Volume 642THE SARCOMERE ANDSKELETAL MUSCLE DISEASE

Editedby NigelG. Laing

Volume 643TAURINE 7

Editedby JunichiAzuma

Volume 644TROPOMYOSIN

Editedby PeterGunning

Volume 645OXYGEN TRANSPORT TOTISSUE XXX

Editedby PerLiss, PeterHansell, DuaneF. Bruley, andDavidK. Harrison

Volume 646EARLY NUTRITION PROGRAMMING ANDHEALTH OUTCOMES IN LATER LIFE

Editedby Berthold Koletzko, Tamas Desci,DenesMolnar, andAnneDe la Hunty

Volume 647THERAPEUTIC TARGETS OFTHETNFSUPERFAMILY

Editedby IqbalGrewal

Volume 648ARTERIAL ANDALLIED CHEMORECEPTORS

Editedby Constancio Gonzalez, ColinA. Nurse, and ChrisPeers

Volume 649MOLECULAR MECHANISMS OF SPONDYLOARTHROPATHIES

Editedby CarlosLopez-Larrea, andRobertoDlaz-Pena

Volume 650V(D)JRECOMBINATION

Editedby PierreFerrier

A Continuation Order Plan is available for this series. A continuation orderwill bring delivery of each newvolume immediately upon publication. Volumes arebilled onlyupon actual shipment. Forfurther informationplease contact thepublisher.

V(D)J RecombinationEditedby

Pierre Ferrier, MD, PhD

Centre d'Immunologie de Marseille-Luminy, Universite d'AixMarseille,

Marseille, France

Springer Science+Business Media, LLCLandes Bioscience

Springer Science+Business Media, LLCLandes Bioscience

Copyright ©2009 Landes Bioscience andSpringer Science+Business Media, LLC

All rightsreserved.Nopartof thisbookmaybereproduced ortransmittedinanyformorbyanymeans, electronic ormechanical, including photocopy, recording, oranyinformation storageandretrieval system, without permissionin writing from thepublisher, withtheexception of anymaterial supplied specifically forthepurpose ofbeingenteredand executed on a computer system; for exclusive use by thePurchaser of thework.

Printedin the USA.

Springer Science+Business Media, LLC, 233Spring Street, NewYork, NewYork 10013, USAhttp://www.springer.com

Pleaseaddress all inquiries to the publishers:LandesBioscience, 1002 WestAvenue, Austin, Texas 78701,USAPhone: 51216376050;FAX: 512/637 6079http://www.landesbioscience.com

The chapters in this bookare available in the Madame CurieBioscience Database.http://www.landesbioscience.comlcurie

V(D)JRecomination, editedby PierreFerrier. LandesBioscience 1Springer Science+Business Media,LLCdual imprint 1Springer series: Advances in Experimental Medicine and Biology.

ISBN: 978-1-4419-0295-5

Whiletheauthors, editors andpublisher believethatdrugselection anddosageandthespecifications andusageof equipment and devices, as set forthin this book,are in accordwithcurrentrecommendationsand practice at the time of publication, they makeno warranty, expressed or implied, with respect tomaterial described in this book. In view of the ongoing research, equipment development, changes ingovernmental regulations andtherapidaccumulation of information relating to thebiomedical sciences,the readeris urgedto carefully reviewand evaluate the information provided herein.

Library of Congress Cataloging-in-Publication Data

V(D)Jrecombination 1editedby PierreFerrier.p. ; em. -- (Advances in experimental medicine andbiology ; v. 650)

Includes bibliographical references and index.ISBN978-1-4419-0295-5I. Genetic recombination. 2. DNA-binding proteins. 1.Ferrier, Pierre, 1951- II. Series: Advances in

experimental medicine andbiology; v. 650.[DNLM: I. Recombination, Genetic.2. VOJRecombinases. 3. DNA-Binding Proteins. 4. Recombinant

Proteins. WI AD559 v.65020091QU475V3932009]QH443.V352oo9571.9'648--dc22

2009011105

PREFACE

v(D)Jrecombination: for thecommunity of immunologists anddevelopmentalbiologists, the molecular routeby whichB andT lymphocytes acquire theiruniquefunction ofaffording adaptive immunity.Yet, formany-from experienced scientiststo trainees-it represents a (rathertoo) sophisticated process whosetrue insightisexcessively demanding. However, whennot simplyconsidered as a privategroundfora fewaficionados, it canbe seenas a wayofunderstanding howmaturelymphocytescarryon theirbasicfunctions. For the groupof aficionados-which includesthis editor-it is an elegant paradigm featuring many fascinating evolutionaryachievements of which the biological world alone has the secret. These include asubtlebiochemical principle most likelyhijacked some470million yearsago froman ancestral gene invaderand since then cleverly adapted by jawed vertebrates toprecisely cleave andrearrange theirantigen receptor (IgandTCR)loci.Thisinvaderwoulditselfhaveassigned the services of the nonhomologous endjoining(NHEJ)DNArepairmachinery as wellasvarious DNApolymerases or transferases toworkinconcert withdevelopmental clues in lymphoid celllineages togenerate animmunerepertoire and efficient host surveillance whileavoiding autoimmunity.

Recently, important newrefinements in these systems haveemerged, continuingtochallenge ourknowledge andbeliefs.These arejust thetopics coveredbytheseniorauthors-all established leaders in thisfield-and theircolleagues, whilst writing thevarious chapters in V(D)J Recombination. They lead us through the latest findingsconcerning thebiochemical properties oftheV(D)J recombinase (Swanson), itsburiedandpotentially harmful transposase and translocase activities (Oettinger; Roth), theincreasing importance of NHEJ, whose dysfunction causes severe forms of immunedeficiencies (deVillartay), andthenumerous facets inthecontrol ofgene rearrangementvianon-coding RNAtranscription andexquisitely regulated changes inchromosomalstructure (Corcoran; Feeney; Jouvin-Marche; Krangel; OltzandSpicuglia).

Burning progress on regulatory aspects has included the large-scale dynamicsandnuclearcompartmentalization ofIg andTCRloci(Singh), theanticipated-butdifficult to ascertain-role of dedicated transcription factors (Zhang), the relationshipsbetween structural properties of the recombination coreapparatus and its cellcycle phase-dependant accumulation/degradation or connection to the chromatin

v

vi Preface

template (Desiderio), the evolution of theseregulatory aspects throughout the phylogeny (Hsu), and how abnormalities in the recombination apparatus/process cancontribute to lymphoid malignancies (Macintyre).

Overall, V(DP Recombination represents a tour over this, in all respects, vitalprocess and I would like to greatlyacknowledge the efforts of these eminent colleagues forconcisely describing its so manyaspects. We believethateveryadvancein thisfieldcontributes to strengthening knowledgeoffundamental importance bothacademically and clinically. Together, we hope that the result is an attractive bookwhich will captivate its readers and encourage some to pursue further digging inthis seemingly inexhaustible mineof biological resources.

ABOUT THE EDITOR...

PIERRE FERRIER is a Principal Investigator and Research Director at theCentre d'Immunologie de Marseille-Luminy (CIML) , France. He has also workedas a Director of Marseille-Nice Genopole, a local consortium of more than twentylaboratories aimed at developing high-throughput research techniques in genomics.Main research interests include the analysis ofthe molecular mechanisms responsiblefor the control ofgene expression and recombination programs during hematopoieticcell development and pathogenesis. He is a member of several national and international scientific organizations including the Institut National de la Sante et de laRecherche Medicale (Inserm) , the Agence Nationale de la Recherche (ANR), theAssociation pour la Recherche sur le Cancer (ARC), the Human Frontier ScienceProgram Organization (HFSPO), and the Universite Virtuelle Medicale de Monaco(UVMM). Pierre Ferrier received his academic degrees from Montpellier (MD) andMarseille (PhD) Universities, France. He was a post-doctoral fellow (1986-90) inthe laboratory of Prof. F.W. Alt at the Columbia University College of Physiciansand Surgeons, New York, NY, USA.

vii

PARTICIPANTS

IratxeAbarrateguiCentrefor EpigeneticsBiotechResearch and Innovation CentreCopenhagenDenmark

Vahid AsnafiINSERM EMIU021 0Hopital Necker-Enfants MaladesAP-HPUniversity Paris-DescartesParisFrance

KheiraBeldjordINSERM EMIU0210Hopital Necker-Enfants MaladesAP-HPUniversity Paris-DescartesParisFrance

DanielJ. BollandLaboratory of Chromatin and Gene

ExpressionBabraham InstituteBabraham Research CampusCambridgeUK

MarieBonnetCentred'Immunologie

de Marseille-LuminyUniversite d'Aix MarseilleCNRS,UMR6102Inserm, U 631MarseilleFrance

Vicky L. BrandtDepartment of Pathology and Program

in Molecular PathogenesisThe HelenL. and MartinS. Kimmel

Centerfor Biology and MedicineSkirball Institute for Biomolecular

MedicineNewYork University School

of MedicineNew York, NewYorkUSA

RobinMilleyCobbDepartment of Microbiology

and ImmunologyVanderbilt UniversityNashville, TennesseeUSA

Anne E. CorcoranLaboratory of Chromatin and Gene

ExpressionBabraham InstituteBabraham Research CampusCambridgeUK

ix

x

Sai'daDadiCentred'Immunologie

de Marseille-LuminyUniversited'Aix MarseilleMarseille, FranceCNRS,UMR. 6102Inserm,U 631MarseilleFrance

StephenDesiderioDepartmentof Molecular Biology

and GeneticsInstitutefor Cell EngineeringThe Johns Hopkins University

Schoolof MedicineBaltimore, MarylandUSA

Jean-Pierre de VillartayINSERM, U768,Unite Developpement

Normalet Pathologique du SystemeImmunitaireandFacultede Medecine Rene DescartesUniversiteParis-DescartesandAP-HP, HopitalNeckerEnfantsMaladesServiced'Immunologieet

d'Hematologie PediatriqueParisFrance

Ann1. FeeneyThe ScrippsResearch InstituteDepartment of ImmunologyLa Jolla, CaliforniaUSA

Pierre FerrierCentred'Immunologie

de Marseille-LwninyUniversited'Aix MarseilleCNRS,UMR. 6102Inserm,U 631MarseilleFrance

Participants

PatriziaFuschiottiDepartment of ImmunologyUniversity of Pittsburgh

Schoolof MedicinePittsburgh, PennsylvaniaUSA

EllenHsuDepartment of Physiology

and PharmacologyStateUniversity of NewYorkHealthScienceCenterat BrooklynBrooklyn, NewYorkUSA

KristenJohnsonDepartment of MolecularGenetics

and Cell BiologyHowardHughesMedicalInstituteThe University of ChicagoChicago,IllinoisUSA

Mary Elizabeth JonesDepartment of ImmunologyDukeUniversity Medical CenterDurham,North CarolinaUSA

EvelyneJouvin-MarcheINSERM, U823Faculte de MedecineInstitutd'Oncologie/DeveloppementAlbertBonniotet InstitutFrancais

duSangUniversite JosephFourier-Grenoble IGrenobleFrance

MichaelS. KrangelDepartment of ImmunologyDukeUniversity Medical CenterDurham,North CarolinaUSA

SushilKumarDepartment of Medical Microbiology

and Immunology,Creighton University Medical CenterOmaha,NebraskaUSA

Pllrticipllnts

Sandrine Le NoirINSERM EMIU0210H6pital Necker-Enfants MaladesAP-HPUniversite Paris-DescartesParisFrance

YunLiuDepartment of Molecular Biology

and GeneticsInstitute for CellEngineeringThe JohnsHopkins University

School of MedicineBaltimore, MarylandUSA

Elizabeth A. MacintyreINSERM EMIU0210H6pital Necker-Enfants MaladesAP-HPUniversite Paris-DescartesParisFrance

PatriceNoel MarcheINSERM, U823Faculte de MedecineInstitutd'Oncologie/DeveloppementAlbertBonniotet InstitutFrancais

du SangUniversite JosephFourier-Grenoble IGrenobleFrance

Adam G.w. MatthewsDepartment of Molecular BiologyMassachusetts General HospitalandDepartment of GeneticsHarvardMedical SchoolBoston, MassachusettsUSA

Marjorie A. OettingerDepartment of Molecular BiologyMassachusetts General HospitalandDepartment of GeneticsHarvardMedical SchoolBoston,MassachusettsUSA

Eugene M. OltzDepartment of Microbiology


PrafullaRavalDepartment of Medical Microbiology

and ImmunologyCreighton University Medical CenterOmaha, NebraskaUSA

KarenL. ReddyDepartment of Molecular Genetics

and CellBiologyHowardHughes Medical InstituteThe University of ChicagoChicago, IllinoisUSA

DavidB. RothDepartment of Pathology and Program

in Molecular PathogenesisThe HelenL. and MartinS. Kimmel

Centerfor Biology and MedicineSkirball Institute for Biomolecular

MedicineNewYork University School

of MedicineNewYork, NewYorkUSA

HarinderSinghDepartment of Molecular Genetics

and Cell BiologyHowardHughes Medical InstituteThe University of ChicagoChicago, IllinoisUSA

xi

xli

Salvatore SpicugliaCentred'Immunologie

de Marseille-LuminyUniversite d'AixMarseilleCNRS, UMR6102Inserm, U 631MarseilleFrance

PatrickC. SwansonDepartment of Medical Microbiology

and ImmunologyCreighton University Medical CenterOmaha, NebraskaUSA

LanceR. ThomasDepartment of Microbiology


PlII1kipants

Andrew L. WoodLaboratory of Chromatin and Gene

ExpressionBabraham InstituteBabraham Research CampusCambridgeUK.

Li ZhangDepartment of Molecular Biology

and GeneticsInstitute for CellEngineeringTheJohnsHopkins University

School of MedicineBaltimore, MarylandUSA

YuanZhuangDepartment of ImmunologyDukeUniversity Medical CenterDurham, NorthCarolinaUSA

CONTENTS

1. EARLY STEPS OF V(D)J REARRANGEMENT: INSIGHTSFROM BIOCHEMICAL STUDIES OF RAG-RSS COMPLEXES .... 1

PatrickC. Swanson, Sushi! Kumarand PrafullaRaval

Abstract 1Introduction 1Assembly and Organization of Single Site and Synaptic RAG-RSS Complexes 3Insights into RAG-Mediated RSS Recognition and Cleavage Mechanisms 5Elements Guiding Enforcement of the 12/23 Rule 8Transcription Factor-Assisted Targeting of Antigen Receptor Loci 10Conclusion and Future Directions 11

2. REGULATION OF RAG TRANSPOSITION 16

AdamG.w. Matthews and Marjorie A. Oettinger

Abstract 16Introduction 16Biochemistry ofV(D)J Recombination 16Overview of RAG Transposition 19Regulation of RAG Transposition 24Current Understanding of Row RAG Transposition Is Regulated 24Additional Potential Regulatory Mechanisms 25Conclusion 27

3. RECENT INSIGHTS INTO THE FORMATION OF RAG-INDUCEDCHROMOSOMAL TRANSLOCATIONS 32

Vicky L. Brandtand DavidB. Roth

Abstract 32Introduction 32Overview of the V(D)J Recombination Reaction 33Potential Mechanisms of RAG-Mediated Translocations 34Mistaken Identities: Substrate Selection Errors 34The Ends That Got Away: Errors in Joining 36

xiii

xiv Contents

4. V(D)J RECOMBINATION DEFICIENCIES 46

Jean-Pierre de Villartay

Abstract 46Introduction 46RAGl and RAG2 Deficiencies 47T-B-SCID with Radiosensitivity 50

5. LARGE-SCALE CHROMATIN REMODELINGAT THE IMMUNOGLOBULIN HEAVY CHAIN LOCUS:A PARADIGM FOR MULTIGENE REGULATION 59

Daniel J. Bolland,AndrewL. Woodand Anne E. Corcoran

Abstract 59Introduction 60Chromatin Remodeling 62Intergenic Transcription 63Intergenic Transcription in the MouseIgh LocusV Region 63AntisenseTranscription 64AntisenseTranscription in the Igh LocusV Region 64Antisenseand Intergenic Transcription in the Igh D Region 66Subnuclear Relocalisation 663-Dimensional Alterations in Chromatin Structure 67Transcription Factories 68BiasedRecombinationFrequency ExplainedbyNumerousMechanisms 68Allelic Choiceand Allelic Exclusion 68Other AntigenReceptor Loci 69Future Directions 69

6. GENETIC AND EPIGENETIC CONTROL OF V GENEREARRANGEMENT FREQUENCy 73

Ann 1. Feeney

Abstract 73Introduction 73SequenceVariation in RSS Can Greatly AffectRecombination 74RSS Is NotAlways Responsible for UnequalRearrangement 75Chromatin as the Gatekeeper ofAccessibility 75Role of Transcription Factors in ControllingRearrangement 77Conclusion 79

7. DYNAMIC ASPECTS OF TCRa GENE RECOMBINATION:QUALITATIVE AND QUANTITATIVE ASSESSMENTSOF THE TCRa CHAIN REPERTOIRE IN MAN AND MOUSE ..... 82

Evelyne Jouvin-Marche, Patrizia Fuschiottiand Patrice Noel Marche

Abstract 82

Contents xv

Introduction 82Complexity of Mouse and Human TCRAn Locus 83Analysis of Human and Mouse TCRA-Chain Diversity 84Comparison between the Frequencies of Rearrangements in Thymus

and Peripheral T-Lymphocytes 8SThe Size of the Mouse and Human TCRa Repertoire 87Conclusion 90

8. GERMLINE TRANSCRIPTION: A KEY REGULATOROF ACCESSIBILITY AND RECOMBINATION 93

lratxeAbarrategui and Michael S. Krangel

Abstract 93Introduction 93A Brief History of Germline Transcription and V(D)J Recombination 94Disruption of Chromatin by Transcription 9SRegulation ofV(D)J Recombination by Transcription 97Future Directions 99

9. DYNAMIC REGULATION OF ANTIGEN RECEPTORGENE ASSEMBLY 103

LanceR. Thomas, RobinMilleyCobband Eugene M. Oltz

Abstract 103Introduction 103Developmental Control ofV(D)J Recombination 104Genetic Control of Recombinase Accessibility lOSChromatin Accessibility Control Mechanisms for V(D)J Recombination 107Control ofV(D)J Recombination by Nuclear Compartmentalization 109Primary Activation of Antigen Receptor Loci for D to J Rearrangement 109Long-Range Control ofV(D)J Recombination 111Allelic Exclusion 111Conclusion 113

10. MOLECULAR GENETICS AT THE T-CELL RECEPTOR lJLOCUS: INSIGHTS INTO THE REGULATIONOF V(D)J RECOMBINATION 116

MarieBonnet, PierreFerrierand Salvatore Spicuglia

Abstract 116Introduction 116Overview of the Tcrb Genomic Structure and Recombination Properties 117Tcrb-RSSsand Rearrangement Efliciency 117Cis-Regulatory Elements at the Tcrb Locus 119Trans-Regulators of Tcrb Locus ExpressionlRecombination 122Chromatin Accessibility 123Allelic Exclusion at the Tcrb Locus 126Conclusion and Future Direction 128

xvi Conte1lls

11. MOLECULAR PATHWAYSAND MECHANISMS REGULATINGTHE RECOMBINATION OF IMMUNOGLOBULIN GENESDURING B-LYMPHOCYTE DEVELOPMENT 133

KristenJohnson, KarenL. Reddyand HarinderSingh

Abstract "" 133Introduction 133B-Cell Fate Specification and the Joining ofD-to-Jn Segments•.••.••.•••••••.....••••••••••_••• 137B-Cell Fate Commitment and V-to-DJn Rearrangement 138The Pre-B-Cell Checkpoint and the Induction of Light-Chain Recombination •••••••• 141Allelic Exclusion 143Perspectives 144

12. REGULATION OF V(D)J RECOMBINATIONBY E-PROTEINTRANSCRIPTION FACTORS 148

Mary Elizabeth Jonesand Yuan Zhuang

Abstract 148Introduction 148Transcriptional Control of Ig and TCR Antigen Receptor and Their

Associated Genes 149Induction of Ig and TCR Gene Rearrangement 150Regulation of the Developmental Window for V(D)J Recombination _ 150Conclusion •••••.••.••.••.••••••••.••••••••••••••.••.......•...••...•.••.••••••••••.....••.•••••••••••••••••••..••.•••••••_.""•.•153

13. TEMPORAL AND SPATIAL REGULATION OF V(D)JRECOMBINATION: INTERACTIONS OF EXTRINSICFACTORS WITH THE RAG COMPLEX••••••••••••••••••••••••••••••••••••••••• 157

Yun Liu, Li Zhangand StephenDesiderio

Abstract 157Functional Organization of RAG-I and RAG-2 157Temporal Regulation ofV(D)J Recombination through Interactions

with the RAG-2 Non-Core Region _••••••••..•• 158Locus Specificity: General Remarks •••••••.•••••••••.....••.•••.•.•.•..•.•••••••...•••..•..•.••••••••.••.••."".. 159Epigenetic Modifications of Possible Relevance to V(D)J Recombination 159DNA Methylation 159Nucleosome Phasing 159Histone Acetylation _.. 160Histone H3 K9 Methylation _.. 160Histone H3 K4 Methylation _•• 160Direct Recognition of Modified Histone H3 by the V(D)J Recombinase .•.••.•..••••••••••• 160Evidence for Allosteric Regulation ofV(D)J Recombinase

Activity by Histone H3 Trimethylated at Lysine 4 161Future Directions: Deposition and Integration of Epigenetic

Signals Controlling V(D)J Recombination _••• 162

Contents xvii

14. V(D)J RECOMBINATION: OF MICE AND SHARKS 166

EllenHsu

Abstract 166Introduction 166V(D)J Rearrangement 168V(D)J Rearrangement Patterns 172Rabbit 172Multiple 19B Loci in Other Vertebrate Species 173Conclusion 176

15. NORMALAND PATHOLOGICAL V(D)J RECOMBINATION:CONTRIBUTION TO THE UNDERSTANDINGOF HUMAN LYMPHOID MALIGNANCIES 180

SaidaDadi,Sandrine LeNoir, Vahid Asnafi, KheiraBeldjordand Elizabeth A. Macintyre

Abstract 180Introduction 180Diagnostic Clonality Analysis 181Recombinase Mediated Oncogenesis 185Conclusion 190

INDEX 195

CHAPTER!

Early Steps ofV(D)J Rearrangement:Insights from Biochemical StudiesofRAG-RSS ComplexesPatrick C. Swanson: Sushi! Kumar and Prafulla Raval

Abstract

\f:D)J recombinationis initiated bythe synapsis and cleavage of a complementary(12123)pair of recombination signalsequences (RSSs) by the RAGI and RAG2 proteins. Ourunderstanding of these processes has been greadyaided by the developmentof in vitro

biochemicalassays of RAG binding and cleavage activity. Accumulating evidencesuggests thatsynapticcomplexassembly occursin a step-wise manner and that the RAG proteinscatalyze RSScleavage by mechanisms similar to those used by bacterial transposases. In this chapter we willreview the molecularmechanisms of RAG synapticcomplexassembly and 12123-regulated RSScleavage, focusingon recent advances that shed new light on theseprocesses.

IntroductionThe antigen-bindingvariable domainsof immunoglobulinsand T-cellreceptorsexhibitgreat

structuraldiversitythat mosdyoriginatesfroma site-specific DNA rearrangementprocess, calledV(D)J recombination, that assembles the exonsencodingthe variable domainsof theseproteinsfrom germlinevariable (V), diversity(D) and joining (J) genesegments during lymphocytedeveloprnenr. ' Adjacentto eachgenesegmentliesa recombinationsignalsequence(RSS);eachRSScontains a conservedheptamer and nonamer motif (consensus heptamer: 5' -CACAGTG-3 ';consensus nonamer: 5 ' -ACAAAAACC-3') separatedby"spacer" DNA, normally 12basepairs(bp) or 23 bp long (12-RSSand 23-RSS, respectively),which displays somesequencepreferencesproximalto the heptarner' but isotherwisenot wellconserved. V(D)J recombinationisgenerallydirected betweentwo genesegments with differentRSSs,a restriction termed the 12123 rule thatservesto facilitate productivereceptorgeneassembly.

The biochemistryofV(D)J recombinationcan be conceptuallydividedinto a cleavage phaseand a joining phase (Fig. 1). To initiate the cleavage phase, two lymphoid cell-specific proteinsencoded by recombinationactivatinggene-I and -2 (RAG1 and RAG2, respectively-"), possiblyassisted byhigh mobilitygroupproteinsofthe HMG-boxfamily(HMGB 1and HMGB2,calledHMGBl/2 henceforth; discussed further below),bring two differentgene segmentsinto closeproximity through interactionswith the adjoining 12- and 23-RSS (forming a "synaptic" complex) and then catalyze a DNA double-strand break (DSB) at each RSSbetween the heptamerand the codingsegment.5.6RAG-mediatedcleavage producestwo typesof DNA ends: blunt and5' -phosphorylatedsignalends containing the RSS and coding ends covalently sealedas DNAhalrpins.Y'Ihese reactionintermediatesoriginatefrom a two-step cleavage mechanismin which

·Corresponding Author: PatrickC. Swanson-Departmentof Medical Microbiologyand Immunology, Creighton University Medical Center, Omaha, Nebraska 68178, USA.Email: pswansonecrelghton.edu

V(D)J Recombination, edited by Pierre Ferrier. ©2009 Landes Bioscienceand SpringerScience+Business Media.

2 V(D)JRecombination

I RAG1/2t HMGB1I2 (1)

I RAG1/2t HMGB1 /2 (1)-- ------D

n-II)AI<AI

u:::aII)

."::sAlenII)

'e--:::s--:::su:::a."::sAlenII)

SignalEnds

Ku70/Ku80

I \ XRCC4/DNA Ligase IVCernunnos(XLF)

~

CodingEnds

~RAG1/2 (1)Artemis/DNA-PKcsTdT, Poill/A--0

CodingJoint

SignalJoint

Figure 1. Overview of V(D)j recombination (adapted from Fugmann et aI6). In the cleavagephase of V(D)j recombination, cod ing segments (filled rectangles), flanked by a 12-RSS or23-RSS(small or large triangles, respectively) are assembled into a synaptic complex by theRAG proteins, possibly assisted by HMGB1/2 (filled ovals). Coupled cleavage by the RAGprote ins yields blunt Signal ends and coding ends sealed as DNA hairpins. In the jo iningphase of V(D)J rearrangement, sealed coding ends are resolved by an Artemis/DNA-PKcscomplex and may be further processed by TdT (if present) and DNA polymerases 11 and/or A(Poll.l/A). Processed coding ends are joined to create imprecise coding joints that may havegained palindromic (P)or nontemplated (N) nucleotides through asymmetric hairpin openingor TdT-mediated addition, respectively, or lost nucleotides through end processing reactions(open rectangle). Signal ends are joined to create signal joints that are typically precise.Alternative, less frequent joining events, such as open-shut and hybrid joints are not shownfor simplicity. Signal and coding joint formation is mediated by the NHEj pathway, whichincludes Ku70, Ku80, XRCC4, DNA Ligase IV and Cernunnos (XLF). Although the processingand joining reactions are shown as sequential processes, these steps may be integrated anditerative for joining of incompatible coding ends, involving single-strand ligation, processingof the unligated strand by Artemis/DNA-PKcs and DNA polymerases and eventual ligationof the second strand resulting in repaired double-stranded DNA.101

Early StepsofV(D}]Rearrangement: Insightsfrom BiochemicalStudiesofRAG-RSS Complexes 3

the RAG proteins first nick the RSSat the 5' end of the heptarnerand then use the resulting3' -OH to catalyze adirect transesterification reactionon the opposingphosphodiesterbond,"Inthe joiningphase, the twosignal endsaretypically ligatedprecisely, formingasignaljoint, and thecodingendsaresubjectedto reactions that resolve the hairpinsand then process and connect theDNA endsto formcodingjoints.Asaresult.codingjointsofienshowevidence ofnucleotidegainor lossat thecodingends.Infrequently. alternative outcomesofV(D)J recombination areobservedin which one genesegmentis joined to the RSS of another genesegment ("hybridjoint") or isseparatedand rejoinedto the sameRSS ("open-shut joint").I0·11 Efficient signaland codingjointformationrequires a competentnonhomologousend-joining(NHEJ) repairpathway. includingKu70.Ku80,XRCC4, DNA Ligase IVandXLF/Cernunnos.P:" Codingjoint formationrequirestwo additionalfactors not strictlyessential forjoiningsignalends.Artemisand DNA-PKcs. whichtogether function asa structure-specific endonuclease responsible foropeningthe DNA hairpinson codingends,"Asymmetric hairpinopeningcangiveriseto palindromic(P) nucleotides beinginserted in coding joints. Terminal deoxynucleotidyl transferase (TdT) and DNA polymerases!J. and/or;" (Pol !J./;") can further diversify these junctional sequences by catalyzing addition ofnontemplaeed (N) nucleotides to coding ends (TdT) and processing incompatible DNA endsto facilitate end-joining(Pol !J./;").16.17A detailed consideration of the proteins involved in theprocessing and repairofV(D)J recombination intermediates is beyondthe scopeof this review.but hasbeendiscussed elsewhere.P'"

Here wereview and discuss the molecularmechanisms ofV(D)J recombination. focusing onthe cleavage phaseof this process and emphasizing newinsights. Readers are referredto previousreviews for more detailed discussion of earlystudies of RAG protein biochemistry, includingthe establishment of cell-free assays ofV(D)J cleavage and joining5.6 and the identification andcharacterization of the various structuraldomainsof the RAGproteins."

Assembly and Organization ofSingle Site andSynapticRAG-RSS Complexes

Cell-free assays ofV(D)J cleavage established usingtruncated.catalytically active "core" formsof RAG1 (full-length 1040 a.a.; core residues 384-1008) and RAG2 (full-length 517 a.a.; coreresidues 1-387)demonstratedthat the RAG1/2 complex isboth necessary and sufficient to mediate RSScleavage? and that RAG cleavage activity exhibitsmetal ion-dependence: Mnl. supportsRAG-mediatedcleavage of a single RSS, whereas Mg2+ is required for coupled cleavage ofRSSpairsabidingby the 12123 rule.22.23 In natural progression. laterstudiesidentifiedand characterizeddiscreteRAG-RSS complexes with increasing complexity, with earlyworkfocused on RAGcomplexes assembled on a single RSS and laterworkanalyzing higher-order RAG synaptic complexes. Most of this workhas been reviewed and discussed elsewhere.5.6.24 Therefore. only salientfeatures willbe highlightedhere.

Core RAG1 containsthreestructurallydistinct regions:" an amino-terminalnonamer bindingdomain (NBD. residues 389-442)that interactswith the RSS nonamer,25.26 a centraldomain(residues 528-760)that recognizes the heptamerand exhibitssingle-strandDNA bindingactivityand a C-terminal domain (residues 761-979) that binds double-stranded DNA nonspecificallyand cooperatively. Core RAG1 aloneexists in solutionprimarilyas a stabledimer7-29and bindsan isolatedRSSwith moderate affinity(Kd -41 nM)28 as a dimer7,28.3o (although higher-orderaggregates aredetectableat elevated RAG1concentrations and conditionsoflowionicstrength")whereas RAG2ispredominantlymonomericin solution" and shows little. if anyDNA bindingactivity.2s.26.32.34 RAG1 and RAG2 interact with one another in the absence of DNA27.29.3S andtogetherbind asingle RSS with greaterspecificity than RAG1alone.32.33.36 Purified coreRAG1/2proteins variably assemble one29.32,33 or two34.37major protein-DNA complexes detectableusingan electrophoretic mobilityshifiassay (EMSA).Therelative abundanceof thesecomplexes, nowgenerally called SC1andSC2(for"single RSS complex"),dependspartlyon howthe RAGproteinsareexpressed andpurified:37.38inour laboratory. individually expressed andpurifiedRAGproteinstend to assemble only SC1, coexpressed RAG proteinspurifiedunder high salt conditionsform

4 V(D)j Recombination

more SCI than SC2 and coexpressed RAG proteins purifiedusingmilderconditionspredominantlyassemble SC2. Both complexes possess similarintrinsiccleavage activity,3oIJ7 but differinRAGprotein stoichiometry. Swanson reponed that both complexes contain a RAGI dimer, butincorporateeitherone (SCI) or two (SC2)RAG2molecules. " Mundyet al reponed comparableresultsfor RAG2 in these complexes, but presentedevidence suggesting SCI and SC2 containthree or more RAGI subunits.r' Possible explanations for this apparent discrepancy havebeendiscussed previously"and will not be revisited here,but wenote that recentdatareponed byDeet al provides corroboratingevidence supporting the contention that RAGI exists as a dimer inanSC (RAG2stoichiometrywasnot determined).39ThetetramericRAGl/RAG2 configurationreponed for SC2 isalsoconsistentwith data publishedbyBailinet al.29

Mutagenesis studies4042revealed that RAGI contains three carboxylate residues (asp-600,asp-708and glu-962)criticalfor catalysis that resemble a "DDE motif" found in manytransposases and integrases." Similar to the TnS transposase,44.45 biochemical studiesestablished that asingle RAGI subunit contributesall threecarboxlare residues to single active sitewhich mediatessequentialnickingandhairpin formationstepsofthe cleavage reaction46.47and that thesereactionsarecatalyzed in trans; that is,bythe subunitof the RAGI heterodimernot bound to the nonamerofthe RSSbeingcleaved."

While the RAG proteins themselves are sufficient for assembling SCI and SC2, HMGBl/2proteins areknown to facilitate RAG-mediatedbinding and cleavage of an isolated23-RSS, butnot a 12-RSS, in vitro.48TheRAGproteinsalsorequirethe presence ofHMGBl/2 to efficientlyassemble a complexcontaininga complementary (12/23) pairofRSSs ("pairedcomplex" or PC)and mediate coupled cleavage at both RSSs adhering to the 12/23 rule in vitro.48•49WhetherHMGBl/2 also assist theRAGproteinsduringV(D)Jrecombination invivo hasnot beenformallyestablished nor entirely ruledout,SO sinceHMGB1/2 exhibitfunctionalredundancyin RAGbindingand cleavage assays," The HMGBl/2 proteins are nonhistone chromosomal DNA bindingproteinsknownto promote DNA bendingand facilitate assembly ofnucleoproteincomplexese"HMGBI further functionsas an alarmin to signalcellular damagein response to inflammatoryprocesses.P HMGBl/2 proteinscontain tandemhomologous HMG -boxdomains(calledA andB) attached to a basiclinkerand an acidictail. HMGBl/2 interactswith the NBD ofRAGI inthe absence of DNA and enhancesthe intrinsic DNA bending activityof the RAG proteins."TheintegrationofHMGBl/2 into RAG-RSS complexes canoften bedetected asa supershifi: byEMSA.51,55Recent structure-function studiesconducted in our laboratory56.57suggest that bothHMG-box domainsmust be competent to bend DNA and physically linked together in eitherorientation (AB or BA) to stimulateRAG-mediated23-RSScleavage in the presence ofMg4.Interestingly, single HMG -boxdomainscan be integratedinto 23-RSS-RAG complexes,56-58 butcannotstimulate23-RSS cleavage unless Mn4 replaces Mg4 in the reaction,57.58 or 12-RSSpartneris added to promote synapsis.57Theseresultssuggest the two HMG-box domainshaveseparablebut potentially redundant rolesin stimulatingRAG binding and cleavage activityin vitro andthat synapsis promotesa conformational changethat bypasses the needfor one of thesedomains.HMGB I lackingthe acidictailstimulates RAGbindingand cleavage activityat lowerconcentrationsthan full-length HMGBl, but promotesaggregation ofRAG-RSScomplexes.56-58 Moreover,loss of the acidic tail enables HMGBI mutants that otherwise fail to suppon RAG-mediatedsynapsis to stimulatePC formation.56These data suggest the acidic tail helpsmaintain the correct oligomerization state of RAG synaptic complexes. The acidictail is alsoknown to facilitateHM GBl-mediated nucleosome repositioning,59.60 which mayhelppromote RSSaccessibility innucleosomal DNA.61-63

Synaptic complex assembly isthought to proceedviainitialformationofSC2 followed bycapture of an appropriatepartner RSS to forma pc. This"capture model"ofassernblywas suggestedinitiallyby biochemical experiments demonstratingthat SC2 can be driven to form the PC byaddingappropriatepartner RSS30I and the observation that RAGcleavage activityisgreaterwhensynapticcomplexes areassembled in step-wise fashionbyaddingfree23-RSS to a 12-RSS-RAGcomplex(or viceversa) than when they are assembled by mixingpreformed 12-RSS-RAG and

Early StepsojV(D)JRearrangement: Insightsftom BiochemicalStudiesojRAG-RSS Complexes 5

23-RSS-RAG complexes together/"This modelhas gained in vivoexperimental support from arecentstudybyCurry et al6Sshowing that nickscanbe detectedat endogenous12-RSSs, but notat 23-RSSs, in lymphoidcells. Thesefindings leadthe authors to proposea model in which RAGproteins bind and nick a 12-RSS first, then captureand nick a 23-RSS and, in rapid succession,finally cleave both RSSs. Thismodelisconsistentwith previous biochemical studiesshowingthatnickingcanoccuron an RSS in the absence of synapsis,66.67but nickingat one RSS isrequiredforefficient cleavage ofits partner.22,66The capturemodelis also consistentwith data this laboratoryand othershavepublishedshowingthat the complementofRAGproteins is the samebetweenaRAG complex bound to a single RSS (asSC2) and the PC.34.37Interestingly, thesestudiesshowthat molecules ofRAG2, but not RAG1,freely re-assertduringPC assembly.34.37Workfrom thislaboratorysuggests that the PC containstwo molecules eachof RAG1 and RAG2 and that thisheterotetramerconfiguration remains the samethrough the cleavage stepsofV(D)J recombinanon." Anotherstudyreportedthesamestoichiometryfor RAG2in the PC,34 but othersconcludethe PC containsthreeor moreRAG1subunits.34.46 Possible scenarios to explainthesediscordantresultshavebeendiscussed elsewhere."

How are the RSSs arrangedin the synaptic complex? Earlyobservations that the efficiency ofin vitro coupledcleavage" and in vivo V(D)J rearrangemenr'" is more sensitive to shorteningofthe intersignal distancewhen the RSSs arepositioned in an inversional configuration than whenthey are positioned in a deletionalconfiguration argued that the RSSs are aligned in a parallel,rather than anti-parallel orientationin thesynapticcomplex. Totest thispossibility moredirectly,Cibutaruet al recently measured levels offluorescence resonance energytransfer(FRET) in RAGsynaptic complexes assembled under various conditions on 12- and 23-RSS oligonucleotidesubstrates labeledwith FAM and TAMRA in differentconfigurarions/" Significant FRET wasdetectedonlywhen the following threeconditionsweremet: (i) the fluorophores wereplacedondifferentRSSs (but not the sameRSS); (ii) the two RSSs containeddifferentlength spacers [i.e.,abiding by the 12123 rule); and, (iii) synaptic complexes wereassembled in binding reactionscontaining Mg2- and the full complementof RAG1/2 and HMGBl/2 proteins. Interestingly,FRET wasobservedin synaptic complexes regardless ofwhich end of a given RSS waslabeled;the only apparent requirementwasthat the two fluorophores wereplacedon differentRSSs (12and 23).Thesedatasuggest that the distancebetweenthe endsof the two bound RSSs in the synaptic complex areapproximately the same. Giventhis constraintand limitationson the maximaldistancebetweenfluorophores to observe FRET, the authors proposethe two RSSs likely adopta bent and crossed configuration in the PC.69

Insightsinto RAG-Mediated RSSRecognitionandCleavage MechanismsInteractionsbetween the RAG proteins and DNA havebeen investigated usinga varietyof

approaches and the insightsfrom thesestudieshavegreatlyimprovedour understandingof howthe RAGproteinsrecognize and cleave their RSStargets.Muchof the earlyworkhasbeenextensively reviewed,s.6.24 soit willnot be covered in depth here.Chemicaland DNaseI protection andmodification interference footprintingassays performedon RAGcomplexes assembled on asingleRSSsuggest RAG1 primarilyinteractswith the nonamer and adjacentspacersequence, whereasRSS contactsin complexes containingboth RAG proteins are overlapping, but more expansive,extendingfrom the nonamer, through the spacerand into the 3' end of the heptamer,with a biasofphosphatecontactstowardoneface of the DNA helix.32.70.71 Photo cross-linkingstudiessuggestRAG1 mediates most of the contact with the RSS, with RAG2-RSS interactionsmore localizedto the junction of the heptamer and coding segrnent.27.36.72.73 Integration of HMGB1/2 into23-RSS-RAGcomplexes enables detectionofheptamer-spacer contactsresembling thoseobservedin 12-RSS-RAG complexes that are not otherwisevisualized in 23-RSS complexes containingRAG1/2 alone,Sl .SS suggesting HMGBI stabilizes RAG association with the heptamer in thesecomplexes. Ethylation interference footprintingsuggests HMGB1/2 contactsthe 23-RSS proximalto the nonamer, expandingthe footprint of the RAG proteins in this region." Although RAGcontactsat the junction of the heptamer and codingsequence are not readily detected in RAG

6 V(D)jRecombination

complexes assembled on a single RSS, this region is protected from DNase I cleavagein synapticcomplexes." Nagawa et al showed that synaptic complexes assembled with nicked RSS substratesshow slight expansion ofthe DNase I footprint relative to precleavage synaptic complexes (from-12 nt to -16 nr), suggesting that RAG-mediated nicking causes more intimate and stable RAGassociation with the coding sequence." Pull-down assaysshowing that nicked RSS substrates aremore readily incorporated into synaptic complexes than intact substrates support thiscontention.Interestingly, two different joining-deficient RAGI mutants (S723C76 and K118/9A77) wereshown to exhibit poor protection of the heptamer-coding junction, leading to speculation thatthe joining defect is caused by poor coding end retention in the postcleavage synaptic complex,"However, closeinspection ofthe mutant RAG 1footprintingpatterns in precleavagecomplexesalsoreveals that these mutants exhibit less protection ofspacer and nonamer sequences compared towild-type RAG l.1his observation argues that these mutations causea global defect in RAG-RSScomplex stability, but can also be interpreted to suggest that the RAG proteins require stablecontact with the coding sequence in order to maintain strong interactions with the RSS (or viceversa) in precleavage complexes.

Direct and interference footprinting experiments suggest RAG-RSS complex formationis accompanied by structural distortions in the spacer region and near the site ofDNA cleavageY·51.70.71 Studies showing that the RAG proteins mediate RSS bending, which is augmentedby HMGB1I2,54 plausibly explain spacer hypersensivity to chemical and enzymatic probes inRAG-RSS complexes. Structural distortions near the cleavage site are likely attributed to baseunpairing mediated by the RAG proteins to promote hairpin formation. which is suggestedby observations that RAG -mediated RSS cleavage is facilitated by incorporating base-pairmismatches78.79or abasic sites80at the coding flank. Clues to how these structural distortionsmay be induced and stabilized are suggested by structural studies ofthe related Tn5 transposase,which, like the V(D)J recombinase, catalyzes DNA hairpin formation (except that hairpins areformed at the transposon end, which is equivalent to the signal end in V(D)J recombination]."Analysis ofa Tn5 postcleavage synaptic complex reveals that the transposase promotes extrusion ofa thymine from the DNA helix, stabilizing the "flipped base" via stacking interactionswith an aromatic tryptophan residue (trp-298).44 Recent studies indicate a similar mechanismis operative in V(D)] recombination. Two lines of evidence suggest the terminal nucleotideon the bottom strand of the coding flank (C 1b, see Fig. 2 inset) is stabilized in an extrahelicalconfiguration by the RAG proteins. First, when thymine is incorporated into the RSS at position Clb, this base exhibits hypersensitivity to permanganate modification under conditionsfavoring RAG -RSS synaptic complex formadon." Second, base removal at Clb potentiateshairpin formation." Both outcomes are consistent with comparable studies of the flipped T2thymine in the Tn5 transposon end.83.84One notable contrast between the two recombinationsystems is that although the base subjected to flipping in the RSS coding flank and the Tn5transposon end are both located opposite the nicking site within the hairpin-formingsequence,they are offset from one another by one nucleotide: in the RSS. the base is at the terminus ofthe sequence; in the Tn5 transposon end . it occupies the penultimate position.

When does base-flipping occur during RSS cleavage?Base-flipping appears to occur after nicking , rather than upon RAG binding to the RSS, as permanganate hypersensitivity is not observedin RAG synaptic complexes assembled on intact substrates." Interestingly,permanganate interference assaysreveal that intact substrates bearing oxidized thymine at Clb and S2b are selectivelybound by the RAG complex relative to unmodified substrates , with the latter modification beingmuch preferred over the formerY·51 Ifthe RAG proteins stabilize base-flipping at Clb during thehairpin-forming step, why is prior modification ofS2b selected over C 1b in interference assays?Since base-flipping is most evident in synaptic complexes assembled on nicked substrates," onepossibility is that a conformational change in the RAG complex occurs after synapsis or nickingthat alters the position ofthymine binding pocket relative to the cleavagesite. Thus, an oxidizedextrahelical thymine at S2b may be preferentially accommodated over Clb in the bindingpockctofa RAG complex bound to an intact RSS. Alternatively, modified S2b may be selected becauseClb is more easily flipped ifthe oxidized base at S2b is already displaced from the DNA helix.

Early Steps ofV(D)JRearrangement: Insightsfom BiochemicalStudiesofRAG-RSS Complexes 7

I 7 I 12 9 I -/ ............C2t CII 511 52t

TIIIIObClb51b52b

HO P:>.<I 7 12 9 I

7 I 23 9 I

HO p

~6 I 12 9 II

ClbHO P~(7 23 9 I

IClb

Figure 2. Integrated model of synaptic complex assembly and coupled RSS cleavage. In thissimplified scheme, RAGl (Rl) contains an active site domain (ASD) that includes the DOEmotif (found within central and C-terminal domains that are not shown") and a nonamer binding domain (NBD). RAG2 (R2) is depicted as a small oval. RAG-RSS complexes are shownat right and reactions catalyzed on the RSSs are diagrammed at left. Nucleotide positions onthe top (t) and bottom (b) of the coding (C) and signal (S) sequence at the heptamer-codingjunction are also indicated (inset, upper left). RAGl and RAG2 form a complex, shown hereas a heterotetramer based on our work" and others" (but see text), that preferentially bindsa 12-RSS. The RAG complex bends and nicks the 12-RSS at the 5' end of the heptamer andthen captures a 23-RSS to form a PC in which both RSSs are bent and cross over one another.HMGB1/2 (H) may assist in this process at the 23-RSS. The RSSs are shown here wrappingaround the outside of the RAG1/2 complex (adapted from Ciubotaru et aI69). An alternativemodel in which the RSSs cross over each other on the same face of the protein complex is notshown for simplicity, but isan arrangementthat meets constraints imposed by FRET data.69 Notethat the bending and crossing angles shown here are not meant to represent angles derivedfrom experimental measurements. The 23-RSSis nicked in and, in rapid succession, the RAGproteins catalyze hairpin formation at both RSSs by a mechanism involving base-flipping atCl b. RAG-mediated cleavage is shown here catalyzed by a single ASD in trans (i.e., the RAGlsubunit bound to the 12-RSS nonamer cleaves the 23-RSS and vice versa) based On studiesof SC1,47 but this configuration has yet to be formally established for the PC. After cleavage,coding ends are likely released first, with the RAG prote ins remaining bound to the signalends until the signal ends complex is disassembled.


Totest whether aromaticresiduesin RAG1participate in base-stacking interactions to promote hairpin formation byanalogyto Tn5 transposition. two differentlaboratoriesperformedsite-directedmutagenesis ofaromatic residuesin RAG1.80,85Lu et al screenedall evolutionarilyconservedaromaticresiduesin the catalyticcoreofRA G1.selectingmutants failingto supportV(D)J cleavage in cells and exhibiting selective impairment of hairpin formation in vitro."The authors identified trp-893 of RAG1 as a plausiblecandidate for mediating base-stackinginteractions, based on the inability of a W893A RAG1 mutant to support hairpin formationand the rescueofthis defectby replacingalaninewith tyrosineat residue893 or byintroducingmismatched basepairs near the RSScleavage site. In contrast. a later. more limited mutagenesis study concluded that trp-893 is unlikely to mediate base-stacking because the cleavagedefect observed with the W893A RAG1 mutant was found to depend on the coding flankcomposition." Specifically, Grundy et al showed that RSSsubstratescontaining "bad" codingflanks (5 ' -GATTC-3' or 5' -TCGAC-3') are cleaved lessefficiently by W893A RAG1 thanbywild-typeRAG1, but wild-typeand W893A RAG1 exhibit similaractivityusingsubstratescontaining "good"codingflanks(5 ' -ACCTG-3 '). Thus,the authors speculatedthat a trp-893mutation affects astepfollowing cleavage. However, because theW893A RAG1mutant supportsmoderate cleavage ofoligonucleotidesubstratesunder conditions favoringsynapsis in trans butpoor nickingandhairpinformationwhenthe sameRSSs areembeddedin cisin aPCR-generatedsubstrate," it isalsopossible that trp-893 mediatesprotein-proteinor protein-DNA interactionsto facilitatesynaptosome assembly and activityon longer,morephysiological substratesthat arelargelydispensablein reactionsperformed on oligonucleotidesubstrates.

Ratherthan trp-893,Grundyet alarguethat trp-956 isa moreplausible candidateforstabilizingbase-flippingbecause althoughaW956A mutant exhibitsdefects in both nickingand hairpinformation in Mg4 (alsoreported by Lu et al85), W956A RAGI cleavage activityis substantiallyrescuedby incorporatingan abasic site at Clb of the RSS substrate." That the W956A RAGImutant is substantially impairedin catalyzing both stepsof the cleavage reactionin Mg4 is nottheoutcomeexpectedbasedon the precedentsetbyanalysis of itspresumedcounterpart,W298ATn5, which exhibits defects in hairpin formation, but not nicking." However, given the closeproximity of trp-956 to glu-962, which is required for catalysis,40·41 a W956A mutation maycause structuralalterations in the active site that prevent the RAGsfrom nickingRSSsubstratesefficiently. Alternatively. theobservation that introducingabasic sitesat Cit and C2t of thecodingflank(seeFig.2,inset)blocksthe nickingstepraises thepossibility that trp-956 isinvolved inbothcleavage stepsofV(D)J recombination, firstto help identifywherethe nickshouldbe introducedand second. perhapsfollowing a conformational change, to help stabilize the extrahelica1 baseatClb in preparationfor hairpin formation.

Elements Guiding Enforcement ofthe 12/23 RuleHow the 12/23 rule is enforcedat the molecularlevel still remains somewhatmysterious.As

discussed previously," the 12/23 ruleislikely enforcedboth at the level ofsynapsis and at thepointwhennicksat both RSSs areconverted to DNA double-strand breaks.At thelevelofsynapsis,Jonesand Gellertdemonstratedthat oncethe RAGproteinsbinda 12-RSS in the presence ofHMGB1,the complex becomes structurallybiasedagainstcapturinganother 12-RSSand insteadexhibitsa strongpreference for capturingand integratinga 23-RSS into a pc.64 However, the oppositeisnot true: RAGproteinsbound to a23-RSS exhibitonlya 5-6foldpreference for incorporatinga12-RSS partner overa 23-RSS partner into a pc. Theauthorsspeculate that due to the length ofthe 23-RSSspacer, the RAG proteins bound to this substratemayundergo rapid isomerizationbetween"12-RSS-like" and"23-RSS-like" RAGcomplexes, enabling thesecondsiteto beoccupiedbyeither typeofRSS,with onlymodestselectivity for a 12-RSS.Althoughthe authorsenvisionedbendingof the 23-RSSspacerasthe meansto achieve lsomerization.f datashowingthat the RAGproteins can aberrantlynick a 23-RSS in the spacerregionat a position equivalent to the 5' -endof the heptamerin a 12_RSS55,l16 raises the possibility that isomerization isalternatively achievedthrough "catch and release" of 23-RSSheptamer and spacersequences. The "conformational

Early StepsofV(n)]Rearrangement: Insightsfrom Biochemical StudiesofRAG-RSS Complexes 9

locking"modelproposed byJones and Gellert wasdevelopedbasedon experiments usingintactRSSsubstrates, but is equallyplausiblefor a scenarioin which the RAGproteins bind and nick a12-RSS beforesynapsis, which,asdiscussed above, is suggested to occur in vivo.65

Once bound to anicked12-RSS, the RAGproteinsmust identifyanaccessible 23-RSSpartnerin a backgroundof available 12-RSSs (intact or nicked) and other randomly nicked DNA. Theconformational locking model providesa framework to discriminate against binding a second12-RSS, but not a mechanismto do so. One possibilityis that the NBD in the RAG1 heterodimer not bound to the 12-RSSmaysampleincomingDNA sequences for nonamer-like elements.Should it find a suitablesequence, it maybind (modestly) to this motif, enablingsequences atthe appropriatedistanceto be interrogatedfor the presenceof a suitableheptamer.Thus,shoulda 12-RSS-RAG complex(as SC2) encounter another 12-RSS, the unoccupied RAG1 subunitcould bind it via NBD-nonamer interactions, but the heptamer's proximity would not allowit to be specifically engagedby the activesite of the RAG complex, causingthe RSS to eventuallydissociate. Alternatively, if the same 12-RSS-RAG complexencountered a randomlynickedsequence, the activesite maybind the nicked DNA weakly, but if the sequencelacksa suitablenonamer-like motif, the DNA wouldnot befullyanchoredto the RAGcomplexviathe NBD andthereforewould not triggertransesterification.Thus,onlywhen hepramerand nonamer elementsare both present and appropriatelyspacedin the partner RSSwould nickingof the partner andsubsequenthairpin formation at both RSSs be initiated. What is the criticalcheckpoint in thisprocess? Nishihara et al showed that base-flippingat C1b is only observedat a nicked 12-RSSwhen its appropriateparmer isbound by the RAG complex," Hence, the decisionto base-flip islikelyacriticalcheckpointin triggeringcoupledcleavage, asthis stepprovidesthe conformationalchanges required to promote transesterification.

What then influences the decision to initiatebase-flipping? Thisdecisionislikely influenced byhowthe RAGproteinsdetectsynapsis, asevidenced bythe recentidentificationofgain-of-functionRAG1mutants that exhibitenhancedin vitro RSScleavage in Mg2+ in the absence of synapsis.82.87

We identified an E649A RAG1 mutant that, relative to wild-type RAG!, exhibits enhancedRAG-mediatedhairpin formationin vitro,but doesnot display increased recombinationactivityof plasmidV(D)J recombination substratescontaining a 12/23 pair of signalsequences in cellculture. However, this mutant does support greater cleavage and recombination of substratescontainingamispairedor unpairedRSS, suggestive ofaselective defectinsensing12123-regulatedsynapsis. Whether the E649ARAG1mutant supportsbase-flippingin the absence of synapsis hasnot been tested,but a RAG1 mutant (calledHA3) with a similarphenotype wasrecentlyfoundto mediatesynapsis-independent base-flipping." It isnotable that in both reports. the mutationsconferringthe gain-of-function phenotype are located proximalto residues of the DDE motif.which suggests that the domain responsible for catalyzingthe stepsofV(D)J cleavage alsoplays akeyrole in sensing12/23-regulatedsynapsis and triggeringbase-flippingat the cleavage site.

Takentogether, thedatasummarized heresupport a modelof RAGsynapticcomplexassemblyand 12123-regulated cleavage shownin Figure2 that involves initial binding. bending and nickingof a 12-RSS by the RAG complex, followed by the selective captureand integration of a free23-RSSinto a synapticcomplexin which the two RSSs adopt a bent and crossed configurationand finally completedby23-RSSnickingand facile conversionof nicksat both RSSs into DNAhairpinsbya mechanismthat involves base-flipping at C1b.Theconformationalchanges requiredto mediate this process on physiological substratesmaybe facilitated in part by mechanisms thatunderwind DNA, assuch substrates are cleaved more efficiently by the RAG proteins in vitro."Basedon data from this laboratory,wespeculatethat the cleavage reactions are mediatedin transby a RAGl/RAG2 heterotetramer, but acknowledge that this organization remains to be fullyvalidated. Geneticand biochemical evidence reviewed elsewhere-"suggests that afrercleavage, thesignaland codingendsareheld transientlyinafour-end"post-cleavage synaptic complex",but codingendsarepoorlyretainedwithin this complex, whereas the RAG proteins remainstablyboundto the signalends.Thisdifferential retention isreflectedin the apparentuncouplingof codingandsignaljoint formation,with the formeroccurringmore rapidlythan the latter.

10 V(D)]Recombination

Transcription Factor-Assisted Targeting ofAntigen Receptor LociFigure2 presentsapictureofthe RAGproteins (with HMGB1/2) asbeingsolely responsible

formediatingsynapsis duringV(D)Jrecombination. However, thisviewisoverlysimplistic.becauseaccumulating evidence suggests that the RAG complexcan be preferentially targetedto specificantigen receptor loci through interactionswith cellular factorsthat mark accessible and activelyrearranging loci (suchasmodifiedhistories ),90-92 or can bind to specific siteswithin particularantigen receptorgenes.93,94 Here wewill briefly review studiesof the latter class of RAG interactionfactorsand discuss the findings as they relateto RAG-RSS complexassembly.

PaxS isaBlineage-specific transcriptionfactorthat regulates manyBlineage-specific genes andisrequiredto support rearrangement ofDwdistal VHgenesegments duringlymphocytedeveloprnenr," Zhang et al recently discovered that 94% of VHcodingregions (whichareall flanked bya 23-RSS)contain two or morepotential PaxS bindingsites.93Theauthorsshowedthat PaxS canindeed bind thesesitesand promote RAG-mediatedcleavage and rearrangement ofdifferentVH23-RSSs when PaxS bindingsitesarepresentin the flankingcodingsequence. Theauthorsfurtherdemonstrated that PaxS directlyinteractswith the RAG proteins; this association requirestheN-terminalpairedDNA bindingdomainofPaxS and isobservedonlywhenboth RAGproteinsarepresent. Based on thesedata, the authorsconcludethat PaxS promotesVWto-DJH rearrangement bystabilizing RAG binding to the VH23-RSS viabridginginteractionsbetweenthe RAGproteinsand the PaxS bindingsite.Whether PaxS binds the codingregionfirstand then recruitsthe RAG complex to the 23-RSS, or, alternatively, whether PaxS stablyinteractswith the RAGproteins beforeRSSengagement and maintains this association after the RAG proteins bind a12-RSSin order to facilitate synapsis with a 23-RSS (containingPaxS bindingsitesin the codingsequence) wasnot directlytested in this study. Ifthe latter weretrue, one mightexpectthat PaxScouldsupershlfi a 12-RSS-RAG complex byEMSA.

How theestablished orderingofTCRf3locusrearrangements (D~-to-J~ recombination preceding V~-to-DJ~ rearrangement) is enforced remains in queseion." To explain this phenomenon,Wanget al94 investigated whether D~ 23-RSSs contain a transcriptionfactor recognitionsite(s)through which its bindingcould direct RAG-mediatedDp-to-]p rearrangement in preference toV~-to-DJ~ recombination.Theauthorsprovideevidence that TCR 3'-Dli 23-RSSs containanAPItranscription factor binding site,which extendsfrom the 6th bp of the heptamer to the 5th bpofthe spacerand that the API component c-Foscan bind to this sequence. c-Fos wasshown topromote RAG association with a 3' D~ 23-RSS and enhanceD~-J~ recombinationin cells, while.conversely,reducingV~-D~ rearrangement. These effectswereabolished iftheputativec-Fosbindingsite wasmutated. Micedeficient in c-Fos wereshown to exhibit impairedTCRf3 rearrangementoverall, but elevated levels ofmis-ordered V~-DJ~ recombination. Whether directV~-to-J~ recombination wasalsoelevated in thesemicewasnot directlytested,but wouldhavebeen interestingto determine because this rearrangement is formally permitted by the 12123rule. The authorsshowedthat c-Fos associates with the coreRAG proteins (primarilycore RAG2),requiringtheDNA bindingdomainand leucinezippermotifofc-Fosfor this interaction. Interestingly. unlikePaxS,93the transcription-activation domainofc-Fos isnot requiredto stimulateV(D)J rearrangemenc." Thus,the authorsconcludethat c-Fosmayfacilitate the selective recruitmentofthe RAGproteinsto the 3'D~ 23-RSS, therebypromotingpreferential D~-J~ rearrangement. Asisthecase forPaxS, the order ofeventsthat leadsto c-Fos association with the RAGsynaptic complex remainsunclear. What is strikingabout the locationof the API bindingsite in the 3'D~ 23-RSSis that itencompasses the sameregioncontactedbythe RAGproteinsin a23-RSS-RAG protein complexassembled in the presence ofHMGBIY Indeed,structuralstudiesofAPI-DNAcomplexes" suggest that API wouldengage this sequence in a mannersimilarto the RAGproteins," interactingprimarilywith the majorgrooveand contactingsomeof the samephosphodieseer bonds in theRSS. Sincethe two protein complexes cannot occupythe samespace. wespeculate that in thesecomplexes, RAG-mediatedinteractionswith the RSS at this locationarcfunctionally replaced byAPI contacts.Theportions of the RAGproteinsnormallymediatingthesecontactsmaybefreedto engage another DNA sequence. One intriguingpossibility is that the displaced RAG DNA

Early StepsofV(D)jRearrangement: Insightsfrom BiochemicalStudiesofRAG-RSS Complexes 11

binding domainscontact the 3' D~-12-RSS and through this engagement, help prevent it frombecominga target for synapsis with an upstreamV~-23-RSS.

Conclusion and Future DirectionsAccumulating evidence supports a capture model of RAG synaptic complexassembly and

coupledRSS cleavage that is initiated byRAG bindingand nickingofa 12-RSS and followed bythe 23-RSS captureandcleavage ofboth RSSs usingabase-flippingmechanism to facilitate hairpinformation. The stoichiometryand organizationof the RAG proteins in the synapticcomplexisstillcontroversial and uncertainand will not likely be resolved until it yields to structuralcharacterization. The base-flipping strategyused by the RAG proteins to mediatehairpin formation isalsousedby the Tn5 transposase during transpositionand represents yet another parallelamongthe many mechanistic similarities between V(D)J recombination and transposition that havebeen recognized over the years," There is little doubt that as years progress. additional featuresheld in commonbetweenthesesystems willbe discovered. One of the moredifficultprocesses tounderstandin thesesystems ishowsynapsis issensed. Forthe RAGproteins. this process remainsmysterious. but the active site itself appears to play an important role. as mutations in RAG1near the DDE motif have recentlybeen identified that enable the RAG complex to mediatebase-flipping and V(D)] rearrangement in violationof the 12/23 rule. The molecularbasisfortheseeffects remains to beelucidated. Recentevidence alsosuggests that the choiceofwhichRSSsto assemble into a synaptic complex maybe guided by interactionsbetween the RAG proteinsand other DNA binding factors. The findingthat core RAG proteins interact with HMGB1I2and. more recently. two different transcriptionfactors. suggests that the core RAGI/2 complexcontainsone or moreprotein interactiondomainspotentiallycompetent to mediateassociationwith a varietyof DNA bindingproteins.This raises the possibility that previously observeddifferences in antigenreceptorgeneusage99•

IOO mayin somecases beexplainedbycellularfactorsthatbind DNA at sites proximal to the RSS and promote RAG-RSS complexformation by directinteractionwith the RAGproteins.

AcknowledgementsThe authors wish to acknowledge support from the National Institures of Health (ROI

AI055599) and fundingfrom the AmericanCancerSociety(RSG-O1-020-0l-CCE).

References1. Bassing CH. Swat W. AIt FW The mechanism and regulation of chromosomal V(D)J recombination.

Cell 2002: 109(Suppl) :S45-55.2. Ramsden DA. Baetz K. Wu GE. Conservation of sequence in recombinat ion signal sequence spacers.

Nucleic Acids Res 1994: 22(10) :1785-1796 .3. Schatz DG. Oettinger MA. Baltimore D. The V(D)J recombination activating gene. RAG-I. Cell 1989:

59(6) :1035-1048 .4. Oettinger MA. Schatz DG. Gorka C ec al. RAG-I and RAG-2. adjacent genes that synergistically

activate V(D)J recombination. Science 1990; 248(4962):1517-1523 .5. Gellert M. V(D)J recombination: RAG proteins. repair factors and regulation . Annu Rev Biochem

2002: 71:101-132.6. Fugmann SD. Lee AI, Shocken PE et al. The RAG proteins and V(D)J recombination: complexes. ends

and transposition. Annu Rev Immunol 2000: 18:495-527.7. Schlissel M. Constantinescu A. Morrow T et al. Double-strand signal sequence breaks in V(D)J

recombination are blunt. 5' -phosphorylaced, RAG-dependent and cell cycle regulated. Genes Dev 1993;7(12B) :2520-2532.

8. Roth DB. Menetski jl', Nakajima PB et al. V(D)J recombination : broken DNA moleculeswith covalendysealed (hairpin) coding ends in scid mouse thymocytes, Cell 1992: 70(6) :983-991.

9. McBlaneJF, van Gent DC. Ramsden DA et al. Cleavage at a V(D)J recombination signal requires onlyRAGI and RAG2 proteins and occurs in two steps. Cell 1995; 83(3) :387-395.

10. Lewis SM. Hesse JE. Mizuuchi K et al. Novel strand exchanges in V(D)J recombination. Cell 1988;55(6) :1099-1107 .

II. Morzycka-Wroblewska E. Lee FE. Desiderio SV. Unusual immunoglobulin gene rearrangement leads toreplacement of recombinational signal sequences. Science 1988; 242(4876):261-263.


12. Lieber MR. Ma Y, Pannicke U et al. The mechanism of vertebrate nonhomologous DNA end joiningand its role in V(D)J recombination. DNA Repair(Arnst) 2004; 3(8-9):817-826.

13. BuckD. Malivert L. de Chasseval R et aJ. Cernunnos, a novel nonhomologous end-joining factor, ismutated in human immunodeficiency with microcephaly. Cell 2006; 124(2):287-299.

14. Ahnesorg P, Smith P, Jackson SP. XLF interacts with the XRCC4-DNA ligase IV complex to promoteDNA nonhomologous end-joining. Cell 2006; 124(2):301-313.

15. Ma Y, Pannicke U, Schwarz K et aJ. Hairpin opening and overhang processing by an artemis/DNAdependent protein kinase complex in nonhomologous end joining and V(D)] recombination. Cell2002; 108(6):781-794.

16. Bertocci B.De SmetA, Berek C et aL Immunoglobulin kappalightchaingenerearrangement is impairedin micedeficient for DNA polymerase mu.Immunity 2003; 19(2):203-211.

17. Bertocci B,De SmetA, WeillJC er aI.Nonoverlapping functionsof DNA polymerases mu,lambdaandterminal deoxynucleotidyltransferase during immunoglobulin V(D)J recombination in vivo. Immunity2006; 25(1):31-41.

18. Weterings E, Chen DJ.The endless tale of nonhomologous end-joining. Cell Res 2008; 18(1):114-124.19. Rooney S. ChaudhuriJ, A1t FW. The role of the nonhomologous end-joining pathway in lymphocyte

development. Immunol Rev2004; 200:115-131.20. LieberMR, Lu H. Gu J ecaI. Flexibility in the order of action and in the enzymology of the nuclease,

polymerases and ligase of vertebrate nonhomologous DNA end joining: relevance to cancer, agingandthe immunesystem. Cell Res2008; 18(1):125-133.

21. De P,Rodgers KK. Putting the pieces together: identification and characterization of structuraldomainsin the V(D)] recombination protein RAG1. ImmunolRev2004; 200:70-82.

22. Eastman QM, LeuTM. SchatzDG.lnitiation ofV(D)] recombination in vitro obeying the 12123 rule.Nature 1996;380(6569):85-88.

23. van Gent DC, Ramsden DA, Gellert M. The RAGI and RAG2 proteins establishthe 12123 rule inV(D)J recombination. Cell 1996;85(1):107-113.

24. Swanson FC, The bounty of RAGs: recombination signalcomplexes and reactionoutcomes. ImmunolRev2004; 200:90-114.

25. DifiJippantonio MJ,McMahan C]. Eastman QM er al. RAGI mediates signalsequence recognition andrecruitmentof RAG2 in V(D)J recombination. Cell 1996; 87(2):253-262.

26. Spanopoulou E, Zaitseva F, Wang FH er aJ. The homeodomain region of Rag-I reveals the parallelmechanisms of bacterial and V(D)] recombination. Cell 1996; 87(2):263-276.

27. Swanson PC. DesiderioS. RAG-2 promotesheptameroccupancy by RAG-I in the assembly of a V(D)Jinitiation complex. Mol Cell BioI 1999; 19(5):3674-3683.

28. Ciubotaru M. Ptaszek LM, BakerGA et al. RAGI-DNA binding in V(D)] recombination. Specificityand DNA-inducedconformational changes revealed by fluorescence and CD spectroscopy. J BioI Chern2003; 278(8):5584-5596.

29. BailinT. Mo X, Sadofsky MJ. A RAGI and RAG2 tetramer complex is active in cleavage in V(D)]recombination. Mol Cdl BioI 1999; 19(7):4664-4671.

30. Rodgers KK. Villey IJ. Ptaszek L er al. A dimer of the lymphoidprotein RAG1 recognizes the recombinationsignalsequence and the complex stably incorporates the high mobilitygroup protein HMG2.Nucleic AcidsRes 1999; 27(14):2938-2946.

31. GodderzLJ,RahmanNS, Risinger GM ecaI.Self-association and conformational properties of RAG1:implications for formation of the V(D)J recombinase. Nucleic Acids Res 2003; 31(7):2014-2023.

32. Swanson PC, Desiderio S.V(D)J recombination signalrecognition: distinct,overlapping DNA-proteincontacts in complexes containingRAGI with and without RAG2. Immunity 1998;9(1):115-125.

33. Hiom K. GellertM. A stableRAG1-RAG2-DNA complex that is active in V(D)J cleavage. Cell 1997;88(1):65-72.

34. MundyCL, Parenge N. Matthews AG et aI.Assembly of the RAG1/RAG2 synaptic complex, Mol CellBioI2002; 22(1):69-77.

35. Santagata S. Aidinis V. Spanopoulou E. The effect of Me2+ cofactors at the initial stages of V(D)]recombination.J BioI Chern 1998; 273(26):16325-16331.

36. Mo X, BailinT, Sadofsky MJ. RAGI and RAG2 cooperate in specific binding to the recombinationsignalsequence in vitro. J BioI Chern 1999; 274(11):7025-7031.

37. Swanson Pc. A RAG-l/RAG-2 tetramersupports 12123-regulated synapsis. cleavage and transpositionofV(D)J recombination Signals. Mol Cell BioI 2002; 22(22):7790-7801.

38. Raval P, Kriatchko AN, Kumar S et al. Evidence for Ku70/Ku80 association with full-length RAG1.Nucleic Acids Res 2008; 36(6):2060-2072.

39. De P,Zhao S. GwynLM et al, Thermal dependency of RAG1 self-association properties. BMCBlochern2008; 9:5.

Early StepsofV(D}jRearrangement: Insightsjom BiochemicalStudiesofRAG-RSS Complexes 13

40. Kim DR. Dai Y. Mundy CL ec al. Mutations of acidic residues in RAG1 define the active site of theV(D)J recombinase. Genes Dev 1999; 13(23):3070-3080.

41. Landree MA. Wibbenmeyer JA. Roth DB. Mutational analysis of RAG1 and RAG2 identifies threecatalytic amino acids in RAGI critical for both cleavage steps of V(D)J recombination. Genes Dev1999; 13(23):3059-3069.

42. Fugmann SD. ViUey IJ. Ptaszek LM er al. Identification of two catalytic residues in RAG 1 that definea single active site within the RAGl/RAG2 protein complex. Mol Cell 2000 ; 5(1):97-107.

43. Haren L. Ton-Hoang B, Chandler M. Integrating DNA: transposases and rerroviral lnregrases, AnnuRev Microbiol1999; 53:245-281.

44. Davies DR. Goryshin IY. ReznikolfWS et al. Three-dimens ional structure of the Tn5 synaptic complextransposition intermediate. Science 2000; 289(5476):77-85.

45. Naumann TA. Reznikoff WS. Trans catalysis in Tn5 transposition. Proc Nat! Acad Sci USA 2000;97(16):8944-8949.

46. Landree MA. Kale SB, Roth DB. Functional organization ofsingle and paired V(D)J cleavagecomplexes.Mol Cell Bioi 2001; 21(13) :4256-4264.

47. Swanson pc. The DDE motif in RAG-1 is contributed in trans to a single active site that catalyzes thenicking and eransesrerification steps ofV(D)J recombination. Mol Cell Bioi 2001 ; 21(2) :449-458 .

48. van Gent DC, Hiom K, Paull TT et al. Stimulation ofV(D)J cleavage by high mobility group proteins.EMBO J 1997; 16(10) :2665-2670.

49. Sawchuk DJ. Weis-Garcia F, Malik S et al. V(D)J recombination: modulation of RAG1 and RAG2cleavageactivity on 12/23 substrates by whole cell extract and DNA-bending proteins . J Exp Med 1997;185(11 ):2025-2032.

50. Calogero S. Grassi F. Aguzzi A et al. The lack of chromosomal protein hmg1 does not disrupt cellgrowth but causes lethal hypoglycaemia in newborn mice. Nat Genet 1999; 22(3):276-280.

51. Swanson pc. Fine structure and activity of discrete RAG-HMG complexes on V(D)J recombinationsignals. Mol Cell Bioi 2002; 22(5) :1340-1351.

52. Thomas JO. TraversAA. HMG1 and 2 and related 'architectural' DNA-binding proteins. Trends BiochemSci 2001; 26(3):167-174.

53. Klune JR, Dhupar R. Cardinal J er al. Hmgb1 : Endogenous danger signaling. Mol Med 2008 ;14(7-8):476-484.

54. Aidinis V. Bonaldi T. Beltrame M et al. The RAG1 homcodomain recruits HMG1 and HMG2 tofacilitate recombination signal sequence binding and to enhance the intrinsic DNA-bending activity ofRAG1-RAG2. Mol Cell Bioi 1999; 19(10) :6532-6542.

55. Yoshida T. Tsuboi A, Ishiguro K et al. The DNA-bending protein, HMG1 . is required for correctcleavage of 23 bp recombination signal sequences by recombination activating gene proteins in vitro.Int Immunol 2000; 12(5):721-729.

56. Bergeron S. Madathiparambil T, Swanson pc. Both high mobility group (HMG)-boxes and the acidictail ofHMGBI regulate recombination-activating gene (RAG) -mediated recombination signal synapsisand cleavage in vitro . J Bioi Chern 2005 ; 280(35):31314-31324.

57. Kriatchko AN. Bergeron S. Swanson PC. HMG-box domain stimulation of RAG 112 cleavage activityis metal ion dependent. BMC Mol Bioi 2008; 9:32.

58. Dai Y, Wong B. Yen YM et al. Determinants ofHMGB proteins required to promote RAG1/2-recombination signal sequence complex assembly and catalysis during V(D)J recombination. Mol Cell Bioi2005 ; 25(11) :4413-4425 .

59. Bonaldi T. Langsr G, Serohner R ec al. The DNA chaperone HMGBI facilitates ACF/CHRAC-dependent nucleosome sliding. EMBO J 2002 ; 21(24):6865-6873.

60. Ueda T, Chou H . Kawase T et al. Acidic Cvrail of HMGB 1 is required for its target binding to nucleosome linker DNA and transcription stimulation. Biochemistry 2004; 43(30):9901-9908 .

61. Kwon J, Imbalzano AN. Matthews A er al. Accessibility of nucleosomal DNA to V(D)J cleavage ismodulated by RSS positioning and HMG1. Mol Cell 1998; 2(6) :829-839 .

62. Nightingale KP, Baumann M. Eberharter A ec al. Acetylation increases access of remodelling complexesto their nucleosome targets to enhance initiation of V(D)J recombination. Nucleic Acids Res 2007;35(18):6311-6321.

63. Baumann M. Mamais A, McBlane F et al. Regulation ofV(D)J recombination by nucleosome positioning at recombination signal sequences. EMBO J 2003; 22(19):5197-5207.

64. Jones JM. Gellert M. Ordered assembly of the V(D)J synaptic complex ensures accurate recombination.EMBO J 2002; 21(15):4162-4171.

65. Curry JD. Geier JK , Schlissel MS. Single-strand recombination signal sequence nicks in vivo: evidencefor a capture model of synapsis. Nat Immunol 2005; 6(12):1272-1279.

66. Eastman QM. Schatz DG . Nicking is asynchronous and stimulated by synapsis in 12/23 rule-regulatedV(D)J cleavage. Nucleic Acids Res 1997; 25(21):4370-4378.


67. Yu K, Lieber MR. The nicking step in V(D)J recombination is independent of synapsis: implicationsfor the immune repertoire. Mol Cell Bioi 2000; 20(21) :7914-7921.

68. Sheehan KM. Lieber MR. V(D)J recombination : signal and coding joint resolution are uncoupled anddepend on parallel synapsis of the sites. Mol Cell Bioi 1993; 13(3):1363-1370.

69. Ciubotaru M, Kriatchko AN. Swanson PC et al. Fluorescence resonance energy transfer analysis ofrecombination signal sequence configuration in the RAGI/2 synaptic complex. Mol Cell Bioi 2007;27(13):4745-4758.

70. Nagawa F. Ishiguro K. Tsuboi A er al. Footprint analysis of the RAG protein recombination signalsequence complex for V(D)J type recombination . Mol Cell Bioi 1998; 18(1):655-663.

71. AkamatsUY, Oettinger MA. Distinct roles of RAG1 and RAG2 in binding the V(D)J recombinationsignal sequences. Mol Cell Bioi 1998; 18(8):4670-4678.

72. Eastman QM. Villey IJ, Schatz DG. Detection of RAG protein-V(D)J recombination signal interactionsnear the site of DNA cleavage by UV cross-linking. Mol Cell Bioi 1999; 19(5):3788-3797.

73. Mo X, Bailin T, Noggle S er al. A highly ordered structure in V(D)J recombination cleavage complexesis facilitated by HMG1. Nucleic Acids Res 2000; 28(5) :1228-1236.

74. Nagawa F. Kodama M. Nishihara T et al. Footprint analysis of recombination signal sequences in the12/23 synaptic complex ofV(D)J recombination. Mol Cell Bioi 2002; 22(20):7217-7225.

75. Nagawa F, Hirose S, Nishizumi H er al.Joining mutants of RAG1 and RAG2 that demonstrate impairedinteractions with the coding-end DNA. J Bioi Chern 2004; 279(37):38360-38368.

76. Tsai CL. Drejer AH . Schatz DG . Evidence of a critical architectural function for the RAG proteins inend processing. protection and joining in V(D)J recombination. Genes Dev 2002; 16(15):1934-1949.

77. QiuJX. Kale SB, YarnellSchultz H et al. Separation-of-function mutants reveal critical roles for RAG2in both the cleavage and joining steps ofV(D)J recombination. Mol Cell 2001; 7(1) :77-87.

78. Cuomo CA. Mundy CL. Oettinger MA. DNA sequence and structure requirements for cleavage ofV(D)J recombination signal sequences. Mol Cell Bioi 1996; 16(10):5683-5690.

79. Ramsden DA. McBlaneJF. van Gent DC et al. Distinct DNA sequence and structure requirements forthe twOsteps ofV(D)J recombination signal cleavage. EMBO J 1996; 15(12):3197-3206.

80. Grundy GJ, Hesse JE. Gellert M. Requirements for DNA hairpin formation by RAGI/2. Proc Nat!Acad Sci USA 2007; 104(9):3078-3083 .

81. Bhasin A, Goryshin IY. Reznikoff WS. Hairpin formation in Tn5 transposition. J BioI Chern 1999;274(52):37021 -37029.

82. Nishihara T. Nagawa F, Imai T er al. RAG-heptamer interaction in the synaptic complex is a crucialbiochemical checkpoint for the 12/23 recombination rule. J Bioi Chern 2008; 283(8) :4877-4885.

83. Ason B. ReznikoffWS. Mutat ional analysisof the base flipping event found in Tn5 transposition. J BioiChern 2002; 277(13) :11284-11291.

84. Bischerour J, Chalmers R. Base-flipping dynamics in a DNA hairpin processing reaction. Nucleic AcidsRes 2007; 35(8):2584-2595.

85. Lu CP, Sandoval H. Brandt Vi er al. Amino acid residues in ragl crucial for DNA hairpin formation.Nat Srrucr Mol Bioi 2006; 13(11):1010-1015 .

86. Bergeron S. Anderson DK. Swanson pc. RAG and HMGBI proteins: purification and biochemicalanalysis of recombination signal complexes. Methods Enzymol 2006; 408:511-528.

87. Kriatchko AN, Anderson DK , Swanson PC. Identification and characterization of a gain-of-functionRAG-l mutant. Mol Cell Bioi 2006; 26(12) :4712-4728.

88. Ciuboraru M, Schatz DG. Synapsis of recombination Signalsequences located in cis and DNA underwinding in V(D)J recombination. Mol Cell Bioi 2004; 24(19):8727-8744.

89. Schatz DG. V(D)J recombination moves in vitro. Scmin Immuno11997; 9(3) :149-159.90. West KL, Singha NC. De Ioannes P er al. A direct interaction between the RAG2 C terminus and the

core histones is required for efficient V(D)J recombination . Immunity 2005; 23(2):203-212.91. Liu Y. Subrahmanyam R, Chakraborty T et al. A plant homeodomain in RAG-2 that binds hyperm

ethylared lysine4 of histone H3 is necessaryfor efficientantigen-receptor-gene rearrangement. Immunity2007; 27(4) :561-571.

92. Matthews AG. Kuo AJ, Ramon-Malques S et al. RAG2 PHD finger couples histone H3 lysine 4 trimethylation with V(D)J recombination . Nature 2007; 450(7172):1106-1110.

93. Zhang Z, Espinoza CR. Yu Z et al. Transcription factor pax5 (BSAP) transactivates the RAG-mediatedV(H)-to-DJ(H) rearrangement of immunoglobulin genes. Nat Immunol 2006, 7(6) :616-624.

94. Wang X. Xiao G. Zhang Y et al. Regulation of tcrb recombination ordering by c-fos-dependene RAGdeposition. Nat Immunol 2008; 9(7):794-801.

95. Cobaleda C. Schebesra A, Delogu A et al. Pax5: the guardian of B-cell identity and function . NatImmuno12007; 8(5) :463-470.

96. Jackson AM. Ktangel MS. Turning T-cell receptor beta recombination on and off: more questions thananswers. Immunol Rev 2006; 209:129-141.

Early StepsofV{D)JRearrangement: InSightsfrom BiochemicalStudies ofRAG-RSS Complexes 15

97. Chen L. Glover IN. Hogan PG et al. Structure of the DNA-binding domains from NFAT, fos and junbound specificallyto DNA . Nature 1998: 392(667I):42-48.

98. Jones JM. Gellert M. The taming of a transposon: V(D)J recombination and the immune system.Immunol Rev 2004: 200:233-248.

99. Feeney AJ. Tang A, Ogwaro KM. Bvcell repertoire formation : role of the recombination signal sequencein nonrandom V segment utilization. Immunol Rev 2000; 175:59-69.

100. Livak F, Burtrum DB. Rowen L et al. Genetic modulation ofT-cell receptor gene segment usage duringsomatic recombination. J Exp Med 2000: 192(8):1191-1196.

101. Ma Y. Lu H, Tippin B et al. A biochemically defined system for mammalian nonhomologous DNAend joining. Mol Cell 2004; 16(5):701-713.

CHAPTER 2

Regulation ofRAG TranspositionAdam G.W. Matrhews" and Marjorie A. Oettinger

Abstract

V (D )J recombination is initiated by the lymphoid specific proteins RAGI and RAG2,which together constitute the V(D)J recombinase. However, the RAG 1/2 complex canalso act as a transposase, inserting the broken DNA molecules generated during V(D)J

recombination into an unrelated piece ofDNA. This process, termed RAG transposition, canpotentially cause insertional mutagenesis, chromosomal translocations and genomic instability.This review focuses on the mechanism and regulation of RAG transposition. We first providea brief overview of the biochemistry ofV(D)J recombination. We then discuss the discoveryof RAG transposition and present an overview of the RAG transposition pathway. Using th ispathway as a framework, we discuss the factors and forces that regulate RAG transposition.

IntroductionDuringlymphoiddevelopment, immunoglobulin and T-cdl receptor genes are assembled from

multiple, nonconsecutive gene segments in a seriesofsite-specific recombination reactions, termedV(D)J recombination.P Bycombinatoriallyjoiningdifferent variable (V), diversity (D) and joining(J) gene segments, V(D)J recombination generates a diverse array ofT-cell receptor (TCR) andimmunoglobulin (Ig)molecules (Fig. 1), thereby enabling the adaptive immune system to recognizean almost limitless number ofantigens and protect us from pathogenic microorganisms.

V(D)J recombination is initiated when the lymphoid specific proteins RAGl and RAG2generate double-stranded DNA breaks at V, D and] gene segments. These breaks are normallyrepaired by the nonhomologous end-joining (NHE]) pathway. However, the same enzyme thatproduces these double-strand breaks-RAG1/2 complex-can also act as a transposase, insertingthe newly generated broken DNA molecules into an unrelated piece ofDNA. This process, termedRAG transposition, can not only cause insertional mutagenesis,' but could also lead to genomicinstability" and the generation ofpotentially oncogenic chromosomal eranslocations.t Therefore,it is important to understand how RAG transposition is suppressed in vivo.

This review willfocus on the mechanism and regulation ofRAG transposition. We will firstprovide a briefoverview of the biochemistry ofV(D)J recombination. We will then discuss thediscovery of RAG transposition and present an overview of the RAG transposition pathway.Using this pathway as a framework, the factors and forces that regulate RAG transposition willbe discussed.

Biochemistry ofV(D)J RecombinationAll recombinationally active V, D and J gene segments are flanked by recombination signal

sequences (RSSS),6which consist ofhighly conserved heptamer (5' -CACAGT G-3') and nonamer(5' -ACAAAAACC-3') sequences separated by a spacer region ofeither 12 or 23 bp ?'10 Efficient

*Corresponding Author: Adam G.W. Matthews-Department of Molecular Biology,Massachusetts General Hospital, and Department of Genetics, Harvard Med ical School,Boston, MA 02114, USA. Email: agwmatthewsegmail.com

V(D)JRecombination, edited by Pierre Ferrier. ©2009 Landes Bioscienceand Springer Sciences-Business Media.

Regulation ofRAG Transposition 17

recombination between genesegments only occurswhen one segment is flanked by a 12-RSSand the other isflanked bya23-RSS, a restrictiontermedthe 12123 rule.61he 12123 ruleensuresthat recombination onlyoccurs betweengenesegments that cangive riseto a functionalantigenreceptorgene.

V(D)Jrecombination requires theexpression oftwolymphoid-specific recombination-activatinggenes, RAGI andRAG2,11·15 whichactrogetherroconstitute theV(D)Jrecombinase that recognizesand cleaves recombination signal sequences,"AlthoughRAGI and RAG2canboth be truncateddown to catalytically active "core" regions, consisting of amino acids 384-1008 out of 1040forRAG117-19 and aminoacids 1-383 out of 527for RAG2,W.21 the "non-core" portionsof RAGI andRAG2, whicharehighlyconserved throughoutevolution,22.23 playkeyregulatory roles in vivo.24-29

V(D)Jrecombination canbeconceptuallydividedinto twostages: generation ofdouble-stranded DNA breaks by the lymphoid-specific proteins RAGI and RAG216and the repairof thosebreaks by nonhomologous end-joining.While DNA double-strand break formation (V(D)J cleavage) requires only the RAG proteins and HMGI (a DNA bending protein), the repairstageof the reactionrequires the ubiquitously expressed nonhomologous end-joining(NHEJ)proteinsKu70,30.31 Ku80,32.33 DNA-pKcs,34.35 Artemis,36.37XRCC4,38 DNA Ligase IV39.40 andXLF(a.k.a. Cernunnosj.v't The RAGproteins alsoplaya role in the repairstageof the reaction.v?'Additionally, other proteinssuchasATM, MreII, Rad50and Nbs1 mayalsobe involved in therepairof RAG-induced double-strand breaks.52s3

During the cleavage stageof the reaction, the RAGI/2 complex first assembles on a 12-RSSand then captures a 23_RSS54·56 to form a synaptic pairedcomplex,'?DNA double-strand breaks

V.' V.134 D.' 0. 13 J .1 J.~ V, , V.8S J. ' J ,S C

~D.'.J.' ~n V.85.J"V.' V..134 OJ J.~ C. V.' VJ J .S C

~n v., J4.{)JV.1 VDJ J~ C

Exp....

~ Express

,( ~ ~<, /

134'1T~ ' 7000 heavy chain. 85' S " ~OO light chain.

>10' distinct .ntlbocU••

Figure 1. Overview of V(D)J recombination. V, D, and J gene segments are depicted as rectangles, constant region genes are depicted as rectangles and recombination signal sequences(RSSs) are depicted as triangles (shaded for the 23-RSS and unshaded for the 12-RSS). As aresult of V(D)J recombination, our bodies generate a diverse repertoire of antigen receptorsfrom a limited amount of genetic material. A color version of this figure is available online atwww.landesbioscience.com/curie.


aregenerated within this pairedcomplex viaa pairof phosphoryl transferreactions (Fig. 2, top).TheRAGproteinsfirstnickthe top strandofeachRSS, just5' of the heptamersequence."Thesenewly liberated3' hydroxyl groupson the top strandof the codingBankthen attackthe bottomstrandviaa direct transesterificadon /" thereby converting theseDNA nicksinto double-strandbreaks andgenerating the cleavage reaction products: two hairpinnedcodingendsandtwoblunt5'-phosphorylated signal ends,"

Duringthe repairphaseof the reaction, Ku70/Ku80heterodirners arethought to bind to thefour cleavage products.DNA-PKcs then binds to Artemisand undergoes autophosphorylation,therebyenablingArtemis to endonucleolytically open the hairpinned coding ends.S9.6o Sincehairpin-opening rarely occurs precisely at the tip of the hairpin,5' or 3' overhangs arecommonlygenerated. These overhangs maythen be trimmedbynucleases or filled in bypolymerases, generatingpalindromic (P) nucleotides. The two processed codingends are then ligatedtogetherin a process requiring the XRCC4-XLF-DNA Ligase IV complex to form the codingjoint.61-63

TheXRCC4-XLF-DNA Ligase IV complex alsorepairs the two signal endsbyprecise heptamer-to-heptamer ligation to formthe signal joint.

1!';;;w;:;DF ~f-": ; ·=t: '.I=: :'"'": : : ......'----!RAGI + RAG2, Me2+

1 ; ; ; ;:~'~-----~ti::J:-:'::~:"': : : :-'-1: I---!RAGI + RAG2. Mel.

Ii:::; :::::=1).-----4fC[.,..,.~.,...,.: J.,..,.: ; : i.,.,: : I--HO

Codin!: Jo int

~H

~G T, nsposltlon

Transposition Product

Figure 2. The biochemistry ofV(D)j recombination and RAG transposition. V gene segments aredepicted as stippled rectangles, Jgene segments are depicted as stippled rectangles, recombination signal sequences are depicted as triangles (for the 12-RSS and for the 23-RSS)and targetDNA is shown in light gray. The RAG1/2 complex initiates V(D)j recombination by first nickingDNA at the border between the coding DNA and the RSS heptamer (hydrolysis). The free 3'hydroxyl (OH) on the coding flank then attacks the opposite strand in a direct transesterificationto form a blunt signal end and a hairpinned coding end. Hairpinned coding ends are repairedvia the nonhomologous end-joining (NHEJ) pathway to form imprecise cod ing joints wh ileblunt signal ends can either be repaired via the NHEJ pathway to form precise signal joints, orthey can inserted into an unrelated piece of DNA via the RAG transposition pathway. A colorversion of this figure is available online at www.landesbioscience.com/curie.


Overview ofRAG Transposition

Discovery ofRAG TranspositionBasedon the palindromicsequence of the RSS's heptamer/" the face that the genomicorienta

tion of 19K recombinationsignal sequences resembles the inverted repeats found at the ends ofprokaryotic transposons'" and the unusualstructure of the mammalian RAG locus{RAG1 andRAG2arecompactlyorganized asadjacent genes andeachgeneisencodedbyasingleexon),'! itwashypothesized that V{D)J recombination maybemechanistically relatedto bacterialtranspositionevents.u.64.65 And indeed, 10years ago,it wasshownthat the RAGproteins can transposesignalends into an unrelatedpieceof DNA, in a process termed RAGtransposition{Fig. 2, bottom).5.66In this reaction, the RAG proteins catalyze another phosphoryl transfer reaction,enablingtheexposed, nucleophilic 3' hydroxyl groupon the bottom strand of the signalend to attacka targetDNA molecule.AlthoughRAGtransposition wasinitiallydiscovered invitro,rareeventsofRAGtranspositionhavealsobeen observed in vivoin human cells,3.67 murinecells' and in engineeredyeast.68Thus,RAGtransposition represents abonafidealternative fatefor thedouble-strand breaksgeneratedduringV{D)J recombination. Bycompetingwith the NHEJ pathwayin vivo, transposition can cause insertionalmutagenesis,' oncogenicchromosomal translocadons? and genomicinstability.'Thepathways leadingto insertionalmutagenesis and chromosomal translocations aredescribed in more detailbelow.

The RAG Transposition PathwayAsdiagrammedin Figure 3, RAGtranspositionproceedsthrough an orderlyseries of steps.69

The RAG proteins firstbind to both a 12-RSS and a 23-RSS to form a synaptic paired complex(PC). TheRAGproteinsthen performcoupledcleavage to generate apairofDNA double-strandbreaks(Fig.3, Step 1), resultingin the cleaved signalcomplex (CSC) which contains two blunt3'-hydroxylated signalends and two hairpinned codingends. Next, codingends are transferredfromthe cleaved signalcomplexto the NHEJ pathway, leaving the RAGproteinsbound to signalendswithin the signal-end complex (SEC) (Fig.3,Step2). Thedecisionto resolve signalendsviaNHEJ (Fig.3, Step3a) or RAG transpositionoccurswithin the signal-end complex. Ifthe RAGproteins bind target DNA and commit to undergoingtransposition (Fig. 3, Step 3b), they firstforma targetcapturecomplex (TCC). Within the targetcapturecomplex, the RAGproteinscancatalyze eithersingle-ended insertionof JUSt one RSS (Fig.3,Step4a), or double-endedinsertionof both RSSs (Fig. 3, Step4b).

Resolution ofBranched Transposition Intermediates Can Leadto EitherInsertionalMutagenesis or Chromosomal Translocations

After double-endedinsertion,the resultingbranched DNA molecule can be resolved in oneof three ways. It can be resolved by DNA repair, resultingin insertionalmutagenesis with thecharacteristic 5 bp target site duplication {Fig. 4a).5 Alternatively, the branched transpositionintermediatecanbe resolved viadisintegration. In this RAG-catalyzed reaction,the nucleophilic3' hydroxyls on the targetDNA attackthe newlyformedphosphodiesrer bondslinkingthe RSSsto the target DNA, therebyregenerating both the cleaved signalend and the target DNA {Fig.4b; Fig.3, Steps4b and 4d).70 Finally, the branched molecule canbe resolved viaRAG-catalyzedhairpinformation. Inthis reaction, whichisanalogous to the formationofhairpinned codingendsduringV{D)J cleavage, the nucleophilic 3' hydroxyls on the targetDNA attackthephosphodiesterbond on the oppositestrandof the targetmolecule, therebygeneratinghairpinnedtargetendsandsignal endswith 3' overhangs of 5 nucleotides {Fig. 4c).70Joining thesehairpinnedtargetends tothe hairpinned codingends would lead to reciprocal chromosomal translocations that couldbepotentiallyoncogenic{Fig. 5).5.70 It is worth noting that translocations generatedin this mannerwould not bear the hallmark target site duplications that are characteristic of traditional RAGtransposition(resultingin insertionalmutagenesis), eventhough they weregeneratedasa resultof RAG-catalyzed transpositionand hairpin formation.


RAG2C-torm.--j Stop 2

SECStop3a =-=

~

~~GTP--jStop 3bt

Singlo-ended InsertionChromosomal translocatlons

Double-ilnded InsertionInsertional mutagenesis

Chromosomal trans locations

Figure 3. The RAGtransposition pathway. V genesegments are depicted as rectangles, Jgenesegments aredepicted asrectangles, recombination signal sequences aredepicted astriangles(for the 12-RSS and for the 23-RSS), the RAG1/2 complex is portrayedasa shaded oval (eithershaded or unshaded) and target DNA is shown in light gray. As described in the text, RAGtransposition initiates with coupled RSS cleavage within the pairedcomplex (PC), therebygenerating the cleaved signal complex (CSC), which consists of all 4 broken DNA endsnoncovalently bound by the RAG1/2 complex (Step 1). Transfer of the cleaved coding endsto the NHE)pathwayresults in the formation of the signal end complex (SEC) (Step 2). Target captureoccurswithin thesignal endcomplex, leadingto the formationof a stabletargetcapturecomplex (TCC)(Step 3b). Transpositional strand-transfer occurswithin the targetcapturecomplex, generatingthe strand transfercomplex (STC) (Steps 4a/4c). The branched DNA molecules present in thestrand transfer complex can be resolved in several different ways (see Fig. 4). A color versionof this figure is availableonline at www.landesbioscience.com/curie.

Hairpin Fonnatlon~


DNA Repair 5'=DBG=3'~

Disintegration

Figure 4. Pathways for resolving branched RAG transposition intermediates. After double-endedinsertion (Fig. 3, Step 4c), the RAG1/2 complex remains bound to the branched transpositionintermediate within the strand transfer complex. There are at least three pathways for resolvingthese transposition intermediates. a) These branched molecules can be resolved by nick repair,leading to insertional mutagenesis with the signal ends flanked by 5 bp target site duplications(shown in dark gray). b) The branched molecules can be resolved via disintegration, wherethe 3' hydroxyls on the target DNA attack the phosphodiester bonds at the RSS-target DNAjunctions, thereby removing the inserted signal ends and rejoining the target DNA. c) Thebranched molecules can be resolved via target cleavage, where the 3' hydroxyls on the targetDNA attack the opposite strand of the target DNA, thereby generating hairpinned target endsand liberating signal ends that contain 3' overhangs of 5 nt. A color version of this figure isavailable online at www.landesbioscience.com/curie.

Single-endedinsertion,followed bytargetDNA transesterifieation, canalso lead to chromosomaltranslocations. Ifthesetranslocations resultfrominsertionof asignal end generatedin a single-sitecleavageevent, theywillbereciprocals If,however, theyresultfrominsertion ofasignal endgeneratedinacoupledcleavageevent,theycanbeeitherreciprocalor nonreciprocalwithlossofgenetic material(Fig. 6).Nonreciprocal chromosomal translocations that areaccompaniedbylossofgeneticmaterialwouldlikely leadto impairedviability of the cellandwouldtherefore bedifficult to detectinvivo. It


~H5' 3'

3' 5'

~RAG Transposlt lcn

5· ......,-----.,;~

3·......----~~-

] GeneA 0

3'

5'

~T''ll.t DNAtran..steriflea tlon

~n.homOIOgOUS end-jolning

Figure 5. Pathway for generating chromosomal translocations from RAG transpositionaldouble-ended insertions. Following RAGtranspositional double-ended insertion, the branchedtransposition intermediate can be resolved via target DNA cleavage.Joining these hairpinnedtarget ends to the previously generated hairp inned coding endsvia nonhomologous end-joiningwould generate reciprocal translocations. If these translocations bring oncogenes (such asGene A or Gene B) into close proximity with the immunoglobulin promoters/enhancers, theycould potentially lead to oncogenic transformation of the cell. It is worth noting that in thesame way that signal ends are normally lost as circular signal joints during canonical V(D)Jrecombination, the modified signal ends generated in the proceses of target DNA cleavagewould also be lost either as linear DNA molecules (if left unprocessed) or as circular signaljo ints (if repaired by NHEJ).

RegulAtion ofRAG Transposition

~H5' 3'

3' 5'

+RAG Transposition

23

+Target DNA transesterlflcatlon

5·---r- - - - n:;::,."'1- - - - -3·........----"-t---r-----

3'

5'

] GeneA 0

~ ~Rn-homologous end-Joining

Figure 6. Pathway for generating chromosomal translocations from RAG transpositionalsingle-ended insertions. Following RAG transpositional single-ended insertion, the branchedtransposition intermediate can be resolved via target DNA cleavage. Joining this hairpinnedtarget end (in this case, Gene A) to one of the previously generated hairpinned coding endsvia nonhomologous end-joining would generate a chromosomal translocation. Joining theremaining hairpinned coding end to the free signal end (in this case, 23-RSS) via either aRAG-dependent process (suchashybrid joining or open-and-shut joining) or nonhomologousend-joining would generate a reciprocal translocation. If, however, the remaining hairpinnedcoding end is not joined to the free signal end (as shown here), then this pathway wouldresult in a nonreciprocal translocation with loss of genetic material, thereby impairing theviability of the cell.


isworth notingthat althoughtranslocations resultingfromsingle-ended insertiondifferfromthoseresultingfromdouble-ended insertion, translocations generated asa resultofsingle-endedinsertionwouldalso lackthe targetsiteduplications that arecharacteristic of traditionalRAG transposition.

Regulation ofRAGTranspositionAsdescribedabove, RAGtranspositioneventsarepotentiallydeleterious.Thus,it isvitallyim

portant that RAGtranspositionbesuppressed indevelopingB-cells and T-cells. Indeed,althoughRAG transposition occurs robustly in vitro,5,66 the frequencyof transposition in vivo is muchlower. One study estimated that in developing T-cells, RAG-mediated insertional mutagenesiscauses phenotypic lossofHPRT genefunction at a frequencyof 1 eventper 107cells,' Sincethisstudy could not detect transposition eventsthat occurred at other locations in the genome,theactual frequencyof RAG transpositionmust be greaterthan 1eventper 107cells. Another studyestimated that in transfected293T-cells, RAG transposition occurred at a frequency of 1 eventper 107plasmidsanalyzed." A third study estimatedthat in pre-Bscells, one RAG transpositioneventoccursper every50,000V{D)Jrecombinations, correspondingto a frequency of2.5 eventsper 105recombinations."Although all three studiesconclude that RAG transpositionoccursata fairlylow frequencyin vivo,our bodiesgenerate_108 new lymphocytes per day. As such, thefrequency of RAG transposition is high enough to causepotentially oncogenicgenomic rearrangements. Therefore, RAG transposition isa biologically relevantpathwayand it is importantto understand the multipleways in which it is regulated

Usingthe pathway in Figure 3 asa framework,wecangain abetter understandingofthe mechanisms involved in the regulation of RAG transposition. In theory, RAG transposition could besuppressed at anyof the fourstepsthat precededonor insertioninto the targetDNA: RSScleavage(Step1); codingend release (Step2); targetcapture(Step3);or donor insertion(Step4).However,whereas manytransposons areregulated eitherat the level of transposase expression or at thestepoftransposonexcision from the host genome(Fig. 3, Step 1),RAG transposition cannot be similarlyregulatedbecause RSS cleavage is crucialfur the assembly of functional antigen receptor genes.Therefore, RAG-mediated transposition mustberegulated at a stepsubsequent to RSSdonor cleavage(Steps 2-4).Below, wereview what iscurrently knownabout the factors and forces that regulateRAG transposition and wespeculate about additionalpotential regulatory mechanisms.

Current Understanding ofHow RAG Transposition Is Regulated

Regulation bythe C-TerminalPortion ofRAG2In vitro studies comparing the transpositional activity of full-length RAG2 (aa 1-527)to that

of core RAG2 (aa 1-387) revealed that RAG transposition can be suppressed by the "non-core"C-terminal portion of RAG2.71.73 Interestingly, full-length RAG2 suppressed transposition ofintact RSSsubstrares,":" but had no effecton transposition ofprecleaved RSS substtates.71.74 Sincefull-length RAG2 only suppressed transposition when codingDNA waspresentin the RAG1/2complex, thisfindingsuggested that the C-terminalportion ofRAG2 blocks transposition ofintactsubstrates bystahlybindingto codingendswithin thecleaved signal complex, therebyoccupyingthetargetDNA bindingsiteandpreventingtargetcapture{Fig. 3,Step2).71 While it ispossible that theC-terminalportionofRAG2also inhibitstransposition at thestepoftargetcapturebythesignal-endcomplex {Fig. 3, Step 3b),72.73 two studiesfound that signal-end complexes containingfull-lengthRAG2werejustasactive in targetcapture" and transposition' V"assignal-endcomplexes containingcoreRAG2.Thus,wefavor amodelwherefull-length RAG2inhibitstransposition bystahilizingthecleavedsignal complexandpreventingsubsequent targetcapture(Fig. 3,Step2).However, eveniftheC-terminalportion ofRAG2 doessuppress transposition in thismanner,codingendsareprocessedmorerapidlythan Signal endsin vivo,?5Therefore, sincethe signal-end complex, whichisdevoidofcodingends,mustpersistfor sometimein the cell,other layers of regulation mustexist.


Regulation by GTPOne of these additional layers of regulation may be inhibition by GTP.73 In vitro experi

ments revealed that at concentrations of 1 mMor higher, GTP (but not ATP, CTP, or UTP)inhibited RAG transposition by blocking target capture within the signal-end complex (Fig.3, Step 3b). This inh ibition was alleviated by introducing substitutions within a putativeGTP-binding domain in RAG 1. Several other transposases-such as the Tn7 transposableelement,76,77 bacteriophage MU78 and Drosophila P element transposase'P-c-arc similarlyregulated by nucleotide-binding. However, since the average intracellular GTP concentration in cells is only 0.5 ± 0.2 mM80 and GTP inhibits RAG transposition very weakly in thisconcentration range," the extent to which this mechanism regulates transposition in vivoremains unclear.

Regulation byDisintegrationAn additional layer of regulation may be the propensity ofthe RAG1/2 complex to resolve

branched transposition intermediates viadisintegration (Fig. 3,Steps4b and 4d).70 As mentionedearlier, the branched DNA moleculesgenerated asa result ofRAG transposition can be resolvedin one ofthree ways:nick repair-leading to insertional mutagenesis (Fig.4a); disintegrationregenerating both the blunt singal ends and the linear target DNA (Fig. 4b); or target DNAtransesterification-generating hairpinned target ends and signal ends with 3' overhangs of 5nucleotides (Fig. 4c). However, at physiologic magnesium concentrations of20-25 mM, disintegration seems to be favored over both target DNA transesterification and nick repair," Thus,by essentially reversing the process of RAG transposition, RAG-catalyzed disintegration mayvery well contribute to the low levelsof transposition observed in vivo.

Regulation by Target Site SelectionTarget site selectivity by the RAG transposase might reduce the frequency ofdeleterious

transposition events by channeling these insertions into relatively safe regions ofthe genome.Initial studies revealed that RAG transposition events are moderately biased towards GC-richtarget sequences .V" Subsequent studies confirmed this preference for GC-rich regions'?and suggested that distorted DNA structures such as DNA mismarches," hairpins81,82 andsingle-strand-double-strand DNA junctions" can also act as preferred sites for RAGtransposition. If RAG transposition events are targeted to these distorted DNA structuresand if these structures are predominantly found within innocuous regions of the genome,then target site selectivity may help to limit the frequency of harmful transposition events.However, it remains unclear whether such distorted DNA structures are predominantlyfound within innocuous regions of the genome. In addition, while this form of regulationmay help to reduce deleterious transposition events, it would not limit the overall frequencyof transposition in vivo.

Additional Potential Regulatory MechanismsAlthough the C-terminus of RAG2, GTP, disintegration and target site selectivity may

help to suppress deleterious RAG transposition events in vivo,additional as-yet-undiscoveredregulatory mechanisms must also exist. That is, at physiological concentrations of 20-25mM MgZ.. 5 I-lM Ca2• and 0.5 mM GTP, the C-terminus of RAG2 inhibits transposition_10 _fold,71 ,73 GTP inhibits transposition -5-fold73 and disintegration inhibits transposition -10-fold.70Taken together, these regulatory mechanisms would suppress transposition-500-fold. However, since RAG transposition occurs so robustly in vitro, this level of suppression is insufficient to explain the low frequency of transposition observed in vivo.3.4.67Here, we will speculate about additional potential regulatory mechanisms for suppressingRAG transposition in vivo.


CodingDNA May Assist in Reducing the Frequency ofInterchromosoma/Transposition

The requirement for both V(D}] cleavage and coding end release prior to target capturesuggests that coding DNA can inhibit RAG transposition. As discussedabove, one waythatcodingDNA can suppresstransposition isbyoccupyingthe non-RSSDNA binding siteoftheRAG1/2 complex, thereby preventing the RAG1/2 complex from binding target DNA andcommitting to the transposition pathway. However, coding ends could also help to preventdeleterious transposition events by temporarily tethering signalends to the antigen receptorloci, thereby reducing the length of time that the signal end complex has to freely diffusethrough the cell and capture interchromosomal target DNA. That is, since it appears thatchromosomes eachoccupytheir owndistinct territorieswithin the nucleus.83.84 newlygeneratedsignal ends would initially be positioned awayfrom other chromosomes. Consequently, theRAG1/2 complex bound to these signal ends would only be able to bind intrachromosomaltarget DNA . Givenenough time, the signalend complexcould randomly diffusethrough thenucleus and potentially encounter interchromosomal target DNA. Yet, since signalends arerepaired to form signaljoints at the G1/S transition," there is a finite window of opportunityfor the signalend complexto diffuseand capture interchromosomal target DNA. Moreover,since the RAGI/2 complex holds on to postcleavage coding ends within the cleavedsignalcomplex, coding end-binding tethers the RAG transposase to the originating antigen receptor locus for a period of time, thereby reducing the length of time that the RAG transposasehas to encounter interchromosomal target DNA. In this way, coding end-binding can help tominimize potentially harmful interchromosomal translocationsbybiasingRAG transpositiontowards intrachromosomal targets.

Intriguingly, it has beenshownthatwhereas XRCC4-I-pS3-1- mice(whicharegenerallydeficientin NHE]) develop progenitorB-celllymphomas harboringinterchromosomal1gH:c-myc translocations,"Artemis-1-pS3-1- mice (whichare specifically deficient in codingend repair) developprogenitorB-celllymphomasharboringintrachromosomal IgH:N-myc translocations.f Althoughithasn't beendeterminedwhetherthesetranslocations arederivedfromRAGtransposition events, thisfindingsuggests that hand-offofcodingendsfromthe RAGI/2 cleaved signal complex to ArtemiscouldbeanimportantstepindeterminingwhethertheRAGproteinsundergointrachromosomalorinterchromosomal transposition.Perhaps, in theabsenceofnormalcodingendrepairbyArtemis, theRAGI /2 cleaved signal complex persistsin the cell, therebygiving the RAG1/2 transposase a muchshorterwindowofopportunity to captureinterchromosomal targetDNA. Ifthisshortwindowofopportunity isnot longenoughfor the RAGtransposase to diffuse throughthe nucleus and comeintocontactwithotherchromosomes (e.g., thec-myc locus), thentheonlytargetDNA thatwouldbereadily available to the RAGtransposase wouldbeintrachromosomal (e.g., theN-myc locus).Thus,in theabsence ofArtemis, persistent codingend-bindingwithintheRAG1/2 cleavedsignalcomplexcouldbiasRAG transposition events towards intrachromosomal targets. But evenin the presenceofArtemis, codingend-binding maygenerally reducethe lengthof timethat the RAGtransposasehasto encounterinterchromosomal targetDNA, therebyhelpingto minimize potentially harmfulinterchromosomal RAGtransposition events.

Mamma/ian NHE]Proteins May Inhibit RAG TranspositionAlthough codingDNA maygenerally blockRAG transposition by preventing target capture

and the tetheringeffect ofcodingendsmayaid in specifically preventing interchromosomal RAGtransposition, codingendsare repaired morerapidly than signal endsin vivo," suggesting that theinhibitoryeffect of codingendsis likely to be transient. However, the requirement for codingendrelease prior to targetcaptureraises the possibility ofsustained inhibitionof RAGtransposition bythe NHE] proteins.Thatis,since the RAG1/2 cleaved signal complex mustinteractwiththeNHE]proteinsduringthe hand-offof codingends,NHE] proteinshave an opportunity to influence thedecision oftheRAG1/2 complex tochannelsignalendstowardssignaljointformation (Fig. 3,Step3a)or targetcapture(Fig.3.Step3b)andsubsequent transposition.Whileinteraetingwith theRAG1/2


cleaved signal complex, perhaps the NHE] proteinsinducea conformational change in the RAGproteins(represented in Fig. 3 asacolorchangefromwhiteto grayat Step2),that favors signal jointformation. Thisconformational change couldeitherclose the target DNA bindingpocketorsimplyinducea conformation that favors subsequent interaction between the signal end complex and theNHE] proteins. In thisway. NHE] proteinscouldhelpto inhibit transposition in vivo.Alongtheselines. it is interesting to note that whenRAG1and RAG2areexpressed in SlUcharomyus cereoisiae,transposition occurs at leastasfrequently assignal joint fOrmation.68 Since yeast lackArtemisandDNA-PKcs homologues, thesetwo proteinsmaysuppress RAGtransposition in mammalian cells.Furthermore, although yeastdo have homologues of mammalian Ku70,87 Ku80.88.89XRCC4,90XLp91.92 andDNA Ligase IY,93 thesefactors areratherpoorlyconserved, with<25% identitybetweenyeast and humans.87.89.93 Given the low degree of conservation between yeast and human NHE]factors. it seems plausible that mammalian Ku70,Ku80,XRCC4. XLF and DNA Ligase IV maysuppress RAGtransposition eventhoughtheiryeasthomologues do not.

MethylatedHistone-BindingMay InhibitRAG TranspositionGiventhat the C-terminalportion ofRAG2 has beendemonstratedto suppress transposition

in vitro,7I.73 it seems reasonable to hypothesize that other RAG-intrinsic regulatorymechanismsmayexistto inhibit transpositionin vivo. Recently, it hasbeendemonstratedthat a plant homeodomain (PHD) fingerpresentin the C-terminalportion ofRAG2 recognizes histone H3 whenit is either trimethylatedon lysine 4 {H3K4me3)29.94 or when it isSimultaneously symmetricallydimethylatedon arginine2 and trimethylatedon lysine 4 (H3R2me2s/K4me3).95 Furthermore,it has been shown that reducingeither the levels of H3K4 methylation or the abilityof RAG2to bind H3K4me3 impairs V{D)] recombination. indicating that recognition of methylatedhistone H3 is important for V{D)]recombinationin vivo." SinceV(D)J recombination isregulated by methylatedhistone-binding. RAG transposition mayalso be regulatedby methylatedH3-binding. MethylatedH3-binding mayallosterically inhibit the transpositionactivityof theRAG transposase by either closing the target DNA binding pocket and blockingtarget capture(Step3b)or byinhibitingstrandtransfer(Steps 4a/4c).Alternatively. methylated H3-bindingmayregulatetarget site selection and direct RAG transposition into regions of the genomethat areenriched for H3K4me3 and/or H3R2me2s/K4me3. Interestingly, many retrotransposons contain chromodomains96-98- a protein module that mediates interactionswith methylatedhistoneproteins-suggesting that regulationby methylatedhistone-binding maybe a generalfeatureofmanydifferenttransposons. In the future.it willbe interestingto seewhetherRAGtranspositionis,indeed,regulatedby recognitionofH3K4me3 and/or H3R2me2s/K4me3.

Regulation byOther Trans-ActingFactorsFinally, it worth noting that in much the samewaythat the NHE] proteins mayact to sup

pressRAG transposition.other. as-yet-unidentified. trans-actingfactorsmayalsoregulateRAGtransposition in vivo.Thesefactors may stably interact with the RAG transposaseto directlyinhibit one of the stepsin the transpositionpathway. Alternatively. theymaytransientlyinteractwith the RAG transposaseto modify one (or both) of the RAG proteins in such a wayas toinhibit transposition. Thesemodificationscould be either covalent(e.g., protein phosphorylation) or noncovalent (e.g.•protein remodelingby an ATP-dependent molecular chaperone).In anycase,it seems likelythat additional regulatory factors maybe involvedin the inhibitionof RAG transposition.

ConclusionAsdescribed above, elucidation ofthe RAGtransposition pathwayhas providedausefulconcep

tualframework withinwhichto understandthe regulation of RAGtransposition. Althoughseveralregulatory mechanisms have already been identified. it seems likely that new forms of regulationwill cometo light in the nextfewyears. In the future.it will be interesting to testhowthesevariousregulatory mechanisms interactwith eachother to suppress RAGtransposition in vivo.


References1. Bassing CH. Swat W; Alt FW. The mechanism and regulation of chromosomal V(D)J recombination.

Cell 2002: 109 Suppl:S45-55.2. Gellert M. V(D)J recombination: RAG proteins. repair factors and regulation. Annu Rev Biochem

2002: 71:101-132.3. Messier TL, O'Neill ]P, Hou SM et al, In vivo transposition mediated by V(D)J recombinase in human

T-lymphocytes. EMBO] 2003; 22(6) :1381-1388.4. Reddy YV, Perkins E], Ramsden DA. Genomic instability due to V(D)] recombinat ion-associated

transposition. Genes Dev 2006: 20(12) :1575-1582.5. Hiom K. Melek M. Gellert M. DNA transposition by the RAG1 and RAG2 proteins: a possible source

of oncogenic translocations. Cell 1998: 94(4) :463-470.6. Tonegawa S. Somatic generation of antibody diversity. Nature 1983; 302(5909) :575-581.7. Akamatsu Y,Tsurushita N. NagawaF et al, Essential residues in V(D)J recombination signals. ] Immunol

1994: 153(10):4520-4529.8. Akira S, Okazaki K, Sakano H. Two pairs of recombination signals arc sufficient to cause immuno

globulin V-(D)-] joining. Science 1987; 238(4830):1134-1138.9. Hesse ]E, Lieber MR, Mizuuchi K et al, V(D)] recombination : a functional definition of the joining

signals. Genes Dev 1989; 3(7):1053 -1061.10. Montalbano A, Ogwaro KM, Tang A er al. V(D)] recombination frequencies can be profoundly affected

by changes in the spacer sequence.] Immunol2oo3: 171(10):5296-5304.11. Oettinger MA. Schatz DG. Gorka C et al. RAG-l and RAG-2, adjacent genes that synergistically

activate V(D)] recombination . Science 1990: 248(4962) :1517-1523.12. Schatz DG. Oettinger MA, Baltimore D. The V(D)] recombination activating gene. RAG-I. Cell 1989:

59(6):1035-1048.13. Mombaerts P. Iacomini],]ohnson RS er aI. RAG-I -deficient mice have no mature B- and Tdymphocyres,

Cell 1992: 68(5):869-877.14. Oettinger MA. Activation of V(D)] recombination by RAGI and RAG2. Trends Genet 1992;

8(12):413-416.15. Shinkai Y,Rathbun G. Lam KP ee aI. RAG-2-deficient mice lack mature lymphocytes owing to inability

to initiate V(D)] rearrangement. Cell 1992: 68(5):855-867.16. McBlane]F, van Gent DC. Ramsden DA et al. Cleavageat a V(D)J recombination signal requires only

RAGI and RAG2 proteins and occurs in two steps. Cell 1995: 83(3):387-395.17. Sadofsky M], Hesse ]E . McBlane]F er al, Expression and V(D)] recombination activity of mutated

RAG-l proteins. Nucleic Acids Res 1993: 21(24) :5644-5650.18. Kirch SA. Sudarsanam P, Oettinger MA. Regions of RAG1 protein critical for V(D)J recombination.

Eur] Immunol 1996; 26(4) :886-891.19. Silver DP. Spanopoulou E, Mulligan RC er al. Dispensable sequence motifs in the RAG-l and RAG-2

genes for plasmid V(D)] recombination . Proc Nat! Acad Sci USA 1993: 90(13):6100-6104.20. Cuomo CA, Oettinger MA. Analysisof regions ofRAG-2 important for V(D)] recombination. Nucleic

Acids Res 1994; 22(10) :1810-1814.21. Sadofsky M]. Hesse ]E, Gellert M. Definit ion of a core region of RAG-2 that is functional in V(D)]

recombination. Nucleic Acids Res 1994: 22(10):1805-1809.22. Litman GW; Anderson MK. Ran ]P. Evolution of antigen binding receptors. Annu Rev Immuno11999;

17:109-147.23. Peixoto BR, Mikawa Y, Brenner S. Characterization of the recombinase activating gene· 1 and 2 locus

in the Japanese pufferfish, Fugu rubr ipes, Gene 2000; 246(1-2) :275-283.24. Akamatsu Y, Monroe R. Dudley DO et al. Deletion of the RAG2 C terminus leads to impaired

lymphoid development in mice. Proc Nat! Acad Sci USA 2003; 100(3) :1209-1214.25. Dudley DO. Sekiguchi ], Zhu C et al. Impaired V(D)] recombination and lymphocyte development in

core RAGl -expressing mice.} Exp Med 2003: 198(9):1439-1450 .26. Elkin SK. Ivanov D. Ewalt M et aI. A PHD finger motif in the C terminus of RAG2 modulates recom

bination activity.] Bioi Chern 2005: 280(31) :28701.28710.27. Kirch SA. Rathbun GA, Oettinger MA. Dual role of RAG2 in V(D)J recombination: catalysis and

regulation of ordered 19gene assembly. EMBO] 1998: 17(16):4881-4886.28. Liang HE. Hsu LY, Cado D et al, The "dispensable"port ion of RAG2 is necessaryfor efficientVeo-D]

rearrangement during B- and T-cell development. Immunity 2002: 17(5) :639-651.29. Matthews AG. Kuo A], Ramon-Maiques S et al, RAG2 PHD finger couples histone H3 lysine 4

trimethylation with V(D)] recombination . Nature 2007: 450(7172) :1106-1110.30. Gu Y, Seidl K], Rathbun GA et al. Growth retardation and leaky SCID phenotype of Ku70-deficient

mice. Immunity 1997; 7(5) :653-665.


31. Ouyang H, Nussenzwetg A. Kurimasa A et al. Ku70 is required for DNA repair but not for T-cellantigen receptor gene recombination In vivo. J Exp Med 15 1997; 186(6):921-929.

32. NussenzweigA. Chen C, da Costa SoaresV et al. Requirement for Ku80 in growth and immunoglobulinV(D)J recombination. Nature 1996; 382(6591) :551-555.

33. Zhu C. Bogue MA. Lim DS et al. Ku86-deficient mice exhibit severecombined immunodeficiency anddefective processing ofV(D)J recombination intermediates. Cell 1996; 86(3):379-389.

34. Gao Y, Chaudhuri J. Zhu C et al. A targeted DNA-PKcs-null mutation reveals DNA-PK-independentfunctions for KU in V(D)J recombination. Immunity 1998; 9(3):367-376.

35. Taccioli GE, Amatucci AG. Beamish HJ et al. Targeted disruption of the catalytic subunit of theDNA -PK gene in mice confers severecombined immunodeficiencyand radiosensitivity. Immunity 1998;9(3):355-366.

36. Moshous D, Callebaur I. de Chasseval Ret al. Artemis, a novel DNA double-strand break repairN(D)J recombination protein. is mutated in human severe combined immune deficiency, Cell 2001 ;105(2):177-186.

37. Rooney S. Sekiguchi J. Zhu C et al. Leaky Scid phenotype associated with defective V(D)J coding endprocessing in Artemis-deficient mice. Mol Cell 2002; 10(6):1379-1390.

38. Gao Y, Sun Y, Frank KM et al. A critical role for DNA end-joining proteins in both lymphogenesisand neurogenesis. Cell 1998; 95(7):891-902.

39. Frank KM, Sekiguchi JM. Seidl KJ er al. Late embryonic lethality and impaired V(D)J recombinationin mice lacking DNA ligase IV. Nature 1998; 396(6707):173-177.

40. Grawunder U, Zimmer D. Fugmann S et al. DNA ligase IV is essential for V(D)J recombination andDNA double-strand break repair in human precursor lymphocytes. Mol Cell 1998; 2(4) :477-484.

41. Ahnesorg P, Smith P,Jackson SP. XLF interacts with the XRCC4-DNA ligase IV complex to promoteDNA nonhomologous end-joining. Cell 2006; 124(2):301-313.

42. Buck D. Malivert L. de Chasseval R er al. Cernunnos, a novel nonhomologous end-joining factor, ismutated in human immunodeficiency wirh microcephaly. Cell 2006; 124(2):287-299.

43. Zha S, Alt FW, Cheng HL et al. Defective DNA repair and increased genomic instability inCemunnos-XLF-deficient murine ES cells. Proc Nat! Acad Sci USA 2007; 104(11):4518-4523.

44. Dai Y,KyselaB, Hanakahi LA et al. Nonhomologous end joining and V(D)J recombination require anadditional factor. Proc Nat! Acad Sci USA 2003; 100(5):2462-2467.

45. Huye LE, Purugganan MM. Jiang MM et al. Mutational analysis of all conserved basic amino acids inRAG-l reveals catalytic, step arrest and joining-deficient mutants in the V(D)J recombinase. Mol CellBioi 2002; 22(10) :3460-3473.

46. Lee GS, Neiditch MB. Salus SS er al. RAG proteins shepherd double-strand breaks to a specific pathway, suppressing error-prone repair, but RAG nicking initiates homologous recombination. Cell 2004;117(2):171-184.

47. Leu TM , Eastman QM , Schatz DG . Coding joint formation in a cell-free V(D)J recombination system.Immunity 1997; 7(2) :303-314.

48. Qiu JX. Kale SB. YarnellSchulrz H et al. Separation-of-function mutants revealcritical roles for RAG2in both rhe cleavageand joining steps ofV(D)J recombination. Mol Cell 2001: 7(1) :77-87.

49. Ramsden DA, Paull IT, Gellert M. Cell-free V(D)J recombination. Nature 1997,388(6641):488-491.50. Tsai CL. Chatrcrji M. Schatz DG. DNA mismatches and GC-rich motifs target transposition by rhe

RAGI /RAG2 cransposase. Nucleic Acids Res 2003; 31(21) :6180-6190.51. YarnellSchultz H. Landree MA, Qiu JX et al.Joining-deficient RAG1 mutants block V(D)J recombina

tion in vivo and hairpin opening in vitro. Mol Cell 2001, 7(1) :65-75.52. Bredemeyer AL, Sharma GG, Huang CY et al. ATM stabilizes DNA double-strand-break complexes

during V(D)J recombination. Nature 2006, 442(7101) :466-470.53. Clatworrhy AE. Valencia-Burton MA, Haber JE et aJ.The MRE11-RAD50-XRS2 complex, in addition

to other nonhomologous end-joining factors, is required for V(D)J joining in yeast. J Bioi Chern 2005;280(21 ):20247-20252.

54. Jones JM. Gellert M. Ordered assembly of the V(D)J synaptic complex ensures accurate recombination.EMBO J 2002, 21(15) :4162-4171.

55. Mundy CL. Patenge N, Matthews AG et al. Assemblyof the RAGl/RAG2 synaptic complex. Mol CellBioi 2002, 22(1) :69-77.

56. Curry JD. Geier JK. Schlissel MS. Single-strand recombination signal sequence nicks in vivo: evidencefor a capture model of synapsis. Nat Immunol 2005, 6(12) :1272-1279.

57. Hiom K, Gellert M. Assembly of a 12123 paired signal complex: a critical concrol point in V(D)Jrecombination . Mol Cell 1998, 1(7):1011-1019.

58. van Gent DC. Mizuuchi K. Gellert M. Similarities between initiation of V(D)J recombination andretroviral integration . Science 1996,271(5255) :1592-1594.


59. Ma Y, Pannicke U. Schwarz K et al. Hairpin opening and overhang processing by an Artemis/DNAdependent protein kinase complex in nonhomologous end joining and V(D)J recombination . Cell2002; 108(6):781-794.

60. Goodarzi AA. Yu Y, Riballo E et al. DNA-PK autophosphorylarion facilitates Artemis endonucleaseactivity. EMBO ] 2006; 25(16):3880-3889.

61. Grawunder U. Wilm M. Wu X et al. Activity of DNA ligase IV stimulated by complex fonnation withXRCC4 protein in mammalian cells. Nature 1997; 388(6641) :492-495.

62. Grawunder U. Zimmer D. Kulesza P er al. Requirement for an interaction ofXRCC4 with DNA ligaseIV for wild-type V(D)J recombination and DNA double-strand break repair in vivo.] BioI Chern 1998;273(38):24708-24714.

63. Lu H. Pannickc U. SchwarzK et al. Length-dependent binding of human XLF to DNA and stimulationofXRCC4.DNA ligase IV activity.] BioI Chern 2007; 282(15):11155 -11162.

64. Sakano H. Huppi K. Heinrich G et al. Sequences at the somatic recombination sites of immunoglobulinlight-chain genes. Nature 1979; 280(5720):288-294.

65. Thompson CB. New insights into V(D)J recombination and its role in the evolution of the immunesystem. Immunity 1995; 3(5):531-539.

66. AgrawalA. Eastman QM. Schatz DG. Transposition mediated by RAGI and RAG2 and its implicationsfor the evolution of the immune system. Nature 1998; 394(6695):744-751.

67. Chatterji M. Tsai CL. Schatz DG . Mobilization of RAG-generated signal ends by transposition andinsertion in vivo. Mol Cell Bioi 2006; 26(4):1558 -1568.

68. Clatworthy AE, ValenciaMA, Haber]E et al. V(D)J recombination and RAG-mediated transpositionin yeast. Mol Cell 2003; 12(2):489-499.

69. Matthews AG. Elkin SK. Oettinger MA. Ordered DNA release and target capture in RAG transposition . EMBO] 2004; 23(5) :1198-1206.

70. Melek M, Gellert M. RAGl/2-mediated resolution of transposition intermediates: two pathways andpossible consequences. Cell 2000; 101(6):625-633.

71. Elkin SK. Matthews AG. Oettinger MA. The C-tenninal portion ofRAG2 protects against transposition in vitro. EMBO] 2003; 22(8) :1931-1938.

72. Swanson PC , Volkmer D, Wang L. Full-length RAG-2 and not full-length RAG-I . specifically suppresses RAG-mediated transposition but not hybrid joint formation or disintegration. ] Bioi Chern2004; 279(6) :4034-4044.

73. Tsai CL. Schatz DG . Regulation ofRAGl/RAG2-mediated transposition by GTP and the C-tenninalregion ofRAG2. EMBO] 2003; 22(8):1922 -1930.

74. Jiang H, Ross AE. Desiderio S. Cell cycle-dependent accumulation in vivo of transposition-competentcomplexes between recombination signal ends and full-length RAG proteins. ] Bioi Chern 2004;279(9):8478-8486.

75. Ramsden DA. Gellert M. Formation and resolution of double-strand break intermediates in V(D)]rearrangement. Genes Dev 1995; 9(19):2409-2420.

76. Peters ]E. Craig NL. Tn7 recognizes transposition target structures associated with DNA replicationusing the DNA-binding protein TnsE. Genes Dev 2001; 15(6):737-747.

n. Srellwagen AE, Craig NL. Mobile DNA clements: cont rolling transpos ition with ATP-dependentmolecular switches. Trends Biochem Sci 1998; 23(12) :486-490.

78. Yamauchi M, Baker TA. An ATP-ADP switch in MuB controls progression of the Mu transpositionpathway. EMBO] 1998; 17(18):5509-5518.

79. Kaufinan PD, Rio DC. P clement transposition in vitro proceeds by a cut-and-paste mechanism anduses GTP as a cofactor. Cell 1992; 69(1) :27-39.

80. Traut TW. Physiological concentrations of purines and pyrim idines . Mol Cell Biochem 1994;140(1) :1.22.

81. Lee GS. Neiditch MB, Sinden RR et al. Targeted transposition by the V(D)J recombinasc. Mol CellBioI 2002; 22(7):2068-20n.

82. Posey]E. Pytlos M]. Sinden RR et al. Target DNA structure plays a critical role in RAG transposition .PLoS BioI 2006; 4(11) :e350.

83. Cremer T, Cremer C. Chromosome territories. nuclear architecture and gene regulation in mammaliancells. Nat Rev Genet 2001; 2(4) :292-301.

84. Parada L. Misteli T. Chromosome positioning in the interphase nucleus. Trends Cell Bioi 2002;12(9):425-432.

85. Gao Y, Ferguson DO. Xic W er al. Interplay of p53 and DNA-repair protein XRCC4 in tumorigenesis.genomic stability and development. Nature 2000; 404(6780) :897-900.

86. Rooney S. Sekiguchi ]. Whitlow S et al. Artemis and p53 cooperate to suppress oncogenic N-mycamplification in progenitor B-cells. Proc Natl Acad Sci USA 2004; 101(8):2410-2415.


87. Feldmann H. Winnacker EL. A putative homologue of the human aucoanngen Ku from Saccharomycescerevisiae, J BioI Chern 1993; 268(17) :12895-12900.

88. Boulton SJ.Jackson SP. Identification of a Saccharomyces cerevisiae Ku80 homologue: roles in DNA double strand break rejoining and in telomeric maintenance. Nucleic Acids Res 1996; 24(23) :4639-4648.

89. Milne GT. Jin S. Shannon KB er al. Mutations in two Ku homologs define a DNA end-joining repairpathway in Saccharomycescerevisiae. Mol Cell BioI 1996; 16(8):4189-4198.

90. Herrmann G. Lindahl T. Schar P. Saccharomyces cerevisiae LIF1: a function involved in DNAdouble-strand break repair related to mammalian XRCC4. EMBO J 1998; 17(14):4188-4198.

91. Callebaut I, Malivert L, Fischer A er al. Cernunnos interacts with the XRCC4 x DNA-ligase IVcomplex and is homologous to the yeast nonhomologous end-joining factor Nejl. J BioI Chern 2006;281(20): 13857-13860.

92. Hentges P, Ahnesorg P. Pitcher RS et al. Evolutionary and functional conservation of the DNA nonhomologous end-joining protein , XLF/Cernunnos. J BioI Chern 2006; 281(49):37517-37526.

93. Schar P. Herrmann G. Daly G et al. A newly identified DNA ligaseof Saccharomyces cerevisiae involvedin RAD52-independent repair of DNA double-strand breaks. Genes Dev 1997; 11(15):1912-1924.

94. Liu Y, Subrahmanyam R, Chakraborty T et al. A plant homeodomain in RAG-2 that binds Hypermethylated lysine4 of histone H3 is necessaryfor efficient antigen-receptor-generearrangement. Immunity2007; 27(4) :561-571.

95. Ramon -Maiques S, Kuo AJ. Carney D er al. The plant homeodomain finger of RAG2 recogn izes histone H3 methylated at both lysine-4 and arginine-2. Proc Natl Acad Sci USA 2007 ;104(48):18993-18998.

96. Hizi A. Levin HL. The integrase of the long terminal repeat-retrotransposon tfl has a chromodomainthat modulates inregrase activities.J BioI Chern 2005; 280(47) :39086-39094.

97. Malik HS. Eickbush TH. Modular evolution of the integrase domain in the Ty3/Gypsy class of LTRretrotransposons . J Vito11999; 73(6) :5186-5190.

98. Nagaki K, Neumann P. Zhang D et aI. Structure. divergence and distribution of the CRR centromericrerrotransposon family in rice. Mol BioI EvoI2005 ; 22(4) :845-855.

CHAPTER 3

Recent Insights into the FormationofRAG-Induced ChromosomalTranslocationsVicky L. Brandtand David B. Roth"

Abstract

Chromosomal translocations are found in many types oftumors, where they may be eithera cause or a result of malignant transformation. In lymphoid neoplasms, however, itis clear that pathogenesis is initiated by any ofa number of recurrent DNA rearrange

ments.These particular translocations typically place an oncogene under the regulatory controlofan Ig or TCR gene promoter, dysregulating cell growth, differentiation, or apoptosis. Giventhat physiological DNA rearrangements (V{D)J and class switch recombination) are integralto lymphocyte development, it is critical to understand how genomic stability is maintainedduring these processes. Recent advances in our understanding of DNA damage signaling andrepair have provided clues to the kinds ofmechanisms that lead to V{D)J-mediated translocations. In turn, investigations into the regulation ofV{D)J joining have illuminated a formerlyobscure pathway ofDNA repair known as alternative NHEJ, which is error-prone and frequentlyinvolved in translocations. In this chapter we consider recent advances in our understandingof the functions of the RAG proteins, RAG interactions with DNA repair pathways. damageSignaling and chromosome biology, all ofwhich shed light on how mistakes at different stagesofV{D)J recombination might lead to leukemias and lymphomas.

IntroductionLymphoid neoplasms are among the most common malignancies in humans; mysteriously,

they have become increasingly common in both adults and children over the past two decades,with the incidence ofnon-Hodgkin's lymphoma alone having doubled.' A number offactors areimplicated in the etiology of these disorders , including ionizing radiation, chemical exposures,viral infection, autoimmune disease and acquired immunodeficiencies. Some ofthese conditionsmight directly create genetic mutations that initiate tumorigenesis; others may simply promotea favorable immune milieu by chronic antigenic stimulation or immunosuppression. It is fairlycertain, however, that many lymphoid neoplasms are born ofchromosomal translocarions involving antigen receptor 10ci.2J Up to 90% of cases of non-Hodgkins lymphoma, for instance, bearsuch translocations.' These aberrant rearrangements most often exert their oncogenic effects byplacing an oncogene under the regulatory control ofa highly expressingIg or TCRgene promoter,thereby dysregulating cell differentiation ,proliferation,or survival.t? Translocations also commonly

*Corresponding Author: David B. Roth-Department of Pathology and Program in MolecularPathogenesis, The Helen l. and Mart in S. Kimmel Center for Biology and Medicine at theSkirbailinstitute for Biomolecular Medicine, New York University School of Medicine, NewYork, NY 10016, USA. Email: [email protected]

V(D)JRecombination, edited by Pierre Ferrier. ©2009 Landes Bioscienceand Springer Science+Business Media.

RecentInsights intotheFormation ofRAG-InducedChromosomal Translocations 33

fuse the codingsequences of two differentgenes,which then encodechimericoncoproteins thatactivateoncogenictranscriptionalprograms," Both typesofeventsfrequentlybearsignsof havingoriginated through some error in V(D)J recombination, the process by which antigen receptorgenesare rearranged.2.3·7.8

V(D)] recombination canbe thought of asa special caseof targeted, stricdyregulatedgenomicinstability.Thereareseven antigenreceptorlocithatencodethe T-cellreceptor(TCR) a, p, yand [)chainsandthe immunoglobulin (Ig)HandL (K andA) chains. GroupsofY,D and] codingsegmentsarearrayed alongthe loci,flanked byrecombination signalsequences (RSS).Thelymphoid-specificrecornbinase, consistingof RAGland RAG2(the proteinproductsof the recombination activatinggenes 1and 2), selects apairof signal sequences that maybe manykilobases apart, cleaves the DNAat the signalsequence borders, and the resulting DNA double-strand breaks arejoinedbythe ubiquitous nonhomologous end joining(NHE]) proteins.Sinceantigenreceptorgenerearrangemententailsbreakingand rejoining the chromosome several timesbeforeacompleteIgor TCR moleculecanbe expressed on the cellsurface, the creationofa diverse repertoire ofantigenreceptors violatesgenomicintegrityasa matterofcourse. It has beenestimatedthat,eachday, the humanbodycreates1x 1011 Bvcells.? Granted,mostof thesenewly generated cells diebecause theyform nonfunctionalor self-reactive antigenreceptors. Even so,an estimated 9 x 109cells survive this process everyday.9Thesenumbersarestaggeringly large. An errorrateoflessthan a thousandthofa percentwouldstillyielda largenumberof cells bearingpotentiallyoncogenic translocations. How is it that leukemiasand lymphomas do not overcome us all? The mechanisms that preserve genomicintegrityduringrearrangement mustbe unusually reliable, multiplyredundant,or both.

In fact, the obviousrisksattendant upon sequentialcutting and pastingof genefragments aremitigatedbynumerousrestrictionson the process, manyofwhichhaveonlyjustbeenappreciated(and many others of which, no doubt, remain to be discovered). Regulation of recombinationrequiresdeft orchestrationof chromatin changes, trans-actingfactors, transcription, selectionofsubstratesfor DNA cleavage and DNA double-strandbreak (DSB) repair machinery. Thereareexcellent reviews in this volumethat do greaterjusticeto the topic of accessibility than wecouldin this chapter (seealsorefs. 10-12). Our focuswill be on recent work elucidatingthe molecularmechanisms for maintaining the fidelityofDSB repair. We will begin the chapter by oudiningthe salientfeatures ofthe V(D)J reaction.Wewill then considerthosestages wheremistakes oftenoccur,with a focus on mechanisms that can lead, in theory at least,to translocations.

Overview ofthe V(D)J Recombination ReactionKeysteps in the reactionare outlined below. For comprehensive and elegantdescriptionsof

the biochemistry, seereferences 7,13 and 14.The recombination signal sequences (RSS) that flank the V, D and] segments consist of

conservedheptamer and nonamer elementsseparatedby an intervening spacerof either 12 or23 nucleotides. Theserecognition sequences are referred to as 12-RSS or 23-RSS, and efficientrecombinationrequiresthat two complementaryRSS(a 12123 pair) be synapsed beforecleavagecan proceed.P'F'Ihe heptamer has the palindromic consensus sequenceCACAGTG, but variations are common and the extent of deviation from the consensus influences the efficiency withwhich a site is cleaved. The AT-richnonamer sequenceis less conservedbut still important forrecomblnarion", and eventhe spacersequences influence the selectionof an RSS.19·22

The RSS are recognizedby the lymphoid-specific proteins RAG1 and RAG2 ("recombination activatinggenes1 and 2"23), which together form a complexwewill referto as the V(D)] orRAG recombinase. HMGB1 (high mobilitygroup box I), a nonspecific DNA bendingprotein,facilitates synapticcomplexformationand cleavage.24.2sThe RAG proteins nick one DNA strandprecisely between the RSSheptamer and the coding segment.This generates a free 3'OH thatis used to attack the opposite strand in a transesterification reaction, forming a double-strandbreak (DSB).The result is that the synapsed pair of RSS/codingsegmentsyields four free DNAends: two covalendysealed(hairpin) coding ends and two signalends that terminate in a flushdouble-strandbreak.26-30

34 V(D)]Ruombination

After coupled cleavage, the RAG proteins hold the DNA ends in a posecleavage complex,aligningthem for proper joining by the nonhomologousend joining (NHEJ) machinery. Theblunt-ended RSS undergo direct ligation (generally with no base loss) to form a signaljoint,whichisusually deletedasan extrachromosomal circularproduct that is lostduringcelldivision.Less frequently, the orientation of the codingsegments necessitates inversional recombination,in whichthe signal joint is retainedin the chromosome. Thereisno knownimmunological function for signal joints, but in cases of inversional recombinationtheir formation is necessary forpreservinggenomicintegrity.Ligationofthe twocodingendsproducesacodingjoint thatencodesthe variable portion of the antigenreceptorprotein. Coding jointsare typically imprecise, asthecodingend hairpinsmust firstbe opened and often undergolossor addition of nueleotides duringprocessing.Thisjunctionalvariability contributesfurther to antigenreceptordiversity and isconsidered characteristic of repairbynonhomologousend-joining.

Potential Mechanisms ofRAG-Mediated TranslocationsErrorsin recombination canbebroadly elassifiedinto twocategories.Thoseoccurringduringthe

earlystageof the reaction(siteselectionand cleavage) can be conceptualized ascases of mistakenidentity: they involve either (1) mixingofauthenticbut inappropriateantigenreceptorloci (e.g.,TCRj3andTCRysegments) in interlocus recombination,or (2) the misappropriation ofsequencesthat fortuitouslyresemble RSS (crypticRSS). One mechanism forpreventingsucherrorsinvolvesregulationofsubstrateaccessibility; wewill discuss thisand relatedregulatorycontrolsrelevanttoeachtypeof substrateselection error in the following section. Errorsthat takeplacein laterstagesof the reaction (joining) can instead be conceived as involving renegade double-strandbreaks.BrokenDNA ends createdin the context ofV(D)J recombinationmight escape normal DNArepairthrough defects in the RAGposteleavage complex, useof an inappropriaterepairpathway,or an impairedDNA damage signaling response. Mechanisms that act to curtail aberrant repairwill be considered in the contextof thesedeficits in subsequentsections.

Mistaken Identities: Substrate Selection Errors

Inter/oeus RecombinationNormalV(D)Jrecombination isrestricted bycelllineage (TCR locirearrange inT-cellsbut not

B-cells), developmental stage(e.g.,TCRj3 beforeTCRn) and, in manycells, to one allele (allelicexclusion). Sincethe RAGproteins, the RSSand the DNA repairmachineryarethe samein eachcase, this complexregulatoryschemedependsin largepart on the degreeofaccessibility allowedthe recombinase to the various lociovertime in differentcells. For this reason, the packaging ofTCR and 19 loci into chromatindiffers in B- and T-cells and varies accordingto the activityofthe loci,which is governed bydevelopmental stage.

Nevertheless, sometemporaloverlap in the sequence of rearrangements doesallowoccasionalinterlocus(trans)recomblnaeion.t' :"These rearrangements, whichcreatea balanced translocationresultingin two derivative chromosomes, can generatefunctional chimericreceptorchainsthatappearin normaltissues.33.34 Aswith recurrentoncogenicrranslocarions, the system seems to favorrearrangements ofparticularsites: forexample, it hasbeenestimatedthat 1 in 10,000normalhuman and mouserhymocyees carries the Db3-J~2.7 rearrangemene.P" Theserearrangements, JUStlike those that occur in as, relyon RSS recognition, RAG-mediatedcleavage and NHEJ repair.TheyarenormalV(D)J reactions simplycarriedout with the wrongpartner. Ineerlocus eventsdo,however, exhibit recurrentbaselossfromsignaljoints31.36 anddifficulty formingcodingjoints.37-39

Thesefeatures suggest that trans rearrangements proceedthrough an abnormalpathway.It is noteworthythat the incidence of interlocus recombination increases dramatically in cells

bearingcertainmutations (suchasATMdeficiency) thatpredispose to lymphoid turnors.32.40-42Theseevents havethe appearance ofsimple substrate selection errors, but at leastsomeof theserearrangementsmight arisefromfailures in DNA damage sensing and repair(seediscussion ofATMdefectsbelow, in the section"Theroleof the DNA damage response in preventing translocations").

RecentInsightsinto theFormation ofRAG-InducedChromosomalTranslocations 35

CrypticRSSThevariabilityofRSSsequence entailsconsiderable flexibility on thepart of the RAGproteins.

Unfortunately, this plasticitymakes it possible for the RAG proteins to bind to fortuitous DNAsequences known as "crypticRSS" that do not border antigen receptor gene segmentsbut aresufficiently closeto the consensus sequenceto allowRAG recognition.43.44 Inone largereviewofoncogenicrearrangements from both B-and T-cellmalignancies, most translocationbreakpointson the nonantigen receptorgenepartner contained RSS-like sequences at or near the breakpoint,supporting "substrate selection error" as the responsible mechanism.' In addition, nontemplatednucleotides arefrequently addedto the junctions, suggestingTdTactivity andthereforethe involvementofV(D)J recombination? Thet(7:9) (q34;q32) translocations foundinT-celllymphoblasticleukemiaprovide the clearest example. Chromosome 7 breakpointsare typicallylocated at theRSSborderingD~ segments, whilebreakpointson chromosome9 are flankedby consensus RSSheptamersequences separatedfrom AT-richnonamer-like sequences by 11or 12basepairs," Thesalientfeatureofsubstrateselectionerrorsis that the V(D)J recombinationreactionproceedsasnormal exceptfor parmering an RSSwith an inauthentic sequence.

PreventingErrors by ControllingAccessibilityAn RSS can deviate quite far from the consensusand still undergo recombination; Lewis

et al definedthe necessary features of cryptic RSSand suggested that evena weak signal, with arecombinationfrequencyof2 x 10-5the canonicallevel, canhaveaphysiological impact." In lightof estimates that the genomecontains 10million potential crypticsites, approximately one every1-2 Kb,46 it is clearthat RAG accessibility to target sitesmust be verytightlyregulated.

In aprescient1985paper,Yancopoulos andAlenoted that rearrangingsegments aretranscribedbefore(or coincidentwith) their activationfor rearrangement and proposedthat generatingthesegermlinetranscriptsalteredchromatinstructuresoasto allowthe recombinase access to asubsetofappropriatesubstrates." Therearealsoother potential mechanisms for regulatinglocusaccessibility that do not relyon rranscriprion.f One approachto controllingaccess is through nucleosomepackaging, which can block cleavage of specific RSS.49 Proteins that enhance RAG interactionwith RSSS48.50.51 could conceivably recruit nucleosome remodelingcomplexes such as Swi/Snfthat alter DNA-histone contactswithin a nucleosome or alter the nucleosomeslocation.52.53Thesecondapproachis through covalentmodifications of the taildomainsof the histone proteins byacetylationoflysines,methylationoflysinesand arginines, polyribosylation, serinephosphorylation and ublqulryladon." Suchposttranslationalmodifications can"open"chromatinbyalteringDNA-histone contactswithin a nucleosome, histone-histone contactsbetween nucleosornes, orinteractions betweenhistonesandotherproteins.Accumulatingevidence suggests that thesereversible,epigeneticmodifications comprisea "histone code" and that they associate with regulatoryproteins known as code readers. Evolutionarily conserveddomains within code-reader proteinsbind to certain histone modifications with such specificity that they can distinguish the samemodificationat differentresidues (for example, trimethylationat K4vs. K9).54

Several recentstudieshaveshownthat the plant homeodomain (PHD) finger, a methyl-lysinebinding domain, serves as a code-reader: it can both promote and repress gene expression byinteractingwith trimethylatedlysine 4 on histone 3 {H3K4).55'58 Evenmore recently, the RAG2PHD fingerhasbeen shownto recognize H3K4 trimethylation.P'" In thesestudies, the bindingofRAG2 to H3K4 enhancedthe selectionand recombinationofchromatinizedgenesegments indeveloping lymphocytes. The RAG complex, then, is not merelysubject to chromatinstructuresdetermined byother factors, but must takean activerole in recognizing substrates.

Other studieshaveshownthat transcriptionalcis-regulatorysequences, suchasenhancersandpromotersspecific to eachlocus,arenecessaryforV{D)Jrecombinarion.PP Furthermore, the RAGgenesareregulateddifferently in B-and T-cells(for example, Foxpl isrequiredfor Bvcell-speclficRAG expression"). Some DNA-binding transcription factors interact with RAGl/RAG2 andguide them to subsetsofRSS: Bvcell-spedfic VH locuscontraction,for instance,requiresPax5tointeract with both the V coding segments and the RAG complex/v" The mechanisms oflocus


contractionand loopingremains poorlyunderstood,but theyareessential forpromotingsynapseformationbetweendistalV and proximalD segments, which can beseparatedbydistances of upto 3 megabases.66 (In this regard, it is interestingto note that coreRAG2knock-inmicehavedifficultywithVto DJ rearrangements at the IgHandTC~ 10ci.67.68)Whether nonantigenreceptorlociare typically constrainedbysuchcomplex regulatoryschemes isnot clear.

Signs 1hata Translocation DidNotArise through Substrate Selection ErrorEven grantingtheoccasional chromatinloophole, threeobservations suggest that substrate selec

tionerrorsdo not accountforthe majorityofRAG-mediated oncogenic translocations.First,manyofthe RSS-like sequences found at translocationbreakpointson the nonantigenreceptorpartnerchromsome contain heptamersthat area poor matchfor the consensus, and a largefraction lackrecognizable nonamerelernenrs.P Previous workhasshown that DNA cleavage in vivorequitesboth heptamerand nonamer;scrambling the nonameror mutatinga singlecriticalnucleotideinthe heptamer decreases cleavage by at leasttwo ordersofmagnitude.15•18.22.69Therefore, the presenceof sequences that deviateso much from the consensus on the partner (nonantigenreceptorlocus) chromosome might be merely coincidental.2.3·7The second argument against the use ofsomecrypticRSSin translocations is that the breakpointsare often not at the heptamer-codingflankborder. This is incompatible with normal RAG-mediatedcleavage, which is a veryprecisereaction. Finally, sometranslocations display short direct repeats,8.70 suggesting that the cleavageeventcreateda shorr single-stranded overhang. This, too, isinconsistentwith normalcleavage bythe V(D)J recombinase.

Thisis not to saythat suchevents did not originate with a mistake in V(D)J recombination. Ifsubstrate selection errorappears unlikely, thereisanalternative modelthatbetterexplainscasessuchasthese. lrisknownasenddonationandpositsthattherecombinasecreatesadouble-strand break(DSB)at an authenticRSS that is then somehow joinedto a random DSBthat hasbeen created throughsomeunrelated process? Until the past fewyears it hasbeendifficult to conceive of a mechanismthat wouldexplain enddonation.but recentworksuggests that brokenDNA endscreatedbyRAGcleavage mightescape theirnormalfatethroughdefects in the RAGpostcleavage complex, useofaninappropriate repairpathway, or an impaired DNA damage signaling response.

The Ends That Got Away: Errors inJoiningDSBsarepotentiallysodamaging that cells have evolved complex networksofproteinsto sense

the presence and precise locationof DNA damage, regulatethe cellcycle and repairthe breaks.Mountingevidence suggests that V(D)Jrecombination enjoys at leasttwolayers ofprotectionthatevenits DNA-rearranging cousin, class switchrecombination,doesnot." an end joiningpathwaythat discourages translocations (classical NHEJ) and the RAG posccleavage complex, which isthought to ensurejoiningthrough thispathwayand exclude other,error-pronerepair. YetanotherlayerofprotectionisprovidedbyATM,part of the DNA damagesignalingmachinery, whichmayhavea role in stabllizingthe postcleavage complex but alsocan leadcells with unrepairedbreaksto undertakeapoptosis.

Genome Guardians: 1he Classical NHE]FactorsThe basicoutline of NHEJ seems simpleenough: a set of enzymes captures the two ends of

the broken DNA molecule, a molecularbridgeis formed to juxtapose the ends.and the break isreligared," Inrealitythe process israthercomplex andmanyaspects remainpoorlyunderstood(seerefs. 72 and 73). A keycomponent ofNHEJ is the DNA-dependent protein kinase(DNA-PK)complex,which comprises the DNA-PKcatalytic subunit (DNA-PKcs) and the Ku70 and Ku80nuclearantigens." Nonhomologousrepairis initiatedwhen the Ku70/80 heterodimerencirclesa brokenend,75.76creatinga scaffold for the recruitment ofother factors. Ku attractsDNA-PKcsto the break,whereit mightserve multipleroles. includingthe formationofasynaptic complex tobring the ends together," ActivatedDNA-PKcsrecruitsXRCC4, DNA Ligase IV and Artemis.DNA-PKcsphosphorylation ofArtemisconverts thelatterfromanexonuclease toanendonucleaseand allows it to open the hairpinnedcodingends.77,78 SinceArtemiscannot process everytype of

RecentInsightsinto theFormationofRAG-Induced Chromosomal Translocations 37

nonligatableend, other typesof end-processing enzymes are alsorecruited. Polymerase activity.for example, is likelysuppliedby the DNA polymerase Mu, which associates with XRCC4, andterminaldeoxynucleotidyl transferase (TdT) addsnonrernplared nucleotides to increase junctionaldiversity.79.80 Finally, XRCC4 andDNA Ligase IVligatethe ends.81-83Themost recentlydiscoveredNHEJ factor,known asCernunnos or XLF (forXRCC4-like factor), isalso recruitedbyKu andinteractswith both XRCC4 and Ligase IV to ligatemismatchedand noncohesive ends.84-88Theorder in which all these factorsare recruited might be flexible, accordingto the specific natureof the break."

GeneticablationofKu,DNA-PKcs,DNA Ligase IV,XRCC4, Artemis,or Cernunnosin micepreventsthe completionofV(D)J recombination,arrestingB-and T-celldevelopmentat an earlystageand leadingto aSCID (severe combinedimmunodeficiency) phenotype.Theoveralldefectin DNA repairalsoproducessensitivity to ionizingradiation,amarkedtendencyto translocarionsand developmentoflymphoma (though in somecases, onlyon a pS3-deficient background).90-97(Bycontrast,NHEJ-proficientmammaliancellsreconstitute their chromosomes with remarkableaccuracy afterbeingexposedto dosesofionizing radiationlargeenoughto inducemassive chromosomefragmentation.98.99) SomeNHE]-deficientlinesdevelopnonlymphoidtumorsaswell.90.loo.lo1Thediscovery that a deficiency ofNHEJ factorspromotes oncogenesis revealed a crucialrolefortheseproteins asgenomeguardians.94.9S

Error-Prone EnJ]oining: AlternativeNHE]Despite their obviousdefectsin DNA repair. NHE]-deficient mice (and humans97.I02.103) can

survive long enough to develop malignancy. The mouse tumors frequently show gene fusionsbetweenthe IgH locusand c-Myc but candisplay manyother nonreciprocaltranslocations. Theremust, then, be alternativemechanisms capable of repairingDSBwithout Ku,DNA-PKcs,LigaseIV,or XRCC4. And, in fact, there is, although it wasnot recognizedas an alternativepathwaywhen it wasoriginally describedin mammalian cells in the 1980s.I04-I06

At the time,it wasknownthat eukaryoticcells areableto repairDNA endsbyboth homologousand nonhomologousmeans.Inthe caseofV(D)J recombinationintermediates,homology-basedmechanisms seemedunlikely, aslittle or no homologyispresentbetweencodingends; moreover,rearranged coding segments underwent a curious addition and lossof nucleotides at the junerion."" The mechanismfor nonhomologous repair,however, had not yet been discovered andthe fieldstruggledto understand how "unrelatedDNA ends are joined together willy-nilly withhigh efficiency~104 Thesimilarityof thesejunctions to codingjoints hinted that the DNA breaksgeneratedbythe V(D)J recombinase might be repairedbythe samemechanism .l'" Within severalyears, studiesofV(D)J recombinationin various radiosensitive celllinesmadeit possible to identify componentsofthe NHEJ pathway.108-112 Our understandingofNHEJ thus grewout ofourunderstandingofV(D)J recombination-and because the wild-typeRAG complexguidesDNAends to the classical pathway, not the alternative pathway(seebelow),the latter settled into quietobscurity. Only recently, in fact,has it been realizedthat the two pathways aredistinct.I13-1IS

The hallmarksof junctions formed by alternativeNHEJ are excessive deletions and a reliance on short sequencehomologies (microhomologies).106.1l 3.11S Evenblunt-ended plasmids inKu80-deficient cells undergoresectionand annealingofmicrohomologous sequences rather thansimplybeingjoinedat the blunt ends.I ISIt isworth noting that thesemicrohomologies arepresentat oncogenictranslocations from NHE]-deficient cells." Therefore, although alternative NHEJprovidesenough repairactivityto allowcellsurvival, it appearsto be error-proneand predisposesthe cellto genomicinstability.

But if alternative NHEJ is relatively efficient, whydoes NHEJ deficiency virtuallyobliterateV(D)J recombination?

The RAG Postcleavage Complex Governs Choice olRepair PathwayThe observation that both nucleotide addition and deletion could occur prior to joiningof

codingendsindicatedthat the DNA endsmust remainin one placelongenoughto allowprocessingbypolymerases and endonucleases,' 16 Thus,evenbeforethe discovery of RAG1and RAG2,it

38 V(D)] R~combjnation

seemedthat a stableprotein-DNA complexmust exist to allowthe ends to be accessible to suchmodifyingenzymesaftercleavage.!" When studiesshowedthat celIs deficientin Ku or DNA-PKcould not resolve V(D)J intermediates. it seemedreasonableto thinkthat. by analogywith theMu transposase, a very stable postcleavage complexwould make DNA ends Inaccessible.!'?Asthe field's understandingofNHEJ repair grew. so did curiosityabout how a RAG postcleavagecomplexmight participate in joining.

Lackingaviablein vitro system to studyjoining,weturned to genetics.Separation-of-functionmutants in RAG-! and RAG-2 that are capableof cleavage but exhibit severe joining defectsprovided compellingevidencethat the postcleavage complexservesa crucialfunction in joiningboth coding and signalends.l18•

120 Thesedata lent support to the notion that the RAG proteinsform a scaffoldthat holds the ends together to facilitate joining. Joining mutants could alterthe architecture of the complex, facilitatingpremature releaseof ends or, conversely, creatingatoo-stablecomplexor hindering the recruitment ofNHEJ factors.118.121Intriguingly. two RAG-!mutants phenocopied NHEJ mutants: the rare joints they did manage to form exhibited theexcessive deletions and short sequence homologiescharacteristicofalternative NHEJ.118Thesemutants led us to propose that the RAG proteins might function as genome guardians withinthe context ofV(D)J recombination.

Wepursued thishypothesis further byexarniningwhetherRAG-generatedendscouldbemadeavailable to repairpathways other than NHEJ. (Althoughhomologousrecombinationand NHEJpredominate at differentphasesof the cellcycle, accumulatingevidencesuggests that theycan actat the same time and even cooperate to repair a DSB.73.122) Usingan in vivosystemto assay forrepair ofsignalends byhomologous recombination,Leeet al showed that two joining-impairedRAG! mutants destabilizethe RAG postcleavage complex, allowingthe DNA ends to be availablefor repair byhomologous recombinadon .F' Wild-type postcleavage complexes, by contrast,stimulatedno homologousrecombination. This led us to propose a model in which the normallyquite stable RAG postcleavage complexactively directs DNA ends to the NHEJ machinery forrepair.123 Thequestion remained: howdo the rarecodingjoints produced in NHE]-deficient celIsmanageto be formed by the alternativeNHEJ pathway?

Since the homologous recombination assaywas unable to map the fate of coding ends andwe had identified mutations in RAG2 that affectedjoining without destabilizingthe postcleavage complex. we again took a genetic approach. We identified a truncated RAG2 allele thatallowssubstantial coding and signal joint formation to occur in cellsdeficient for DNA-PKcsor XRCC4.124Junction sequencesrevealeda tendency toward largedeletions and microhomology use. Surprisingly, this RAG2 mutant also revealed alternative NHEJ to be active even inwild-type cells.124Thesestudies, alongwith work from the Alt and de Villartaylabsstudying theuseof alternativeNHEJ in class switch recombination,125.126 makeit clearthat alternativeNHEJisquite robust, albeit error-prone.Thus, wehavecome full circle:V(D)J recombination allowedthe discoveryof classical NHEJ and now has brought attention back to alternative NHEJ.

Why is classical NHEJ less prone to translocations than the alternative pathway? Perhapscomponents of the classical NHEJ pathway interact with chromatin (or chromosome) components to maintain the chromosomalidentity of broken ends (seebelow). In addition. studies ofNHEJ haverevealedthat repair is biphasic:most repair occursquite rapidlyupon induction ofa DSB, but there is a slowcomponent that might correspond to alternativepathwaysand whichcontinues at the samelevelwhen the classical pathwayis disabled.J27 Thus, it seemsthe rapidityofclassical NHEJ repairensuresthat most DSBsarehealed within a fewhours; those lesionsthatcannot be repaired in this time will be subject to alternativeend joining. It is conceivable thatdifficult-to-repairDSBslingeringin the nucleusmight, over time, separateor drift to a differentchromosome territory in the courseof other cellularprocesses (but seebelow).

How Do Chromosome Ends Meet?Mammalianchromosomes occupydiscretethree-dimensional regions in the nucleus known as

chromosometerritories. Theseterritoriesare not fixed. but arespecific to differentcell rypesy8 In

RecentInsightsinto theFormation ofRAG-InducedChromosomal Translocations 39

order for a translocationto occur,there must be DSBsin (at least)two chromosomes at the sametime; the DSBsmust haveescapedthe normal repairmechanisms; the broken chromosomeendsmust physically meet and theymustbe illegitimately repaired.An obviousquestion arises: do theDSBsroam the nucleus. lookingfor a partner, or do they stayput?

Two hypotheses havebeen put forth. The breakage-first model posits that breaksare able totraverse the nuclearspace, searching for potential partners. and cometogether to produce translocations. The contact-firstmodel,on the other hand, proposes that sincechromosomes occupyterritories in the nucleus, breakson distinct chromosomes willmeet only if they occupynearbyor interminglingdomains.!" To test these possibilities, Soutoglou et al developeda cell systemin which they could induceone DSBat a definedsite and followthe fate of eachof the damagedDNA ends in real time by observingspecific fluorescent tagson either sideof the break.129The

authors demonstratedthat a single DSBin mammaliancellsispositionally stable,with onlyslightmotion of the DNA break.129This stability required the end-bindingKu80/Ku70 heterodimerbut, surprisingly, wasindependentofother DNA repairfactors, the structuralproteins H2AX andSMC 1,the cohesincomplexand eventhe Mrel l complex, whichhasbeen stronglyimplicatedinanchoringends.Whether other factorswill turn out to be necessary for this immobilizationof abreak-or whether the causeof the breakage, or the number of breaksinduced at the sametime,influence this positionalstability-remains to be seen.

Theseresults havestrikingimplications forunderstandinghowtranslocations forminvivo.First,they demonstrate that chromosomalpositional stability is related to genomicstability. (At leastin mammals; yeastdo not havechromosometerritories.DSBsin yeastmigrate to any of severalsmallnuclearsites that act as damagerepair centers.P') Second, the data support a contact-firstmodel in mammalian cells and areconsistentwith the emergingmotion that nonrandom, higherorder spatial organization of chromosomes accounts in largepart for the recurrenceof specifictranslocations.Tenyears ago,experiments showedthat y-irradiation ofnormalhumanlymphocytesinduces translocations in chromosomepairs that have been observed in leukemias, suggestingthat thesechromosomes are near neighborsin lymphocytes.131•m Several frequent translocationpartners, includingMyc-Igh and BCR-ABL,havebeenfound to existin closespatialproximitytoeachother in normalcellsbeforethe formationof translocarions.Pl Themisjoiningof proximallypositioned chromosomeregionssupports the observedcorrelationbetween the degreeof chromosome intermingling and the likelihoodof translocations.P" The frequencyof translocationsinvolving antigen receptor loci likely reflects the fact that more gene-rich chromosomes undergoless compactionand more lnrermingling.l"

The Role ofthe DNA Damage Response in Preventing TranslocationsTheDNA damagesensingpathwaywasnot initiallythought to beinvolved in V(D)) recombi

nation,asdamagecheckpointsarenot activatedduringthe process; in fact,it wasassumedthat theRAGpostcleavage complex sequestered the DSBfromthe DNA damagesensingmachinery.It thuscameasasurpriseto findthat ATM,y-H2AXand the Mrel l complexlocalize to RAG-mediatedDNA breaks.I34.135The mystery wasdeepenedbythe firststudiesto investigate whether thesefactors had any role in V(D)) recombination: the answer, apparently, wasno.136.137 Further probingunearthed a greatertendency to TCR alb interlocusrecombinationin micedeficientfor ATM,MrelI ,Nbs1,or S3BP1.42,138.141 Micedeficient in ATM,RadSO.or H2AXdevelopthymiclymphomas,asdo H2AX-andMrel l-deficientmiceon apS3nullbackground,136-139Manyofthesetumorsharbor translocations thought to derivefromerrorsin V(D)J recombination,and tumorigenesis isreducedor delayed in micewhenATMdeficiency iscoupledwith RAG1or RAG2deficiency.142,143Mutationsin ATM,Nbs1and Mrel l causeAtaxia-Telangiectasia, NijmegenBreakage syndromeand Ataxia-Telangiectasia-Likedisorder,respectively; like the mice, patients with these diseaseshavea predispositionto lymphoidmalignancies and harbor frequent translocations between theTCR and Igloci.

Recentstudiesprovideinsight into the roleplayedbyATM (and perhaps,byextension,otherdamage sensors) in V(D)) recombination and why this role is virtually invisible under normal


circumstances. In additionto its newly discovered rolein stabilizing DSBcomplexes duringV(D)Jrecombinarlon.lf ATM has a checkpoint function to prevent the propagation ofDSBs causedeither by RAG or low-dose gammairradiation to daughter cells.l" Callen and colleagues positthat ATM-I- lymphocytes that fail primaryV(D)J assembly, leaving a DSBon one allele, canstillachieve productiverearrangement through independent recombinationof the secondallele. Thepresence of the DSBin ATM-deficient cells would not preventpre-B-cells from undergoingthematurationalprocess. Therefore,DSBsproducedinprecursorcellswouldpersistin matureB-cellsin peripherallymphoidtissues, wheretheywould then undergoclass switchingand be subjecttofurther (AID-mediated) DNA breakage.!"The initial RAG-mediatedbreak could persist formanydays, ultimately to be joined to another chromosome in a progenycell.

Thismodelputs an interestingtwist on extant modelsof how chromosome ends meet in thenucleus and undergomisrepair, forminga translocation. Theworkof Callenand colleagues supports acontact-firstmodelbut suggests that a DSBcouldmigratefromits originalpositionin thechromosome territoriesand participatein a repaireventwith another chromosome broken in aprogenycell.14s One might think ofthisasdiachronicend donation.With regardto physiologicalrelevance, it is strikingthat up to 50%of mantle celllymphomas havemutationsor deletionsinATM.l46 Callenet al suggest that ATM mutation is likely to be an earlyevent in the malignanttransformation.l"

The foregoing studiesemphasize that creating (or preventing) a translocation is a complexprocess. One has to considernot only the natureofrepair factorsand the orderedassembly anddisassemblyofDNA-proteincomplexes, but the factthat theseprocesses take placein threedimensionsand over time. Understandingthe spatiotemporal regulationof these repairprocesses andtheir coordination with chromosome dynamics, changes in chromatin structure, DNA damagesignaling, the cellcycle and other physiological processes represents one of the majorchallengesto unraveling the puzzleof aberrant V(D)J recombinationevents. Indeed, the recent discoverythat over700 proteins interact with ATM and ATR in the DNA damage response!" indicatesthat this story is likely to get much more complicated.

References1. Fisher SG , Fisher RI. The epidemiology of nonHodgkin's lymphoma. Oncogene 2004:

23(38) :6524-6534 .2. Tycko B. Sklar ]. Chromosomal translocarions in lymphoid neoplasia: a reappraisal of the recombinase

model. Cancer Cells 1990: 2:1-8.3. VanasseG]. Concannon P. Willerford DM. Regulated genomic instability and neoplasia in the lymphoid

lineage. Blood 1999: 94(12) :3997-4010.4. Kirsch lR. Morton CC. Nakahara K et al. Humanimmunoglobulin heavy chain genes map to a region

of translocations in malignant B lymphocytes. Science 1982; 216(4543) :301-303.5. Dalla-Favera R. Bregni M. Erikson] et al. Human c-myc onc gene is located on the region of

chromosome 8 that is translocated in Burkitt lymphoma cells. Proc Nad Acad Sci USA 1982:79(24):7824-7827.

6. Look AT. Oncogenic transcription factors in the human acute leukemias. Science 1997;278(5340) :1059-1064.

7. Lewis SM. The mechanism ofV(D)] joining : Lessons from molecular. immunological and comparativeanalyses. Adv Immuno11994; 56:27-150.

8. Kuppers R. Dalla-FaveraR. Mechanisms of chromosomal translocations in B-celllymphomas. Oncogene2001; 20(40) :5580-5594.

9. Saada R, Weinberger M. ShahafG et al. Models for antigen receptor gene rearrangement: CDR3length.Immunol Cell Bioi 2007; 85(4) :323-332.

10. Krangel MS. Gene segment selection in V(D)] recombinat ion : accessibility and beyond. Nat Immunol2003; 4(7) :624-630.

11. Schlissel MS. Regulating antigen-receptor gene assembly. Nat Rev Immuno12003: 3(11):890-899.12. lung D, Giallourakis C. Mostoslavsky R er al. Mechanism and control ofV(D)] recombination at the

immunoglobulin heavy chain locus. Annu Rev Immunol 2006; 24:541-570.13. Gellert M. V(D)] recombination: RAG proteins. repair factors and regulation. Annu Rev Biochem

2002; 71:101-132.

Recent Insights intotheFormation ofRAG-InducedChromosomal Translocations 41

14. Fugmann SD. Lee AI. Shocken PE et al. The RAG proteins and V(D)J recombination: complexes. endsand transposition. Annu Rev Immunol 2000; 18:495-527.

15. Steen SB. Gomelsky L. Roth DB. The 12/23 rule is enforced at the cleavage step of V(D)J recombinationin vivo. Genes to Cells 1996; 1(6):543-553.

16. Eastman QM . Leu TMJ. Schatz DG. Initiation of V(D)J recombination in vitro obeying the 12/23rule. Nature 1996; 380:85-88.

17. van Gent DC. Ramsden DA. Gellert M. The RAGI and RAG2 proteins establish the 12/23 rule inV(D)J recombinat ion. Cell 1996; 85:107-113.

18. Hesse JE. Lieber MR. Mizuuchi K et al. V(D)J recombination: a functional definition of the joiningsignals. Genes Dev 1989; 3:1053-1061.

19. Nadel B. Tang A. £SCUro G et al. Sequence of the spacer in the recombinat ion signal sequence affectsV(D)J rearrangement frequency and correlates with nonrandom Vkappa usage in vivo. J Exp Med 1998;187(9) :1495-1503.

20. Bassing CH. Alr FW; Hughes MM er al. Recombination signal sequences restrict chromosomal V(D)Jrecombination beyond the 12/23 rule. Nature 2000; 405(6786) :583-586.

21. Feeney AJ. Goebel P. Espinoza CR. Many levelsof control of V gene rearrangement frequency. ImmunolRev 2004; 200:44-56.

22. Swanson Pc. The bounty of RAGs: recombination signal complexes and reaction outcomes. ImmunolRev 2004; 200:90-114.

23. Oettinger MA. Schatz DG . Gorka C et al. RAG-I and RAG-2. adjacent genes that synergisticallyactivate V(D)J recombination. Science 1990; 248:1517-1523.

24. van Gent DC. Hiom K. Paull TT et al. Stimulation ofV(D)J cleavageby high mobility group proteins.EMBO J 1997; 16(10):2665-2670.

25. Dai Y.Wong B. YenYM et al. Determinants of HMGB proteins required to promote RAGl/2-recombination signal sequence complex assembly and catalysis during V(D)J recombination. Mol Cell Bioi2005; 25(11) :4413-4425.

26. Roth DB. Nakajima PB. Menetski JP et al. V(D)J recombination in mouse thyrnocyres. Double-strandbreaks near T-cdl receptor delta rearrangement signals. Cell 1992; 69:41-53.

27. Roth DB. Menetski JP. Nakajima PB et al. V(D)J recombination : Broken DNA molecules withcovalently sealed (hairpin) coding ends in scid mouse rhymocyres, Cell 1992; 70:983-991.

28. Roth DB. Zhu C. Gellert M. Characterization of broken DNA molecules associated with V(D)Jrecombination. Proc Nat! Acad Sci USA 1993; 90:10788-10792.

29. Schlissd M. Constantinescu A. Morrow T et al. Double-strand signal sequence breaks in V(D)J recombinat ion are blunt . 5'-phosphorylated, RAG-dependent and cell cycle regulated. Genes Dev1993; 7:2520-2532.

30. McBlaneJF, van Gent DC. Ramsden DA et al. Cleavage at a V(D)J recombination signal requires onlyRAGI and RAG2 proteins and occurs in two steps. Cell 1995; 83:387-395.

31. Tycko B. Palmer JD. Sklar J. T-cell receptor gene trans-rearrangements: chimeric gamma delta genes innormal lymphoid tissues. Science 1989; 245:1242-1246.

32. Kobayashi Y. Tycko B. Soreng AL er al. Transrearrangemenrs between antigen receptor genes in normalhuman lymphoid tissues and in ataxia telangiectasia. J Immuno11991 ; 147:3201-3209.

33. Davodeau F. Peyrar MA. Gascher J et al. Surface expression of functional T-ccll receptor chains formedby interlocus recombination on human T-lymphocytes. J Exp Med 1994; 180(5):1685-1691.

34. Bailey SN. Rosenberg N. Assessing the pathogenic potential of the V(D)J recornbinase by interlocusimmunoglobulin light-chain gene rearrangement. Mol Cell Bioi 1997; 17(2) :887-894.

35. Marculescu R. Le T, Simon P et al. V(D)J-mediated translocations in lymphoid neoplasms: a functionalassessment of genomic instability by cryptic sites. J Exp Med 2002; 195(1):85 -98.

36. Garcia IS. Kaneko Y,Gonzalez-Sarmiento R er al. A study ofchromosome lip13 translocations involvingTCR beta and TCR delta in human T-cell leukaemia. Oncogene 1991; 6(4 ):577-582.

37. Han ]-0. Steen SB. Roth DB. Intermolecular V(D)J recombination is prohibited specifically at thejoining step. Mol Cell 1999; 3:331-338.

38. 'Ievelev A. Schatz DG. Intermolecular V(D)J recombination. J Bioi Chern 2000; 275(12 ):8341-8348.39. Agard EA. Lewis SM. Posccleavage sequence specificity in V(D)J recombination. Mol Cell Bioi 2000;

20(14):5032-5040.40. Lipkowitz S. Stem MH. Kirsch IR . Hybrid T-cell receptor genes formed by interlocus recombination

in normal and ataxia-telangiectasia lymphocytes. J Exp Med 1990; 172(2) :409-418.41. Kirsch IR. Lipkowitz S. A measure of genomic instability and its relevance to lymphomagenesis,Cancer

Res 1992; 52:5545s-5546s.42. Theunissen jw; Kaplan MI . Hunt PA et al. Checkpoint failure and chromosomal instability without

lymphomagenesis in Mrdl(ATLDl/ATLDl ) mice. Mol Cell 2003; 12(6):1511-1523.


43. Lewis SM. AgardE. Suh S et al. Cryptic signals and the fideliry ofV(D)J joining. Mol Cell Bioi 1997;17(6):3125-3136.

44. Zhang M, Swanson PC. V(D)J recombinase binding and cleavage of cryptic recombination signalsequences identified from lymphoidmalignancies. J Bioi Chern 2008:283(11):6717-27.

45. Tycko B. Reynolds TC, SmithSD et al. Consistentbreakage between consensus recombinasc heptamersof chromosome 9 DNA in a recurrentchromosomal translocation of humanT-celllcukcmia. J Exp Med1989; 169(2):369-377.

46. Cowell LG, Davila M, Yang K er al. Prospective estimation of recombination signal efficiency andidentification of functional cryptic signals in the genome by statistical modeling. J Ezp Med 2003;197(2):207-220.

47. Yancopoulos GD. Alt FW. Developmentally controlled and tissue-specific expression of unrearrangedVH gene segments. Cell 1985;40(2):271-281.

48. Stanhope-Baker P. Hudson KM. Shaffer AL et al. Cell type-specific chromatinstructuredetermines thetargetingofV(D)J recombinase activity in vitro. Cell 1996; 85(6):887-897.

49. Baumann M, Mamais A, McBlane F et al. Regulation ofV(D)J recombination by nucleosome positioning at recombination signal sequences. EMBOJ 2003; 22(19):5197-5207.

50. Muegge K, West M. Durum SK. Recombination sequence-binding protein in thymocytes undergoingT-cell receptorgene rearrangement. Proc Nat AcadSci USA 1993; 90:4151-4155.

51. Kwon J, Imbalzano AN, Matthews A et al. Accessibility of nucleosomal DNA to V(D)J cleavage ismodulatedby RSS positioningand HMG1. Mol Cell 1998; 2(6):829-839.

52. Oettinger MA. How to keepV(D)J recombination under control. ImmunolRev2004; 200:165-181.53. Saha A. WittmeyerJ, Cairns BR. Chromatin remodelling: the industrial revolution of DNA around

histones. Nat RevMol Cell Bioi2006; 7(6):437-447.54. Rurhenburg AJ, Li H. Patel DJ et al. Multivalent engagement of chromatin modifications by linked

bindingmodules. Nat RevMol Cell Bioi2007; 8(12):983-994.55. Li H. Ilin S,WangW er al. Molecular basis for site-specific read-out of histoneH3K4mc3by the BPTF

PHD finger of NURF. Nature 2006; 442(7098):91-95.56. Pena PY, Davrazou F. Shi X et al. Molecular mechanism of histone H3K4mc3 recognition by plant

homeodomain ofING2. Nature 2006; 442(7098):100-103.57. Shi X. Hong T. WalterKL et al. ING2 PHD domain links histone H3 lysine 4 methylation to active

gene repression. Nature 2006; 442(7098):96-99.58. WysockaJ, Swigut T, Xiao H er al.A PHD finger ofNURF couples histoneH31ysine4 trimethylation

with chromatin remodelling. Nature 2006; 442(7098):86-90.59. Liu Y, Subrahmanyam R. Chakraborty T et al. A plant homeodomain in RAG-2 that binds hyperm

erhylared lysine 4 of histoneH3 is necessary for efficientantigen-receptor-gene rearrangement. Immunity2007; 27(4):561-571.

60. Ramon-Maiques S, Kuo AJ, Carney D ecal. The plant homeodomain finger of RAG2 recognizeshistone H3 methylated at both lysine-4 and arginine-2. Proc Nat! Acad Sci USA 2007;104(48):18993-18998.

61. Matthews AG. Kuo AJ, Ramon-Maiques S et al. RAG2 PHD finger couples histone H3 lysine 4trimethylation with V(D)J recombination. Nature 2007; 450(7172):1106-1110.

62. Krangel MS. T-celldevelopment: better livingthrough chromatin. Nat Immunol2007; 8(7):687-694.63. Hu H, Wang B, Borde M et al. Foxpl is an essential transcriptional regulatorof B-cell development.

Nat Immuno12006; 7(8):819-826.64. FuxaM, SkokJ, SouabniA et al. Pax5 induces V-to-DJ rearrangements and locuscontraction of the

immunoglobulin heavy-chain gene. GenesDev 2004; 18(4):411-422.65. Roldan E, Fuxa M, Chong W et al. Locus 'decontraction' and centromeric recruitment contribute to

allelic exclusion of the immunoglobulin heavy-chain gene. Nat Immunol 2005; 6(1):31-41.66. SkokJA, Gisler R. Novatchkova M et al. Reversible contraction by loopingof the Tcra and Tcrb loci

in rearranging thymocytes. Nat Immuno12007; 8(4):378-387.67. LiangHE, Hsu LY, Cado D er al. The"dispensable' portion of RAG2is necessary for efficient V-to-DJ

rearrangement during B- and T'-cell development. Immuniry 2002; 17(5):639-651.68. Akamarsu Y. Monroe R. Dudley DD et al. Deletion of the RAG2 C terminus leads to impaired

lymphoid development in mice. Proc Nat! Acad Sci USA 2003; 100(3):1209-1214.69. SteenSB. Gomelsky L, Speidel SLer al. InitiationofV(D)J recombination in vivo: roleof recombination

signal sequences in formation of single and paired double-strand breaks, EMBO Journal 1997;16(10):2656-2664.

70. Bakhshi A, Wright JJ. Graninger W et al. Mechanism of the t(14; 18) chromosomal translocation:structural analysis of both derivative 14 and 18 reciprocal partners. Proc Nat! Acad Sci USA 1987;84:2396-2400.

RecentInsights intotheFormation ofRAG-Induced Chromosomal Translocations 43

71. Posey]E, Brandt VL, Roth DB. Paradigm switching in the germinal center. Nat Immunol 2004;5(5):476-477.

72. Weterings E. Chen D]. The endless tale of nonhomologous end-joining. Cell Res 2008; 18(1):114-124.73. ShrivastavM, De Haro LP.Nickoloff]A. Regulation of DNA double-strand break repair pathway choice.

Cell Res 2008; 18(1):134-147.74. Gottlieb TM. Jackson SP. The DNA-dependent protein kinase: Requirement for DNA ends and

association with ku antigen. Cell 1993; 72:131·142.75. Walker ]R, Corpina RA. Goldberg J. Structure of the Ku heterodimer bound to DNA and its

implications for double-strand break repair. Nature 2001; 412(6847):607-614 .76. Roberts SA. Ramsden DA. Loading of the nonhomologous end joining factor, ku,on protein-occluded

DNA ends.} Bioi Chern 2007; 282(I4):10605-10613.77. Ma Y.Pannicke U. Schwarz K et aI. Hairpin opening and overhang processing by an Artemis/DNA-de

pendent protein kinase complex in nonhomologous end joining and V(D)] recombination. Cell 2002;108(6):781-794.

78. Leber R. Wise TW; Mizuta R et al. The XRCC4 gene product is a target for and interacts with theDNA-dependent protein kinase, ] Bioi Chern 1998; 273(3) :1794-1801.

79. Mahajan KN. Gangi-Peterson L. Sorscher DH er al. Association of terminal deoxynucleotidyl transferasewith ku. Proc Nat! Acad Sci USA 1999; 96(24) :13926-13931.

80. Purugganan MM. Shah S, Kearney ]F et al. Ku80 is required for addition of N nucleotides toV(D)J recombination junctions by terminal deoxynucleotidyl transferase. Nucleic Adds Res 2001;29(7):1638-1646.

81. Critchlow SE, Bowater RP,]ackson SP. Mammalian DNA double-strand break repair protein XRCC4interacts with DNA ligase IV. Current Biology 1997; 7:588-598.

82. Grawunder U, Wilm M. Wu X ec al. Activity of DNA ligase IV stimulated by complex formation withXRCC4 protein in mammalian cells. Nature 1997; 388:492-495.

83. Modesti M, Hesse]E, Gellert M. DNA binding ofXrcc4 protein is associatedwith V(D)] recombinationbut not with stimulation of DNA ligase IV activity. EMBO] 1999; 18(7):2008-2018 .

84. Dai Y, KyselaB. Hanakahi LA et al. Nonhomologous end joining and V(D)] recombination require anadditional factor. Proc Natl Acad Sci USA 2003; 100(5):2462-2467.

85. Buck D. Malivert L. de Chasseval R et aJ. Cernunnos, a novel nonhomologous end-joining factor, ismutated in human immunodeficiency with microcephaly. Cell 2006; 124(2):287-299.

86. Ahnesorg P, Smith P, Jackson SP. XLF interacts with the XRCC4-DNA ligase IV complex to promoteDNA nonhomologous end-joining. Cell 2006; 124(2):301-313.

87. Callebaut 1, Malivert L. Fischer A er al. Cernunnos interacts with the XRCC4 x DNA-ligase IVcomplex and is homologous to the yeast nonhomologous end-joining factor Nejl.] Bioi Chern 2006;281(20) :13857-13860.

88. TsaiC]. Kim SA. Chu G. Cemunnos/Xl.F promotes the ligation of mismatched and noncohesive DNAends. Proc Nat! Acad Sci USA 2007; 104(19):7851·7856.

89. Lieber MR. Lu H, Gu] et al. Flexibility in the order of action and in the enzymology of the nuclease.polymerases and ligase of vertebrate nonhomologous DNA end joining: relevance to cancer. aging andthe immune system. Cell Res 2008; 18(1):125-133.

90. ]happan C. Morse HC, Fleischmann RD et al. DNA -PKcs: a Tvcell tumour suppressor encoded at themouse scid locus. Nat Genet 1997; 17:483-486.

91. Custer RP, Bosma GC. Bosma M]. Severe combined immunodeficiency in the mouse: pathology.reconstitution. neoplasms. Am] Pathol 1985; 120:464-477.

92. Gu Y, Seidl K]. Rathbun GA er al. Growth retardation and leaky scm phenotype of Ku70-deficientmice. Immunity 1997; 7:653-665.

93. Li GC. Ouyang H, Li X et al. Ku70: a candidate tumor suppressor gene for murine T-cell lymphoma.Mol Cell 1998; 2(I):1-8.

94. Difilippanronio M]. ]Z.]T C et al. DNA repair protein Ku80 suppresseschromosomal aberrations andmalignant transformation . Nature 2000; 404:510-514.

95. Gao Y, Ferguson DO. WX er al. Interplay of p53 and DNA-repair protein XRCC4 in tumorigenesis,genomic stability and development. Nature 2000; 404:897-900.

96. Zhu C. Mills KD, Ferguson DO et al. Unrepaired DNA breaks in p53-deficient cells lead to oncogenicgene amplification subsequent to translocations, Cell 2002; 109(7):811-821.

97. Moshous D. Pannerier C, de Chasseval R et aI. Partial T and B lymphocyte immunodeficiency andpredisposition to lymphoma in patients with hypomorphic mutations in Artemis. ] Clin Invest 2003;111:381-387.

98. Ferguson DO, Ale FW. DNA double strand break repair and chromosomal translocation : lessons fromanimal models. Oncogene 2001; 20(40) :5572-5579.


99. Ferguson DO. Sckiguchi JM. Chang S er al. The nonhomologous end-join ing pathway of DNA repairis required for genomic stability and the suppression of translocations. Proc Natl Acad Sci USA 2000;97(12):6630-6633.

100. Sharpless NE. Ferguson DO. O'Hagan RC et al. Impaired nonhomologous end-joining provokes soli:tissue sarcomas harboring chromosomal translocations, amplifications and deletions. Mol Cell 2001 ;8(6):1I87-1I96.

101. Gladdy RA. Taylor MD. Will iams CJ et al. The RAG-I/2 endonuclease causes genomic insrability andcontrols CNS complications of lymphoblastic leukemia in p53/Prkdc-deficient mice. Cancer Cell 2003;3(1) :37-50 .

102. Riballo E. Critchlow SE. Teo SoH et al. Identification of a defect in DNA ligase IV in a radiosensitiveleukaemia patient. Curr Bioi 1999: 9:699-702.

103. Toita N. Hatano N. Ono S et al.Epstein-Barr virus-associated B-cell lymphoma in a patient with DNAligase IV (LIG4) syndrome. AmJ Med Genet A. 2007: 143(7):742-745.

104. Wilson JH. Berget PB. Pipas JM . Somatic cells efficiently join unrelated DNA segments end-to-end.Mol Cell Bioi 1982; 2(10):1258-1269.

105. Roth DB. Porter TN. Wilson JH. Mechanisms of nonhomologous recombination in mammalian cells.Mol Cell Bioi 1985; 5:2599-2607 .

106. Roth DB. Wilson JH. Nonhomologous recombination in mammalian cells: role for short sequencehomologies in the joining reaction. Mol Cell Bioi 1986: 6:4295-4304.

107. Alt FW; Baltimore D. Joining of immunoglobulin heavy chain gene segments : implications from achromosome with evidence of three D-JH fusions. Proc Natl Acad Sci USA 1982: 79:41I8-4122.

108. Pergola F,Zdzienicka MZ. Lieber MR. V(D)J recombination in mammalian mutants defective in DNAdouble-strand break repair, Mol Cell Bioi 1993: 13:3464-3471.

109. Taccioli GE. Rathbun G. Oltz E et al. Impairment of V(D)J recombination in double-strand breakrepair mutants. Science 1993; 260:207-210.

110. Getts RC. Stamato TD. Absence of a Ku-like DNA end binding activity in the xes double- strand DNArepair-deficient mutant. J Bioi Chern 1994: 269:15981-15984.

Ill. Taccioli GE. Cheng H -L. Varghese AJ et al. A DNA repair defect in chinese hamster ovary cells affectsV(D)J recombination similarly to the murine scid mutation. J Bioi Chern 1994: 269:7439-7442-

112. Taccioli GE. Gottlieb TM. Blunt T et al. Ku80: Product of the XRCC5 gene and its role in DNArepair and V(D)J recombinat ion. Science 1994; 265:1442-1445.

113. Kabocyanski EB. Gomelsky L. Han J-O et al.Double-strand break repair in Ku86- and XRCC4-dcficientcells. Nucleic Acids Res 1998; 26(23 ):5333-5342.

114. Baumann P. West Sc. DNA End-joining catalyzed by human cell-free extracts. Proc Natl Acad Sci USA1998; 95:14066-14070.

lIS. Verkaik NS. Esveldt-van Lange RE. van Heemst D er al. Different types ofV(D)J recombination andend-joining defects in DNA double -strand break repair mutant mammalian cells. Eur J lmmunol2oo2:32(3):701-709.

116. Lewis S. Gellert M. The mechanism of antigen receptor gene assembly. Cell 1989; 59:585-588.1I7. Zhu C. Bogue MA. Lim D-S et al. Ku86-deficient mice exhibit severe combined immunodeficiency

and defective processing ofV(D)J recombination intermediates. Cell 1996; 86:379-389.118. Huye LE. Purugganan MM. Jiang MM et al. Mutational analysis of all conserved basic amino acids in

RAG-l reveals catalytic. step arrest and join ing-deficient mutants in the V(D)J recombinase, Mol CellBioi 2002: 22(10) :3460-3473.

1I9. Yarnall Schultz H. Landree MA . Q iu JX et al. Joining-deficient RAGI mutants block V(D)Jrecombination in vivo and hairpin opening in vitro. Mol Cell 2001: 7(1):65-75.

120. Qiu JX. Kale SB. Yarnell Schultz H et al. Separation-of-function mutants reveal critical roles for RAG2in both the cleavage and joining steps ofV(D)J recombination. Mol Cell 2001; 7(1):77-87.

121. Tsai CL. Drejer AH. Schatz DG. Evidence of a critical architecrural function for the RAG proteins inend processing. protection and joining in V(D)J recombination. Genes Dev 2002: 16(15) :1934-1949.

122. Richardson C. Jasin M. Coupled homologous and nonhomologous repair of a double-strand breakpreserves genomic integrity in mammalian cells. Mol Cell Bioi 2000 ; 20(23):9068-9075.

123. Lee GS. Neiditch MB. Salus SS et al. RAG prote ins shepherd double -strand breaks to a specific pathway. suppressing error-prone repair. but RAG nicking init iates homologous recombination. Cell 2004;117(2):171-184.

124. Corneo B. Wendland RL. Deriano L et al. Rag mutations reveal robust alternative end joining. Narure2007: 449(7161):483-486.

125. Yan CT. Boboila C. Souza EK et al. IgH class switching and translocarions use a robust nonclassicalend-joining pathway. Nature 2007: 449(7161 ):478-482.

126. Soulas-Sptauel P. Le Guyader G. Rivera-Munoz P ec al. Role for DNA repair factor XRCC4 inimmunoglobulin class switch recombinat ion. J Exp Med 2007 ; 204(7):1717-1727 .

Recent Insights intotheFormation ofRAG-Induced Chromosomal Translocations 45

127. DiBiase S]. Zeng ZC. Chen Ret al. DNA.dependent protein kinase stimulates an independently active.nonhomologous. end-joining apparatus. Cancer Res 2000; 60(5) :1245-1253 .

128. Meaburn KJ. Misteli T. Soutoglou E. Spatial genome organization in the formation of chromosomaltranslocations, Sernin Cancer Bioi 2007; 17(I):80-90.

129. Soutoglou E. Dorn JF, Sengupta K et al. Positional stability ofsingle double-strand breaks in mammaliancells. Nat Cell Bioi 2007; 9(6) :675-682.

130. Lisby M. Mortensen UH. Rothstein R. Colocalization of multiple DNA double-strand breaks at a singleRad52 repair centre. Nat Cell Bioi 2003; 5(6):572 -577.

131. Lukasova E. Kozubek S. Kozubek Met al. Localisation and distance between ABL and BCR genes ininterphase nuclei of bone marrow cells of control donors and patients with chronic myeloid leukaemia.Hum Genet 1997; 100(5-6):525-535.

132. Kozubek S. Lukasova E. Ryznar L et al. Distribution of ABL and BCR genes in cell nuclei of normaland irradiated lymphocytes. Blood 1997; 89(I2):4537-4545.

133. Branco MR. Pombo A. Intermingling ofchromosome territories in interphase suggests role in translocations and transcription-dependent associations. PLoS Bioi 2006; 4(5) :d38.

134. Chen HT. Bhandoola A. Difilippanronio MJ et al. Response to RAG-mediated VOJ cleavageby NBSIand gamma-H2AX. Science 2000; 290(5498):1962-1965 .

135. Perkins EJ. Nair A. Cowley DO et al. Sensing of intermediates in V(D)J recombination by ATM. GenesDev 2002; 16(2):159-164.

136. Bender CF. Sikes ML. Sullivan Ret al. Cancer predisposition and hematopoietic failure in Rad50(S/S)mice. Genes Dev 2002; 16(17):2237-2251.

137. Celeste A. Petersen S. Romanienko PJ et al. Genomic instability in mice lacking histone H2AX. Science2002; 296(5569):922-927.

138. Liyanage M. Weaver Z. Barlow C ct al. Abnormal rearrangement within the alpha/delta T-cell receptorlocus in lymphomas from Arm-deficient mice. Blood 2000; 96(5):1940-1946.

139. Kang J. Bronson RT. Xu Y. Targeted disruption of NBSI reveals its roles in mouse development andDNA repair. EMBO J 2002; 21(6):1447-1455.

140. Difilippaneonio S. Celeste A. Fernandez-Capetillo 0 et al. Role of Nbs1 in the activation of the Atmkinase revealed in humanized mouse models. Nat Cell Bioi 2005; (7):675-685 .

141. Ward 1M. Difilippantonio S. Minn K et al. 53BPI cooperates with p53 and functions as a haploinsuf.ficienr tumor suppressor in mice. Mol Cell Bioi 2005; 25(22) :10079-10086.

142. Liao M]. Van Dyke T. Critical role for atm in suppressingV(D)J recombination-driven thymic lymphoma.Genes Dev 1999; 13(10):1246-1250.

143. Petinior LK. Weaver Z. Vacchio M et al. RAG-mediated V(D)J recombination is not essential fortumorigenesis in arm-deficient mice. Mol Cell Bioi 2002; 22(9) :3174-3177 .

144. Bredemeyer AL. Sharma GG. Huang CY et al. ATM stabilizes DNA double-strand-break complexesduring V(D)J recombinat ion. Nature 2006; 442(7101):466-470.

145. Callen E. Jankovic M. Difilippanronio S er al. ATM prevents the persistence and propagation ofchromosome breaks in lymphocytes. Cell 2007, 130(I):63-75.

146. Rosenwald A. DNA microarrays in lymphoid malignancies . Oncology (Williston Park). ec 2003 :17(I2):1743-1748 ; discussion 1750. 1755. 1758-1749 passim.

147. Matsuoka S. Ballif BA. SmogorzewskaA et al. ATM and ATR substrate analysisrevealsextensiveproteinnetworks responsive to DNA damage. Science 2007; 316(5828):1l60-1166.

CHAPTER 4

V(D)] Recombination DeficienciesJean-Pierre de Villartay*

Abstract

V D)] recombination not only comprises the molecular mechanism that insures diversityof the immune system but also constitutes a critical checkpoint in the developmentalprogram ofB- and T-Iymphocytes .The analysisofhuman patients with SevereCombined

Immune Deficiency (SCID) has contributed to the understanding of the biochemistry of theV(D)J recombination reaction. The molecular study V(D)] recombination settings in humans.mice and in cellular mutants has allowed to unravel the process ofNon Homologous EndJoining(NHE]). one ofthe key pathway that insure proper repair ofDNA double strand breaks (dsb) ,whether they occur during V(D)] recombination or secondary to other DNA injuries. TwoNHE] factors. Artemisand Cernunnos, were indeed discovered through the study of humanV(D)J recombination defective human SCID patients.

IntroductionForeign antigens are recognized by the immune systemofvertebrate through specialized recep

tors expressed on the cell surface ofT- and B-Iymphocyces; the T-cell receptors (TCR) and theB-cell receptors (BCR) or immunoglobulins respectively.

V(D)jRecombinationImmunoglobulins and TCRchains are composed oftwo domains: one constant region. which

insures effector function and one highly polymorphic antigen recognition domain. or Variabledomain. The Variable domain can befurther subdivided into three separate segments known asVariable(V) . Diversity (D) and joining(J) elements. whose respectiveencodinggenesare dispersedon the chromosome (Fig. lA). The fusion of these various elements. at the DNA level. by a sitespecific rearrangement process results in the formation ofa functional V(D)J gene unit that willencode the Variabledomain.1 The combinatorial association ofV,D and] elements thus enforcesthe required diversity ofantigenic receptors. The V(D)J recombination reaction (Fig. lB) is initiated by the lymphoid specific factors Ragl and Rag2,2.3which specifically recognize recombination signal sequences (RSS) that flank all ¥, D and] gene units and introduce a DNA-dsb at theborder ofthe RSS.4 The resulting D NA-dsb is resolved by the ubiquitous DNA repair machineryknown as nonhomologous end joining (NHE]). Asdiscussed below. the V(D)] recombinationprocess not only enforces the diversity of the immune system, it also can be considered as a critical checkpoint in the development ofB- and Tdymphocyres as a faulty V(D)] reaction leads toan arrest in the differentiation ofthese two lineages (Fig. 2) causing a Severe Combined ImmuneDeficiency (SCID) phenotype.

*Jean-Pierre de VilIartay-1NSERM U768, Hopital Necker Enfants Malades, 149 rue de Sevres,75015 Paris, France. Email: devillar@neckedr

V(D)] Recombination, edited by Pierre Ferrier. ©2009 Landes Bioscienceand Springer Sciences-Business Media.

V(D)JRecombination Deficiencies 47

A

Y(D)J recomblnlltlon

Y(D)J recombInation

B

SCIDwith Ixscce, DNA·UgIYmlcroc~Phaty-t+ Cemunnoe

oT·B·SCID--tj Ra t/Rag2

o merm g.....".,.

~I

+

ASS ASS

~··.·-·· <llIlHlr--t~-

NHEJ ---------'

Figure 1. V(D)J recombination. A) Organization of IgH locus and rearrangement process.B) The V(D)J recombination reaction and faulty steps in SODs.

Human Primary Immune DeficienciesSevere combinedimmunedeficiencies (SCIDs) compriseabout eleveninherited raredisorders

whichhaveincommonablockofT-celldifferentiation/function associatedwithadirector indirectimpairmentofB-cellimmuniry,'Asaconsequence of their moleculardefects, the clinicalpresentation of SCID patients is rather uniform and mainlycharacterized bythe earlyonset of infectionsaffecting the respiratorytractand thegut.Commonopportunisticorganisms (Pneumocystis cariniiand Aspergillus) as wellas viruses (Cytomegalovirus for example) causerecurrent infectionsandfailure to thrive. About30%ofhumanscmcases arisefromadefectinV(D)J recombination (Fig.2, Table1), leadingto an earlyarrestof both B- and T-lyrnphocyte maturation. ThisT-B-SCIDcondition can be either the resultof deleterious mutations in the Rag]and Rag2genes" affectingthe initiation of the V(D)J recombination,or impinge on the DNA repair phase of the V(D)Jrecombinationreaction.In the latter casethe immunedeficiency isaccompaniedbyan increasedcellular sensitivity to ionizing radiation (RS-SCIDs), a condition resembling the murine scidsituation. Inaddition to theserather straightforwardclinicalpresentations, several other immunedeficiencies causedbyvariable defectsin V(D)J recombinationhavebeendescribedmorerecently,which areassociated with additionaldevelopmental anomalies suchasa facial dismorphyand microcephaly.The molecular analyses of thesehumanandmousepathologies werehighlyinstrumentalin definingsomeof the actorsof the nonhomologousend-joiningpathway.

RAGl and RAG2 Deficiencies

T-B-SClDsThefirstevidence foracriticalroleofV(D)J recombinationfor theproperdevelopmentofboth

B- and T-lyrnphocytes carnefrom the analysis ofRag]and Rag2KO mice." Both mousestrains

48

Precurs~

Pro B

V(D)J recombinationdefects

V(D)jRecombination

LyNK

T y~

TCDS

TCD4

IgG. A, E

19M

Figure 2. Specific block of S- and T-cell maturation in V(D)J recombination deficiencies.

are characterized by a complete absence ofperipheral mature B- and Tdymphocytes owing toa defect in the initiation of the V(D)J recombination. Rag] and Rag2deficiency thymocytesaccumulate as quiescent cells at the CD4-CD8- double negative (DN) stage, just prior theonset ofTCR-lJ rearrangement. Similarly, B-cells development is arrested at the preB-cell stagein the bone marrow. Except for their immunological phenotype, Rag] and Rag2KO mice donot develop any other functional anomaly. A V(D)J recombination defect was subsequentlyidentified in a group ofhuman patients presentingwith the same clinical/biological condinon.v'?which turned out to be caused by mutations in either one ofthe Rag genes.I IThe Rag]andRag2deficiencies , MIM (mendellan Inheritance in Man) #179615 and #179616 respectively, areautosomal recessivediseases. Both genes are located on human chromosome 11p 13 and carriersofheterozygous mutations are healthy without any immunological disturbance. Apart from thefinding ofa complete alymphocytosis in the blood, no simple functional assaysare available toreveal a V(D)J recombination defect caused by Rag] or Rag2mutations.

RAgl/2 Structure andFunctionThe biochemistry ofthe initial steps ofthe V(D)J recombination and the precise function of

the Rag]/2 proteins are detailed in other chapters ofthis book. The identification ofa whole seriesof mutations in either genes from the molecular analysis ofhuman T-B-SCID patients over theyears was highly instrumental in drawing structure/function relationships that help defining thevarious functional domains of these two proteins. '?

Omenn SyndromeOmenn syndrome (OS) was first described in 1965 as a rare autosomal recessivedisease (MIM

#2603554) characterized by an immunodeficiency accompanied by a severeerythroderma causedby skin infiltrating activated lymphocytes (Fig. 3), eosinophilia, hepatosplenomegaly, lymphadenopathy, high levelofIgE but very low levelsofthe other Ig lsorypes," The existence ofboth T-Bscm and Omenn syndrome in the same familyl4 suggested that as could result from a V(D)Jrecombination defect caused by Rag]/2 mutations. Mutations in the Rag]and Rag2genes wereindeed identified in severalcasesofOS. 12.15.16These mutations are by essencehypomorphic as theyallow V(D)J recombination and hence the development ofB- and T-Iymphocytes, to proceed tosome extent. Consistent with the observation that as appears in the context ofa faulty, althoughnot complete failure, ofT- and sometimes B-cell development, particular mutations in Artemis(MIM #605988),17the a chain ofthe IL7 receptor (MIM #600802)18and in the MitochondrialRNA-processingendoribonuclease (RMRP; MIM #157660) gene'? were reported to cause as.Lastly, as like phenotype was also noted in DiGeorge syndrome." Conversely, leaky V(D)Jrecombination is not always associated with the development ofOSI2and recent reports identified hypomorphic Rag] and Rag2mutations in patients characterized by an elevated count ofTCR-y/b expressing T-Iymphocytes in the peripheral blood secondary to CMV infection in the


context ofalmost complete T-B-SCID.21.22Altogether,

... r-, these reports suggest that additional genetic or envi-GI ::: lI'l lI'l O r-, r-, N

ronmental factors may be required for the oligoclonal~ ~ ~~ ~ r-, eo

expansion of few T-cell clones that emerge as a resultofdrastically reduced output ofT-cells.

c0

Physiopathology..' iii

One characteristic feature ofas is the infiltration

'" of the skin and gastrointestinal mucosa with activated:sGI T-Iymphoeytes causing a skin rash resembling that oc-..ll.

curring during the graft versus host reaction (GVH)...GIv Activated T-Iymphocytes with a h ighly restrictedc

0 0 '" 0 '"" ,... ~ ~ TCR heterogeneity are also present in the blood ofU Z Z Z

as patients." These activated Tvcells, which areskewed toward the TH2 phenotype, are probably

-; respon sible for the associated eosinophilia and hyper.:0- IgE secretion." Such an autoimmune like diseaseGI Q)V :0 strongly suggests th at T- cell tolerance do es not oc-e .!l1v

0 0 o 0 0~ '" cur properly in this condition. A recent study has

~ Z Z zz Z ~ tackle the idea that AlRE deficiency in as could beat the base ofthe autoimmune manlfesranons." AlRE

~ (Autoimmune regulator element) is a transcriptional"ii activator expressed by medullary epithelial cells in theQ.: thymus. AlRE regulates the ectopic expression ofa set..~ oftissue-specificproteins,which are normally expressed0 0 0 o 0 0 '" '"..

~ ~ in the periphery, thus driving central tolerance towardsc Z Z zz Zthese proeelns," Mutations ofAlRE causeautoimmune.. polyendocrynopathy-candidiasis-ectodermaldystrophyv

~ (AP ECED). Since AlRE expression in the thymus re-GIQ quire s normal T-cell development, it wasproposed thatGI 0 00c 0 c

00e 0 0 thymic AlRE expression in as could somehow be im-= c c

E V'l Q) V'l V'l Q) 0 0 paired resulting in faulty negative selection and survival.E c!l E ~ J1 E V'l II(

~ 0 ~~ 0 ~ ::l. ofautoimmune T-Iymphocytes. Indeed two thymi from

>-as pat ients showed a strongly reduced AlRE expres-

til .. sion both at the protein and mRNA level." AlthoughQ .;;o ',0: this find ing providesan important breakthrough in the' iiiII) c understandingofasphysiopathology, severalquestionsII) GI

'"Cll: 0 remain unanswered. The recent design ofanimal models"a :s '" '" '" '" ."

c:: " 0 0~~ ~ ~ ~ mimicking some ofthe keyas featureswill clarify some

~ Z ZIIIofthese questions.Qo ,... Murine Models ofOmenn SyndromeII) u u u u U

I 1: 1: 1: 1: 1: Marrella et al. Introduced the Rag2 R229Q muta-Q:l 0. e- e- e- o...:. c 0 0 0 0 0 tion, known to be associated with as and atypical0.5 ~

E E E E E0 0 0 0 0 scmin humans, by means ofkno ck-in on a full Rag2

~.. :; 0. = 0. 0. 0. 0.= >- :J >- >- >- >- KO background." About half of the resulting mice~ Z I ZI I I I

<,;; developed alopecia , erythroderma, wasting syndrome~ and colitis by the age ofthre e months. This phenotype,

Q.I

~ ::::.resembling human as,is also accompaniedwith T-cell

"l s infiltration of the skin and the gut. These mice alsoQ() :ao~ .!!! c:experienced a severe depletion of their B-cell count in...: .... c:

GI E:~

::>~ c i ~ E spleen and a partial block of thymocyte development.c GI ...

~~

o ll<: "( a at the DN stage, prior to TCR-13 rearrangements. Like

50

A GVH like disease B

V(D)JRecombination

CD3 Teells

Figure 3. GVH like disease in Omenn syndrome. A) Skin erythroderma B) T-cell infiltrationof the skin and the intestine.

in the human as condition, the thymic expression of AlRE is severely reduced in these mice.Interestingly, the two particular sub population ofT-lymphocytes regulatory T-cells (Treg) andinvariant natural killer cells (iNKT) are also strongly reduced in Rag2R129Q mice. The absence ofiNKT cells in human as was recently demonstrated and may well participate in the physiopa thology ofas condition." During a systematic survey oftheir mouse colony for the appearanceofanomaly in the development ofT-cells, Khiong et al identified a spontaneous mutant mousewith an increased number ofso called "memory" T-cells.28 It turned out that these mice carried aRag1 R972G mutation and presentedwith many characteristic features ofhuman as in whom theequivalentRag1 R975 amino acid was found mutated. Based on the introduction ofthis mutationinto a CD4 KO background, the authors propose that the abnormal homeostasis ofCD4+ T-cellscould participate in the onset ofosmanifestations. These two as mouse models will certainlybe very helpful in the future for the better understanding ofthe as condition, in particular withregard to the possible impact ofenvironmental factors on the development of the autoimmunemanifestations.A third interesting model ofas came out from studies in the WE Paul laboratory.Milner and col. showed that reconstitution oflymphopenic (Rag2 KO) mice with suboptimalnumbers ofT-lymphocytes results in a multiorgan inflammatory disease resembling OS.29.30 Theseauthors show that reconstitution ofRag2 KO mice with a small number ofT-cells, in contrast toa large number ofT-cells, results in the onset ofas like phenotype. Indeed it is not the absolutenumber ofT-cells per se that causes this phenotype but rather the reduced TCR diversity of thetransplanted T-cells. They could further link this phenotype to the reduced heterogeneity oftheTCR repertoire expressed by the limited numbers of Treg in their inoculums. This study verynicely complements the data obtained with the two as mice and the observation gathered fromhuman as condition.

T-B-SCID with Radiosensitivity

'Ibe ScidMouse and the CHO-XRCCMutantsThe description ofthe scid mouse," a natural mutant mouse characterized by a lackofcircu

lating B- and T-Iymphocytes, as resulting from a general DNA repair defect accompanied by anincreased sensitivity to ionizing radiation or other agents causing DNA dsb, provided the initiallink between V(D)J recombination and general DNA-dsb repair.32'34The faulty V(D)J recombination in scid mice can be demonstrated in pre-T and pre-B-cells usingexcrachromosomal V(D)Jrecombination substrates" as well as on endogenous TCR loci in exvivo isolated rhymocyees."

V(D)JRecombination Deficienaes 51

The design ofV(D)J recombination substrates was at the baseof the strategy for the identification of the Rag] and Rag22J genesand are still in use in many laboratories to assess variousaspectsofV(D)J recombination. Another major breakthrough came from the very cleverideaof performing V(D)J assays in mutagenizedChinese hamster ovarycells(CHO) that had beeninitially selectedfor their DNA repairdefect revealedbyan increasedX raysensitivity(XRCCmutants). Severalofthesemutants happened to be V(D)J recombination defective.37.38 Tosummarizeyearsofintensive workin manylaboratories. theseexperimentsentitled the definition oftwo important protein complexes at playduringNHEJ; The DNA-PK complexformedbyKu70(XRCC6). Ku80 (XRCC5) and the DNA-PKcs catalyticsubunit (XRCC7. Mu-scid) on onehand and the XRCC4/DNA-LigaseIV on the other hand. The precise function and activitiesof these NHEJ factors havebeen thoroughly reviewedin recent years" Briefly. the DNA-PKcomplexidentifies the Ragl /2 generatedDNA-dsb (Fig.1B)and recruits the processingenzymeArtemis (seebelow)whilethe XRCC4/DNA-LigaseIVcomplex, together with Cernunnos (seebelow), terminates the reaction by rejoining the broken DNA ends.

ArtemisSomeB-T-SCIDpatientsdo not harbormutationineitherRag]orRag2genes, yettheypresent

the sameclinical/biological features asRag]/2defective T-B-SCIDs, i.e., a completeabsence ofcirculatingmature B- and T-Iymphoeytes. The alymphocytosis in thesepatients is accompaniedbyan increased sensitivity to ionizingradiationsofbone marrowcells (CFU-GM) and skinfibroblasts." This characteristic, also sharedby the scid mice,led to the hypothesisof a generalDNArepair defect in RS-SCID patients. The RS-SCID phenotype is also found with high incidenceamongAthabascan-speakingNativeAmericanIndians.Consistentwith their generalDNA repairdeficiency, they present a V(D)J recombination defect which can be demonstrated in vitro. infibroblasts, usingV(D)J recombinationsubstratesand ectopicexpression ofboth Rag]and Rag2genes.41.42Despite the strongsimilarityofRS-SCID patients and scid mice,DNA-PK activityisnormal in thesepatientsand the implicationof the DNA-PKcs genehasbeen ruled out bygeneticmeansin several famllies.v The disease-related locus in RS-SCID wasassigned to the short armof the chromosome10.41.44Giventhe locationofthe RS-SCID locuson human chromosome10,genomicDNA sequences covering this regionwereanalyzedin silico for the presenceof putativegenes,leadingto the identification ofanewDNA repairfactorcalledArtemis." Functionalcomplementation studiesand mutation analyses certifiedthat Artemiswasindeed the genedefective inRS-SCID. Consistent with its function during V(D)J recombinationand DNA repair,Artemisis ubiquitouslyexpressed and is localizedin the nucleus. Artemis mutations, which account forthe RS-SCID condition, areprimarilylocalizedin the N-terminushalfof the protein, thought toharbor the catalyticdomain.Thesemutationsinvolve nonsenseand misense substitutionsaswellas splicingdefectsleadingto severely truncated proteins. The inactivationof the Artemisgeneinmice recapitulates the clinical and biological features of RS-SCID patients.46•

47 HypomorphicArtemismutationshavebeen identifiedin patientspresentinga leakySCID phenotype17•48aswellas in one patient characterized by a progressive combined immune deficiency resultingfrom anelevatedlymphocyteapoptosisbut a delayed celldeath ofIR treated fibroblasts in vitro."

Artemis Structure and FunctionIn depth in silicoArtemis sequence analysis highlighted significant similaritiesof the first

150 amino acids to well-established members of the metallo-Bdactarnase superfamily.vThemetallo-Bvlactamase fold is adopted by various metallo-enzymeswith a widespread distribution and substrate specificiry." It consistsof a four-layered ~ sandwichwith two mixedg sheetsflanked by a helices. Biochemical studies demonstrated that Artemis does indeed exert anintrinsic 5 ' to 3' exonuclease activity in vitro." A similar exonuclease activity has also beenrecognizedin Apollo/SNM IB, aprotein related to Artemis that functions in the protection oftelomeres.F" When Artemis is associatedwith and phosphorylated by DNA-PKcs it switchesits catalytic activity to a DNA endonuclease capable of opening Ragll2 generated hairpinstructuresduringV(D)J recombination." Consistent with this hairpin openingactivity,Artemis


and DNA-PKcs are required for efficient adeno-associated virus (AAV) infection. the processofwhich goes through resolution ofhairpin loops at the AAV inverted terminal repeat (ITR)extremity of the viral DNA genome.t' Sequence analysis. secondary structure prediction andmutagenesis studies clearly indicated the conservation of motifs (HxHxDH) typical of themetallo-Bdactamase fold. participating in the metal binding pocket and representing the catalytic site of the metallo-p-lactamases.v'" Following the rnetallo-Bdactamase domain. Artemisshares several conserved features with other metallo-Bvlactamases acting specifically on nucleicacids and involved in DNA repair (Artemis, SNMl, PS02) and RNA processing (CPSF).This new domain was called ~_CASP.55 The ~-CASP domain, which is always appended to ametallo-f-lacramase domain, is strictly required for Artemis function.56The three-dimensionalstructure of several RNA-specific ~-CASP members has recently revealed the general organization of these proteins into two domains: a metallo-fl-lactamase domain and a ~-CASP

domain, with the active site being located at the interface between the two domains.59,60 SeveralSerine residues, mostly located in the C-terminus half of the protein, have been identified invitro and in vivo as targets of phosphatidylinositol 3- kinase like kinases (PIKK), includingDNA-PKcs.61-68 Unexpectedly however, Artemis function in V(D)J recombination does notrely upon the phosphorylation ofthese sites. Moreover. a truncated version ofArtemis lackingthe C-terminus half is still proficient in V(D)J recomblnadon.W' One current hypothesis isthat, in the absence ofDNA-PKcsArtemis would adopt a particular conformation by which itsC-terminus domain masks the ~-Lact/~-CASPcatalytic site. Artemis would then gain its fullenzymatic activity through conformational changes upon DNA-PKcs interaction.56.65 Anotherproposed function for DNA-PKcs would be to facilitate the access ofArtemis to DNA damage. DNA-PKcs is indeed required for the proper loading ofArtemis on damaged chromarln."However, although DNA-PKkinase activity prevents Artemis dissociation from the DNA-PKIDNA complex, it is the autophosphorylation of DNA-PKcs and not that ofArtemis whichis critical for the ultimate activation of Artemis endonuclease activity.68 which suggests thatconformational changes triggered by DNA-PKcs autophosphorylation expose DNA ends forfurther processing by Artemis.

Artemis and the DNA Damage ResponseRS-SCID patients and Artemis KO mice present, in addition to their V(D)J recombination

defect. a general increased cellular sensitivity to DNA damaging agents, arguing for an Artemisfunction during the repair of these damages. Indeed, the repair ofabout 10%ofDNA lesionsinflicted by ionizing radiations rely on Artemis as shown by the retention of yH2AX foci, amarker ofDNA breaks.?"on a fraction ofcells at late time points following IR. 6

2,71 Artemis wasfound to process 3'-phosphoglycolare terminally blocked DSB in vitro. DNA modificationsknown to be induced by IR or bleomycin in vivo.72 Artemis thus appears to be one constituentofthe DNA damage response (D DR). The DD R is orchestrated by a seriesofbiochemical eventsamong which protein phosphorylation by the PIKK kinases, ATM and ATR, playa centralrole." Like many DNA repair factors, Artemis is hyperphosphorylated in an ATM dependentmanner after IR.61.66.68.74 The exact role ofATM dependent phosphorylation ofArtemis duringDNA repair is not fully understood as mutations ofthe posphorylation sites do not impact onthe capacity ofthese Artemis mutants to complement the radiosensitivity ofArtemis deficientfibroblasts." In addition to the DNA repair per se, cell cycle checkpoints constitute anotherkey feature ofthe DD R. Following DNA damage, the cells arrest their cycling at the G1/S andthe GUM boundaries to allow DNA repair to proceed. In the case ofIR induced DNA damage, these cell cycle checkpoints depend on ATM. Whether Artemis participates in cell cyclecheckpoints remained a matter ofdebate. Although it is clear that Artemis deficient cells arrestnormally in Gl following IR, the maintenance and/or recovery from the GUM checkpointfollowing IR was found altered.63.66.75.76 Whether this reflects a direct function ofArtemis oncell cycle through the regulation ofCdkl-cyclin B63 or the impaired repair ofa subset ofdamageafter IR75.76 remains an open issue.

V(D)j Recombination Deficiencies 53

DNA-LigaseIVDNA-LigaseIVmutationswerefirstidentifiedin patientspresentingdevelopmentalanomalies

and immunodeficiency" In contrast to RS-SCIDs these patients are not completelydevoid ofB- and T-lyrnphocytes, although their numbers can be drasticallyreduced.Severalother reportsofDNA-LigaseIVdeficiency further demonstrated the high heterogeneityof this syndrome forits impact on immunodeficiency (from no deficiency to SCID) as well as on its developmentalconsequences (with or without microcephaly) and cancerlncidence.P'" Inthe moresevereforms,the V(D)J recombinationisstronglyaffectedboth quantitatively andqualitatively asaconsequenceof the DNA rejoiningdeficiency.Whatever the nature ofhuman DNA-LigaseIVmutations, theyall result in partial lossof function alleles.

CernunnosAnother seriesof five patients characterized by severecombined immunodeficiencyassoci

ated with growth delayand microcephalywas reported.P Ihe clinical and cellular phenotypesofthesepatients (including increasedradiosensitivity,defectiveV(D)J recombination, impairedin vitro NHEJ activity)wasstrikinglyreminiscent to that observedin DNA-LigaseIVcondition(see above). However, neither DNA-LigaseIVnor the other known NHEJ factors were foundmutated, suggesting that these patients suffered from a novel NHEJ defect." A new NHEJfactor, named Cernunnos, was indeed identified through eDNA functional complementationof patients' fibroblasts. The same NHEJ factor, named XLF (for XRCC4-like factor) , wasindependently identified through a yeast two hybrid screen usingXRCC4 as a bait.84Recentlydeveloped murine Cemunnos-deficienr ES cellspresent a phenotype similar to that of humandeficient cells (increased radiosensitivity,genomic instability, DNA repair defect), except forV(D)J recombination." Although the efficiencyofV(D)J recombination is highly compromised, the fidelity of signal joins is not altered in Cemunnos-deficiene ES cells, contrastingwith the human situation from which more than half of the signal joins are imprecise, withvarious lengths of nucleotide deletions.82.86The nature of the mutation engineered in ES cells(the deletion of Cernunnos exons4 and 5 could result in the lowlevelexpressionofa truncatedCernunnos protein created by an in-frame splicing from exon 3 to exon 6) maypartly accountfor these dlfferences."

Deleterious mutations of the Cernunnos gene were found in all patients and the ectopic expressionof a wild type Cernunnos complemented the DNA repair defect observed in patients'cells.82.84Whether these mutations lead to a complete lossof function or representhypomorphicalleles isnot yet known with certainty.Giventhe structural and functional relationshipsbetweenCernunnos/Xl.F and XRCC4 (seebelow),one wouldexpectacompletelossoffunction allele notto be compatiblewith lifeas is the casefor XRCC4 KO mice.The development of a Cernunnoscomplete lossof function mouse model willcertainlyhelp to addressthese issues.

Cernunnos StructureThe human Cemunnos gene,composedof eight exons,is located on the long arm of chromo

some2 (2q35) and isexpressed asa 2063 nucleotideslong cDNA.82.84The Cernunnos/XLF protein is299 amino acid longwith an apparent weightof about 33kDa. Cemunnos isubiquitouslyexpressed and localizedpredominantly in the nucleus.Sequenceanalysis revealedthat Cernunnosshares structural features with XRCC4 revealing the existence of a new protein family.55.84.87Basedon the XRCC4 strucmre88.89 one can predict a similar conformation for Cernunnos, i.e.,a globular head domain followedby a coil-coiledtail.84.87•90 Cernunnos/XLF,like XRCC4, canbind DNA in a sequence-independent manner.87.91Cernunnos/XLF and XRCC4 can homodimerizeor participate in the same complextogether with DNA-LigaseIV.84.87.90.92Their globularhead domains could drive their direct association. Both Cernunnos/XLF and XRCC4 appearto directly interact with DNA-LigaseIVbut the Cernunnos/XLF-DNA-LigaseIV interactionisveryweak.87.92The exactnature of the complex(es) formed between XRCC4, DNA-LigaseIVand Cernunnos/XLF remains to be clearlyestablished, but one can anticipate that differential


complex formation may have important regulatory function for the DNA-end ligation reactionduring the NHE] process.

Lastly, sequence analysisrevealed that CernunnoslXLF, although highly divergent, is the genuine orthologofNej Ip/LiD.,90 a NHE] factor described in the yeast S. Ctrevisiae:THSCernunnos orthologs (referenced as NejI p or XLFl) havefurther been found in many eukaryotes demonstratingthat Nej I p and Cernunnos/XLF belong to the same protein family.87.90.96 Nej I p in yeast interactswith the XRCC4 ortholog Liflp, suggesting that Nej Ip and Cernunnos/Xl.F have conserved ananalogous function throughout evolution.

Cernunnos FunctionLike XRCC4 and several other factors that participate in the DNA damage response

(DDR), Cernunnos/XLF and its yeast ortholog Nejlp are phosphorylated upon DNA damage.97

•98 However, the recruitment of Cernunnos to the site of DNA breaks does not require

this DNA-PK dependent phosphorylation event." Although XRCC4 and Cernunnos sharestructural characteristics and are part ofthe same complex, the over expression ofXRCC4 cannotfunctionally complement Cernunnos deficient cells." suggesting that these two factors participate to the DNA-end ligation activity in a cooperative manner. Moreover, the defects ofXRCC4or Cernunnos have different impact on the DNA-LigaseIV protein. Whereas DNA-LigaseIVprotein is destabilized in the absence ofXRCC4,99.1oo this is not the case in Cernunnos deficientcells.84.90 Although the XRCC4/DNA-LigaseIV complex exerts DNA-end ligation in vitro,'?'Cernunnos/Xl.F further potentiates this activity.87.91The presence ofCernunnos, which seemsparticularly important for the ligation of mismatched or non cohesive DNA ends but not ofcompatible DNA ends in vitro102,I03would suggest that it may potentiate the ligation activity ofthe XRCC4/DNA-LigIV complex on specific DNA end structures. Although the informationconcerning the role ofCernunnos are still scarce, the attractive hypothesis that XRCC4 stabilizesDNA-LigaseIV whilst Cernunnos switches-on the ligaseactivity ofthe XRCC4/DNA-LigaseIVcomplex can however be drawn. Hence, several corollaries follow this hypothesis: (I) Cernunnosmight be a crucial regulator of the NHE] pro cess (as is the case for its S. cerevisiae orthologNej lp, see below) and (2) The Cernunnos ability to interact with the DNA-LigIVIXRCC4complex and/or to associate with the DNA breaks and/or to potentiate the ligase activity shouldbe tightly regulated (either transcriptionally as is the case for Nejlp, or posttranscriptionallyor both). These hypotheses will be certainly tested in the next future and the structural analysisofCernunnos crystal alone or in association with XRCC4 and DNA-LigaseIV will also be ofgreat interest to unravel the specific role of Cernunnos.

V(D)] Recombination in NHE]DeficientAnimalModelsIn addition to the scid mouse, deficient animal models were developed for the various NHE]

factors. All these models have in common an impact on V(D)J recombination and consequentlyon lymphocyte developmental arrest. thus recapitulating the human RS-SCID condition." Inthe case ofXRCC4and DNA-LigaseIVKOmice, the immunological phenotype is accompaniedby embryonic lethality owing to a massiveapoptosis ofpostmitotic neurons,IM.IOS the corollary ofwhich in humans could be the microcephaly observed in DNA-LigaseIVand Cernunnos patients.Another very interesting aspect came out from the analyses of these models. When the NHE]defect is crossed onto aP53KO background, this invariably leads to the earlyonset ofvery aggressive Pro-B-celllymphomas bearing chromosomal translocations, thus demonstrating that NHE]factors are genetic caretakers.P'

AcknowledgementsI thank all present and past members ofmy team who worked on the molecular basisofhuman

SCID casesover the yearsaswell as the colleagues in our Institute. Ourwork issupported by grantsfrom Institut National de la Sante et de laRecherche Medicale (INSERM), the Agence Nationalede la Recherche (ANR), the Ligue Nationale centre le Cancer (Equipe labelliseeLA UGUE 2(05),The Commissariat al'Energie Atomique (LRC-CEA N°40V) and the INCa/Canceroplole IdE


References1. Tonegawa S. Somatic generation of antibody diversity. Nature 1983;302(5909):575-81.2. Schatz DG. Oettinger MA. Baltimore D. The V(D)] recombination activating gene. RAG-I. Cell 1989;

59(6) :1035-48.3. Oettinger MA. Schatz DG . Gorka C et at. RAG-I and RAG-2. adjacent genes that synergistically

activate V(D)] recombination. Science 1990; 248(4962) :1517-23 .4. Dudley DD. Chaudhuri ], Bassing CH et at. Mechanism and Control ofV(D)] Recombination versus

Class Switch Recombination: Similarities and Differences. Adv Immunol 2005; 8643-112 .5. Fischer. A. Human primary immunodeficiency diseases: a perspective. Nat Immunol2004; 5(1):23-30.6. Villa A. Sobacchi C. Notarangelo LD er aI. V(D)] recombination defects in lymphocytes due to RAG

mutations : severe immunodeficiency with a spectrum of clinicaI presentations. Blood 2001; 97(1):81-8.7. Shinkai Y. Rathbun G. LamKP er aI. RAG-2-deficient mice lack mature lymphocytes owing to inability

to initiate V(D)] rearrangement. Cell 1992; 68(5) :855-67.8. Mombaerts P. Iacomini j .johnson RS et aI. RAG-I-deficicnt mice have no mature B- and T-lymphocytes.

Cell 1992; 68(5) :869-77.9. Schwarz K. Hansen-Hagge TE. Knobloch C et at. Severe combined immunodeficiency (SCI D) in man:

Bvcell-negative (B-) SCID patients exhibit an irregular recombinat ion pattern at the ]H locus. ] ExpMed 1991; 174(5) :1039-48.

10. Abe T. Tsuge 1. Kamachi Y ec at. Evidence for defects in V(D)] rearrangements in patients with severecombined immunodeficiency.] Immuno11994; 152(11) :5504-13.

11. Schwarz K. Gauss GH. Ludwig L ec at. RAG mutations in human B-cell-negative SCID. Science 1996;274(5284):97-9.

12. Sobacchi C. Marrella V, Rucci F et at. RAG-dependent primary immunodeficiencies. Hum Murat 2006;27(12) :1174-84.

13. Omenn GS. Familial Reticuloendorhellosis with Eosinophilia. N Engl] Med 1965; 273:427-32.14. de Saint-Basile G. Le Deist F. de Villartay ] P et at. Restricted heterogeneity of T-lymphocytes

in combined immunodeficiency with hypereosinophilia (Omena's syndrome). ] Clin Invest 1991;87(4) :1352-9.

15. Villa A. Santagata S. Bozzi F et al, Partial V(D)] recombination activity leads to Omenn syndrome. Cell1998; 93(5) :885-96.

16. Corneo B. Moshous D. Gungor T et at. Identical mutations in RAG I or RAG2 genes leading to defective V(D)J recornbinase activity can cause either T-B-severe combined immune deficiency or Omennsyndrome. Blood 2001; 97(9) :2772-6.

17. Ege M. Ma Y. Manfras B et at. Omenn syndrome due to ARTEMIS mutations. Blood 2005;105(11):4179-86.

18. Giliani S. Bonfim C. de Saint BasileG et at. Omenn syndrome in an infant with IL7RA gene mutat ion.] Pediatr 2006; 148(2):272-4.

19. Roifman CM. Gu Y. Cohen A. Mutations in the RNA component of RNase mitochondrial RNAprocessing might cause Omenn syndrome. J Allergy Clin Immunol 2006; 117(4) :897-903.

20. Markert ML. Alexieff M]. Li ] et at. Complete DiGeorge syndrome: development of rash. lymphadenopathy and oligoclonaI T-cells in 5 cases. J Allergy Clin Immunol2004; 113(4):734-41.

21. Ehl S. Schwarz K. Enders A er at. A variant of SCID with specific immune responses and predominanceof gamma delta Tvcells, J Clin Invest 2005; 115(I1):314O-8.

22. de Villartay ]P. Lim A. AI-Mousa H et aI. A novel immunodeficiency associated with hypomorphicRAGI mutations and CMV infection.] Clin Invest 2005; 115(11):3291-9.

23. Schandene L. Ferster A. Mascart-Lemone F et at. T-helper type 2-like cells and therapeutic effects ofinterferon-gamma in combined immunodeficiency with hypereosinoph ilia (Omenns syndrome) . EurJ Immunol 1993; 23(1) :56-60.

24. Cavadini P. Vermi W; Facchetti F et aI. AIRE deficiency in thymus of2 patients with Omenn syndrome.J Clin Invest 2005; 115(3):728-32.

25. Anderson MS. Venanzi ES. Klein L et at. Projection of an immunologicaI self shadow within the thymusby the aire protein. Science 2002; 298(559 7):1395-401.

26. Marrella V. Poliani PL. Casati A ec at. A hypomorphic R229Q Rag2 mouse mutant recapitulates humanOmenn syndrome. ] Clin Invest 2007; 117(5) :1260-9.

27. Matangkasombut P, Pichavant M. Saez DE et at. Lack of iNKT cells in patients with combined immunedeficiency due to hypomorphic RAG mutat ions. Blood 2007 [epub ahead of print.]

28. Khiong K. Murakami M. Kitabayashi C et at. Homeostatically proliferating CD4 T-cells are involvedin the pathogenesis of an Omenn syndrome murine model.] Clin Invest 2007; 117(5):1270-81.

29. Milner ]D. Ward ]M. Keane-Myers A er at. Lymphopenic mice reconstituted with limited repertoireT-cells develop severe. multiorgan, Th2-associated inflammatory disease. Proc Nat! Acad Sci USA 2007;104(2) :576-81.


30. Milner ], Ward], Keane-Myers A et aI. Repertoire-dependent immunopathology. ] Autoimmun 2007;29 (4):257-61.

31. Bosma GC, Custer RP, Bosma MJ. A severecombined immunodeficiency mutation in the mouse. Nature1983; 301(5900):527-30.

32. Fulop GM, Phillips RA. The scid mutation in mice causes a general defect in DNA repair. Nature 1990;347479-482.

33. Biedermann KA, Sum ]R, Giaccia A] et al. Scid mutation in mice confers hypersensitivity to ionizing radiation and a deficiency in DNA double-strand break repair. Proc Nat! Acad Sci USA 1991;88:1394-1397.

34. Hendrickson EA, Qin XQ, Bump EA er al. A link between double-strand break related repair andV(D)J recombination: the scid mutation. Proc Nat! Acad Sci USA 1991; 88:4061-4065.

35. Lieber MR, Hesse ]E, Lewis S er al. The defect in murine severe combined immune deficiency: joiningof signal sequences but not coding segments in V(D)J recombination. Cell 1988; 55(1):7-16.

36. Roth DB, MenetskiJP, Nakajima PB et al. V(D)J recombination: broken DNA moleculeswith covalentlyscaled (hairpin) coding ends in scid mouse thymocytes. Cell 1992; 70(6):983-91.

37. Taccioli GE, Rathbun G, Oltz E et al. Impairment of V(D)] recombination in double-strand breakrepair mutants. Science 1993; 260(5105):207-10.

38. Pergola F, Zdz ienicka MZ, Lieber MR . V(D)] recombination in mammalian cell mutants defective inDNA double-strand break repair. Mol Cell Bioi 1993; 13(6):3464-71.

39. Weterings E, van Gent DC. The mechanism of nonhomologous end-joinlng: a synopsis of synapsis.DNA Repair (Arnst] 2004; 3(11):1425-35.

40. Cavazzana-Calvo M, Le Deist F, de Saint Basile G et al. Increased radiosensitivity of granulocytemacrophage colony-forming units and skin fibroblasts in human autosomal recessiveSevere CombinedImmunodeficiencj. ] Clin Invest 1993; 91:1214-1218.

41. Moshous D, Li L, de Chasseval Ret al. A new gene involved in DNA double-strand break repair andV(D)J recombination is located on human chromosome lOp. Hum Mol Genet 2000; 9(4):583-588.

42. Nicolas N, Moshous D, Papadopoulo D et al. A human scmcondition with increased sensitivity toionizing radiations and impaired V(D)] rearrangements defines a new DNA Recombination/Repairdeficiency] Exp Med 1998; 188:627-634.

43. Nicolas N, Finnic N], Cavazzana-Calvo M et al. Lack of detectable defect in DNA double-strand breakrepair and DNA-dependant protein kinase activity in radiosesitive human severe combined immunodeficiency fibroblasts. Eur] Immuno11996; 26:1118-1122.

44. Li L, Drayna D, Hu D er al. The gene for severe combined immunodeficiency disease in Athabascanspeaking Native Americans is located on chromosome lOp. Am] Hum Genet 1998; 62(I):136-44.

45. Moshous D, Callebaut I, de Chasseval Ret al. ARTEMIS, a Novel DNA Double-Strand Break Repair/V(D)] Recombination Protein, is Mutated in Human Severe Combined Immune Deficiency. Cell 2001;105:177-186.

46. Rooney S, Sekiguchi ], Zhu C er al. Leaky Scid phenotype associated with defective V(D)] coding endprocessing in Artemis-deficient mice. Mol Cell 2002; 10(6):1379-90.

47. Li L, Salido E, Zhou Y et al, Targeted disruption of the Artemis murine counterpart results in scmanddefective V(D)] recombination that is part ially corrected with bone marrow transplantation.] Immunol2005; 174(4) :2420-8.

48. Gennery AR, Hodges E, Williams AP er al. Omenns syndrome occurring in patients without mutationsin recombination activating genes. Clin Immunol 2005; 116(3) :246-56.

49. EvansPM, Woodbine L, Riballo E er aI. Radiation-induced delayed cell death in a hypomorphic Artemiscell line. Hum Mol Genet 2006; 15(8):1303 -11.

50. Aravind 1. An evolutionary classification of the merallo-bera-lactamase fold proteins. In Silico Bioi1999; 1(2):69-91.

51. Ma Y, Pannicke U, Schwarz K er al. Hairp in opening and overhang processing by an Artemis/DNA·dependent protein kinase complex in nonhomologous end joining and V(D)] recombination. Cell2002; 108(6) :781,94.

52. Lenain C, Bauwens S, Amiard S et al. The Apollo 5 ' exonuclease functions together with TRF2 [0

protect telomeres from DNA repair. Curr Bioi 2006; 16(13) :1303-10.53. van Overbeck M, de Lange T. Apollo, an Artemis-related nuclease, interacts with TRF2 and protects

human telomeres in S phase. Curr Bioi 2006; 16(13):1295-302 .54. Inagaki K, Ma C, Storm TA et aI. A Role ofDNA-PKcs and Artemis in Opening Viral DNA Hairpin

Termini in Various Tissues in Mice.] Virol 2007; 81(20) :11304-2155. Callebaur I, Moshous D, Momon ]P et al. Metallo-Bdactamase fold within nucleic acids processing

enzymes: the fl-CASP family. Nucl Acid Res 2002; 30:3592-3601.56. Poinsignon C, Moshous D, Callebaut I er al. The Metallo-fl-Lactamase/fl-CASP Domain of Artemis

Constitutes the Catalytic Core Required for V(D)] Recombination.] Exp Med 2004; 199:315-321.

V(D)j RecombinationDeficiencies 57

57. Pannicke U. Ma Y, Hopfner KP et al. Functional and biochemical dissection of the srructure-specificnuclease ARTEMIS. EMBO] 2004; 23(9):1987-97.

58. Niewolik D, Pannicke U. Lu H et al. DNA-PKcs dependence of Artemis endonucleolyric activity. differences between hairpins and 5 ' or 3 ' overhangs.] Bioi Chern 2006; 281(45) :33900-9.

59. Ishikawa H. Nakagawa N. Kuramitsu S ee al.Crystalstructure ofTTHA0252 from 'Ihermus thermo philus HB8, a RNA degradation protein of the rnetallo-bera-Iactamase superfamily.] Biochern (Tokyo)2006; 140(4):535-42.

60. Mandel CR. Kaneko S. Zhang H er al. Polyadenylation factor CPSF-73 is the pre-mRNA 3'-endprocessing endonuclease. Nature 2006; 444(7121) :953-6.

61. Poinsignon C, de Chasseval R, Soubeyrand S er al. Phosphorylation of Artemis following irradiationinduced DNA damage. Eur] Immunol 2004; 34(11) :3146-55.

62. Riballo E. Kuhne M. Rief N ee al. A pathway of double-strand break rejoining dependent upon ATM,Artemis and proteins locating to gamma-H2AX foci. Mol Cell 2004; 16(5):715-24.

63. Zhang X. Succi], Feng Z er al. Artemis is a phosphorylation target of ATM and ATR and is involvedin the GUM DNA damage checkpoint response. Mol Cell Bioi 2004; 24(20):9207-20.

64. Chen L. Morio T. Minegishi Y et al. Ataxia-telangiectas ia-mutated dependent phosphorylation ofArtemis in response to DNA damage. Cancer Sci 2005; 96(2) :134-41.

65. Ma Y. Pannicke U. Lu H er al. The DNA-dependent protein kinase catalytic subunit phosphorylationsites in human Artemis.] Bioi Chern 2005; 280(40):33839-46.

66. Wang] . Pluth ]M. Cooper PK et al. Artemis deficiencyconfers a DNA double-strand break repair defectand Artemis phosphorylation status is altered by DNA damage and cell cycle progression. DNA Repair(Arnst) 2005; 4(5) :556-70.

67. Soubeyrand S. Pope L, De ChassevalRet al. Artemis phosphorylated by DNA-dependent protein kinaseassociates preferentially with discrete regions of chromatin . ] Mol Bioi 2006; 358(5): 1200-11.

68. Goodarzi AA. Yu Y, Riballo E et al. DNA -PK autophosphorylation facilitates Artemis endonucleaseactivity. EMBO J 2006; 25(l6):3880-9.

69. Drouer ] , Frit P. Delteil C et al. Interplay between Ku, Artemis and the DNA-dependent protein kinasecatalytic subunit at DNA ends.] Bioi Chern 2006: 281(38) :27784-93.

70. Rogakou EP. Pilch DR. Orr AH et al. DNA double-stranded breaks induce histone H2AX phosphorylation on serine 139.] Bioi Chern 1998; 273(10) :5858-68.

71. Darroudi F. Wiegant W; Meijers M et al. Role of Artemis in DSB repair and guarding chromosomalstability following exposure to ionizing radiation at different stages of cell cycle. Murat Res 2007;615(1-2) :111-24.

72. Povirk LF. Zhou T. Zhou R et al. Processing of 3 ' -phosphoglycolace-rerminared DNA double strandbreaks by Artemis nuclease.] Bioi Chern 2007; 282(6) :3547-58.

73. Sancar A. Lindsey-Boltz LA. Unsal-Kacmaz K er al. Molecular mechanisms of mammalian DNA repairand the DNA damage checkpoints. Annu Rev Biochem 2004; 7339-85.

74. Geng L. Zhang X. Zheng S et al. Artemis links ATM to GUM checkpoint recovery via regulation ofCdkl -cyclin B.Mol Cell Bioi 2007; 27(7):2625-35 .

75. Krernpler A. Deckbar D. Jeggo PA ec aI. An imperfect G2M checkpoint contributes to chromosomeinstability following irradiation of Sand G2 phase cells. Cell Cycle 2007: 6(14):1682-6.

76. Deckbar D. Birraux], Krempler A et al, Chromosome breakage alter G2 checkpoint release.] Cell Bioi2007; 176(6):749-55.

77. O'Driscoll M, Cerosaletri KM. Girard PM et al, DNA ligase IV mutations ident ified in patientsexhibiting developmental delay and immunodeficiency. Mol Cell 2001; 8(6):1l75-85.

78. Ben-Omran TI. Cerosalerri K. Concannon P et al, A patient with mutations in DNA LigaseIV: clinicalfeatures and overlap with Nijmegen breakage syndrome. Arn J Med Genet A 2005: 137(3):283-7.

79. Buck D. Moshous D. de Chasseval R et al, Severe combined immunodeficiency and microcephaly insiblings with hypomorphic mutations in DNA ligase IV. Eur] Immunol2006; 36(1) :224-35.

80. Enders A. Fisch P. Schwarz K ec al. A severe form of human combined immunode6ciency due to mutations in DNA ligase IV.] Immunol 2006; 176(8) :5060-8.

81. van der Burg M. van Veelen LR. Verkaik NS et al, A new type of radiosensitive T-B-NK+ severe combined immunode6ciency caused by a LIG4 mutation. J Clin Invest 2006: 116(1):137-45.

82. Buck D. Malivert L. de Chasseval R et aJ. Cernunnos, a novel nonhomologous end-joining factor, ismutated in human immunode6ciency with microcephaly. Cell 2006; 124(2):287-99.

83. Revy P. Buck D. le Deist F et al. The repair of DNA damages/modi6cations during the maturation ofthe immune system: lessons from human primary immunode6ciency disorders and animal models. AdvImmunol 2005: 87237-95.

84. Ahnesorg P. Smith P,Jackson SP. XLF interacts with the XRCC4-DNA ligase IV complex to promoteDNA nonhomologous end-joining. Cell 2006; 124(2):301-13.


85. Zha S, Alt FW; Cheng HL er aI. Defective DNA repair and increased genomic instability inCernunnos-XLF-deficient murine ES cells. Proc Natl Acad Sci USA 2007; 104(11):4518-23.

86. Dai Y,Kysela B, Hanakahi LA er aI. Nonhomologous end joining and V(D)] recombination require anadditional factor. Proc Natl Acad Sci USA 2003 ; 100(5) :2462-7.

87. Hentges P. Ahnesorg P, Pitcher RS er aI. Evolutionary and functional conservation of the DNA nonhomologous end-joining protein, XLF/Cernunnos.] BioI Chern 2006; 281(49):37517-26.

88. Sibanda BL, Critchlow SE, Begun] et al. Crystal structure of an Xrcc4-DNA ligase IV complex. NatStruct Bioi 2001; 8(12) :1015-9.

89. ]unop MS, Modest i M, Guarne A et al,Crystal structure of the Xrcc4 DNA repair protein and implications for end joining. EMBO] 2000; 19(22) :5962-70.

90. Callebaut I, Malivert L, Fischer A er al. Cernunnos interacts with the XRCC4 x DNA-ligase IVcomplex and is homologous to the yeast nonhomologous end-joining factor Nej1.] BioI Chern 2006;281(20):13857-60.

91. Lu H. Pannickc: U.Schwarz K et aI. Length-dependent binding of human XLF to DNA and stimulationof XRCC4.DNA ligase IV activity.] Bioi Chern 2007; 282(15):11155-62.

92. Deshpande RA, Wilson TE. Modes of interaction among yeast Nejl, Lifl and Dnl4 proteins andcomparison to human XLF, XRCC4 and Lig4. DNA Repair (Arnst) 2007; 6(10):1507-16.

93. Frank-Vaillant M. Marcand S. NHE] regulation by mating type is exercised through a novel protein.Lifzp, essential to the ligase IV pathway. Genes Dev 2001 ; 15(22) :3005-12.

94. Kegel A, Sjostrand]O. Astrom SU Nejlp, a cell type-specific regulator of nonhomologous end joiningin yeast. CUrt BioI 2001 ; 11(20) :1611-7.

95. Valencia M, Benrele M. Vaze MB er aI. NE]1 controls nonhomologous end joining in Saccharomycescerevisiae. Nature 2001; 414(6864):666-9.

96. Cavero S. Chahwan C, Russell P. Xlfl is required for DNA repair by nonhomologous end joining inSchizosaccharomyces pombe . Genetics 2007; 175(2):963-7.

97. Wu PY, Frit P. Malivert L er al, Interplay between Cernunnos/Xl.F and nonhomologous end-joiningproteins at DNA ends in the cell] BioI Chern 2007; 282(44):31937-43.

98. Ahnesorg p. Jackson SP. The nonhomologous end-joining protein Nejlp is a target of the DNA damagecheckpoint. DNA Repair (Amsr) 2007 ; 6(2) :190-201.

99. Grawunder U. Zimmer D. Leiber MR. DNA ligase IV binds to XRCC4 via a motif located betweenrather than within its BRCT domains. Curt BioI 1998; 8(15) :873-6.

100. Teo SH,]ackson SP. Liflp targets the DNA ligase Lig4p to sites of DNA double-strand breaks. CurrBioI 2000 ; 10(3) :165-8.

101. Grawunder U. Wilm M. Wu X er aI. Activity of DNA ligase IV stimulated by complex formation withXRCC4 protein in mammalian cells. Nature 1997; 388(6641):492-5.

102. Tsai C]. Kim SA, Chu G. Cernunnos/XLF promotes the ligation of mismatched and noncohesive DNAends. Proc Natl Acad Sci USA 2007 ; 104(19) :7851-6.

103. Gu ], Lu H. Tsai AG er aI. Single-stranded DNA ligation and Xl.F-sdmulared incompatible DNA endligation by the XRCC4-DNA ligase IV complex : influence of terminal DNA sequence. Nucleic AcidsRes 2007 ; 35(17):5755-62.

104. Gao Y,Sun Y,Frank KM et aI. A critical role for DNA end-joining proteins in both Iymphogenesis andneurogenes is. Cell 1998; 95(7) :891-902.

105. Frank KM. Sekiguchi ]M. Seidl K] et al. Late embryonic lethality and impaired V(D)J recombinationin mice lacking DNA ligase IV. Nature 1998; 396(6707):173-7.

106. Franco S. Ale FW; Manis ]P. Pathways that suppress programmed DNA breaks from progressing tochromosomal breaks and translocations. DNA Repair (Arnst) 2006; 5(9-10) :1030-41.

107. Moshous D, Pannerier C. Chasseval Rd R et al. Partial T- and Bvlymphocyte immunodeficiency andpredisposition to lymphoma in patients with hypomorphic mutations in Artemis. J Clin Invest 2003 ;111(3):381-7.

CHAPTERS

Large-Scale Chromatin Remodelingat the Immunoglobulin HeavyChain Locus:A Paradigm for Multigene RegulationDaniel]. Bolland, Andrew L. Wood and Anne E. Corcoran"

Abstract

V <D )J recombination in lymphocytes is the cutting and pasting together of antigenreceptor genes in cis to generate the enormous variety of coding sequences requiredto produce diverse antigen receptor proteins. It is the key role of the adaptive immune

response, which must potentially combat millions of different foreign antigens. Most antigenreceptor loci have evolved to beextremely large and contain multiple individual V,D and] genes.The immunoglobulin heavy chain (Igh) and immunoglobulin kappa light chain (Igle) loci are thelargest multigene loci in the mammalian genome and V(D)] recombination is one ofthe mostcomplicated genetic processes in the nucleus . The challenge for the appropriate lymphocyteis one of macro -management-to make all of the antigen receptor genes in a particular locusavailable for recombination at the appropriate developmental time-point. Conversely, theselarge loci must be kept closed in lymphocytes in which they do not normally recombine, to

guard against genomic instability generated by the DNA double strand breaks inherent to theV(D)] recombination process . To manage all of these demanding criteria, V(D)] recombination is regulated at numerous levels. It is restricted to lymphocytes since the Rag genes whichcontrol the DNA double-strand break step of recombination are only expressed in these cells.Within the lymphocyte lineage. immunoglobulin recombination is restricted to B-Iymphocytesand TCR recombination to Tvlymphocyres by regulation oflocus accessibility, which occursat multiple levels. Accessibility of recombination signal sequences (RSSs) flanking individualV,D and] genes at the nucleosomallevel is the key micro-management mechanism, which isdiscussed in greater detail in other chapters. This chapter will explore how the antigen receptor loci are regulated as a whole, focussing on the Igh locus as a paradigm for the mechanismsinvolved. Numerous recent studies have begun to unravel the complex and complementaryprocesses involved in this large-scale locus organisation. We will examine the structure of theIgh locus and the large-scale and higher-order chromatin remodelling processes associated withV(D)J recombination, at the level ofthe locus itself, its conformational changes and its dynamiclocalisation within the nucleus .

·Corresponding Author: Anne E. Corcoran-Laboratory of Chromatin and Gene Expression,Babraham Institute, Babraham Research Campus, CambridgeCB22 3AT, UK.Email : anne.corcoranebbsrc.ac.uk

V(D)]Recombination, edited by Pierre Ferrier. ©2009 Landes Bioscienceand Springer Science+Business Media.


Introduction

B-CellDevelopmentIn order to generate the primary repertoire of immunoglobulins and T-cell receptors, an

tigen receptor loci undergo variable, diversity and joining (V(D)J) recombination in B- andT-Iymphocytes. This involves generationof DNA double strand breaksat recombinationsignalsequences (RSSs) flankingindividual genes, followed by removal of the intervening DNA andjuxtapositioningandligationof the recombininggenesegments. Thisprocess isregulatedat severallevels.First,recombinationiscatalyzed byarecombinase complex containingtheprotein productsof the recombinase activatinggenesRag] and Rag2.1 RAG expression is restricted to precursorlymphocytes, thereby restrictingV(D}Jrecombination to these cells. Second,within precursorlymphocytes, this processis strictlylineage-specific with heavy(Igh) and light (Igk and Igl') immunoglobulin locionlyfullyrecombiningin B-Iymphocytes and T-cellreceptorloci (Tera, Tcrb,Tergand Terti) onlyrecombininginT-cells.Third, within lineages, lociarerecombinedinapreciseorder. Recombinationof the 19b locus in pro-Bvcells is the earlieststep in the generationof themature antibody repertoire in B-Iymphocytes and isfollowed byIgk and then Iglrecombinationin prell-cells, Fourth, the order is alsostrictlymaintained within loci: DH-to-JH recombinationoccurson both Igh alleles beforeVwto-DJHrecombinationtakesplace? Finally, RAG activityistargeted to RSSs flankingindividualV,D andJ genes. Apart from restriction to lymphocytes byrestrictedRAGexpression, thisorderedregulationiseffectedbyseveral levels ofimmunoglobulinlocusaccessibility.

Description ofthe IghLocusThisneedfor multiplelevelsof regulation isboth necessitated andcomplicatedbytheenormous

sizeof the antigenreceptorloci.The mouseIgh and Igk loci are the largestmultigenelociknown,with sizes of3Mband 3.2Mb respectively.3.4 The19b locusof the C57BL/6 mousehasrecentlybeencompletelyassembled and annotated. It comprises 195 VHgenesspanning2.5Mb, 10 DHgenes(-60kb), 4JHgenes(2kb)and 8 constant (CH) genes(200kb) (Fig.1}.3,5The Vgenesareorganizedinto 16 families of varyingsizes, basedon sequencehomology.The majorityare functional, but alargeproportion (85)areclassed aspseudogenes,someofwhichnevertheless recombine, althoughthey do not makefunctional Igpolypeptides. All of the functional V, D and J genesare used inmultipledifferentcombinations and thislargechoiceofV,D andJ recombination partnersprovidesthe first step in immunoglobulindiversity. However there is a bias in recombination frequencybetween the 3' and 5' ends of the V region i.e., the 3' end is recombinedmore frequentlyin fetalliverand in the earliestbone marrowB-cells. The extent of the biasvaries betweenmousestrainsand recombination frequencynormalises in later B-cells.6-8 Large-scale mechanisms which maycontribute to the biaswill be discussed below.

Each V and D gene has its own promoter and all genesare transcribed in the sameorientation (Fig. 1), although this is not the casefor all antigen receptor loci. Promoters haveseveralfeatures in common, but also family-specific differences which may be a factor in observedfamily-specific differences in recombination frequency.3The human Igh locus is smaller(1Mb)and contains only 123 V genes,79 of which are pseudogenes,? The V regions of the Igh andother antigen receptor loci are believedto haveevolvedfrom much smallerVgeneclustersthatwere frequently duplicated, possiblydue to ability of the Ragenzymesto act as general transposases.P'" Consequently evenwithin species there are significantdifferences in numbers andfamilydistribution ofV genes,particularly in the mouse .P? For examplethe 7183 gene familyat the 3' end of the V region has 21 V genesin the C57BL/6 strain and 49 V genesin the 129strain. This is an extremelyimportant consideration when comparing Igh locus recombinationbetween mouse strains. In the future it is likely that studies on the C57BL/6 strain will predominate as this is the strain in which the mousegenome wassequencedand thus contains allother relevant sequenceinformation.

I::" ~ c:..,

l:: ~ 9 <! ~ g. ~ ~ '" 1:} l::':'

~ l:> ... ~ " ti' ~ ~ '" ~ !f ~ l::i ..s 9 l::i is' t'-< '" l'l ...

44-

oto

JR

ecom

bina

tion

JD

Il-

vto

OJR

ecom

bina

tion

!V

DJ

_

195

Vge

nes

120

4JB

cons

tant

gene

sI

J55!

V36

09fa

mili

esI

mid

dle

fam

ilies

fl83 ,

Q52

fam

ilier

I

~176~

09.12.17

4H-

rrf

f7

A16

5044

Imer

gen

lcd

lmn

ceIn

kb

Mr

rc. d.b.a.

Fig

ure

1.S

che

ma

ticd

raw

ing

ofth

em

ouse

C57

BL/

6Ig

hlo

cus

and

no

nco

din

gR

NA

tra

nsc

rip

tion

du

rin

gV

(O)J

reco

mb

ina

tio

n.

a)A

pp

roxi

mat

esc

ale

dra

win

gof

the

V,

0an

dJ

gene

fam

ilies

,w

ith

diff

eren

tco

lors

/sha

des

repr

esen

ting

dif

fere

nt

Vge

nefa

mili

es

.A

pp

rox

imat

edi

stan

ces

be

twe

en

gene

sar

ede

pic

ted

,a

lbe

itfo

rcl

ari

tyth

ege

nes

occ

up

ya

dis

pro

po

rtio

na

tely

larg

elin

ea

rsp

ace

.F

ore

xam

ple

the

J558

/360

9re

gio

nw

hic

ho

ccu

pie

sth

e5'

1 .5

Mb

con

tain

sw

ide

lysp

aced

j55

8ge

nes

inte

rspe

rsed

wit

h3

60

9ge

nes

inth

e5'

1Mb

(see

zoo

me

do

utp

ictu

re),

but

hig

hly

clus

tere

dj5

58

gene

sw

ith

noin

terv

en

ing

36

09

gene

sin

the

3'5

00

kb.

Tra

nscr

ipts

(sen

seab

ove

,an

tisen

seb

elo

wth

ege

nes)

are

pres

ente

dby

ho

rizo

nta

lar

row

s.b)

Ma

gn

ifie

dp

ictu

reof

the

germ

line

Igh

locu

s,sh

ow

ing

non

codi

ngR

NA

tran

scri

pts

.N

um

be

rsab

ove

the

Vge

nes

ind

ica

tege

neid

en

tity

(see

ref.

3).

c)M

ag

nif

ied

pic

ture

of

the

OJ

reco

mb

ine

dlo

cus,

asab

ove.

d)M

ag

nifi

ed

pic

ture

ofth

eV

OJ

reco

mb

ine

dlo

cus,

asab

ove.

~


Chromatin RemodelingHow is such an enormouspieceof DNA manipulated in the nucleusto ensurethat its many

genesareaccessible forV(D)Jrecombination in pro-Bvcells, but not inT-cells or laterstage B-cells?While recombinationitselfis a geneticprocessi.e.,alterationsaremade in the DNA sequenceofthe locus,it is regulatedbya multitude ofepigeneticprocesses i.e.,heritablechanges in chromatinstructure that do not involve a change to the primary sequence. It is important to bear in mindthat structurallythis extremely long DNA sequenceisnot simplya pieceof string,but occupiesa3-dimensionalspacein the nucleus.It isestimatedthat the linearlength of DNA helixcontainedin a mammaliangenomeis3 metersand this mustbe accommodatedin acellnucleuswith adiameter of5-1Oprn, Thisisachievedat the basiclevelbywrappingthe DNA helixaround the histoneocramerin the nucleosome, followed byseveral levels ofhigherorder foldingofnucleosomes overeachother, in ways that are not wellunderstood (Fig.2). To facilitategeneralgenetranscription,this higherorder chromatinmust firstbe unravelled to achieve a more open and ultimatelysinglenucleosomal structure, Thiskind ofmulti-tieredregulation alsocontrolsV(D)J recombination'?and recent studies haveexplored the extent to which these mechanisms are involved in V(D)Jrecombination.Thischapterwill exploreseveral aspects in detail-noncodingRNA transcription,nuclearlocalizationand regulatoryelements, whileplacingthesein contextwith other processesincludinghistone modification, whichwill be exploredin detail in other chaptersin thisvolume.We will explorewhat is currentlyknown, what current studiesmaypredict and what the futuredirections are likelyto be.

NoncodingRNA TranscriptionContrary to the 'centraldogma'that 'DNA makes RNAmakes protein:numerous genome-wide

transcriptional analyses have estimated that over half of all transcribed mammalian genomicsequences are nonprotein-coding" and some of this transcription is predicted to playkey rolesin gene regulation. Notably, over 20 years ago, the Igh locuswas the one of the first loci shownto express noncoding RNAs. This transcription was originallytermed 'sterile' or 'germline' todistinguishit from codingtranscription from V(D)J recombinedgenes. Inthe Igh locus,the firstgermline transcripts occur before DwtO-JH recombination and initiate from two regions; theintronic enhancerEp (I, transcript)!5 and from apromoter,PDQ52, immediatelyupstreamof themost 3' DHgene segment,DQ52 (pO rranscripe)" (Fig. 1). Following Dwto-1H recombination,the DJH gene segmentproduces D, transcriprs'? and sensegermline transcription initiates overthe VHgenes (Fig. 1).!8.19 Subsequently, noncoding RNA transcriptshavebeen identified in allantigen receptor loci across genesegments competent for recombinadon." Thediscovery of VHgenegermlinetranscriptionformed the basisofthe accessibility hypothesis, whichproposed that

intergenic transcript

Inaccessible genes

" . . '. ". ....., " I • ••• •• '

PeG

Figure 2. Model of intergenic transcription. The RNA Pol II complex, depicted as a sphere,with associated smaller spheres denoting transcription factors of the basal complex, is picturedprocessing through closed chromatin, recruiting activating chromatin remodelling factors andpromoting egressof PcG (Polycomb) proteins. HAT: histone acetyltransferase; Setll2: membersof Trithorax family of histone H3 HMTs; SWI/SNF: SWitch/sucrose nonfermentable.

Large-Scale Chromatin Remodelingat theImmunoglobulin HeavyChainLocus 63

lineage-and stage-specificity of recombinationare regulatedbydifferential chromatin accessibility ofantigen receptorgenesegments to the recombinase machinery, with gerrnlinetranscriptionassociated with open chromatin.18.21 However, a function for VHgermlinetranscription has notbeen formallydemonstrated and it has been argued that it maybe a secondaryeffectof the VHgene promoters becomingaccessible for Vwto-DJH recombination. Neither havefunctions yetbeen assignedto the pO and Iptranscripts.However, quantitative RNA-FISH visualization of I,transcription22.23haveclassed this transcript as a 'supergene'i.e.,a gene that is transcribedalmostcontinuouslyfromboth alleles in an individualnucleus," Thisproperty appliesto surprisinglyfewgenes, ~-globin amongthem. Ipisthe firstnoncoding'supergene' to be identifiedand wouldmorecorrectlybe termed a 'super-transcription unit',sinceit is a noncoding, intergenictranscript.Thepossible implicationsof this high leveltranscriptionfor the roleof noncoding RNA transcriptionin the 19h locuswill be discussed below.

Intergenic TranscriptionRecent studies suggest that intergenic transcription may playa role in opening up the 19h

locus. In the largeV region,the relatively smallV genes (SOObp) areseparatedbyenormousintergenicdistances(10-20kb).3Thechromatinremodelingprocesses previously discovered arelargelyconfined to V genes(germlinetranscription above, histone modifications, discussed in detail inother chapters).Suchfocusedalterationsareunlikelyto besufficient to open the closedchromatinconformation of the locus,the default state in nonll-cells" and additional large-scale processeswereinvestigated. In numerousloci,including~-globin and the MHC complex, intergenictranscriptiondelineates domainsof modifiedchromatin that surround activegenesand their regulatory elements.26-29 RNA polymerase II (PolII) recruitsa widerangeof chromatin remodelingandhistone-modifyingfactors, includinghistoneacetyltransferases (HATs)and Set1and Set2histonemethyltransferases (HMTs),requiredforhistonemodifications associatedwithgeneactivation.30-34Furthermore, transcription triggers histone turnoverand the depositionof varianthistone H3.3,enriched with active modifications." Collectively these activities suggest several mechanisms bywhich the processingactivityofelongatingPolII complexcanachieve chromatin accessibility.36.37Accordingly, intergenic transcription has been proposed to drive through repressive chromatinin several multigeneloci, recruitingremodelingfactorsand openingup largechromatindomainsinto a poised state, thus facilitatingfurther focused chromatin opening over genes to regulategeneexpression (Fig. 2).38 In several largedevelopmentally regulated10ci,28 this isbelieved to occur by transcription-dependenr'v" higher order chromatin remodelingand loopingout of theirchromosometerritories.t'r"

In manycases, intergenictranscriptionmayonlyneed to drivethroughonceor twiceto open upthe chromatin.However, in other instances, includingthe Drosophilahomeoticbithoraxcomplex,continuousintergenictranscriptionisrequiredto preventbindingof repressive Polycomb proteinswith H3K27 HMTactivity and to recruit activatingTrithoraxH3K4 HMTs.43 Furthermore, manyenhancers and Locus Control Regions undergo transcription, which is essential for activationof their target genesand thus transcription from an intergenic regulatory region can influenceexpression of a distal gene.r'

Intergenic Transcription in the Mouse 19h Locus V RegionAnalysis of transcription from genesand intergenic regions throughout the 19h VHregion,

usingRT-PCR to measuresteady-state levels and RNA-FISH to visualize primary transcriptsonindividualalleles in singlecellsrevealed that intergenic transcription occurs throughout the IghV region. It isabsenton germlinealleles that havenot yet recombinedDHtoJHin earlyBvcells, isexpressed on the majorityofDftlHrecombinedalleles and disappears onceV to DJ recombinationhas occurred. This tightly developmentally regulated pattern of expression is characteristic of alarge-scale functionalprocess. Furthermore,patterns of transcriptsdetected by RNA-FISH wereextendedoverlargeregions, suggesting extensive transcription on individualalleles (Fig. 3).

64 V(D)j Ruombination

Figure 3. Visualization of antisense transcription in the Igh locus by RNA-FISH. Nuclei fromex vivo wild-type bone marrow sorted for Fraction B-cells (the majority of cells are DJrecombined). I. sensetranscripts, hybridized with a single-stranded antisense probe, are detected byTexas Red (red/light grey punctate signals). J558 gene family antisense transcripts, hybridizedwith a single-stranded sense probe, are detected by fluoroisoth iocyanate (FITC) (green/darkgrey extended signals). Nuclei are counterstained w ith 4', 6-diamidino-2-phenylindole (blue).(adapted with permission from ref. 22). A color version of this figure is available at www.landesbioscience.com/curie.

Antisense TranscriptionSurprisingly inrergenictranscriptiononlyoccurredon the antisense strand, wheretranscription

alsooccurredovertheVH genes. Antisense transcriptionhasclassicallybeenassociatedwith transcription repression in imprinted loci.in which it appearsto silencegeneexpression in cisnom the alleleon which it is expressedThe beststudied example isAir (Antisense to Igf2receptor) transcription(108kb transcript), which silences expression ofthe Igf2R.with which it partially overlaps, in cis.45

Antisensetranscription has alsobeen documented in several lowereukaryoticsystems to generatedsRNA and heterochromatinformation.46.47 However, it is now thought that the majorityofmammaliantranscriptionunitsdisplayoverlappingsenseandantisense transcription.f Thishighincidenceand co-ordinate regulation ofmany sense-antisense pairs, indicatesthat antisense transcription isinvolvedin mechanisms other thanits classical association with transcriptionalrepression.4S.4~ Forexample, antisense transcription across the yeastPHOS gene promoter is required to increase therate of transcription and is believed to evict histones to enablegreateraccess ofRNAPolII to thegene." In the mammalianHOXA duster, antisense intergenictranscriptionis required to activateneighboringH OX genes. in part bydisrupting interaction with repressive PeG complexes," Theseexamples maybe the firstofmanyin which antisense transcriptionplays an activatingrole.

Antisense Transcription in the 19b Locus V RegionIn theIgh locus.the absence ofV regionantisense transcriptionon germlinealleles arguesagainst

this transcription keepingthe VH region dosed , since it would haveto be present before DH to JH

recombinationto do this. Rather,it it isconsistentwith a rolefor intergenictranscriptionin openingup the VH region and thus it doesn't appear on germline alleles, since the VH region must be keptcloseduntil DHJH recombination hastaken place. Furthermore antisensetranscription is biallelic,arguing against a monoallelic mechanism of silencingone alleleto prevent recombination. Theexpressionpattern ofantisensetranscription in the Igh locus thus arguesin favorofits havinganactivatingrather than a repressive role in V(D)J recombination. Further, this transcription is notcontrolled byVH genepromoters and thus cannot be regardedasa by-productofthe activationof

Larg~-Scal~ Chromatin Remodelingat th«Immunoglobulin H~avy Chain Locus 65

thesepromoters forV(D)J recombination.Thisisthe firstevidence in support ofa functional rolefor germlinetranscriptionin19h V(D)J recombination.Weproposedthis large-scale transcriptionremodels theVHregionto facilitate accessibility forVH-to-DJH recombination, perhapsbydirectingchromatin remodelingfactorsto direct other changes in chromatinstructure that precedeV(D)Jrecombination (Fig.4)YTheseoccur mostlyoverthe VHgenesand includelossof histone H3K9

!

!, •• t

• ••' • • II ... • • •• • " .

4

, ' . . ' . ". : .' iO . t','

»,

------.....~VDJ recombined••• 1!' . lu ~l)se

..• 0 ' ••• , - '

Figure4. Model of role of antisenseintergenic transcription in IghV(DlJrecombination. Schematicof order of events, depicting alterations in chromatin structure. Key: Multiple red/light greyboxes: V genes; yellow/light grey boxe: D genes; blue/dark grey boxes: Jgenes; E~: green oval;large rectangular box : constant region; black arrows : sense/antisense transcripts; Me in redcircle: repressive histone modifications; Ac in green circle: activating histone modifications.A color version of this figure is available online at www.landesbioscience.comlcurie.


methylation. acetylationof histones H3 and H4, markers of accessible chromatin, histone H3.3exchange and methylationofH31ysine 27 (H3_27).2s,s2.55

Antisense and Intergenic Transcription in the 19b D RegionThe discoveryofintergenicantisensetranscription over the 19h V region beforeVwto-D.JH

recombination raised the question of whether similar transcriptional processes precede otherV(D)J recombination events.Antisenseintergenic transcription alsooccursthroughout the DH(60kb) andJHregionsof the mouse19h locus in pro-B-cells poised for Dwto-JHrecombinationand isthus awidespreadprocessduringV(D)J recombination.56 It isactivatedon germlineallelesbefore Dwto-JH recombination. Notably, it initiates near to and is regulated by the intronicenhancer Ew56 E~ wasoriginallyproposed to regulateVHto DJH recombination.57.58However,recent studieshaveshownthat targeteddeletionofE~ causes adefectin Dwto-JHrecombination,suggestingthat E~ primarilyregulatesthis processand that defectsin VHto DJHrecombinationmay be secondary to this earlier defect.59.60 It is not yet understood how E, regulatesDHto JHrecombination. Transcription ofthe I~ 'supergene'initiates immediatelydownstream. DeletionofE, results in lossof both I~ sense'"and D region antisense transcription, up to SO kb away.56This suggeststhat E~ controls DwtO-JH recombination at least in part by activatinggermlineIgh transcription and that in particular, the processivityof the antisense transcription rendersthe DHandJH regionsaccessible for Dwto-JH recombination (Fig.4).

Thismodel issupported byconcomitant increases in DNase I sensitivity, histone H3 and H4acetylation,H3K4 methylation and nudeosomeremodeling enzymes overDHandJH genes in pro-Bcells.52.6l.62 Histone acetylationiswidespread throughout the DHregion,52 but ishighestovertheJHregionand the DQS2 gene.52.61which ispreferentially usedin earlyDwto-JHrecombinarion.PThis model is in agreementwith a recent suggestion that the region encompassing DQS2. thefourJ genesand E, formsa separatechromatindomain to the restof the DHregion.62Strikingly,DQS2 is the only DHgene that expresses both senseand antisense germlinetranscriptsand thistranscriptionoverlapextendsinto theJHregion.561hesedata suggest stronglythat the transcriptsdo not producedsRNAsthatlead to heterochromatin. Indeed,theyarecoordinatelyup-regulatedby Ew Additionally, there is no sensegermline transcription in the remainderof the DHregion.precludingdsRNA formation.56.64Nevertheless. a recent report of active retention of repressivehistone marksover the middle DHgeneshas led to the opposite hypothesis that antisense transcriptionmaycontribute to repression of thesegenes,byformationofdsRNAand Dicer-mediatedheterochromatinization,albeitno dsRNAwasdetected.64 Definitiveresolutionof theseopposingmodelsmustawaitclarification of thefunctionalroleofantisense transcriptionbytargetedremovalof this transcriptioninvivo. Similargenetargetingstudieshaveshownthat intergenictranscriptionis functionally required for V(D)J recombinationat the Tcra locus, but in this caseit originatesfrom the sensestrand.651his suggests that the strand origin isnot important. whichsupports themodel that the processing activityis the keyfunction of this transcription.

DHantisense transcriptsinitiate on germlinealleles and VHtranscriptson DJ recombinedallelesand DHand VHantisense transcriptsare rarelyassociated on individualalleles.561hus thereis a stepwise progression of antisense intergenic transcription, in a strikingly similarpattern tothe stepwise progression of activehistone modifications during19h V(D)J recombination.Theseoccur first over the D.JH region.then sequentially over the 3' end. the middle regionand the S'end of the VHregion.53.66-68 Thusantisense intergenictranscription mayfacilitatethe exchange ofrepressive histone marksassociated with thelocusin nonB-cells with active histonemarks,perhapsby histone exchange in favorof active histories e.g.• H3.3 (Fig. 4).25 Notably in the Tcra locus,intergenic transcription hasbeen shown to increase activehistone marksovergenes,"

Subnuclear RelocalisationIn addition to these localized and large-scale epigenetic changes over the 19h locus, the

location of the locus in the nucleus has an enormous impact on its recombination potential.In nonB-lymphoid cells, the 19h and 19k loci are maintained at the nuclearperiphery,generally

Large-Scale Chromatin Remodelingat theImmunoglobulin HeavyChainLocus 67

regardedasarepressive chromatinenvironment,although it isnot clearwhether the 19h isspecificallyassociated with repressive chromatin at this location.69The DRfHdistalJ558 VHgenesareoriented towards the nuclearenvelopeand the locus is effectively 'tethered' at the periphery viatheJ558 genes,while the DRfH region is oriented towards the centre of the nucleus, which maycontribute to DRfH occurringbefore VHto DRfH recombination." In earlyB-cells undergoingV(D)J recombination, both 19h and 19k alleles are repositioned to the euchromatic interior ofthe nucleus,a regionpermissive for transcription.69The relocationisdependent on interleukin-7receptor signalling,but isindependent ofRAG69 or Pax57! expression.Thisnuclearrepositioningappearsto besufficient for DRfHrecombinationand VHto DRfHrecombinationofDH-proximalVHgenesin the 19h locus.

3-Dimensional Alterations in Chromatin StructureHowever, to achieve recombination of middle and DH-distal VHgenes, central nuclear re

positioning is not sufficient, presumably due to the enormous size of the locus.An additionalprocess, termed locuscontraction, is required. Thisjuxtaposes the distalVHgeneswith the DRfHrecombinedgenesegmentin pro-B-cells and is mediated by higher-order chromatin loopingofindividualIgH subdomains.Y" It is regulatedby the transcription factor Pax5(Fig. 5).71 Pax5isthe pivotal transcription factor that regulates establishment and maintenanceofB-lymphocyteidentity and its absence preventsrecombinationofmiddle and D-distal genes." Looping is alsoregulated by the multifunctional transcription factor,YYI, which binds EW

75 It is unclear how

Figure 5. Nuclear organisation of the Igh locus. The sequential stagesof IghV(D)J recombinationare represented in the context of the spatial location of the Igh loci in the nucleus and theirlarge-scale conformation changes. The locus is initially tethered at the nuclear periphery viathe 5' end of the V region. Key: Multiple red/light grey lines: V genes; yellow/light grey box:D region; blue/dark grey box : J region; E~: green oval; large rectangular box : constant region;short (blue) squiggles: sense transcripts; long (purple) squiggles: antisense transcripts . A colorversion of this figure is available online at www.landesbioscience.com/curie.

68 V(D)] R~combination

either PaxS orITI enableDNA looping.Neither isrequiredforgermlinetranscriptionor histoneacetylationofthesegenes," suggesting that theseprocesses areeither necessary but not sufficient.or independent oflooping. One possibilityis that ITI mayrecruit other parts ofthe locusto theenhancer and future studieson the role ofE, in loopingwill be informative. ITI binds Ezh2,apolycombgroup protein with H3K27 HMTase activity, although this binding has not yet beenshown in B-cells. Ezh2 is required also for recombinationof distal V genes," Its mechanismofaction is currentlyunclear. but intriguinglyit alsoappearsto be requiredfor DNA loopingof theIgh VHregion (ATarahkovsky. personalcommunication).

Transcription FactoriesA largebody of recentevidencehas shown that transcriptiondoesnot occur homogeneously

throughout the nucleus. but appearsto be concentrated in sub-nuclearfociof active RNA PolIIcomplexes. termed transcription factories,?6.77 Individual transcription factories are believedtocontain up to ten RNA PolII complexes and to transcribeseveral genesslmuleaneously/f Thesegenescanbe up to 40MBapart on the samechromosomeand evenon separatechromosomes.P"Theseare dynamicinteractions that reflect the frequencyof transcription of individualgenes.71l

Most genesare not transcribedcontinuously. but rather switchedon and offstochastically.79TheI, 'supergene' is transcribed almost all of the time in both proB and mature B-cells and is thusalmost continuouslyassociated with a transcription factory,22.23 It has recentlybeen shown thatenhancerscan relocate genes awayfrom the nuclear periphery by recruiting them to a transcription factory," In a similarmanner.F,. maypromote nuclearrelocationbyrecruitingthe D]region to a transcription factory in the nuclear interior. where E, facilitated transcription maythen keep the D0H region in the transcription factory. providinga relatively stablefocalpointfor DNA looping.

Biased Recombination Frequency Explainedby Numerous Mechanisms

Thestudiesaboveprovideseverallarge-scale contributingreasons forpreferential recombinationof3' VHgenesin earlyB-cells. First. the 19b is tethered at the nuclearperipheryin nonB-cells viathe]SS8 genesat the S'end. thus the 3' genesareoriented towardsand relocatedinto the centraleuchromatin first. Following relocation,it appearsthat proximalVHgenes are less dependent onDNA loopingof the VHregionfor recombination.presumablydue to their proximityto the D0Hregion.7l.72 Furthermore. all the factorsthat regulatelooping(PaxS. ITI. Ezh2)areonlyrequiredfor recombinationofdistalVHgenes.Theinterleukin7 receptorisalso requiredfor recombinationofS' genes.but not 3'VHgenesin the bone marrow," Sinceit activates germlinetranscriptionoverS'VHgenes. but not 3' VHgenes. it wasproposedthat it increasedVregionchromatinaccessibilityto the recombinase."Subsequentstudieshaveidentifiedother contributorymechanisms regulatedby the IL7R. It is requiredfor relocationfrom the nuclearperiphery" and histone acetylationofS'VHgenes.66•67

Allelic Choice and Allelic ExclusionUltimately the goal of the B-lymphocyteis to express a VHD0H recombined19b gene from

only one allele at the cell surface. Surface expression of the immunoglobulin polypeptide isbelieved to lead to a feedback signaling cascadethat silences the second allele.a mechanismtermed allelic exclusion.lThis ensures that each lymphocyte produces monoclonal antibodies that recognize a single antigen with high specificity. Severalprocesses contribute to thismonoallelicexpression. In the 19b locus,VHto D0H recombination is asynchronous-Le.•onealleleundergoes recombination first. This reducesthe danger ofsimultaneouslyproducing twoproductive recombination events.However. unlike the 19k (seebelow), it isunclear how this allelicchoiceisachievedin the 19b locus.Relocationand antisenseintergenictranscription appearto be biallelic. It iscurrentlyunclearwhether locuscontraction ismono or biallellicand furtherstudies are required to revealwhether it playsa role in allelicchoice,?2.73 However. it isclear that

Large-Scal«Chromatin R~mod~lingat tbe Immunoglobulin H~avy Chain Locus 69

the opening mechanisms required for V(D)] recombination are mirrored by a set ofopposingprocesses designed to stop further V(D)J recombination once a productive recombinationevent has yielded a protein product. Histone acetylation is reduced over VH genes,54.81 sense andantisense germline transcription is lost ," locus de-contraction occurs ." These processes occuron both alleles. An additional mechanism occurs specifically on the second allele that haseitheryielded a nonproductive VHDwH rearrangement or has not yet managed to rearrange the VHgene (DHlH rearranged allele). In either case, the allele is believed to be recruited to repressivepericentromeric heterochromatin, which may preclude further V to D] recornbinarion.P" It isrecruited via the 5' end of the V region and silencing of the locus is not complete. The 1",22 D]rearranged" and sense germline transcripts from 3' V genes" continues to be transcribed. Thisis presumably because D to] recombination has already occurred on both alleles and thus onlythe V region needs to be prevented from further recombination.

Other Antigen Receptor LociWe have focused on the 19h locus, which has proved to be a useful paradigm for other antigen

receptor loci, since,with some exceptions, processesdiscovered in the 19b locus, also occur in otherantigen receptor loci. For example, noncoding sense RNA transcription over V genes has beenobserved in most other antigen receptor loci.20 Similarly relocation from the nuclear peripheryandlocus contraction by DNAlooping has been reported in the1gk, Tcra and Tcrb loci.69•72.85 However,it isnot yet known how widespread the process ofantisense and/or intergenic transcription is.Thebiggest difference between recombination ofIgb and 19k is the order and narure ofthe events thatensure monoallelic expression. In contrast to the 19h, in which this appears to be controlled afterV(D)J recombination, the 19k loci undergo several monoallelic processes before Vto] recombination, which render one allele preferentially more available for the initial recombination event.One allele is preferentially DNAdemerhylated'" and acquires active histone marks before V to]recombination, while the second allele remains DNAmethylated and is recruited to heterochromatin before V to] recombinarion."

Future DirectionsFurther studies are required to unequivocally determine the function ofantisense intergenic

transcription in the 19h locus in vivo. Furthermore is it the processivity ofthe transcription thatis important, its strand-specificity,or indeed the transcripts themselves? These are also importantconsiderations for other antigen receptor loci.

There is alsolittle known about other chromatin remodelingprocessesin 19b intergenic regions.It is unclear whether non coding RNA transcription is regulated by the same histone modificationsas coding transcription. Since there are now more than 150 known histone modificarions.f it willbe important to explore the possibility that recombination may have a unique histone code whichdoes not correspond to the code for transcription.

There is as yet no regulatory element defined for the 19h V region. However, a novel pro-B-cellspecific HS site has recently been identified 5' ofthe V region." It will be interesting to see ifthiselement regulates V to D] recombination, albeit initial characterization indicates a repressiverole.How might this or another regulatory element function? It might activate V region antisensetranscription or enable DNAloopingby interactingwith elements close to the D] region. Further,the large sizeofthe V region and the differences in recombination timing and dependence on theIL7R , PaxS, Ezh2 and ITl in different domains, suggest that there may be boundary elementsseparating different regions. Furthermore there is 90kb ofuncharacterized sequence between thelast VH and first DH gene and it will be interesting to see ifit contains any enhancers, or insulatorelements to prevent the V region recombining before the D] region.

AcknowledgementsThe authors thank members of the Corcoran group for helpful discussions. This work was

supported by the Biotechnology and Biological Sciences Research Council.


References1. Hesslein DG, Schatz DG. Factors and forces controlling V(D)J recombination. Adv Immunol 2001;

78:169-232.2. Corcoran AE. Immunoglobulin locus silencing and allelic exclusion. Sernin Immunol 2005;

17(2):141-154.3. Johnston CM, Wood AL, Bolland OJ et aI. Complete sequence assembly and characterization of the

C57BLl6 mouseIgheavy chain V region. J Immunol 2006; 176(7):4221-4234.4. Brekke KM, Garrard wr. Assembly and analysis of the mouse immunoglobulin kappa gene sequence.

Immunogenetics 2004; 56(7):490-505.5. Ye J. The immunoglobulin IGHD gene locus in C57BLl6 mice. Immunogenetics 2004; 56(6):399-404.6. Atkinson MJ, Michnick DA, PaigeC] et aI. Iggenerearrangements on individual alleles of Abelson murine

leukemia cell lines from (C57BL/6 x BALB/c) Fl fetal livers. J Immunoll991; 146(8):2805-2812.7. Jeong HD, Komisar JL, KraigE er aI. Strain-dependentexpression ofVH genefamilies.J Immuno11988;

140(7):2436-2441.8. MalynnBA, Yancopoulos GO, BarthJE et aL Biased expression of JH-proximal VH genesoccursin the

newlygeneratedrepertoire of neonataland adult mice. J Exp Med 1990; 171(3):843-859.9. MatsudaF, IshiiK, Bourvagnet P eeal. The complete nucleotidesequence of the human immunoglobulin

heavy chainvariable regionlocus. J Exp Med 1998; 188(11):2151-2162.10. Hiom K, MelekM, Gellert M. DNA transposition by the RAGI and RAG2 proteins: a possible source

of oncogenic translocarions, Cell 1998; 94(4):463-470.11.JonesJM, GellertM. Thetamingof a transposon: V(D)Jrecombination and the immunesystem. Immunol

Rev 2004; 200:233-248.12. Retter I, Chevillard C, Scharfe M et al. Sequence and Characterization of the IgHeavyChain Constant

and PartialVariable Region of the MouseStrain 129S1. J Immunol2oo7; 179(4):2419-2427.13. McBlane F, Boyes J. Stimulation of V(D)J recombination by histone acetylation. Curr Bioi 2000;

10(8):483-486.14. MattickJS. The functional genomics of noncoding RNA. Science 2005; 309(5740):1527-1528.15. Lennon GG, Perry RP. C mu-containing transcripts initiate heterogeneously within the IgH enhancer

region and contain a novelS'snontranslatable exon. Narure 1985;318(6045):475-478.16. Thompson A, Timmers E, Schuurman RK et al. Immunoglobulin heavy chain germ-line JH -C mu

transcription in human precursor B.lymphocytes initiates in a unique region upstreamof DQ52. Eur JImmuno11995; 25(1):257-261.

17. Reth MG, AleFW:Novel immunoglobulin heavy chainsareproduced&om DJH genesegment rearrangements in lymphoid cells. Nature 1984; 312(5993):418-423.

18. Yancopoulos GO, Alt FW:Developmentally controlledand tissue-specific expression of unrearranged VHgenesegments. Cell 1985; 40(2):271-281.

19. Corcoran AE, Riddell A, Krooshoop 0 et al. Impaired immunoglobulin gene rearrangement in micelacking the IL-7 receptor. Nature 1998; 391(6670):904-907.

20. Krangel MS. Gene segment selection in V(D)J recombination: accessibility and beyond. Nat Immunol2003; 4(7):624-630.

21. Stanhope-Baker p, Hudson KM, Shaffer AL et al, Cell type-specific chromatinstructure determines thetargetingofV(D)J recombinase activity in vitro. Cell 1996; 85(6):887-897.

22. BollandOJ, Wood AL, Johnston CM et al. Antisense intergenic transcription in V(D)J recombination.Nat Immunol2004; 5(6):630.637.

23. Osborne CS, Chakalova L, MitchellJA et al. Mycdynamically and preferentially relocates to a transcription factoty occupiedby lgh, PLoS Bioi2007; 5(8):e192.

24. Fraser P. Transcriptional control thrown for a loop. CUrt Opin Genet Dev 2006; 16(5):490-495.25. Johnson K, POugh DL, Yu0 et al. B-cell-specific lossof histone3 lysine 9 methylation in the V(H) locus

depends on Pax5. Nat Immunol2004; 5(8):853-861.26. GribnauJ, Diderich K, Pruzina S et al. Intergenic transcription and developmental remodeling of chromarin

subdomains in the human beta-globin locus. Mol Cell 2000; 5(2):377-386.27. DrewellRA, BaeE, BurrJ et al. Transcription defines the embryonic domainsof cis-regulatory activity at

the Drosophilabithoraxcomplex. Proc Nad AcadSci USA 2002; 99(26):16853-16858.28. MasternakK, Peyraud N, Krawczyk M et al. Chromatin remodeling and extragenic transcription at the

MHC class II locuscontrol region. Nat Immunol2oo3; 4(2):132-137.29. Bernstein BE, KamalM, Lindblad-Toh K et al. Genomicmapsand comparative analysis of histone modi

fications in human and mouse. Cell 2005; 120(2):169-181.30. Wilson C], Chao OM, Imbalzano AN er al. RNA polymerase II holoenzyme containsSWI/SNF regula.

tors involved in chromatinremodeling. Cell 1996; 84(2):235·244.31. Cho H, OrphanidesG, SunX et al.A humanRNA polymerase II complex containingfactors that modify

chromatinstructure. Mol Cell Bioi 1998; 18(9):5355·5363.

Large-Scale Chromatin Remodelingat theImmunoglobulin HeavyChain Locus 71

32. WittschiebenBO, Otero G, de Bizemont T et aI. A novelhistone acetyltransfcrase is an integralsubunitof elongating RNA polymerase II holoenzyme. Mol Cell 1999;4(l):123-128.

33. KroganN], Kim M, TongA et al. Methylation of histoneH3 bySet2 in Saccharomyces cerevisiae is linkedto rranscriptional elongation by RNA polymerase II. Mol Cell Bioi2003; 23(12):4207-4218.

34. Ng HH, Robert F, Young RA et aLTargetedrecruitmentof Set! histone methylase by elongating Pol IIprovides a localized mark and memoryof recent transcriptional activity. Mol Cell 2003; 11(3):709-719.

35. Schwartz BE,AhmadK. Transcriptional activation triggers deposition and removal of the histonevariantH3.3. GenesDev 2005; 19(7):804-814.

36. Mito Y,Henikoff]G, HenikoffS.Genome-scaleprofiling of histoneH3.3 replacement patterns.Nat Genet2005.

37. Orphanides G, Rcinberg D. RNA polymerase II elongation through chromatin . Nature 2000;407(6803):471-475.

38. Chakalova L, Debrand E, Mitchell]A et al. Replication and rranscription: shaping the landscape of thegenome. Nat RevGenet 2005; 6(9):669-677.

39. Mahy NL, PerryPE, Bickmore WA. Genedensityand transcription influence the localization of chromatinoutsideof chromosome territories detectableby FISH.] Cell Bioi2002; 159(5):753-763.

40. MullerWG, Walker D, HagerGL et aI. Large-scale chromatin dccondcnsation and recondensation regulatedby transcription from a natural promoter.] Cell Bioi2001; 154(1):33-48.

41. Volpi EY, Chevret E, Jones T et aI. Large-scale chromatin organization of the major histocompatibilitycomplex and other regions of humanchromosome 6 and its response to interferonin interphase nuclei. ]Cell Sci 2000; 113 ( Pt 9):1565-1576.

42. Chambeyron S, Da Silva NR, Lawson KA et al, Nuclear re-organisation of the Hoxb complex duringmouseembryonic development. Development 2005; 132(9):2215-2223.

43. Schmitt S, Prestel M, ParoR. Ineergenic transcriptionthrough a polycomb group response elementcounteracts silencing.GenesDev 2005; 19(6):697-708.

44. Ho Y, ElefantF,Liebhaber SAer al. Locusconrrol regiontranscription plays an active role in long-rangegene activation. Mol Cell 2006; 23(3):365-375.

45. Sleutels F, Zwart R, Barlow DP. The noncodingAir RNA is required for silencing autosomal imprintedgenes. Nature 2002; 415(6873):810-813.

46. Grewal SI, ]ia S. Heterochromatin revisited. Nat Rev Genet 2007; 8(1):35-46.47. SlotkinRK, Martienssen R. Transposable elements and the epigenetic regulation of the genome. Nat Rev

Genet 2007; 8(4):272-285.48. Katayama S,Tomaru Y,Kasukawa T et aI. Antisense transcription in the mammalian transcriptome, Science

2005; 309(5740):1564-1566.49. Rinn]L, Kertesz M, Wang]K et al. Functional Demarcation of Active and SilentChromatin Domains in

Human HOX Loci by NoncodingRNAs. Cell 2007; 129(7):1311-1323.50. Uhler ]P, Hertel C, Svejstrup ]Q, A role for noncoding transcription in activation of the yeast PH05

gene. Proc Nat! AcadSci USA 2007;104(19) :8011-8016.51. Sessa L, Breiling A, Lavorgna G et al. NoncodingRNA synthesis and lossof Polycomb group repression

accompanies the colinear activation of the human HOXA cluster. RNA 2007; 13(2):223-239.52. Chowdhury D, Sen R. Stepwise activation of the immunoglobulin mu heavychain gene locus. EMBO ]

2001; 20(22):6394-6403.53. Maes], O'Neill LP, Cavelier P et aLChromatin remodeling at the Ig loci prior to V(D)J recombination.

J Immunol2001; 167(2):866-874.54. Johnson K. Angelin-Duclos C, ParkS et aI. Changes in histone acetylation areassociated with differences

in accessibility of V(H) genesegments to V-D] recombination during B-cellontogenyand development.Mol Cell Bioi 2003; 23(7):2438-2450.

55. SuIH, Basavaraj A, Krurchinsky AN et aI. Ezh2controlsB-cc11 development throughhistoneH3 methylation and Igh rearrangement. Nat Immunol2oo3; 4(2):124-131.

56. BollandD], Wood AL, AfsharRet al. Antisense intergenic transcription precedes Jgh Dvro-] recombination and is controlled by the intronic enhancerEmu. Mol Cell Bioi2007; 27(l5):5523-5533.

57. SakaiE, BottaroA, Davidson L et al. Recombination and transcription of the endogenous Ig heavy chainlocus is effected by the lg heavy chain inrronic enhancercore region in the absence of the matrixattachment regions. Proc Nat! AcadSci USA 1999; 96(4):1526-1531.

58. Serwe M, Sablitzky F.V(D)] recombination in B-cells is impairedbut not blocked by targeteddeletionof the immunoglobulin heavy chain intron enhancer. EMBO] 1993; 12(6):2321-2327.

59. PerlorT, Alr FW;Bassing CH et al. Elucidation ofIgH intronicenhancerfunctions viagerm-line deletion.Proc Nat!AcadSci USA 2005; 102(40):14362-14367.

60. AfsharR, Pierce S, BollandD] et al. Regulation of IgH Gene Assembly: Roleof the Intronic Enhancerand 5'DQ52 Regionin Targeting DH]H Recombination.] Immunol2006; 176(4):2439-2447.


61. Morshead KB.Ciccone DN. Taverna SD et aI. Antigen receptor locipoisedfor V(D)J rearrangement arebroadly associated with BRG1 and Aanked by peaks of histone H3 dimethylated at lysine 4. Proc NatlAcad Sci USA2003: 100(20):11577-11582.

62. Maes ], Chappaz S. Cavelier P et aI. Activation ofV(D)] Recombination at the IgHChain]H Locus Occurswithina 6-Kilobase ChromatinDomainand IsAssociated with Nucleosomal Remodelmg.] hnmunol2006; 176(9):5409-5417.

63. Bangs LA. SanzIE. Teale ]M. Comparison of D.]H and junctionaldiversity in the fetal. adult and agedBvcell reperroires.] hnmuno11991: 146(6):1996-2004.

64. ChakrabortyT. Chowdhury D. Keyes A er al. Repeat organization and epigenetic regulation of theDH-Cmu domain of the immunoglobulin heavy-chain genelocus. Mol Cell 2007;27(5):842-850.

65. Abarrategui I. Krangel MS. Regulation ofT-cell receptor-alpha generecombination by transcription. NatImmunoI2006: 7(IO):1109-11I5.

66. Chowdhury D. Sen R. Stepwise activation of the immunoglobulin mu heavy chaingenelocus. EMBO]2001: 20(22):6394-6403.

67. Hesslein DG. PAugh DL. Chowdhury D et aI. paxs isrequired for recombination of transcribed. acerylared,5' IgH V genesegments. Genes Dev 2003: 17(1):37-42.

68. Maes ]. Chappaz S. Cavelier P ee aI. Activation of V(D)] recombination at the IgH chain]H locusoccurs within a 6-kilobase chromatin domainand is associated with nucleosomal remodeling.] Immunol2006; 176(9):5409-5417.

69. Kosak ST. Skok]A. MedinaKL et aI. Subnuclear compartmentalization of immunoglobulin loci duringlymphocyte development. Science 2002: 296(5565):158-162.

70. Yang Q, Riblee R. Schildkraut CL. Sites that direct nuclear compartmentalization are near the 5' end ofthe mouse immunoglobulin heavy-chain locus. Mol Cell Bioi2005; 25(14):6021-6030.

71. Fuxa M. Skok], Souabni A et al. Pax5 induces V-to-D] rearrangements and locus contraction of the immunoglobulin heavy-chain gene. Genes Dev 2004; 18(4):411-422.

72. Roldan E. Fuxa M. ChongW et aI. Locus 'decontracrion' and centromeric recruitment contribute to allelicexclusion of the immunoglobulin heavy-chain gene. Nat Immunol 2005;6(1):31-41.

73. Sayegh C. ]hunjhunwala S. Riblet R et al. Visualization of looping involving the immunoglobulinheavy-chainlocus in developing Bscells. Genes Dev 2005; 19(3):322-327.

74. Cobaleda C. Schebesta A. DeloguA er aI. Pax5: the guardian of Bvcell identityandfunction. Nat hnmunol2007: 8(5):463-470.

75. Liu H. Schmidt-Supprian M. Shi Y et aI. Yin Yang 1 is a critical regulator ofB-cell development. GenesDev 2007:21(10):1179-1189.

76. Ibarra F],Pombo A.Jackson DA eeal.Active RNApolymerases arelocalizedwithindiscrete transcription'factories' in humannuclei.] Cell Sci 1996: 109 (Pt 6):1427-1436.

77. Faro-Trindade I. Cook PRo Transcription factories: structures conserved duringdifferentiation and evolution. Biochem SocTrans 2006: 34(Pt 6):1133-1137.

78. Osborne CS. Chakalova L. Brown KE ecaI. Active genes dynamically colocalize to sharedsites of ongoingtranscription. Nat Genet 2004:36(10):1065-1071.

79. TrimbomT. Gribnau]. Grosveld F et aI. Mechanisms of developmental control of transcription in themurinealpha- and beta-globin loci, Genes Dev 1999: 13(1):112-124.

80. Ragoczy T. Bender MA. Telling A et aI. Thelocus controlregion is required for association of the murinebeta-globin locus with engaged transcription factories during erythroid marurarion. Genes Dev 2006;20(11):1447-1457.

81. Chowdhury D. Sen R. Transient IL-7/IL-7Rsignaling provides a mechanism for feedback inhibition ofimmunoglobulin heavy chaingenerearrangements. Immunity 2003: 18(2):229-241.

82. Skok]A. Brown KE. Azuara V et al. Nonequivalenr nuclear location of immunoglobulin alleles inB-lymphocytes. Nat ImmunoI2001: 2(9):848-854.

83. Daly]. Licence S. NanouA et aI. Transcription of productive and nonproductive VOJ-recombined allelesafterIgH allelic exclusion. EMBO] 2007:26(19):4273-4282.

84. HainesBB. Brodeur PH. Accessibility changes across the mouse Igh-Vlocus duringB-ceU development.Eur] Immuno11998; 28(12):4228-4235.

85. Skok]A. Gisler R. Novatchkova M et al. Reversible contraction by loopingof the Tcra and Tcrbloci inrearranging thymocytes, Nat ImmunoI2007; 8(4):378-387.

86. Mosroslavsky R. SinghN. Kirillov A ecaI. Kappachain monoallelic demethylarion and the establishmentof allelic exclusion. Genes Dev 1998: 12(12):1801-1811.

87. Goldmit M. ]i Y, Skok] er aI. Epigenetic ontogeny of the Igk locus duringBvcell development. Nat ImmunoI2005; 6(2):198-203.

88. Kouzarides T. Chromatinmodifications and their function. Cell 2007; 128(4):693-705.89. Pawlirzky I. Angeles Cv. Siegel AM et al. Identification of a candidate regulatory dement within the

5' Aanking region of the mouse Igh locus defined by pro-Bvcell-specific hypersensitivity associated withbindingof PUI. paxs and E2A.] Immunol 2006: 176(11):6839-6851.

CHAPTER 6

Genetic and Epigenetic ControlofVGene Rearrangement FrequencyAnn J. Feeney*

Abstract

T heantibodyrepertoireisenormousand reflects the powerofcombinatorialand junctionaldiversity to generateavastrepertoirewithamoderatemunberofV,D andJgenesegments.However, although there are manyVH and VIC genesegments, the usageof thesegenesis

highlyunequal.In this chapter,wesummarize our studiesducidating manyofthe factorsthat contribute to thisunequalrearrangement frequency ofindividualgenesegments. Firstly, thereismuchnatural variation in the sequence of the RecombinationSignalSequences (RSS) that flank eachrecombininggene.Thisgeneticvariationcontributesgreatly to unequalrecombination frequencies.However, other factors also playa major role in recombination frequencies, as evidencedby thefact that somegenes with identicalRSSrearrangeat verydifferentfrequencies in vivo.Analysis ofthese genesegments by chromatin imrnunoprecipitation (ChIP) suggests that differences in thestructure of the chromatin associated with each gene is alsoa major factor in differential accessibilityfor rearrangement. Finally, transcriptionfactorscandirect accessibility for recombination,possiblybyrecruitingchromatin-modifyingenzymes to thevicinityof the genesegment. Together,these factorsdictate the compositionofthe newlyformed antibody repertoire.

IntroductionThevastantibodyrepertoireiscreatedbyacombinationofjunctionaldiversity and combinato

rial diversity. Eachantibody heavychain is encoded bya heavychain and a light chain, the latterbeingencodedbyeitherthe kappalocusor the lambdalocus.Theheavy chainvariable regionisitselfcomposedof three segments, V,D and], while the light chainvariable region has two segments,V andJ. Combinatorialdiversityisgeneratedthrough the useofone eachof the manyV,D andJgenesegmentsto encode the heavyand light chain exonsand junctionaldiversityis generatedbythe ddetion of a variable smallnumber ofnucleotidesfrom the ends of each recombininggenesegment and the random addition of a few nucleotides to the junction by 'IdT,' The BALBlcIgH locus contains -50-100 functional VH genes, 13 functional DH genesand 4 funcrionalj.,genes.2-4 In the mouse,the random association of one VH' one DH and oneJH would theoreticallycreate-75 x 13 x 4 differentH chains and the random association of light chain genesegmentswould similarlycreate -50-100 VIC x 4 JIC kappa light chains and 4 different lambda chains.Further random association of heavyand light chainswould thus createovera million differentantibodieson the basisofcombinatorialdiversityalone.Thistheoreticaldiversityofcombinationsof genesegments hasbeen thought to be a major factor in the sizeof the repertoireand giventhelargenumber of genesegmentsin the Ig loci,combinatorialdiversitydoes contribute greatlytothe antibody repertoire. However, we and others haveshown that the rearrangementfrequency

*Ann J. Feeney-The Scripps Research Institute, Departmentof Immunology, IMM22, 10S50N.TorreyPines Rd, LaJolla,CA 92037, USA. Email: [email protected]

V(D)JRecombination, edited by Pierre Ferrier. ©2oo9 Landes Bioscienceand SpringerScience+Business Media.


ofthe differentgenesegments isveryunequaland thus the contribution ofsomeVgenesegmentsto the repertoireis muchsmallerthat that of other V genes,"! Someofthe geneticand epigeneticreasonsfor this difference in geneusagearesummarizedhere.

Sequence Variation in RSS Can Greatly Affect RecombinationEach genesegmentis flankedby a recombinationsignalsequence(RSS)which is composed

of a conservedheptamer and nonamer,separatedby a spacerof -12 or -23 bp.9.10 The heptamerand nonamer have consensus sequences, but there is great natural variation in the sequencesfound in the 19and TCRloci.ThepioneeringworkofGellertand colleagues usingplasmid-basedrecombination substrates containing two RSSwhich can be varied in sequenceclearly showedthat the sequenceof the heptamer and nonamer ofthe RSSwereveryimportant in determiningthe frequencyof recombinadon," Byvaryinga nucleotideat each position of the heptamer andnonamer and varyingthe spacerlength,generalruleswereestablished whichshowedthat the firstthree bp of the heptamer werecriticaland changes in those positions away from the consensusalmost abolished recombination. In contrast, variation in other positions showeda wide rangeofdecreasedrecombination. Thesestudieswerecomplementedby the RSSdatabaseanalysis byRamsden and Wu of all published Ig and TCR RSSas of 1994,11 They demonstrated that thefirst three basepairs of the RSS, CAC, wereessentially invariant, whereas other positionsoftheheptamer and nonamer had morevariability.

In order to assess whether the natural variationin RSScould be responsible for the unequalrearrangementfrequency,wefirstidentifiedthe frequencywith whichspecific Vgenesrearrangedin vivobeforeanybiological selection couldoccur.Weanalyzed rearrangementofmurineVH genesin J.lMT mice, in which the mutation in the transmembraneexonof the heavychain preventeddifferentiationpast the pro-B-cell stageandwealsoanalyzedrearrangement OfVIC genesin humancord blood cells,7·8.12-15In both cases we identifiedwhich genesrearrangedmore ofien than others in vivo. Then, usinga modificationof the recombinationsubstrateapproach,wedeterminedif the RSS could be responsible for this nonrandom rearrangement. We designed"competitionrecombinationsubstrates" in which,for example, two VIC genes competedfor rearrangement to a]IC gene,as shown in Figure 1,13 In this way, smalldifferences in recombination could beassayedby determining the relative frequencywith which the jx gene rearrangedto each of the two VICgenes. Eachof the RSSin our plasmidsweremadeby PCRso that they included-SO-loo bp offlankingDNA on either sideof the RSS.

Theanalysis of two VK alleles providesa cleardemonstration of the abilityof a singlebasepairin the RSSto significantly affectrecombinationfrequency.TheVKA2 geneisusedin the majorityofanti-HaemophilusinfluenzaeTypeb (Hib) antibodies." Navajos and genetically relatedNative

CATPtac Vr:. ext V.,. Int termination signal

5'~ .........J1C1

c:::J1- 1-- 3' Comp2O

~P4

Urvecomblned

V.,.lntemaJ • J1l:1

V.,. external· J1l:1

1580bp

630bp

370bp

Figure 1. Competition recombination substrate. The top panel shows the basic design of theplasmid-based recombination substrate and the bottom panel shows the PCR assay used todetermine the relative rearrangement of the JKgene to the internal or external VK. This basicdesign was used for all of our studies on the efficiency of various RSS. This figure is reproducedwith permission from the Journal of Experimental Medicine, 1998, 187:1495-1503 . Copyright1998, The Rockefeller University Press.

Genetic and EpigeneticControlofV GeneRearrangement Frequency 75

Americans haveahigh incidenceofHib disease17•18 and wediscovered that theyhada uniqueallele

of the VKA.2 gene,with one changefrom the predominant VICA2a alleleat the 6th position in thehepramer," Peripheral bloodDNA fromVKA.2a!b heterozygotes showedthat theVKA2a allelewasrearranged-S times more often than this new NavajoVKA.2b allele," Byplacingthe two VKA.2alleles in competitionforaJIC geneRSSina recombinationsubstrate, wewereableto showthat thissinglebasepair changein the RSSwasresponsible for the difference in rearrangement frequency,"In this particular case, we hypothesizedthat this single nucleotidepolymorphismwas likelytoplayan important role in the increased incidenceofHib disease in Navajos, sinceimpairedrearrangementof this VIC genewoulddecrease the frequencyofprotectiveanti-Hib antibodies.ISThiswould be one of the raresituationsin which there wasagenetic"hole"in the antibody repertoirewith severe biological outcome: susceptibUity to potentiallyfatal Hib disease.

RSSIs Not Always Responsiblefor UnequalRearrangementWefound other examples wherethe rearrangement frequencyin vivowasalsorecapitulatedin

the recombinationsubstrate,demonstratingthat the geneticbasisfor rearrangementdifferenceswasdue to changes in the sequence of the RSS. Forexample, the smallVHS 107 famUy has3 functional VHgenesthat rearrange at verydifferentfrequencies in ViVO.12 In pro-B-cells, the VI generearranges S times more often than VII and 40 times more often than VB. Usingcompetitionrecombination substrates,wedemonstratedthat theVI genehas an RSSthat supports3 timesmorerearrangementthan theVII RSS,thus accountingin largemeasure for the difference in rearrangement frequencyin vivo." However, VII and V13 haveverydifferentrearrangementfrequenciesin vivo, yet their RSSare identical. Recombinationsubstrateassays with -100 bp fragments ofVII and V13showedthat the S'and 3' DNA flanking the RSSalsodid not affectrecombinationfrequency," Hence, factorsother than the RSScontrol the rearrangement frequency of thesetwoV genes, aswill be discussed later in this chapter.

In another example of geneswith identicalRSSrearrangingat differentfrequencies, we analyzedthe 20-memberVH7183 gene famUy. This is the most proximalVHfamUy, along with theVHQS2 famUy that is interspersedwith it in the 2S0 kb at the 3' end of the VHlocus.The most3' functional VHgene in this family, 8IX, has been shown by several groups to rearrangeat anextremely high frequency,19.2Q but the frequencyof rearrangementof the other membersof thefamUy had not beendetermined.Weanalyzedthe rearrangement frequencyofthe entireVH7183family in pro-B-cells and showedthat the genes rearrangedwith a widerangeoffrequencies,"Wecloned and sequencedeach of the genesin the famUy and the RSSfell into two major groups.One group, which we termed Group I , had an RSSthat wascloserto the consensus than GroupII and in competition recombinationsubstrates, we showed that the Group I RSSsupported ahigher frequencyof rearrangement than the Group II RSS, aswould be predlcred."However, therearrangementfrequencyofVHgeneswith identicalRSSwasquite differentin vivoin manycasesand the Group I genesdid not rearrangeat a higher frequencythan the Group II genes. Thus,factors other than the RSSweremore important than the differences in the efficiency of the RSSincontrollinggenerearrangement frequencyfor this VHgenefamily. Wemappedallof the VHgenesin the famUy and found a much higher correlationbetween chromosomallocation and V generearrangementfrequency" The genes closest to 81X at the 3' portion of the locusrearrangedmorethan the VHgenesin the middle of the locusand the genesin the S' third of the locusrearrangedvery poorly, with the exceptionof the last VHgene in the family, 61-1P. We propose that thechromatin structure may be different at these different portions of the VH7183 part of the VHlocus,resultingin the observeddifferentrearrangement frequencies for genes with identicalRSSscatteredthroughout this 2S0kb region.

Chromatinas the Gatekeeper ofAccessibilityThe processof gene rearrangement is lineage-specific, in that TCR genesdo not rearrange

in B-cells and Ig genesdo not rearrange in T-cells. other than some DwJH rearrangements."Furthermore,this processofV(D)J rearrangement ishighlyordered: DHtoJH, followed byVHto


DJH. followed by kappa rearrangement and lastlylambda rearrangement. The sameorder isobservedinTvcells, withTC~ rearrangement occurringbeforeTCRn.Over2decades ago.Altandcolleagues proposedthe "accessibilityhypothesis" to explaintheseobservations.22Thishypothesisstated that accessibility to recombinationwould be limited to only certainsmallportions oftheIgor TCR lociin anygiven lymphocyteprecursorpopulation.e.g.•the DH andJH sublociin earlypro-B-cells. This hypothesis was supported by the observation that germline transcription ofunrearrangedgenes precedes generearrangement. thussuggesting that this transcriptionreflectedthe inducedaccessibility forRAGbindingand rearrangement.23Themechanismbywhichregionsweremaintainedin inaccessible statusuntil thepropertimefortheir rearrangementwasnot dear atthat time. but it isnow generally agreedthat chromatinstructure is likely to be the key factor.24

The tails of histone proteins protrude from the core nucleosome and they can be posttranslationally modified by acetylation. methylation. phosphorylation and ubiquitinylation.2S.26 Ingeneral. lysines on tailsof histones H3 and H4 are acetylated on active genes. Methylations aremore complexand methylationof specific Iysines, suchas lysine 9 (K9me)or lysine 27 (K27me)on H3. are associated with repressed genes in general. while methylationoflysine 4 (K4me)onH3 isassociated with active genes. It hasbeenshownthat V.D andJgenes that arerearranging aremorehighlyassociated with acetylated H3 and H4 and less frequently associated with repressivemodifications such as H3K9me. than genes that are not rearranging at that particular stage inlymphocytedevelopment.F''" Thus.the statusofhistone posttranslationalmodifications (PTM)maycontrol the accessibility of V, D andJ genes.

Sincehistoneacetylation appears to affect the accessibility ofv' D andJgenes. wehypothesizedthat perhapsthe V genes that did not rearrange aswellwereassociated with histonesthat did nothaveashigh an extentof this positive PTM and werehigher in negative PTM suchasH3K9me.Wethereforeanalyzed the VHSI07 genefamily bychromatin immunoprecipitation (ChIP) withantibodiesagainstacerylaeed H3 and H4. Wewereparticularlyinterestedin determiningiftherewereanydifferences betweenVII and VB. sincethey had identicalRSSyet rearrangedat suchdifferent frequencies. Indeed.therewasanexcellent correlation betweenthe relative rearrangementfrequencyof the three VHS107 genesand their enrichment in acetylared H3 and H4 (Fig. 2).31Furthermore. there wasan inverse relationship between the level of the repressive modification

Rearrangement Frequency AcH3

1. 1.

O.

0.

O.

O.

O.

Vi Vii Vi3

AcH41.0

o.a

0.6

0.4

0.2

0.0

Vi Vii Vi3 Vi Vii Vi3

Figure 2. Relative rearrangement frequency in vivo of the three functional VHS107 genescorrelates positively with the extent of histone acetylation and negatively with the extent ofhistone K9 methylation .

Genetic and Epigenetic ControlofV Gene Rearrangement Frequency 77

H3K9me2 and the rearrangement frequency{Fig. 2).32 Thus,the histone PTM statusaccuratelyreflects the relative accessibility for rearrangement of thesegenes.

We further investigated the histone PTM status of the 81X gene.This gene rearranges at avery high frequencyin fetal life and also rearranges at a high frequency in adult bone marrow,although not as frequentlyas in fetal liver. We compared the histone acetylationstatus of thisgene as compared to the rest of the VH7183 family, usingan 81X-specific primer and a primerthat amplifies allVH7183 genes except81X. 81X wasmorehighlyenrichedin acerylated histonesthan the remainderof the VH7183 familyand the extent ofenrichment wasgreater in fctallifethan in adult life, correlatingwith the relatively higherrearrangement of81X in fetallife{Fig. 3),31As with the VHS 107 genes, there wasa reciprocal relationshipof thesegeneswith the repressiveH3K9me2 PTMY

Further evidence that histone PTM mayinfluencerearrangement frequencyisdemonstratedbyanalysis of micedeficientin the histonemethyltransferase Ezh2,whichaddsthe repressive K27methylationPTM. The pro-B-cells from these micedo not rearrange the VHgenesin the distalhalfof the locus,although theproximalhalfrearranges at nearnormalfrequency,"Wehaveshownthat the H3K27me PTM is found on the proximalVHgenesin pro-Bvcells (C.-R. Xu and A]F,unpublisheddata) and thus weproposethat the presenceof this repressive PTM on the proximalVHgenesisnecessary for the distalVHgenesto rearrange at normal frequency.

Role ofTranscription Factors in Controlling RearrangementAlthough there clearly seems to be a good correlationbetween histone PTM patterns and

accessibility for recombination, it isnot clearwhat determinesthe histone modificationstatusofgenes. Histone acetylases, deacetylases and methylases areoften recruitedinto largemulti-proteincomplexes and it is likely that the specificity of these complexes derives from DNA-bindingtranscription factors. We have investigated the role of transcription factors in inducing accessibilityfor recombination.Micethat aredeficientin EBF, PaxS, or E2A, arealldevoid in B-cells,demonstratingthe essential roleof thesetranscriptionfactorsin B-cell differentiation.34-37Thefirsttwo factorsare B-cellspecific and areessential for B-celldevelopment. E2A isa widelyexpressedtranscription factor,but it is only in B-cells that it is present as a homodimer and this probablyexplainsthe specific lossofB-celis in the E2A-deficientmice,"

Usinganovelsystem, devised byour collaborator Dr ComelisMurreinwhichexpressionvectorsfor E2Aor EBFweretransiently transferred, alongwith expression vectors for RAG1 and RAG2,into a nonlymphoidcellline,the abilityof transcription factors to induceaccessibility of genes forrecombination was revealed." Transient transfection withEBFresulted inthe inductionofrearrangement ofVA3genes, but not of anykappagenes. Conversely, the ectopicexpression ofE2A resultedin recombination of many VId genes in this cell line. Importantly, although the three major VI(

AcH3 H3K9me2

Figure 3. The frequently rearranging 81X gene is more highly enriched for histone acetylationand less enriched for H3K9 methylation than the remainder of the VH7183 family.


families areinterspersed,onlytheVxl genes areinducedto rearrange, but not the neighboringVdIorVdII genes {Fig.4).40Thus, this meansthat theVI(locusisnot madeaccessible asawholeunit bytheactionofE2A,but thatindividualV genes, or relatedV genes suchasmembers ofaVI(family.areinducedon alocalized level to become accessible.Sincemembers ofaVI(or VH family arosebygeneduplication, their codingand flanking sequences areverysimilar.Therefore,weproposedthat therearetranscription factor bindingsitesin thevicinityofallfunctionalV genes and thatthe bindingoftheappropriate transcriptionfactorcouldthenrecruitchromatinmodifyingenzymessuchashistoneacetyltransferases or deacerylases, histone methyltransferases or demethylases, or ATP-dependentchromatinremodeling complexes,whichwouldthen changethe chromatinstructureof theVgene.makingit accessible, or inaccessible, for recombination."

Wehypothesized that the expression ofE2A wouldincrease the histoneacetylation of the Vdgenes, but not of theVdI andVdII genes whichwerenot inducedto undergorecombination atterectopic E2Aexpression. Similarly, wehypothesizedthat expression ofEBF wouldincrease the histone acetylation ofVA3genes specifically. Weassessed this byChIP, usingprimersthat flanked theRSSandwefound that this wasindeedthe case (P GoebelandAlF, unpublisheddata).Surprisingly,however, we found that the extent of acetylation of the appropriate genes wasverymodest. We

A

§bj I:=~500

~

300

200

100

0

BVKl VKlI VKlU VKlV

1000

900 lJRRlE 12

800 lii2RR1E47.RRlEBF

700

800

500

400

300

200

100

0

vxr \?JI ViJII

Figure 4. Ectopic expression of E2A in a non lymphoid cell li ne induces preferential rearrangement of VICI genes (A), while ectopic expression of EBF induces preferential rearrangement ofV>..31 (B). This f igure is reproduced w ith permission from the/oumal ofExperimen tal Medicine,2001 , 194:645-656. Copyright 2001, The Rockefeller Un iversity Press.

Genetic andEpigenetic Control ofV Gene Rearrangement Frequency 79

therefore proposedthat onlya small fraction of the VIC genes areinducedto becomeacerylared byE2Aand that thesegenes arepreferentially usedfor recombination. Similarly, ourdatashowing thatwithin the VHS I07 family, the VI geneismosthighly enrichedin acetylated histonesand VI3 theleast, most likdy reflects the factthat moreVI genes within the populationof pro-Bvcells that weinvestigated wereassociated withacerylated histones.F'Ihis maysuggest that the limitingfactorforrecombination isthe numberofVgenes that areacerylared at anygiven time.

PaxS alsohasan important rolein V(D)] recombination, in addition to its rolein controllingtheexpression ofhundredsofgenes criticalforB-cell function." Micedeficient in PaxS areblockedat the latepro-Bvcell stageof differentiation.42 Although the proximalVH7183 family rearrangesat almost normal frequency, the distal VH1SS8 genes seldom rearrange and the VHfamilies inbetween these two families rearrange at intermediate levels."An explanation for this could bethat the VHlocus in PaxS-deficient pro-Bvcells fails to undergo the compaction that appears tobe criticalto bring the distalVHgenes closerto the D-]H 10cus.44 In this extendedconfiguration.the distalVH1SS8 genes, whicharelocated 1-2.SMb from the D-] region,would be too far awayfrom the D]Hgenes to undergorearrangement. In addition, it hasbeen shown that VHgenes inPaxS-deficient B-cells areenrichedin the repressive modification H3K9me2 and it hasbeensuggested that PaxS is requiredfor the histone exchange necessary to makethe VHgenes associatewith acetylated histone H3 and not K9 methylatedH3.4s

We havedescribed another function for PaxS which is important for V(D)J recombination.Although transcriptionfactor bindingsitesare traditionallyfound in promoters and enhancers,wesearched for PaxS bindingsiteswithin VHcodingregionssincewehypothesized that the RAGcomplex maybind PaxS.The reasonfor this hypothesis wasthe fact that the coreRAG2knock-inmice had a defect in V to D] rearrangement, although D] and kappa rearrangement was notimpaired." SincePax'i-deficient micewereoriginally reported to havethe samegeneralized defectin VHrearrangement. but not D] rearrangement.f we hypothesized that perhaps the non-coreregionofRAG2 might bind to PaxS whichwouldstabilize its interactionwith the RSS. A searchofVHgenesequences with thesequences of the PaxS bindingsitesinKI,KII, RAG,CD19,showedseveral potential matches (AF],unpublisheddata) and EMSAanalysis showedPaxS did bind toseveral of thesesites.with varying affinities.32.47TheVHS I07 geneVI had the highestaffinitysiteand the VH7183 genes had strongPaxS bindingsitesalso. TheVH1SS8genes alsohad PaxS binding sites.although their affinity estimatedby cold target competition waslower. ChIP analysisshowedthat PaxS wasbound to VHgenes in pro-B-cells.47 Our collaborators. Zhixin ZhangandMaxCooper, showedthat PaxS interactedwith RAG complex, although they showedthat PaxSalso bound to complexes made with the core RAG1I2,47 thus renderingour initial hypothesisthat PaxS maybind to the non-coreportion ofRAG2 unlikely. Usingan in vitro assay, our collaboratorsfound that PaxS increased recombination. suggesting that the interactionofPaxSwiththe RAG complex did stabilize the interaction." In addition to this role of PaxS in interactingwith the RAG complex,weproposethat thesePaxS siteslocatedthroughout the IgHVlocus maybe the reason that PaxS-deficient micecannot undergolocuscontraction and thus the functionof these PaxS sitesin VHgenes maybe to initiate IgHV locuscontraction. Micedeficient in thetranscriptionfactorITI alsohaveadefectin rearrangingdistalVHgenes andalsodo not undergolocuscompaction" and thus a complex containing PaxS and ITl complex maybe involved inthe contractionof the locus.

ConclusionThe antibody repertoire derives part of its sizefrom the combinatorialdiversity generated

when differentV,D and] genes areused to encodethe two chainsof the receptorheterodimers,However, all V,D and] genes are usedat verydifferentfrequencies. We havesummarized workshowingthat part of this unequal representation is due to the natural variationin the sequencesof the RSS flanking eachgene. Sincethe RSSis the DNA bindingsitefor the RAG recombinase,the mechanism for the influence of thesegeneticvariations is clear. However. the differences inthechromatinstructureofnucleosomes associated with individualV genes canoverride thesimple


direct effectof the geneticvariationin RSS efficiency in recruitingand stabilizingRAG binding.Wefound that the extentofpositive or negative histone PTM can affectthe abilityofindividualV genesto undergorearrangement. One of the important unanswered questionsis to determinewhatdirectsthe histonemodifications to occuron specific genes withintheVloci.Wehypothesizethat specific transcriptionfactorsbind to sitesnear theV genes, in the promoter or evenin codingregionsaswehaveshownfor PaxS.Theseproteinsmaythen recruit histone-modifying enzymes,chromatin remodeling complexes and DNA methyltransferases. These epigenetic modificationswould then rendera genemoreor less accessible or inaccessible to undergorearrangement.

AcknowledgementsThiswassupported byNIH grants RO1AI29672and AIS2313.I gratefully acknowledge the

manycontributionsof present and former lab membersto this research and in particular wouldlike to acknowledge theworkofDrsCheng-RanXu,BertrandNadel,PeterGoebel,CeliaEspinozaand G Stuart Williams.

References1. Tonegawa S. Somatic generation of antibody diversity. Nature 1983; 14:575-581.2. Retter I. Chevillard C, Scharfe M et al. Sequence and characterization of the Ig heavy chain constant

and partial variable region of the mouse strain 129S1.J Immuno12007; 179(4) :2419-2427.3. Johnston CM. Wood AL. Bolland D] er al. Complete sequence assembly and characterization of the

C57BLl6 mouse Ig heavy chain V region.] Immuno12006; 176(7) :4221-4234 .4. Feeney A]. Ribler R. Dm : a new and probably the last. functional DH gene in the BALB/c mouse.

Immunogenetics . 1993; 37:217-221.5. Brezinschek HP. Brezinschek RI. Lipsky PE. Analysis of the heavy chain repertoire of human peripheral

B-cells using single-cell polymerase chain reaction. J Immuno11995; 155:190-202.6. Foster SF, Brezinschek H-P. Brezinschek RI ec al. Molecular mechanisms and selective Influences that

shape the kappa gene repertoire ofIgM+ cells.] Clin Invest 1997; 99:1614-1627.7. Feeney A]. Lugo G. Escuro G. Human cord blood I( reperroire.] Immuno11997; 158:3761-3768 .8. WUliams GW; Martinez A. Montalbano A et al. Unequal VH gene rearrangement frequency within the

large VH7183 gene family is not due to RSS variation and mapping of the genes shows a bias of rearrangement based on chromosomal location. ] Immuno12001; 167:257-263.

9. Hesse ]E. Lieber MR. Mizuuchi K et al. V(D)] recombination: A fimctional definition of the joiningSignals. Genes Dev 1989; 3:1053-1061.

10. Akira S. Okazak i K. Sakano H. Two pairs of recombination signals are sufficient [0 cause immuno globulin V-(D) -] joining. Science 1987; 238 :1134-1138 .

11. Ramsden DA, Baetz K. Wu GE. Conservation of sequence in recombination signal sequence spacers.Nucl Acids Res 1994; 22:1785-1796.

12. Love VA. Lugo G. Merz D ec al. Individual promoters vary in strength. but the frequency of reartangement of those VH genes does not correlate with promoter strength nor enhancer independence. MolecularImmunol 2000; 37:29-39.

13. Nadel B. Tang A. Lugo G et al. Decreased frequency of rearrangement due to the synergistic effect ofnucleotide changes in the heptamer and nonamer of the recombination signal sequence of the V kappagene A2b. which is associated with increased susceptibility of Navajos to Haernophilus influenzae Typeb disease.] Immunol1998; 161(11):6068-6073.

14. Montalbano A. Ogwaro KM. Tang A ec al. V(D)] recombination frequencies can be profoundly affectedby changes in the spacer sequence.] Immuno12003; 171(10):5296-5304.

15. Feeney A]. Atkinson M], Cowan M] er al. A defective VKA2allele in Navajos which may playa role inincreased susceptibility to Haemophilus inAuenzae Type b disease.] Clin Invest 1996; 97:2277-2282 .

16. Scott MG. Crimmins DL. McCourt DW er al. Clonal characterization of the human IgG antibodyrepertoire to Haernoph ilus inAuenzae Type b polysaccharide III. A single Vldl gene and one of several]1( genes are joined by an invariant arginine to form the most common L chain V region. ] Immunol1989; 143:4110-4116.

17. Coulehan jl., Michaels RH. Hallowell C et al. Epidemiology of Haemophilus inAuenzaeType b diseaseamong Navajo indians. Pub Hlth Rep 1984; 99:404.

18. Petersen GM. Silimperi DR. Rotter ]1 er al. Genetic factors in Haemophilus inAuenzae Type b diseasesusceptibility and antibody acquisition.] Pediatrics 1987; 110:228-233.

19. Perlmutter RM. Kearney]F, Chang SP et al. Developmentally controlled expression of immunoglobulinVH genes. Science 1985; 227:1597-1600.

Genetic andEpigenetic ControlofV Gene Rearrangement Frequency 81

20. Yancopoulos GD. Desiderio Sv. Paskind M et al. Preferential utilization of the most Jwproximal VHgene segments in pre-Bvcell llnes. Nature 1984; 311:727-733.

21. Alt FW; Blackwell TK . Yancopoulos GD. Development of the primary antibody repertoire. Science1987; 238:1079-1087.

22. Yancopoulos GD. Alt FW; Regulation of the assemblyand expressionofvariable-region genes. Ann Revlmmuno11986; 4:339-368.

23. Yancopoulos GD. Malynn BA. Alt FW; Developmentally regulated and strain-specific expression ofmurine VHgene families. J Exp Med 1988; 168:417-435.

24. Krangel MS. Gene segment selection in V(D)J recombination: accessibilityand beyond. Nat Immunol2003; 4(7):624-630.

25. Goldberg AD. Allis CD. Bernstein E. Epigenetics. a landscape takes shape. Cell 2007; 128(4):635-638.26. Kouzarides T. Chromatin modifications and their function. Cell 2007; 128(4):693-705.27. Chowdhury D. Sen R. Stepwise activation of the immunoglobulin mu heavy chain gene locus. EMBO

J 2001; 20(22) :6394-6403.28. McMurry MT. Krangel MS. A role for histone acetylation in the developmental regulation of VOJ

recombination . Science 2000; 287(5452):495-498.29. Johnson K. Angelin-Duclos C. Park 5 et al. Changes in histone acetylation are associatedwith differences

in accessibility ofVHgene segments to V-DJ recombination during Bvcell ontogeny and development.Mol Cell Bioi 2003; 23(7):2438-2450.

30. Morshead KB. Ciccone DN. Taverna SD et al. Antigen receptor loci poised for V(D)J rearrangementare broadly associated with BRG1 and flanked by peaks of histone H3 dimethylated at lysine 4. ProcNatl Acad Sci USA 2003; 100(20):1l577-11582.

31. Espinoza CR. FeeneyAJ. The extent of histone acetylation correlates with the differential rearrangementfrequency of individual VHgenes in pro-Bscells. J Immunol 2005; 175:6668-6675.

32. Espinoza CR. FeeneyAJ. Chromatin accessibility and epigenetic modifications differ between frequentlyand infrequently rearranging V(H) genes. Mol Immunol 2007; 44(10) :2675-2685.

33. Su IH. Basavaraj A. Krutchinsky AN et aJ. Ezh2 controls Bvcell development through histone H3methylation and 19b rearrangement. Nat Immunol2oo3; 4(2):124-131.

34. Urbanek P, Wang ZQ, Fetka I et aJ.Complete block of early Bvcell differentiation and altered patterningof the posterior midbrain in mice lacking Pax5/BSAP. Cell 1994; 79(5) :901-912.

35. Bain G. Maandag ECR, Izon D er al. E2A proteins are required for proper B-cell development andinitiation of imunoglobulin gene rearrangements. Cell 1994; 79:885-892.

36. Lin H. Grosschedl R. Failure of Bvcell differentiation in mice lacking the transcription factor EBF.Nature 1995; 376(6537):263-267.

37. Zhuang Y. Soriano P. Weintraub H . The helix-loop-helix gene E2A is required for B-cell formation.Cell 1994; 79:875-884.

38. Bain G. Gruenwald S. Murre C. E2A and E2-2 are subunits of Bvcell-specific E2-box DNA-bindingproteins. Mol Cell Bioi 1993; 13(6):3522-3529.

39. Romanow W]. Langerak AW;Goebel P et aJ.E2A and EBF act in synergy with the V(D)] recombinaseto generate a diverse immunoglobulin repertoire in nonlymphoid cells. Mol Cell 2000; 5(2):343-353.

40. Goebel P. Janney N. Valenzuela]R et al. Localized gene-specific induction of accessibility to V(D)]recombination induced by E2A and early Bvcell factor in nonlymphoid cells. ] Exp Med 2001 ;194(5):645-656.

41. Schebesra A. McManus S. Salvagiotto G et aJ. Transcription factor Pax5 activates the chromatin ofkey genes involved in B-cell signaling. adhesion. migration and immune function . Immunity 2007;27( 1):49-63.

42. Nutt LS. Urbanek P, Rolink A et aJ. Essential functions of Pax5 (BSAP) in pro-Bvcell development:difference between fetal and adult B lymphopoiesis and reduced V-rooD] recombination at the IgHlocus. Genes Dev 1997; 11:476-491.

43. Hesslein DG. Pflugh DL. Chowdhury D et aJ. Pax5 is required for recombination of transcribed .acerylared, 5' IgH V gene segments. Genes Dev 2003; 17(1):37-42.

44. Fuxa M. Skok ], Souabni A et al. Pax5 induces V-to-D] rearrangements and locus contraction of theimmunoglobulin heavy-chain gene. Genes Dev 2004; 18(4):411-422.

45. Johnson K. Pflugh DL. Yu D et al. Bvcell-speclfic loss of histone 3 lysine 9 methylation in the VHlocusdepends on Pax5. Nat Immunol 2004; 5(8):853-861.

46. Liang HE. Hsu LY, Cado D et aJ.The "dispensable"portion of RAG2 is necessary for efficient V-to-D]rearrangement during B and Tvcell development. Immuniry 2002; 17(5):639-651.

47. Zhang Z. Espinoza CR. YuZ et al. Transcription factor Pax5 (BSAP) transactivates the RAG-mediatedVHro-D]H rearrangement of immunoglobulin genes. Nat Immuno12006; 7(6) :616-624.

48. Liu H . Schmidt-Supprian M. Shi Y er aJ.Yin Yang1 is a critical regulator of Bvcell development. GenesDev 2007; 21(10) :1179-1189.

CHAPTER 7

Dynamic Aspects ofTCRa GeneRecombination:Q!!.alitative and Q!!.antitative Assessmentsofthe TCRa Chain Repertoire in Man and MouseEvelyne jouvin-Marche," Patrizia Fuschiotti and Patrice Noel Marche

Abstract

Most T-Iymphocytes expressa highly specificantigen receptor (TCR) on their cellsurface,consisting ofa clonotypic n~-heterodimer. Both u- and ~ chains are products ofsomaticrearrangements of V, (D) and} gene segments encoded on the respective loci. The

qualitative, quantitative and dynamic aspects ofthe TCRn chain repertoire ofhumans and micehave been difficult to estimate, mainly due to locus complexity. Analyses ofthe T-cell repertoirewere first performed at the transcriptional level using classical cloning and sequencing strategiesand then later at the genomic level using sensitive multiplex PCR assaysthat allow surveying theglobal rearrangement ofthe TCRAD locus. These all converge and support the conclusion thatthe V-}recombination pattern in both human and mouse thymus is not random but depends onthe reciprocal V and} positions within the locus, thereby limiting the combinatorial diversity ofthe TCRn chain repertoire. The recombination profile is compatible with a sequential openingofthe V region with progressive tracking along the two regions in opposite directions starting fromthe nearest and then moving towards the most distant V and} gene segments. In this chapter, wereport new insights into the degree ofhuman and mouse TCRn chain diversity in thymic andperipheral T-Iymphocytes. Since the comparison ofhuman and mouse V-}recombination showsa similar pattern ofrearrangement, we suggest that spatial and temporal synchronization on theaccessibilityofV and} gene segments are general features ofV-} rearrangements that are conservedthroughout evolution.

IntroductionT -cell function relies on the specific recognition of foreign antigens. The majority of

T-Iymphocytes from humans and rodents express a clonotypic n~ TCR, which is a membrane-boundheterodimer composed ofn and ~ chains that specificallyrespond to peptides derivedfrom pathogens and bound to self-MHC molecules.' Each chain contains a constant domain anda variable domain, the latter being responsible for MHC and peptide recognition via interactionwith highly diverse complementary-determiningregion (CDR) loops.iIhese chains are producedin differentiating lymphocytes by a series ofsomatic, site-specific DNA recombination reactionsofmultiple gene segments encoding TCR V,D and} domains,'

*Corresponding Author: Evelyne louvin-Marche-c-Unlversite Joseph Fourier-Grenoble I, Facultede Medecine, Institut d'Oncologie/Developpement Albert Bonniot et Institut Francais du Sang,UMR-S823, Grenoble, France. Email: evelyne .marche@cnrs-did

V(D)JRecombination, edited by Pierre Ferrier. ©2009 Landes Bioscienceand Springer Science+Business Media.

DynamicAspects ofTeRa Gene Recombination 83

Lymphocytes haveevolved sophisticatedmechanisms forgeneratingadiverse TCRrepertoire.Multipledifferentcopiesofthe V, (D) andJ genesegments areeachcapableof contributing to aTCR antigen recognition domain and different combinationsof gene segments can be used inindependentrearrangement events. Inaddition to combinatorialdiversity, variability isintroducedbyrandom removaland addition ofnucleotidesat the V-Jor V-DJjunctions.t Thisnontemplatedmechanismconsiderably increases the repertoire.A further diversifying factor is the pairingof aand ~ chains5.6to form TCR heterodimers. The potential diversitygeneratedby random V(D)Jrecombinationhas beenestimatedat 1015a~ TCRs.2However, this number is much higher thanthe actualsizeof the peripheralT-cellcompartment, estimatedat around 108 in mouseand 1012

in human. Furthermore,at leastsomecellsexpress the sameTCR specificity.7·8 Consequently, atanygiventime,onlya fractionof the potential repertoire[i,e., accordingto the random model) isachievedimplyingthat other mechanisms must governimmunediversity.

In retrospect,the theoreticaldiversityof a~T-lymphoeytes hasbeen overestimated in severalways. Firstly. the T-cellrepertoirehasbeenevaluatedassumingthat anyVgenecanrearrangewithanyJ genein the TCRA locus. However, severalsetsofdataon the mouse(thymus)modelindicatethat the numberof V-J combinations isconsiderablylowerdueto apreferential association betweenV andJ genesegments whichdependson their position within the locus.9-14 Secondly, the pairingofa and ~ chainsto formtheTCRheterodimerisconstrainedbystructuralcompatibilitybetweenthe subunits, further limiting the reperroire.v Thirdly, within the thymus, the newlygeneratedrepertoireispositively selectedl5.16viainteractionswith selfMHC molecules expressed on stromalcells. reducing the sizeof the generated repertoireby approximately lOO-fold. Furthermore, theestablishmentof aperipheralT-cellrepertoiredependsnot onlyon the interactionsofeachT-cellwith their respective ligandsbut alsoon complexhomeostaticmechanisms ensuringthe maintenanceof numbersand immunefunctionsoflymphoeyte populations.'?

Clearly, the sizeof the available peripheral TCRaf} diversityis difficult to determine and isopen to debate.While the total number of lymphocytes in the blood can be measureddirectly.the diversityof the lymphocyte compartment on which immunocompetence is based cannot.Despite considerable knowledge of the determinants and profileof theTC~ chain repertoire,very little is known about human and mouseTCRa chain diversitylikelydue to the TCRADlocuscomplexityand the limited number of anti-VAD antibodiesavailable. Thus, we haveonlya partial viewof the entire TCRA repertoire. Molecular measurements of TCR diversityusingCDR3length analysis" estimatedabout 0.5 x 106differenta chains and 106different ~ chainsexpressed in human blood lymphocytes," However, this calculationwasbasedon the analysis ofTC~ transcripts expressed in af3 T-cellclonesusingsomeV genesand with the following twoassumptions: 1) the probabilityof reartangementbetweenanyV geneand] geneisequal;and 2)the V families areexpressed at the samelevel.

Evaluation of the TCR repertoireisan important measure of the immunological competenceofan individual.Animalmodelshavebeen moreextensively studiedbut the degreeto whichtheseresultsapplyto the human modelhasyet to be established. Bymakingcomparisons betweenspecies,wehope to learnabout the generalprinciplesin operation aswellastheir specific originsandwhat this mayimplyabout the evolutionof immunity.

Complexity ofMouse andHuman TCRADLocusThe mapsof both mouseand human TCRAD locihavebeenelucidatedin the lastdecadeand

areupdatedbyIMGT.2O-22 Briefly, thehumanTCRAD locusspansabout 1000kb andconsists ofS4V genes belonging to 41 families including8 to 10pseudogenes, 61J genesegments, aswell as 12Jpseudogenes, giving49 functionalJsand auniqueC gene.20,23-25 Similarly, the mouseTCRAD locusiscomposed of70 to morethan 100Vgenes dependingon thehaplotype,regroupedinto23families,60J genesegments including16pseudogenes (namelyji, 3,4,8,14,19.20,25 ,29,36,41, 46,51,55.59 and 60) giving 44 fUnctional JSI4.20 and a uniqueC gene. Inconclusion, the humanJ regioncontainsmorefunctionalJ segments ableto rearrange than its mousehomologue(49 functionalJsin human against44 in mouse), providingmore combinatorypossibilities for the human V genesand compensating in part for the lowernumberof V genes comparedto that in mice.

84 V(D}]Recombination

AnalysisofHuman andMouseTCRA-ChainDiversityOur previous studieson the V2genefamUy of the mouseTCRAD locusindicatedthat rather

than beingstochastic. V2-Jgene rearrangements depend on the respective locationof the geneand occur in concentricwavesY·26 DuringT-celldevelopment. J usage moves fromJ genes whicharethe closest to the VgeneregiontoJ geneslocatedfarthestfrom this region;similarly, V2usagemoves fromV2genes closest to theJ generegionto V2genes locatedat the extremityof the locus.In other words. the mostproximalV2genestarget the mostproximalJ genesegments whereas themost distalV2 genes rearrange preferentially with the most distalJs.However. thesestudieswerefocused on V2genes and considered them representative of allV genes. Furthermore. the analysisofV2-J gene combinationswas conducted at the mRNA level. One cannot thereforeexcludevariedtranscriptionefficiency betweendifferentV2genesthat mayaffectthe distributionof theV2-Jcombinations. Toobtain a more accurateviewof the V-J diversity. we must analyze all V-Jcombinationeventsat the genomiclevel. As already mentioned. the diversity ofthe mammalianTCRrepertoireisgenerated bygenerearrangement. Wethereforedeveloped aPCR assayallowingvisualization at the DNA level ofseveral contiguousrecombinationeventsbetweenagiven Vgeneor V genefamily and severalJ genes segments of the TCRAD locus. Asdescribed in Figure1.ineachPCR assay.] primerswerechosento hybridize adownstream sequence allowingamplificationoffour to seven differentJ genes.Thus.apanelof nine to elevenJ primersallowed thedescriptionof the rearrangement status of all functional mouse and humanJ genes and provided a globalvisualization ofrearrangement patterns (Fig. 2).

Genomic multiplexPCR analysis of mouseTCRa chain diversity confirms previous data atthe rearrangement level, in that V-Jrearrangements are not random but depend on the V and]positionswithin the locus. For example. in the mousethymus. V families locatedclosest to theC codingregion.suchasVI9 and V20.rearrange predominantlywith the most proximalJs(J60toJ48) and rarelywith theJ segments locatedin the mid-section or the distalpart of theJ region(shownin Fig. 3). Reciprocally, VI and V2 situated in the most distalpan of the V generegionpreferentially rearrange with theJ segments found in the mid-section or distalpartsoftheJ regionbut not with theJsfound moreproximally. Thus, the TCRAD locusisaccessible from the 3' endofthe V regionand from the 5' end ofthe] regionand consequently the proximalV and] genesare the firstgenesegments accessible for recombinationfollowed lateron bymoredistalV andJsegments. In addition,wereportedthat dependingon its locusposition.eachVgenedifferentiallyrearranged withasetofcontiguous]swith agaussian-like distribution." Forinstance, the realtimePCR quantification ofVI and V21 rearrangements revealed that the proximalV21 geneusedasmallsetof] genes, less than 10,but with a 6 foldhigherfrequency than distalVgenes whichusedalargerpanelofJ genes (morethan32).These preferential associations betweenVandJ geneswereobservedwith differentV genes locatedat differentpositions in the TCRAD locus, suggestingthat eachV genetargetedparticularsetsofJ segments.

A similarmultiplexPCR experimental approachhas been used to characterize the a chainrepertoire in human thymi.Byfocusing the analysis on single member families to correlate theposition of each V genewith its rearrangement pattern (Fig. 3. top panels). it can be observedthat the two V genesmost distant from theJ region (VI. V2, locatedat -925 and -835kb fromthe C gene.respectively) rearrange with the centraland 3' end of theJ region.whereas the threeJ-proximalV genes, namelyV38.V40 and V41, locatedbetween-267 to -227 kb with respecttothe C gene,mainlyrearrange with the mostproximalJs. Finally. the membersofthe multigenicV8famUy locatedin the middleparrofthe locus. includingmembers locatedat -701,-653and -569kbrespectively fromthe C region, rearrange to approximately the sameextentwith alltheJ segmentsthroughout the locus. Takentogether, thedatashowpreferential distributionofrecombinationofparticularV families to certain] genesegments dependingon their localization within the locus.Thesefindings are consistentwith the modelof synchronized waves ofaccessibility movingin aconcentricmanneracross both V and] generegions. Thesewaves ofrearrangement movefromJgenes locatedproximal to theVregiontowardsJ genes locatedcloser to theC geneandfromVgeneslocatedproximalto] regiontowards moredistallylocatedV genes, supportingthe bi-directional

DynamicAspects ofTCRa Gene Recombination 85

A V-J rearrangements In T lymphocytes

B

c

Amplification

Ident ification lind ~antificatlonof the V.J I"llarrangemoms

1666 bp

646bp

Figure 1. Schematic representation of multiplex PCRanalysis of TCRA gene rearrangements.Briefly, by using two specific primers, one upstream of a given V gene and another downstream of a given Jgene, the PCR will amplify all rearrangements involving both of thesegenes. This multiplex assayallows the detection of a V-J rearrangement as well as that of a setof four to seven upstream Jgenes with a maximum amplicon size of approximately 8 kb. Thespecificity of TCR rearrangement products can be assessed both by successive hybr idizationwith internal V and Jprobes and by an accurate measure of the length of the PCR productscompared to the known position of the genes in the IOCUS. 21

and coordinatedmodelpostulatedin the mouse.13•14 Inconclusion, the comparisonof human and

mouseTCRA V-J recombination in the thymusshows asimilarpatternofrearrangement suggestingthat this mechanisticregulationof the processis conservedthroughout evolution.

Comparison between the Frequencies ofRearrangements in Thymusand Peripheral T-Lymphocytes

In order to gain further insight into the frequendes of V-J combinations. we set up a precisequantification of rearrangements byreal-time genomic quantitativePCR (qPCR). Particular VandJ genes wereselected asrepresentative of several locationsin the TCRAD locusand qPCR wascarriedout with DNA fromthymi(Fig.4A)and fromperipheralbloodlymphocytes (PBLs) (Fig.4B).WhilethepatternsofV-Jcombinations appearsimilaramongindividuals andfollow thegeneral rules,somediscrete differences in recombination frequencies aredetectedwhen comparingthe patternsobtainedin the thymusandperipheralT-Iymphocyte DNA. Several observations emerge fromthesedetailedanalyses. Firstly,someV-J combinations (l,e.,VI-J56.VI-J53,V4O-J10andV41-J10)arcnot

Mo

use

V12

-Jre

arr

an

ge

me

nt

Inth

ymu

s

~·'

I~..

f!!!I~

I:!ti

I~III!

llJ5

6J.

C8.M

OJ3

3.m

mJI

GJi

Jre

gio

n

Hum

anV

8-J

rea

rra

ng

em

en

tIn

thym

us

6060

4030

2010

1

,I,'u

::tl'I

II'I

IIIII

IIIIII

IIIII

IIII'

IIIII,

!.11

1111,

1-~~

---.

-J5

6J5

3J.

C8J4

1J3

3J2

9J2

,4J1

8J1

0J5

Jl

Jre

gio

n

~

47 4561

J~1521

1 44 43

__ 421~

1130

1-1

L1

18~1

11'

58 57

9

56

Fig

ure

2.M

ult

iple

xP

CR

assa

ysof

V-J

rear

rang

emen

tsof

the

mou

sean

dhu

man

TC

RA

Dlo

cus.

Ge

no

mic

DN

Aw

ase

xtra

cte

dfr

omth

ymus

and

am

plif

ied

usin

gpr

imer

ssi

tuat

edd

ow

nst

rea

mof

dif

fere

ntJ

gene

san

dsp

read

over

the

Jge

nere

gion

,to

ge

the

rw

ith

ap

rim

er

spe

cific

for

eith

er

the

mou

seV

12(le

ftpa

nel)

orhu

man

V8

(rig

htpa

nel)

gene

fam

ilies

.M

ult

iple

xP

CR

was

pe

rfo

rme

dw

ith

the

Exp

end

Hig

hF

ide

lity

PC

Rsy

stem

(Roc

he,

Me

yla

n,

Fra

nce)

asfo

llow

s:5

min

at94

·C,

follo

we

dby

26cy

cles

cons

istin

gof

1m

inat

94·C

,1

min

at58

·C,

6m

inat

n·c

and

fin

ally

10m

inat

n·c.

The

PC

Rs

wer

ein

the

linea

rph

ase

ofa

mp

lific

ati

on

wit

hin

the

26cy

cles

.P

CR

prod

ucts

wer

em

igra

ted

on1.

6%ag

aros

ege

lsan

dp

rob

ed

byS

outh

ern

hyb

rid

iza

tion

.T

hepr

obes

used

tod

ete

ctth

eT

CR

Are

arra

ngem

ents

are

spec

ific

for

the

resp

ect

ive

Vge

nes.

;s. ,2 <.;;;:; ~ 0: li .... r ~.

DynamicAspects ofTCRa Gene Recombination 87

detectableeither in the thymusor the PBL,presumably because theyareveryinfrequent.Thisresultconfirms the combinatorial pattern described in Figure3, dependenton the reciprocal position ofthe V andJ genes within the locus. Secondly, somecombinationsarefavored in the peripherywithrespectto others(forinstanceVI-J33 canbefound at ahighfrequency inallsamples tested).1hirdly,somerearrangements arequantitatively less abundant in the peripherywith respectto the thymus.Inparticular, proximalV-J rearrangements, like V4O-J56or V40-J53, areweakly found (at 6 to 8 cyclesofqPCR) in theperipherycomparedto theirhighfrequency in thymussamples.Severalpossibilitiesmayaccountfor thesedifferences, including:(I) variationin the numberofT-cellsbetweenthymusand PBL samples; (2) the contribution of rearrangements occuringon excision circles (thesemaybe more frequently amplified in thymus than in peripheralT-cells in which excision circles havebeen diluted); (3) the occurrence of secondaryrearrangements in the thymusor receptor revisioneventsin the peripherywhichwould replace the most-proximal and accessible V-J rearrangementsbyjoiningbetweenmoredistal V andJ genesr" (4) positiveand negative selectionevents.28 Finally,the expansion/contraction ofspecificrearrangements (Le.,V4O-J41,VI-J41,VI-JIO) canbe identifiedin certainindividuals.Takentogether, this analysis demonstrates that, whilethe recombinationpattern isquantitatively similarin thymussamples of several individuals, moreheterogeneityofV-Jcombinationisobserved in theperipheralT-cdlTheseobservations mayindicatethesharingamongstindividuals of thymicselection eventswith similarimpacton V-J combination,whereas adivergenceamongstindividuals in the peripheryregarding someV-J combinations could reflectexpansions ofparticularclonorypes inducedbyimmuneresponses or homeostaticmaintenanceforces.

The Size ofthe Mouse and Human TeRu RepertoireDependenton thelocuspositiontogetherwiththedifferentialexpression ofV families, preferential

V-Jrecombinationleadsto a restrictionof the potential combinatorialTCR a chain repertoire. Byanalyzingheterogeneityin CDR3 sequences, the diversity of the humana chainrepertoirewasestimated at around0.5 x ICf chainsin the blood," However, in thiscalculation, all the humanTCRAV-Jcombinations wereconsidered asequallylikely.Thetheoreticalnumberof combinationsifall54V genescould rearrange to eachof the 61J genesegments within the locusis3294. However, only46 human V genes and 49J segments areavailable for rearrangement. Takinginto account (i) thatthe recombinationof proximalV genes includingV1.1 to V7 is restrictedto the closest halfof theJregioncorresponding to approximately 32Js; (u) that the centralV genes rearrange with about45Jgenesegments; and (iii) that the distalVgenes, (i,e.,V31to V41)do not rearrange with the first9Jsgiving 9 functionalV genes rearranging with 40Js,then the numberofpossible V-Jcombinations isless than 2000 (8Vx 32J + 29Vx45 + 9Vx 4OJ).Thissuggests that the actualnumberofcombinations corresponds to less than 60%of the estimated0.5 x 106 total combinatorialpossibilities, i.e.,0.3 x 106 TCRa chains. Thisvalueis also likely overestimated as it does not take into account thedifferentfrequencies of utilizationof V andJ genesegments within the locus. Concerning mouse,the number of differenta chainshavebeen estimatedas around 1.2 x 1()4 in the C57Bl/6 or BIOTCRAD haplotype." It isworth noting that the number of V genes varies from I to 3 fold amongdifferenthaplorypes, for instancethe C57Bl/6 haplotypepossesses 113less V genes comparedtothe Balblc haplorype' P? leadingto an estimated0.8 x 104 TCRa chainsin the Balblc haplotype.In addition, multiple roundsof V geneduplications mean that most V families contain between2and 10members, in somecases perhapsdifferingbyonlyone to threepunctual mutationsscatteredthrough the V genes." Thispreventsa precise determination of the number of] segments usedbyV genes. In the Balblc TCRA haplotype, (i) the most proximalV genes (V21 to V23) are foundrearranged to less than !OJ genes(thosebetweenJ60 toJ48), (ii)the middleV genesusea panelofabout 35Js and finally (iii) the distalV genes (VI to V3.l) usea panelofless than 30J segments.Usingthisinformation,weestimateda reductionof around 30%in the numberofV-] combinationsin Balb/c micecomparedto the theoriticalnumber of combinations(re£14 and our unpublishedresults) yieldinganestimated0.6x I()4 differenta chains.Takentogether, thesefindings indicatethatwhilstremaininglargeenoughto maintain a high functionaldiversity, limitationsof combinatorialdiversity reducethe sizeof the available human and mouseTCR a chain repertoires.

DIS

TA

LM

IDD

LEP

RO

XIM

AL

J6i

Jgg

664

84

03

32

72

31

89

V1

58

"40

3327

2316

9

V2

=

•

6648

4033

272

31

89

66

"40

33

27

2318

9

V12

V19

--

V8-

VB

53"

4129

18

10t5

---

---

-

V2

----- -'- ----

-:

.-----

•

Vi.

i

MO

US

E

HUM

AN

MID

DLE

--

PR

OXI

MA

L--

Vi

V2

V12

Vi9

V20

V2i

J6i

J

Fig

ure

3.V

-Jre

arra

ngem

ent

depe

nds

onV

and

Jlo

caliz

atio

nw

ithin

the

TR

AD

locu

s.V

-Jsp

eci

ficre

arra

ngem

ents

wer

ean

alyz

edby

mu

ltip

lex

PC

Ron

DN

Aex

trac

ted

from

tota

lth

ymus

ofa

6d

ay-

old

hum

an(u

pper

pane

l)an

dfr

omto

tal

thym

usof

a6

we

ek

old

Bal

b/c

mou

se(l

ow

er

pane

l).

Pro

duct

sof

PC

Rw

ere

sepa

rate

dby

gel

elec

trop

hore

sis,

Sou

ther

nbl

otte

dan

dhy

brid

ize

dw

ithra

dio

act

ive

lyla

bele

dpr

obes

spe

cific

for

Jge

nes.

Spe

cific

Vge

nepr

imer

sar

ein

dica

ted

onth

eto

por

the

bo

tto

mof

each

pane

l.T

hey

wer

eus

edin

con

jun

ctio

nw

ith

seve

nre

pre

sent

ativ

eJp

rim

ers,

asin

dic

ate

dfo

rea

chla

neof

the

elec

trop

hore

sis

gels

.Eac

hba

ndco

rres

pond

sto

are

arra

ngem

ente

vent

asd

ete

rmin

ed

bydi

stan

cem

igra

tion

.The

sche

mat

icre

pres

enta

tion

ofth

eT

CR

AD

locu

s(n

otdr

awn

tosc

ale)

indi

cate

sth

ere

lativ

epo

sitio

nof

each

Vge

nein

hum

an(t

op)

and

mou

se(b

otto

m).

For

s im

plic

ity

on

lyth

est

udie

dV

gene

san

da

part

ofth

eJg

ene

regi

onar

ein

dica

ted.

~ ~ <-=::; ~ 0: ~ ~. ~.

DynamicAspects ofTCRa GeneRecombination

A

89

12

o~ C) ...- s s s C)...... ::.2-3_ .:s .:s.:s.:l. C) ... ...

- - .:s ;: ~ ~ ... .... -1J56 J53 - - __ .:s ;: .:l. ~ ~

J41 - _ .:s .:sJ33 J10 -

B

J41 J33J10 -

Figure 4. Relative abundance of V-J specific rearrangements among healthy individuals asdetermined by quantitative genomic PCR analysis.The investigated rearrangementsinvolvedV1, 40, 41 and J56, 53, 41, 33, 10 of the human TCRAD locus. The results are expressedin arbitrary units (AU) ind icating the differences in cycle numbers at which the productswere first detected, therefore reflecting the relative quantities of PCR products for eachV-J rearrangement in different individuals. The figure depicts examples of representativeresults for 3 thymi (A) and 3 PBl (B) DNA samples belonging to 6 healthy individuals agedbetween 25 and 55 yrs (numbered 1 to 6). Normalization for the DNA content of eachsample was performed by amplification of the G3PDH gene. Data are representative ofthree different experiments.

90

V region (-1000 Kb)_ DISTAL _ _ DOLE _ _ PROXIMAL _

_ ------------ OPENING

V(D)j Recombination

J region (-60 Kb)

Figure 5. Schematic representation of the V-] combinations in TCRA rearrangements. V and] genes were respectively categorized according to their respective relative locations in theTCRA locus as distal (white dashed boxes), middle (grey boxes) and proximal (black boxes).Combinations of V to Jgenes are indicated by arrows, with the same color code, where thethickness of the line is indicative of the relative frequencies of V-] combinations. The openingof the TCRA locus to V-] gene rearrangement appears as concentric, from the closest to themost distant V and] genes.

ConclusionThe fact that V and} gene segments combine preferentially according to their position in the

TCRA locus suggests a control of rearrangements depending mostly on the strict regulation ofchromatin accessibility in both the V and} gene regions (Fig.5). Cis-acting elements, particularlyenhancers and promoters. have been proposed as being involved in chromatin remodelling.v-" Inthe murine TCRA locus, accessibility of the} region is controlled by the En enhancer located 3'of the C coding region.33 In addition. two promoters contribute to the control of}a rearrangements, namely the T earlya (TEA) at the 5' end ofthe} region and a second promoter located 15kb downstream ofTEA before the }49 coding region. Both promoters can be activated by Ea.34.35

The TEA promoter has been shown to spatially regulate} gene utilization36 and drive noncoding transcription to positively and negatively instruct the activity ofdownstream} promoters."Interestingly, TEA transcription has been proposed to target V rearrangements to the S' end ofthe} region and consequently determines the rearrangement profile of this region by promotingthe activation ofproximal} promoters (J58 to}56) while repressing that ofmore distal} promoters (see chapter by Abarrategui and Krangel) .These recent data on the role ofTEA transcriptionon} gene accessibility support the recombination profiles discussed in this report. Whilst we arebeginning to gain a better understandingofthe mechanisms contributing to the use of} segments,the process ofV gene accessibility to rearrangement and the control of their uses remain to beelucidated.

The evaluation of the TCR repertoire is an important measure of the immune competenceof an individual. It is assumed that the larger the number of distinct immune T-cells, the moreefficient the protection against infectious diseases. Consequently, the size and diversity of theavailable repertoire are crucial in shaping the immune response to a given antigen. Our studiesstrongly suggest that although it remains large enough to maintain a high functional diversity,the TCR repertoire ofhuman and mouse a chains is smaller than that predicted by the randomrearrangement model. Detailed knowledge about the extent and diversity ofthe TCR repertoireused in specific immune responses will facilitate the ability to understand the role of the TCRgenes in normal and diseasestates.Whereas clonal populations are hallmarks ofmalignancy,clonalor oligoclonal populations ofT- and B-Iymphocytes may also arise in nonmalignant conditions,including normal individuals (responses against some pathogens such as HIV and EBV), elderlypatients and patients suffering from auto immunity or immunodeficiency. Our straightforwardexperimental approach enables a qualitative and quantitative description of the overall TCRachain diversity in humans and offersa unique opportunity to characterize and track the repertoirefor each individual in healthy and diseased states.

DynamicAspects ojTCRa Gene Recombination 91

AcknowledgementsTheauthorswishto thank] Conway, MATheluand HL Dougierfor their helpfuldiscussions

and comments. Thisworkwassupported by institutionalgrants from the Institut Nationalde IaSanteet de IaRecherche Medicale and from the Commissariat a l'EnergieAtomique and partlybyspecific grantsfrom the Agence Nationalepour laRecherche sur leSIDA et lesHepatitesandthe LigueContre le Cancer. P Fuschiottiwas recipientof a fellowship from the Centre Nationalde laRecherche Scienrifique.

References1. Davis MM, Bjorkman PJ. T-cell antigen receptor genes and Tvcell recognition. Nature 1988:

334(6181 ):395-402.2. Davis MM , Boniface JJ, Reich Z et al. Ligand recognition by alpha beta T-cell receptors. Annu Rev

Immuno11998: 16:523-544.3. Cobb RM, Oestreich K]. Osipovich OA et al. Accessibility control of V(D)J recombination. Adv

Immunol2006: 91:45-109.4. Cabaniols JP, Fazilleau N, Casrouge A et al. Most alpha/beta T-cell receptor diversity is due to terminal

deoxynucleotidyl transferase. J Exp Med 2001; 194(9) :1385-1390.5. Malissen M, Trucy J, Letourneur F et al. A T-cell clone expresses two T-cell receptor alpha genes but

uses one alpha beta heterodlmer for allorecognition and self MHC-restricted antigen recognition . Cell1988: 55(1) :49-59.

6. Saito T, Sussman JL, Ashwell JD et al. Marked differences in the efficiency of expression of distinctalpha beta T-cell receptor heterodirners, J Immuno11989: 143(10):3379-3384.

7. Mason D. A very high level of crossreactivity is an essential feature of the T-cell receptor. ImmunolToday 1998: 19(9):395-404.

8. Reiser JB, Darnault C, Gregoire C et al. CDR3 loop flexibility contributes to the degeneracy ofTCRrecognition . Nat Immunol2003: 4(3) :241-247 .

9. Thompson SD, Pelkonen J, Hurwitz JL. First T-cell receptor alpha gene rearrangements during T-cellontogeny skew to the 5' region of the J alpha locus. J Immuno11990; 145(7):2347-2352.

10. Roth ME, Holman PO, Kranz DM . Nonrandom use of J alpha gene segments. Influence of V alphaand J alpha gene location. J Immuno11991 ; 147(3):1075-1081.

11. Rytkonen MA, Hurwitz JL, Thompson SD er al. Restricted onset of T-cell receptor alpha gene rearrangement in fetal and neonatal thymocytes. Eur J Immuno11996: 26(8) :1892-1896.

12. Jouvin-Marche E, Aude-Garcia C, Candeias S et al. Differential chronology of TCRADV2 gene use byalpha and delta chains of the mouse TCR. Eur J Immuno11998: 28(3) :818-827.

13. Huang C, Kanagawa O. Ordered and coordinated rearrangement of the TCR alpha locus: role ofsecondary rearrangement in thymic selection. J Immunol 2001; 166(4):2597-2601.

14. Pasqual N, Gallagher M, Aude-Garcia C ec al. Quantitative and qualitative changes in V-J alpha rearrangements during mouse thymocytes differentiation: implication for a limited Tvcell receptor alphachain repertoire. J Exp Med 2002: 196(9) :1163-1173 .

15. Sant'Angelo DB, Waterbury PG, Cohen BE et al. The imprint of intrathymic self-peptides on the matureT-cell receptor repertoire . Immunity 1997; 7(4) :517-524.

16. van Meerwijk JP, Marguerat S, Lees RK et al. Quantitative impact of thymic clonal deletion on theT-cell repertoire. J Exp Med 1997; 185(3) :377-383.

17. Corre ia-Neves M, Waltzinger C, Mathis D et al. The shaping of the T-cell repertoire. Immunity 2001:14(1):21-32.

18. Pannetier C, EvenJ, Kourilsky P. T-cell repertoire diversity and clonal expansions in normal and clinicalsamples. Immunol Today 1995; 16(4):176-181.

19. Arstila TP, Casrouge A, Baron V et al. A direct estimate of the human alphabeta T-cell receptor diversity.Science 1999; 286(544l) :958-961.

20. Lefranc MP. IMGT, the international immunogrnetics database. Nucleic Acids Res 2001; 29(1):207-209.21. Baum TP, Pasqual N, Thuderoz F et al. IMGT/GeneInfo: enhancing V(D)J recombination database

accessibility. Nucleic Acids Res 2004: 32(1) :D51-54.22. Baum TP, Hierle V, Pasqual N et al. IMGT/GeneInfo: T-cell receptor gamma TRG and delta TRD

genes in database give access to all TR potential V(D)J recombinations . BMC Bioinformatics 2006:7(1) :224.

23. Koop BF, Rowen L, Wang K er al. The human T-cell receptor TCRAC/TCRDC (C alpha/C delta)region: organization, sequence and evolution of 97.6 kb of DNA. Genomics 1994: 19(3):478-493.

24. Glusman G, Rowen L, Lee I et al. Comparative genomics of the human and mouse T-cell receptor loci.Immunity 2001; 15(3):337-349.


25. Fuschiotti P, Pasqua!N. Hierle V et at Analysisof the TCR alpha-chain rearrangement profile in humanT-lymphoeytcs. Mol Immunol 2007; 44(13):3380-3388.

26. Aude-Garcia C. Gallagher M. Marche PN et al. Preferential ADV-AJ association during recombinationin the mouse T-eell receptor alpha/delta locus. Immunogenetics 2001; 52(3-4):224-230.

27. Mostoslavsky R. Alt FW Receptor revision in T-cells: an open question? Trends Immunol 2004;25(6) :276-279.

28. Bandeira A. Faro J. Quantitative constraints on the scope of negative selection: robustness and weaknesses.Trends Immunol2oo3; 24(4):172-173.

29. Jouvin-Marche E. Hue I. Marche PN er al. Genomic organization of the mouse T-cell receptor Valphafamily. EMBO J 1990; 9:2141-2150.

30. Gahery-Segard H. Jouvin-Marche E. Six A et at Germline genomic structure of the BI0. A mouseTcra-V2 gene subfamily. Immunogenetics 1996; 44(4) :298-305.

31. Krangel MS. Gene segment selection in V(D)J recombination: accessibility and beyond. Nat Immunol2003; 4(7) :624-630.

32. Mostoslavsky R. Alt FW, Bassing CH. Chromatin dynamics and locus accessibility in the immunesystem. Nat Immunol 2003; 4(7) :603-606.

33. Slcckman BP. Bassing CH. Bardon CG et al. Accessibility control of variable region gene assemblyduring T-cell development. Immunol Rev 1998; 165:121-130.

34. Villey I. Caillol D. Selz F et al. Defect in rearrangement of the most 5' TCR-J alpha following targeted deletion ofT early alpha (TEA) : implications for TCR alpha locus accessibility. Immunity 1996;5(4):331-342.

35. Hawwari A. Bock C. Krangel MS. Regulation of T-cell receptor alpha gene assembly by a complexhierarchy of germline J alpha promoters . Nat Immunol 2005; 6(5) :481-489.

36. Mauvieux L. Villey I. de Villartay JP. T early alpha (TEA) regulates initial TCRVAJA rearrangementsand leads to TCRJA coincidence. Eur J Immuno12001; 31(7):2080-2086.

37. Abarrategui I. Krangel MS. Noncoding transcription controls downstream promoters to regulate T-cellreceptor alpha recombination. EMBO J 2007; 26(20) :4380-4390.

CHAPTER 8

Germline Transcription:A Key Regulator ofAccessibility and Recombinationlratxe Abarrategui and MichaelS. Krangel*

Abstract

T he developmental control ofV(D)J recombination is imposed at the level ofchromatinaccessibility of recombination signal sequences (RSSs) to the recombinase machinery.Cis-acting transcriptional regulatory elements such as promoters and enhancers playa

central role in the control ofaccessibility in vivo. However. the molecular mechanisms by whichthese elements influence accessibilityare still under investigation.Although accessibilityfor V(D)Jrecombination is usually accompanied by germline transcription at antigen receptor loci. thefunctional significance ofthis transcription in directing RSS accessibility has been elusive. In thischapter. we review past studies outlining the complex relationship between V(D)J recombinationand transcription as well as our current understanding on how chromatin structure is regulatedduring gene expression . We then summarize recent work that directly addresses the functionalrole oftranscription in V(D)J recombination.

IntroductionV(D)J recombination at antigen receptor loci takes place within the complex nucleoprotein

environment of chromatin. An extensive body of literature supports the notion that chromatin-embedded recombination signal sequences (RSSs) must be made accessibleto the recombinasefor the V(D)J recombination reaction to proceed and that the regulation of RSS accessibilityprovides an important layer ofdevelopmental control to V(D)J recombination in vivo.' Studiesofantigen receptor loci have implicated promoters and enhancers as developmental regulators ofboth chromatin structure and V(D)J recombination. However. the detailed mechanisms bywhichthese elements stimulate recombination are not well understood. Enhancers and promoters serveas docking sites for the recruitment of factors that initiate changes in chromatin structure. Theyalso serve ascritical regulators oftranscription. Studies ofantigen receptor loci havedemonstratedthat unrearranged gene segments typically become transcriptionally active at the developmentalstage at which they undergo V(D)J recombination. Nevertheless. whether transcription plays adirect role in providing the recombinase machinery access to RSSs. or is simply an unrelated consequence oflocus accessibility.has remained obscure for two decades. Resolution ofthis issue hasrequired an experimental approach that can discriminate and independentlyevaluate the individualdownstream consequences of enhancer and promoter activity as they relate to the stimulationofV(D)J recombination in vivo. Recent studies have provided important steps forward in thisregard and implicate germline transcription as a key developmental regulator ofaccessibility forV(D)J recombination.

·Corresponding Author: Michael S.Krangel-Department of Immunology, Duke UniversityMedical Center, Durham NC 27710, USA.Email: krangOO1 @mc.duke.edu

V(D)JRecombination. edited by Pierre Ferrier. ©2009 Landes Bioscienceand Springer Science+Business Media.


A BriefHistory ofGermline Transcription and V(D)J RecombinationItwasfirstobservedmore than 20 years ago byYancopoulos and Alt that the developmental

activation ofVHsegment recombination at the Igh locus coincidedwith the appearance ofVHgermlinetranscription.i Germlinetranscriptsinitiatingfrom promotersassociated with V,D andJ gene segmentshavesincebeen documented at allantigen receptor loci and havebeen shownto coincidedevelopmentally with the onset ofV(D)J recornbination.i? In addition to these examplesofsensetranscriptionacross antigenreceptorgenesegments, recentstudieshavedescribeda developmentalrelationshipbetween antisense intergenictranscription across the VHlocusandrecombination of VHgenesegmenrsf On the basisof such correlations, germlinetranscriptionhas long been proposed to playa role in the establishmentof an open chromatin configurationthat stimulatesRSSaccessibility.

A linkage between germline transcription and recombination competence was reinforcedover the years by a variety of experimental approaches. For example, stable transfection ofpreB-cells with a recombinationsubstratecontainingan exogenous promoter demonstratedthatactively transcribed substrates underwent DH-JH recombination? Lipopolysaccharide treatmentofpreB-cellsinduced bothIgklocustranscriptionand VIC-JIC rearrangemene,"StabletransfectionofpreB-cells with recombinationsubstratesshowed that an enhancerpromoted both recombination and transcription." Likewise, the introduction of a strong promoter into the Ig/locusbyhomologous recombination caused a dramatic increase in both JA. germline transcription andVA.-JA. recombination.'?

Several transcription factorshavealsobeen shownto coordinatelyregulateboth trancriptionand V(D)J recombination.Overexpression ofEZAin recombinase-expressingnonlymphoidcellsinduced both germlinetranscriptionand recombinationof Igk, Tcrg and Tcrd genesegmenrs.l':"Micedeficientfor the transcriptionfactorOcaBdisplayed defective transcriptionand recombination ofa subsetof VIC genes." In addition, StatSwasshown to be required for transcriptionandrecombinationof distalVHsegments andJy segments,14.15 in response to IL-? receptor signalingand for transcription and recombinationofthe VyS genesegmentin response to IL-l S receptorsignaling," Consistent with allof the above, gene targetingexperiments haveshowndeletion ofenhancersand promotersat endogenouslocito inhibit both transcriptionand V(D)J recombination oflinkedgenesegments.I However, none of the abovestudieshad the powerto critically testa causal relationshipbetween transcription and V(D)J recombination.

In contrast, several other studieshaveindicatedthat V(D)J recombinationand germlinetranscription are not invariably linked. In some instances, transcription through genesegmentswasshown to be insufficient to promote recombinase activity. For example, distalVHgenesegmentsaretranscribedat high levels inPaxS-I- pro-B-cells but failto undergorecombmadon."However,these transcribedVHsegments might retain a permissive chromatin configurationin the absenceofPaxS, but might fail to rearrangedue to additional PaxS functions that are needed for recombination. For example, PaxS hasbeen shownto regulatean Igh locusconformationalchangethatis required to bringdistalVHand DJHsegments into proximityfor VHto DJHrecombinationandto recruit RAG proteins to VHsegmenrs.P :" Several studiesusingversions of a transgenicTCRflminilocusrecombinationsubstratehavealsoprovidedexamples of transcriptionin the absence ofrecombination.In one case, E~ and Ell- wereshown to stimulatesubstrateD~ toJ~ but not V~ toD~J~ recombination in Bvcells, eventhough theseenhancerscouldpromote germlinetranscription ofV~ and D~J~ segments in those cells.20 In other instances, minimalformsofE~ or 4 thatlackedbinding sitesfor specific nuclearfactorsefficiently stimulatedminilocustranscription butcould not support recombination.21,22Theseresultssuggest that there maybe enhancerfunctionsthat promote recombination independent of enhancer effects on transcription,but do not ruleout a role for enhancer-directed transcription.

Studiesof the Tcrb locushavealsodescribedcircumstances in whichactivetranscriptionisnotpredictiveofrecombination. For example, germlinetranscriptionofV138.2 occurson both allelesin allCD4-CD8- double negative (DN) thymocyres even though V138.2 usually rearranges ononly a singlealleleand in only a fractionofthesecells." Ectopicintroduction ofEa downstream

Germline Transcription: A Key RegulatorofAccessibility andRecombination 95

ofVpl2 greatlyenhancedthe transcriptionof this segmentin DP thymocytes but did not induceVPI2-DP1P recombination." Similarly. a large Tcrb locus deletion that placed Vp segmentsunder the influence ofEP stimulated high levelVp transcription but not recombination in DPthymocyres." However. the failure to rearrangein these examples maybe explainednot by anylackofVP segmentaccessibility but byinappropriatenuclearlocalization or locusconformation.Alternatively, theremaybeconstraintsimposedbyunknownfactorsthat mightpromoteor inhibitVp to D1Precombinationat the appropriatedevelopmentalstage.

In other instances. recombinationhas been documented to occur in the absence of germlinetranscription. A study usingisolated nuclei from RAG deficientcells showed that endogenousRSSs could be cleaved by the addition of RAG proteins in vitro in the absenceof ongoing transcription." However.chromatinmodifications introducedbytranscriptionprior to the isolationofnucleicouldhavebeen sufficient to providegenesegmentaccessibility duringsubsequentin vitrocleavage reactions.AnotherstudyidentifiedendogenousVH segmentrearrangements in pro-B-cellsthat did not detectablytranscribethoseVH segments." In thissystem. rearrangement wasinducedby transfectionofRAG-deficient pro-B-cellcloneswith RAG expression plasmids. However, thetranscription status of VH segments at the time of recombinationcould not be analyzed, leavingopen the possibilitythat VH transcriptiondid occur in the smallfraction of cells that underwentrearrangement during the three-week culture period following RAG transfection.

Several studieshaveshown that localizedchromatin remodelingat promoters is sufficient tostimulaterecombinationat adjacentgenesegments in the absence of read-throughtranscription.In one example. an RSSwastightlyassociated with the induciblemousemammarytumor viruslong terminal repeat." When propagated as a chromatinized episomal substrate, nucleosomeorganizationat the promoter precluded protein access. However. the mobilizationof promoternucleosomes by treatment with dexamethasone wasfound to promote recombinationin the absenceof measurable transcription.At the endogenousTcrb locus,accessibility at Dp1requirestheconcertedactionofEp and apromoter tightlyassociated with this segment,PDpI,29.32 A physicalinteraction between Ep and PDpl is required to deliverthe SWl/SNF chromatin-remodelingcomplex to PDpI, resulting indecreasednucleosome occupancyat Dp1,31.32 Aseries ofexperimentsmakinguse of a TCRp minilocushaveargued that Dpi accessibility depends on local targetingofSWl/SNF by PDpl but can occur independent of PDp l-derived transcription. For example,minilocusrecombination requiresthat PDpl is situated immediately adjacent to the Dpl RSS,but can be supported bya versionof PDp1that doesnot stimulatetranscription through the DpandlP segments." Moreover, PDp1function could be substituted bycontrolled targetingofthecatalyticsubunit ofSWI/SNF to D~1.34Thesestudiesarguepersuasively that transcriptionisnotan absoluterequirementfor accessibility, particularlywhen apromoter and RSSaretightlyassociated. However. these studiesdo not discount the possibilitythat transcription could contributesubstantially to accessibility at endogenousantigen receptor loci in vivo.

Disruption ofChromatin byTranscriptionThegeneticmaterial ispackedin theeukaryotic nucleusin ahighlyorganizedfashion.35Thefirst

level ofcompaction isachievedbywrapping146 bpofDNA aroundthehistoneoctamer(composedof two copies eachof histoneH2A, H2B, H3 and H4) to form the nucleosome parricle."A lineararrayof nucleosomes, with 20-60bp of internucleosomallinker DNA, formsthe 10nm fiberthathas the appearance of'beads on a string'whenviewed under an electronmicroscope. Formationofthe more compact30 nm fiberdependson the binding of histone HI to linkerDNA and on theestablishment ofinternucleosornal interactions. However, the mechanisms that govern compactionof the30nm fiberinto higherorderstructures,ultimately resulting in theassemblyofchromosomes,remainelusive.F'Ihehighlycompactchromatinorganization inhibitsaccess ofproteinsto theunderlyingDNA, therebyimposing anobstacle to transcription. Eukaryotic cells useavarietyof strategiesto dynamically modulatechromatinstructureto achieve regulatedgeneexpression.

All four histones are subjected to a variety of posttranslational modifications that includeacetylation, methylation.phosphorylation and ublquicylation,38 Thesemodifications aretargeted


both to the extended amino terminal tails ofhistones that project awayfrom the nudeosome surface and to the globular domains ofhistones . Specificpatterns ofhistone modifications correlatewell with the activation status ofa gene. For example, active genes typically display high levelsofhistone H3 and H4 acetylation and histone H3lysine 4 (H3 K4) methylation, whereas repressedgenes are typically enriched for histone H3 K9 and K27methylation. Histone modifications arehighly dynamic . For instance, the introduction ofhistone acetylation by histone acetyltransferases(HATs) is reversed by histone deacetylases (HDACs),39 and the introd uction ofhistone methyla tion by histone methyltranferases (HMTs) is reversed by histone demeehylases/? Moreover, theintroduction of histone marks can be influenced by chromatin context, in the sense that priormodifications to neighboring amino acids can either promote or inhibit the introduction of asubsequent modification.41•42 The function of these modifications is twofold: on one hand, theyloosen or decornpact chromatin structure." and on the other hand, they recruit multiprotein effector complexes that directly regulate chromatin structure and function.44-47 These effectors aretargeted to chromatin based on the properties ofthe various chromatin binding modular proteindomains that they contain. For example, bromodomain-containing proteins bind acetylatedhisrones, whereas chromodomain and PHD finger motif-containing proteins recognize histonesmethylated at different lysine residues.39•42 Proteins or protein complexes with combinations ofchromatin binding domains may be preferentially recruited to nudeosomes displaying morecomplex patterns ofhistone rnodificarion."

Chromatin structure can also be modulated through the activity ofATP-dependent chromatin-remodeling complexes, which use the energy derived from ATP hydrolysis to disrupthistone-DNA contacts.49

•50 These remodeling complexes can be categorized into three major

families (SWI2ISNF2, ISWI and Mi21 CHD) based on the structures oftheir ATPase catalyticsubunit. The function and the recruitment of these complexes are determined by the domainstructures oftheir ATPase subunit and ofadditional subunits within the complex. For example,the BRM and BRG 1 catalytic subunits ofthe SWI2ISNF2 complex contain a bromodomain thattargets the complex to acerylated histones to positively or negatively regulate gene expression.The SANT and SLIDE domains ofISWI recognize histone tails and linker DNA, respectively.ISWI complexes playa central role in the ordering and spacingofnucleosomes to promote generepression. However, when ISWI is part ofthe PHD finger motif-containing NURF complex,its recruitment to H3 K4 trimethylared histones is required for proper gene activation." TheATPase subunit Mi-2 contains a chromodomain that directs binding to nucleosomes; Mi-2complexes generally contain HDAC subunits that contribute to transcriptional repression.

The exact mechanisms by which these complexes induce nucleosome remodeling are not fullyelucidated.50 Several srudies have shown that changes in nucleosome structure can bepromotedthrough an initial DNA translocation within the nucleosome which is then propagated as a DNAbulge around the histone octamer, This leads to nucleosome sliding with respect to the DNAsequence and , in some cases, to nucleosome disassembly," Ultimately, nucleosome remodelingcauses an increase in accessibility ofthe underlying DNA to transcription factors.

Gene expression is initiated by the binding of transcriptional activators to promoter regions.41.52 This is then followed by the sequential recruitment ofhistone-modifying enzymes andchromatin-remodeling complexes to the promoter. These chromatin regulators act cooperativelyto disrupt promoter nucleosomes and to allow the formation ofa stable preinitiation complex atthe promoter. For example, acetylation ofpromoter nucleosomes can help to recruit the SWI/SNFchromatin-remodelingcomplex, resulting in nucleosome displacement or disassemblyprior to geneactivadon." Promoter-engaged RNA pol II complexes subsequently transit through chromatinto mediate transcriptional elongation.

Transition into the elongation phase oftranscription (clearance ofRNA pol II from the promoter) requires phosphorylation ofthe RNA pol II carboxy-terminal domain (CTD).54 The RNApol II CTD represents a platform for the recruitment ofhistone modifyingcomplexesand elongation factors that allow the polymerase to efficiently transcribe through chromatin. For instance,the PAF complex is required for the recruitment ofSet P? (responsible for H3 K4 methylation)

Germline Transcription:A Key RegulatorofAccessibility andRecombination 97

and Rad6/Bre56 (responsible for H2B monoubiquitination) to elongatingRNA pol II, whereasSet257 (responsible for H3 K36 methylation) binds directly to the phosphorylated RNA pol IICTD. Among the chromatin modifications introduced by thesecomplexes, H3 K4 methylationrecruits additional chromatin regulators that in turn remodel nucleosomes within transcribedregions.45-47 Histone acetylationand monoubiquitination stimulate transcription by promotingnucleosomedisassembly aheadof the polymerase.58.59 Finally, H3 K36 methylationand histonedeacerylation reestablish the chromatinstructure byrepositioningnucleosomes afterRNA pol IIpassagef'()·61 Several histonechaperones that travelwith RNApolII alsomediatehistoneH2A/H2Band H3/H4 evictionaheadof the polymerase and the subsequent reassembly of nucleosomes inits wake.62.63 In addition, chromatin-remodeling complexes are targeted to transcribedcodingregionsto facilitatethe passage of RNApol II through the nucleosomal barrier.64.65Thus, chromatindisruption is an intrinsicfeatureof the process of transcription.

Regulation ofV(D)J Recombination by TranscriptionWehaverecentlyaddressedthe roleoftranscriptionalelongationin the controlofaccessibility

for recombinationat the mouseTcra locus.TheJa regionnear the 3' end of the Tcra locuscontains61Ja segments that span 65 kilobases (Fig.1A).66 AllVa toJa recombinationeventsdepend onthe Tcra enhancer (Ea) locatedat the extreme3' end of the locus." PrimaryVa to Ja rearrangements are targeted to the most 5'Ja segments by the activityof two germlinepromoters that areactivatedby Ea.68.69These promoters,TEA at the 5' end of the Ja arrayand theJa49 promoter15kb downstream, drivethe expression ofgermlinetranscriptsthat extendacross the entireJa-Caregion.Subsequently, secondaryVa toJa rearrangement eventsreplace theprimaryrearrangementsresultingin the useof progressively more 5'Va and 3'Ja segments.P" Secondaryrearrangementsare thought to be targeted toJa segments downstreamof a primaryVo]« rearrangement by thepromoter of the rearrangedVa genesegment."

Studiesperformedon TEA-deleted micehaveshown that TEA controls rearrangements andchromatin structure across a 12 kb region encompassing the most 5' Ja segments(Ja61-Ja52)(Fig. 1B).68.69.73 Within the TEA-dependent accessibility region,Ja58, Ja57 andJa56 segmentsare associated with promoters whose activation depends on TEA.69 However, Ja61 , Ja53 andJa52 lack their own promoters and are located at a distancefrom the nearestupstreampromoters.Ja61 is situated 1.7kb downstreamof TEA and Ja53 and Ja52 are located 3.4 and 6.1 kb,respectively, from the nearestupstreampromoter atJa56. Wewonderedwhether accessibility ofthesepromoter-distalJa segments might depend on transcription from upstreampromoters.

Tocritically test the regulatoryfunction of transcriptionin Ja segmentrecombination,weusedhomologousrecombinationto introduceastrongtranscriptionterminatorcassette downstream ofJa56 (Fig. 1C),?4The terminator iscomposedof asetof fourpolyadenylation sitesfollowed byanarrayof twelve bacteriallac operators.Thelacoperatorsarethought to functionasstrongpausesitesfor RNA pol II,whichwouldincrease the efficiencyofboth polyadenylation and termination.Theintroduced transcriptionterminatorwasshownbyreverse transcriptase polymerase chainreactionand bynuclearrun-on to imposean effective blockto RNA pol II passage through theJa53-Ja52region in vivo,virtuallyeliminatinggermlinetranscription across these segments. Moreover, thetranscriptionalblock causedan 87% reduction in recombination at Ja53 and a 67% reductionat Ja52. As expected,recombinationof upstreamand downstreamJa segmentswasunaffected,due to the presenceof additionalpromotersassociated with thesegenesegments. Transcriptionalblockadealsoledto veryspecific alterationsin chromatinstructureatJa53 andJa52.AcetylationofhistonesH3 and H4 wasunaffected. However, dimethylationand trimethylationofhistone H3K4wereboth reducedat theseJa segments.Thesefindingsprovidedthe firstdirect experimentalproofthat transcriptionalelongationacross promoter-distalgenesegments isrequiredfor normalratesof recombination in vivoand suggested a potential link between transcription-dependentH3 K4 methylationand accessibility to the recombinase.

Thetranscription terminatorapproachalsoprovidedinsightsinto the mechanisms that regulategermlineJa promoter activity. TheJa regioncontainsa series of crypticpromoters locatedin the


A Ja Ca Ea

TEA 61 58 5756 53 52 504948 47464544 43 42 40 393837

- .. ~ ~~ I Itl III I I I III/~

B

-{

C

-' 1)2 ' 1 Itl ~I~ I I ~ II ~/~

0 ITenninator I

'-----'2 kb

Figure 1. Regulation of Tera locus Ja segment recombination by transcription. A) Wild-typelo-Cc region, identifying active promoters (bent arrows) and enhancer Ea(filled oval). B)TEApromoter-deleted Ie-Cc region.73.7S Shading identifies region of reduced transcription, reducedhistone acetylation and methylation and reduced recombination. C) Terminator introductiondownstream of Ja56.74 Shading identifies region of reduced transcription, reduced histonemethylation and reduced recombination. D) Terminator introduction downstream of TEA/SShading identifies region of reduced transcr iption, reduced histone acetylation and methylation and reduced recombination.

central portion ofthe array (Ja47-Ja37) that become activated when the TEA promoter isdeleted(Fig. lB) .69 Like TEA deletion, transcriptional blockade also caused the de-repression of thesecrypticJa promoters, showing that their activity is normally suppressed through a transcriptionalinterference mechanism (Fig. 1C),74 Suppression ofthese promoters by transcriptional interferenceis likely to be important becausewhen these promoters are active they target rearrangement eventsto the central]« segments and lead to disordered usage ofJa segments.69

More recently, we extended this analysis by introducing the transcription terminator downstream ofthe TEA promoter (Fig. ID).75 In this location the terminator almost completelyeliminared recombination of]a61 throughJa52, mimickingthe phenotype ofTEA-deleted mice.Thisoccurs in part because TEA transcription is required for the activity ofpromoters associated with]a5S,Ja57 andJa56, even as it suppresses the activity ofcenrral]« promoters.

In conclusion, transcriptional elongation can stimulate accessibility for recombination intwo different ways. On one hand, transcription can directly provide long-range accessibility atpromoter-distal RSSs, probably through alterations in chromatin structure associated with theprocess of transcriptional elongation. On the other hand, transcription can regulate the activityofadditional promoters which themselves can provide accessibility to nearby RSSs.

We think that transcription islikely to contribute to the regulation ofRSS accessibilityat otherimmunoglobulin and T-cell receptor loci, but that the extent ofcontribution may depend on theproximity ofthe RSS to the nearest promoter. D~1 RSS accessibility may occur independent of

GermlineTranscription: A Key RegulatorofAccessibility and Recombination 99

transcription because the D~1 RSSis closeenough to the promoter that it can be influencedbychromarin remodelingcomplexes that are targeted to the promoter. V segment promoters aregenerally located three to four nucleosomes away from the V segmentRSS.Therefore, accessibility at these RSS might be influencedby both transcription-dependent and promoter-targetedchromatin remodeling. At the Tcrb locus.jp l segments aresituated 1-2kb from PD~1. AnalysisofPD~ l -deletedmiceindicatedthatJ~1.6accessibilitydependsat leastin part on PD~1,32 andwesuggest that this islikely to beaconsequence of transcriptionfromPD~1.Longantisense intergenic19h transcripts may similarly regulate chromatin remodelingevents that influenceaccessibilityacross the VHand DWH regions.6.76

Future DirectionsThestudiesoutlinedabove fallshortofclarifyingtheprecise mechanisms bywhichtranscription

can stimulateaccessibility for V(D)J recombination. It will be important to address thesemechanistic issues in furure studies.Specific histone modifications introduced during transcriptionalelongationarelikely to stimulateaccessibility in avarietyofways. Forexample, wenoted adramaticsuppression ofH3 K4trimethylationatJaS3 andJaS2 asaconsequence of transcriptionalblockade. H3 K4 trimethylationcan recruit PHD fingermotif-containingproteins to active genes,",Sincethis domain ispresent in several histone-modifyingand chromatin remodelingcomplexes.the recruitment of theseadditional activitiesonto chromatin might in turn be important for accessibility.46.47.59 It isalsonotablethat the C-terminus ofRAG2 containsa PHD fingermotif thatis important for recombinase activityand that mediates interactions with histones.77,78 Hencethere couldbe a direct mechanistic link between transcription-dependentH3 K4 trimethylationof nucleosomes and recombinase recruitment. H3 and H4 acetylationwas not suppressed byblockadeof transcription at JaS6 , indicaringthat histone acetylationis not sufficient for accessibility. However, acetylationwassuppressed when transcription wasblockedacross the entire S'regionbyterminatorinsertiondownstream ofTEA,75 Transcription-dependenthistoneacetylationmight influence accessibility bypromoting generalchromatin opening, sincehistone acetylationis known to reducethe compactionof nucleosome arrays," Several histone modifications introduced during transcription, includingacetylationand H2B monoubiquirylarion, are thought tofunction together with histone chaperones to promote the transient disassembly ofnucleosomes(for example,evictionofH2A/H2B dimersor ofH3 and H4) that isassociated with RNA PolIIpassage.58.79 Transientnucleosome disassembly couldalsoplaya rolein RAG accessibility, particularlyifRAG itselfcouldassociate with RNA pol II and couldbedelivered at the appropriatetime.Regardless, transient nucleosome disassembly might allowaccess ofother relevantfactorsto theircognateDNA sequences. Forexample, PaxS binding to VHsegments canmediatethe recruitmentof RAG to VHsegments and promote VHto DWH recombinatlon.'? Transcription-dependentchromatin remodelingmight directlystimulate PaxS access to VHsegments and thus indirectlypromote RAGrecruitment.Morethan likely, transcriptionimpactsrecombinationinseveralways.Thus although at one level our studiesclosea longstandingdebate in the accessibility field, theyalso raisemany new questions and suggest avenues for additional research that mayultimatelycontribute to a detailedmechanisticunderstandingofaccessibility.

AcknowledgementsThe authors acknowledge support of the National Institutes of Health.

References1. Cobb RM. Oestreich KJ. Osipovich OA et al. Accessibility control of V(D)J recombination. Adv

Immuno12006: 91:45-109.2. Yancopoulos GD. Alt FW Developmentally controlled and tissue-specific expression of unrearranged

VH gene segments. Cell 1985: 40:271-281.3. Fondell JD, Marcu KB. Transcription of germ line Va segments correlates with ongoing T-cell receptor

a-chain rearrangement. Mol Cell Bioi 1992: 12:1480-1489.4. Goldman JP, Spencer DM, Raulet DH. Ordered rearrangement of variable region genes of the T-cell

receptor y locus correlates with transcription of the unrearranged genes. J Exp Med 1993; 177:729-739.


5. Schlissel MS. Corcoran LM. Baltimore D. Virus-transformed prell-cells show ordered activation but notinactivation of immunoglobulin gene rearrangement and transcription.] Exp Med 1991; 173:711-20.

6. Bolland D]. Wood AL. Johnston CM et al. Antisense intergenic transcription in V(D)] recombinat ion.Nat Immunol 2004; 5:630-637.

7. Blackwell TK. Moore MW; Yancopoulos GD et al. Recombination between immunoglobulin variableregion gene segments is enhanced by transcription. Nature 1986; 324:585-589.

8. Schlissel MS. Baltimore D. Activation of immunoglobulin kappa gene rearrangement correlates withinduction of germline kappa gene transcription. Cell 1989; 58:1001-1007.

9. Olrz EM. A1t FW; Lin WC et al. A V(D)J rccombinase-inducible B-celliine: role of transcriptionalenhancer elements in directing V(D)] recombination . Mol Cell Bioi 1993; 13:6223-6230.

10. Sun T. Storb U. Insertion of phosphoglycerine kinase (PGK)-neo 5' of]A.l dramatically enhances V]A.lrearrangemenr, ] Exp Med 2001; 193:699-712.

II. Romanow W]. Langerak AW;Goebel P et al. E2A and EBF act in synergy with the V(D)J recombinaseto generate a diverse immunoglobulin repertoire in nonlymphoid cells. Mol Cell 2000; 5:343-353.

12. Ghosh ]K. Romanow W] . Murre C. Induction of a diverse Tvcell receptor 1jI/6 repertoire by thehelix-loop-helix proteins E2A and HEB in nonlymphoid cells. ] Exp Med 2001; 193:769-775.

13. Casellas R. Jankovic M. Meyer G et al. OcaB is required for normal transcription and V(D)J recombination of a subset of immunoglobulin kappa genes. Cell 2002; 110:575-585.

14. Bertolino E. Reddy K. Medina KL et al. Regulation of interleukin 7-dependent immunoglobulinheavy-chain variable gen e rearrangements by transcription factor STAT5. Nat Immunol 2005;6:836-843.

15. YeSK. Agata Y, Lee HC et al. The IL-7 receptor controls the accessibility of the TCRy locus by Stat5and histone acetylation. Immunity 200 1; 15:813-823.

16. Zhao H. Nguyen H. Kang J. Interleukin 15 controls the generation of the restricted T-cell receptorrepertoire of y6 intestinal intraepitheliallymphocytes. Nat Immunol 2005; 6:1263-1271.

17. Hesslein DG. Pflugh DL. Chowdhury D et al. Pax5 is required for recombination of transcribed.acerylated, 5' IgH V gene segments. Genes Dev 2003; 17:37-42.

18. Fuxa M. Skok ] , Souabni A et al. Pax5 induces V-to-D] rearrangements and locus contraction of theimmunoglobulin heavy-chain gene. Genes Dev 2004; 18:411-422.

19. Zhang Z. Espinoza CR. Yu Z et al. Transcription factor Pax5 (BSAP) transactivates the RAG-mediatedVH-to -D]H rearrangement of immunoglobulin genes. Nat Immunol 2006; 7:616-624.

20. Okada A. Mendelsohn M. Alt F. Differential activation of transcription versus recombination oftransgenic T-cell receptor ~ variable region gene segments in B and T lineage cells. JExp Med 1994;180:261-272.

21. Fernex C. Capone M. Ferrier P. The V(D)J recombinational and transcriptional activities of theimmunoglobulin heavy-chain intronic enhancer can be mediated through distinct protein-binding sitesin a transgenic substrate. Mol Cell Bioi 1995; 15:3217-3226.

22. Tripathi RK. Mathieu N. Spicuglia S er al. Definition of a T-cell receptor ~ gene core enhancer ofV(D)J recombination by transgenic mapping. Mol Cell Bioi 2000; 20:42-53.

23. ]ia ] . Kondo M. Zhuang Y. Germline transcription from T-cell receptor V~ gene is uncoupled fromallelic exclusion. EMBO J2007; 26:2387-2399 .

24. Jackson A. Kondilis HD. Khor B et al, Regulation ofT-cell receptor ~ allelic exclusion at a level beyondaccessibility. Nat Immuno12005; 6:189-197.

25. Senoo M. Wang L. Suzuki D er al. Increase ofTCR V~ accessibility within E~ regulatory region influences its recombination frequency but not allelic exclusion.] Immuno12003; 171:829-835.

26. Stanhope-Baker P, Hudson KM. Shaffer AL et al. Cell type-specific chromatin structure determines thetargeting ofV(D)] recombinase activity in vitro. Cell 1996; 85:887-897.

27. Angelin-Duclos C, Calame K.Evidencethat immunoglobulin V"D] recombination docs not require germline transcription of the recombining variable gene segment. Mol Cell Bioi 1998; 18:6253-6264.

28. Cherry SR. Baltimore D. Chromatin remodeling directly activates V(D)] recombination. Proc Nad AcadSci USA 1999; 96:10788-10793.

29. Whitehurst CEo Chattopadhyay S, Chen J. Control ofV(D)] recombinational accessibility of the D~1

gene segment at the TC~ locus by a germline promoter. Immunity 1999; 10:313-322.30. Mathieu N. Hempel WM . Spicuglia S et aI. Chromatin remodeling by the Tscell receptor (TCR)-~ gene

enhancer during early T-cell development : Implications for the control ofTCR-~ locus recombination.J Exp Med 2000; 192:625-636.

31. Spicuglia S. Kumar S. Yeh]H et al. Promoter activation by enhancer-dependent and -independent loadingof activator and coactivator complexes. Mol Cell 2002; 10:1479-1487.

32. Oestreich K], Cobb RM. Pierce S et aI. Regulation ofTC~ gene assembly by a promoter/enhancerholocomplex. Immunity 2006; 24:381-391.

Germline Transcription:A Key RegulatorofAccessibility andRecombination 101

33. SikesML, Meade A, Tripathi Ret al. Regulation ofV(D)J recombination: a dominant role for promoterpositioning in gene segment accessibility. Proc Nat! Acad Sci USA 2002; 99:12309-12314.

34. Osipovich 0 , Cobb RM, Oestreich KJ et al. Essential function for SWI-SNF chromatin remodelingcomplexes in the promoter-directed assembly of Tcrb genes. Nat Immunol 2007; 8:809-816.

35. Misteli T. Beyond the sequence: cellular organization of genome function . Cell 2007; 128:787-800.36. Khorasanizadeh S. The nucleosome: from genomic organization to genomic regulation . Cell 2004;

116:259-272-37. Tremethick DJ. Higher-order structures of chromatin : the elusive 30 nm fiber. Cell 2007;

128:651-654 .38. Kouzarides T. Chromatin modifications and their function . Cell 2007; 128:693-705.39. Lee KK. Workman JL. Histone acetyltransferase complexes: one size doesn't fit all. Nat Rev Mol Cell

Bioi 2007; 8:284-295.40. Shi Y,Whetstine JR . Dynamic regulation of histone lysine methylation by demethylases, Mol Cell 2007;

25:1-14.41. Li B. Carey M. Workman JL. The role of chromatin during transcription. Cell 2007; 128:707-719.42. Ruthenburg AJ, Allis CD, Wysocka J. Methylation of lysine 4 on histone H3 : intricacy of writing and

reading a single epigenetic mark. Mol Cell 2007; 25:15-30.43. Shogren-Knaak M, Ishii H , Sun JM et al. Histone H4-KI6 acetylation controls chromatin structure and

protein interactions . Science 2006; 311:844-847.44. Lachner M. O'Carroll D. Rea S er al, Methylation of histone H3 lysine 9 creates a binding site for HP 1

proteins . Nature 2001; 410:116-120.45. Shi X, Hong T, Walter KL et al. ING2 PHD domain links histone H3 lysine 4 methylation to active

gene repression. Nature 2006; 442:96-99 .46. Taverna SD. llin S, Rogers RS et al. Yngl PHD finger binding to H3 trimethylaced at K4 promotes

NuA3 HAT activity at K14 of H3 and transcription at a subset of targeted ORFs. Mol Cell 2006;24:785-796.

47. Wysocka J. Swigut T. Xiao H et al, A PHD finger of NURF couples histone H3 lysine 4 trimethylationwith chromatin remodelling. Nature 2006; 442:86-90.

48. Li B, Gogol M, Carey M et al. Combined action of PHD and Chromodomains directs the Rpd3SHDAC to transcribed chromatin. Science 2007 ; 316:1050-1054.

49. Saha A. Wittmeyer J. Cairns BR. Chromatin remodelling : the industrial revolution of DNA aroundhistories. Nat Rev Mol Cell Bioi 2006; 7:437-447.

SO. Smith CL, Peterson CL. ATP-dependent chromatin remodeling. Curr Top Dev Bioi 2005;65:115-148 .

51. Bruno M. Flaus A. Stockdale C et al, Histone H2A/H2B dimer exchange by ATP-dependent chromatinremodeling activities. Mol Cell 2003; 12:1599-1606.

52. Workman JL. Nucleosome displacement in transcription. Genes Dev 2006; 20:2009-2017.53. Hassan AH. Neely KE. Workman JL. Histone acetyltransferase complexes stabilize SWI/SNF binding

to promoter nucleosomes. Cell 2001; 104:817-827.54. Sims R] 3rd, BelotserkovskayaR. Reinberg D. Elongation by RNA polymerase 11 : the short and long

of it. Genes Dev 2004; 18:2437-2468.55. Krogan N], Dover ]. Wood A et al. The PafI complex is required for histone H3 methylation by

COMPASS and Dotlp: linking transcriptional elongation to histone methylation. Mol Cell 2003;11:721-729.

56. Wood A. Schneider ]. Dover] et al. The PafI complex is essential for histone monoubiquitination bythe Rad6-Brel complex, which signals for histone methylation by COMPASS and Dorlp. J Bioi Chern2003; 278:34739-34742.

57. Xiao T, Hall H , Kizer KO ec al. Phosphorylation of RNA polymerase 11 CTD regulates H3 methylationin yeast. Genes Dev 2003; 17:654-663.

58. Govind CK, Zhang F,Qiu H et al . Gcn5 promotes acetylation, eviction and methylation of nucleosomesin transcribed coding regions. Mol Cell 2007; 25:31-42.

59. Pavri R, Zhu B. Li G et al. Histone H2B monoubiquitination functions cooperatively with FACT toregulate elongation by RNA polymerase 11. Cell 2006; 125:703·717.

60. Carrozza M]. Li B. Florens L et al. H istone H3 methylation by Set2 directs deacetylation of codingregions by Rpd3S to suppress spurious intragenic transcription. Cell 2005; 123:581-592.

61. Joshi AA, Strohl K.Eaf3 chromodomain interaction with methylated H3-K36 links histone deacerylationto Pol II elongation. Mol Cell 2005; 20:971-978.

62. Belotserkovskaya R, Oh S, Bondarenko VA er al. FACT facilitates transcription-dependent nucleosomealteration . Science 2003; 301:1090-1093.

63. Schwabish MA, Struhl K. AsfI mediates histone eviction and deposition during elongation by RNApolymerase 11. Mol Cell 2006; 22:415-422 .


64. Carey M, Li B, Workman JL. RSC exploits histone acetylation to abrogate the nucleosomal block toRNA polymerase II elongation. Mol Cell 2006; 24:481-487.

65. Corey LL, Weirich CS, Benjamin IJ et al. Localized recruitment of a chromatin-remodeling activity byan activator in vivo drives transcriptional elongation . Genes Dev 2003; 17:1392-1401.

66. Krangel MS, Carabana J, Abarrategui I et al. Enforcing order within a complex locus: current perspectives on the control of V(D)J recombination at the murine T-cell receptor a/~ locus. Immunol Rev2004; 200:224-232.

67. Sleckman BP, Bardon CG. Ferrini R ec al. Function of the TCRa enhancer in ap and y~ Tscells.Immunity 1997; 7:505-515.

68. Villey I, Caillol D, Selz F et al. Defect in rearrangement of the most 5' TCR-Ja following targeted deletion ofT early alpha (TEA) : implications for TCRa locus accessibility. Immunity 1996; 5:331-342.

69. Hawwari A, Bock C, Krangel MS. Regulation ofT-cell receptor-a gene assemblyby a complex hierarchyof germline Ja promoters. Nat Immuno12005; 6:481-489.

70. Buch T, Rieux-Laucat F, Forster I et al. Failure of Hr-speclfic rhymocytes to escape negative selectionby receptor editing . Immunity 2002; 16:707-718.

71. Wang F. Huang CY, Kanagawa O. Rapid deletion of rearranged T-cell antigen receptor (TCR) Va-Jasegment by secondary rearrangement in the thymus : role of continuous rearrangement of TCRachain gene and positive selection in the Tvcell repertoire formation . Proc Nad Acad Sci USA 1998;95:11834-11839.

72. Hawwari A, Krangel MS. Role for rearranged variable gene segments in directing secondary T-cellreceptor-a recombination. Proc Nat! Acad Sci USA 2007; 104:903-907.

73. Mauvicux L, Villey I, de Villartay J-P. TEA regulates local TCR-Ja accessibilitythrough histone acetylation . Eur J Immunol 2003; 33:2216-2222.

74. Abarraeegui I. Krangel MS. Regulation of Tvcell receptor-o gene recombination by transcription. NatImmuno12006; 7:1109-1115 .

75. Abarrategui I. Krangel MS. Noncoding transcription controls downstream promoters to regulate T-cellreceptor a recombination. EMBO J 2007; 26: 4380-4390.

76. Bolland DJ, Wood Ai, Afshar R et al. Antisense intergcnic transcription precedes Igh D-co-J recombination and is controlled by the intronic enhancer Ell. Mol Cell BioI 2007; 27:5523-5533 .

77. Elkin SK, Ivanov D, Ewalt M et al. A PHD finger motif in the C terminus of RAG2 modulates recombination activity. J BioI Chern 2005; 280 :28701-28710.

78. West KL, Singha NC, De Ioannes P et al. A direct interaction between the RAG2 C terminus and thecore histones is required for efficient V(D)J recombination. Immunity 2005; 23:203-212.

79. Tanny lC. Erdjumenr-Bromage H. Tempst P et al. Ubiquitylation of histone H2B controls RNApolymerase II transcription elongation Independently of histone H3 methylation. Genes Dev 2007;21:835-847.

CHAPTER 9

Dynamic Regulation ofAntigenReceptor Gene AssemblyLance R. Thomas, RobinMilleyCobbandEugene M.Ohz"

Abstract

Ahallmark featureof adaptiveimmunity is the production oflymphocytes bearingan enormous repertoireofreceptorsfor foreignantigens.Thisrepertoireisgeneratedearlyin BandT-cell development by the processofV(D)] recombination, which randomly assembles

functional immunoglobulin (Ig) and T-cell receptor (TCR) genes from large artays of DNAsegments.Precursor lymphocytes must target then retarget a singleV(D)] recombinaseenzymeto distinct regionswithin antigen receptor loci to guide lymphocyte development and to ensurethat each mature Band T-cellexpresses only a singleantigen receptor specificity. Proper targetingofV(D)J recombinaseisalsoessentialto avoidchromosomalaberrations that result in lymphoidmalignancies. Earlystudies suggestedthat changesin the specificity ofV(D)J recombination areachievedbydifferentiallyopeningor closingchromatin associatedwith Igand TCRgenesegmentsat the proper developmental time point. This accessibility model has been extended significantlyin recent yearsand it has becomeclear that control mechanismsgoverningantigen receptor geneassembly are multifaceted and varyfrom locusto locus.In this chapterwe reviewhow geneticandepigeneticmechanismsaswellaswidespreadchangesin chromosomalconformation synergize toorchestrate the diversification ofgenesencoding B and T-cell antigen receptors.

IntroductionOne triumph ofvertebrateevolution is the developmentofan adaptiveimmune system,which

recognizes and eliminates a continually changing spectrum of pathogens. To accomplish thisfeat,mammalshaveevolveda "brute-force" approach to adaptiveimmunity in which developinglymphocytes assemble an enormous repertoire ofantigen receptor genes(>108) . Theseimmunoglobulin (Ig) and T-cellreceptor (TCR) genesaregeneratedfrom largearraysofgenesegmentsbya unique processofsomaticDNA recombination, calledV(D)] recombination. As a result, eachprecursor B- or T-Iymphocytebears a unique antigen receptor that, followingnegativeselectionto delete autoreactiveclones,will recognizeits signature spectrum offoreign antigens. However,receptor diversification byaprocessthat altersthe genomeofsomaticcellscomesat a cost. Geneticdefects that compromiseanystep of the complexV(D)] recombination mechanismblock antigenreceptor gene assembly and lymphocyte development, resulting in a severe combined irnmunodefidency.'fAlternatively, aberrant targeting of the recombination apparatus (recombinase) toregionsofthe genome harboring oncogenesleadsto chromosomal translocations that underlie amajority oflymphoid tumors (e.g.,leukemiasand lymphomas).' Although the basicmechanismsof the V(D)J recombination reaction havebeen worked out in great detail,weare still unraveling

·CorrespondingAuthor: Eugene M. Dltz-Department of Microbiology and Immunology,Vanderbilt University, Nashville TN, 37232, USA. Email: [email protected]

V(D)JRecombination, edited by Pierre Ferrier.©2009 Landes Bioscienceand Springer Science+BusinessMedia.


the geneticand epigenetic strategies employed bydeveloping lymphocytes to differentially targetrecombinase activityat specific regionswithin antigen receptor loci. In this chapter,we reviewour current understandingof these regulatorystrategies. One emerging theme from recentstudiesis that only a subsetof control mechanisms is broadlyemployed at all antigen receptorloci.'Indeed,apotpourri of regulatorystrategies governs the targetingofdistinctgenesegmentclustersby recombinase. Manyofthe lessons learnedthrough the studyof antigenreceptorgenecontrolare broadly applicable to the dynamic regulationof complex geneticloci during development,cellularactivation and cellular differentiation.

Developmental Control ofV(D)J RecombinationThe enzymatic components ofV(D)J recombinase are products of the Recombination

Activating Genes1and2 (RAG-II2), whichareexpressed specificallyinprecursorlymphocytes.5,6

The RAG-l 12 complex targetsrecombinationsignalsequences (RSSs) that flankV, D andJ genesegments within all antigen receptor loci.' The RSSs are composedof a palindromichepramer,which abuts each codingsegment, a nonconservedspacerof 12 or 23 basepairs (bp) in lengthand an AT-richnonamer,'The molecularconstraintsofV(D)J recombinationincludea strict requirement for synapsis betweenthe RAG-l 12 complex and two RSSs, oneofwhichmustharbora 12 bp spacerand the other a 23 bp spacer (the 12/23 rule). Once the recombinase formssucha synapse, it introducesdouble-stranded DNA breaksprecisely at the junction betweenthe twocompatibleRSSs and the codingregionoftheir adjacentgenesegments.RAG-mediatedcleavagethus generates two signal ends, which contain the RSSs and interveningDNA and two codingends,which contain the genesegments and the remainderof the brokenchromosome. The twotypes of ends are then sealedseparately by the cellular DNA repairmachineryto createa signaljoin-usually in the formofan extrachromosomal circle containingthe two RSSs-and a codingjoin, which fuses the two selected genesegments irreversibly in the cellulargenome,"

Seminalstudiesfrom the Alelaboratoryin the 1980sshowedthat a single recombinase, nowknown to be RAG-II2, could rearrange both Ig and TCR genesegments in transformedcells,"However, in vivo, the process ofV(D)J recombinationis tightlyregulatedat threemajorlevels.4

•10

Thefirstandmostobvious level istissue-specificity.Althoughbothprecursor B-andTdymphocyresexpress RAG-l/2, thymocytes only target TCR genes for recombination while Ig genes arespecifically targeted by developing B-cells. Second, all antigen receptor genes are assembled viaa stepwise process that is intimatelycoupledto the lymphocytedevelopmental program. Uponlineagecommitment, pro-Bvcells first target recombinase to the DHJH clusterwithin the Igheavychain (IgH) locus (Fig. lA). Following DH-+JH recombination, a pro-B-cell clonewillthen retargetrecombinase to the upstream VH clusterto generate aVHD HJH join." Ifthis coding join is in-frame, the cellexpresses IgH protein, which in combinationwith a surrogate lightchain forms a preB-cell receptor (pre-BCR).J2 If the first IgH allele is rearranged out-of-frame,the second allele is then targeted for VH-DHJH recombination. Once formed, the pre-BCRsignals for developmental progression of the pro-B-cell to the pre-Bstage, whichthen completesIg light chain (IgL) geneassembly in an ordered manner (firstVIC-JIC then VA.-JAo ifboth Igxalleles areour-of-frame].'!Thymocytes undergoan analogous developmental programto produceTCRa/~ T-cells usingthe following ordered rearrangement process (Fig. IB): (i) D~""'J~ then(ii)V~-D~J~ in CD4-CD8- (double-negative,DN) pro-T-cells and (iii) Va-Ja in CD4+CD8+(double-positive, DP) pre'I-cells.v'"

The third levelofcontrol is imposedon V(D)J recombination at the developmentaltransition between pro-BIT and pre-By'Ivcells. Followingexpression of the pre-BCR or pre-TCR,each developingprecursor cell must shut down further rearrangement at IgH or TCR~ loci,respectively, despite continued expression of recornbinase. Thisprocess, calledallelicexclusion,precludes the generation of two productive IgH or TCR~ rearrangements in a given cell.I4-16

Allelicexclusionis essentialfor ensuring the monospecificiry of antigen receptors on maturelymphocytes,a cardinal featureof the adaptiveimmune response. Together, the three levels ofcontrol imposed on V(D)J recombination assure expression of the proper antigen receptor by

DynamicR~gulation ofAntigenReceptor Gen~ Assembly 105

CLPA Early Late Large Small ImmaturePrOoB PrOoB Pre-B Pre-B B cell

(Q)..C6)..(6) prollferml;n(6)..(6)..(Q).J Pre-BCR IgU

G II 0 J V OJexpression V"-J,, expression

arm ne - H ~ HI V -JI . I ),. >..JIgH assembly Allelic IgLassembly

exclusion

OP (C04+COB+) SInglePre-T Positive

aM g/ OR""""':'08

Pre-TCRexpression V J, n- a,

TCRn assembly TCRexpression

ON(C04"COB-)PrOoT

CLPt:::::\ t:::::\ Prollt rallon

QI"QI ~Germllne 01l-.J1I .J

VlI-OIIJllI ,

TCR~ All lieassembly exclusion

B

Figure 1. Control of Ant igen receptor gene assembly during lymphocyte development. A)Precursor B-cell development and Iggene assembly.Mouse B-cell development proceeds in thebone marrow initiating from a common lymphoid progenitor (CLP), proceeding through to thepro-B, pre-B and immature B-cell stages. Eachdevelopmental step is guided by stage-specificrecombination of the B-cell receptor Ig genes and results in the expression of a signature IgMprotein on the surface of the immature B-cell. The order in wh ich specific IgH and IgL chaingene rearrangements occur is indicated below each cell type. Pre-BCR expression results ina pro liferative burst and feedback inhibition of further IgH gene rearrangement (allelic exclusion). B) Precursor T-cell development and TCR gene assembly. After migration of CLPsfromthe bone marrow, T-cell development in the thymus proceeds in a stepw ise fashion throughthe pro -T, pre-T and immature T-cell stages. As indicated, each stage is marked by the ordered rearrangement of TCR genes and the surface expression of CD4 and CDB coreceptormolecules. Pre-TCRreceptor expression signifies progression to the preT-cell stageand blocksfurther TCR~ gene rearrangement.

B versus Tvcells, coordinate lymphocyte development programs and maintain a single, uniquebinding signature for each lymphocyte clone.

Genetic Control ofRecombinase AccessibilityOnce it was discovered that all precursor lymphocytes employ a single recombinase to target

indistinguishable RSSs,a keyquestion became how clusters ofIg and TCR gene segments couldbe targeted with the observed tissue-, stage- and allele-specificity. An important clue came fromthe observation that unrearranged Ig gene segments are transcribed at the precisedevelopmentaltime points they are targeted for rearrangemenr.'P! Sincethis originalobservation, the correlationbetween germline transcription ofgene segments and their recombination has been extended toall antigen receptor loci4•19 and gavebirth to the accessibilitymodel for control ofV(D)J recombination. In its simplest form, this model invokeschanges in the accessibilityofgene segments tothe recornbinaseas the key mechanism by which precursor lymphocytes target specificportionsof an antigen receptor locus." For example, upon commitment to the T-cell lineage, the DP1Pcluster "opens up" and becomes accessible to RNA polymerase and RAG complexes.whereas the


TCRn and Igloci remainhidden from thesenuclearfactors (Fig.2).Theaccessibility model hasbeenconfirmedbynumeroussubsequentstudies,includingthosebySchlissel and colleagues whoshowed that infusion of recombinant RAG complexes into nuclei from precursor lymphocytescleaves onlygenesegments that are targeted byendogenousrecombinase at a particulardevelopmental stage,"

The connection betweengermlinetranscriptionand recombinationled AIt and colleagues totest whether common geneticelementscontrol both processes, perhaps by alteringlocus accessibility.22,23 At the time, it wasknown that transcription of allantigen receptor loci is regulatedbyenhancerelements, usually located at the 3' end of the locusand a series of promoters situatedupstreamofindividualgenesegments or 5' to a clusterofrelatedsegments (Fig.2).Targeteddeletion ofknown enhancersfrom Igand TCR lociseverely inhibits both germlinetranscriptionandrecombination of genesegments in cis.4.24-27 Thesefindings support a dual role for enhancersastranscriptionalregulatoryregionsand asaccessibility control elements(ACEs)for V(D)J recombination. Subsequentstudiesdemonstrated a similarrole for germlinepromoters, which haveamore localizedimpact on the accessibility of neighboringgenesegments (seebeloW).2S'31

ON Pro-T cell

OP Pre-T cell

~1:'!~11 ".,>:\ 1~

~~T -, ."\;-,-'1 (~-~'JI-......A_r __ "'",---'~J"'\_ .., 1~-A.-

Figure 2. Stage-specific activation of TCR genes. Schematic representation of TCRf\, TCRal6and IgH genes are depicted. Block arrows indicate promoter regions and circles representenhancer elements. The chromatin accessibility status of these regulatory elements at the ON(double-negative)pro-T-ell stageand the OP(double-positive) preT-celi stageisspecified (white"open," gray - "closed"). The shaded boxes overlying the loci indicate a closed configuration.

Dynamic Regulation ofAntigen Receptor Gene Assembly 107

The precise mechanisms by which ACEs control V(D}J recombination are still under studyand may differ significantly between regions within each antigen receptor locus. For example, anobvious mechanism by which promoters might serve as ACEs is via transcriptional readthrough ofdownstream gene segments, which is known to impact the configuration ofassociated chromatin.This appears to be the case at theJa cluster ofgene segments, which harbors a series ofpromotersdirecting transcription ofsegments up to several kb away.32 Disruption ofreadthrough transcription by placinga strong terminator within aJa cluster leaves recombination ofupstream gene segments intact but severely impairs rearrangement ofJa segments downstream ofthe transcriptionalterminator.33 In contrast, the D~1 germline promoter regulates recombinase accessibility of thesmallD~ IJ~1 cluster (1-2 kb) independent oftranscriptional readthrough ofthe gene segments.29

Taken together, these studies suggest that transcription may be dispensible for promoter-directedaccessibility ofproximal gene segments but may significantly augment recombinase accessibilityat more distal elements.

Chromatin Accessibility Control Mechanismsfor V(D)J Recombination

A major strategy ofeukaryotic cells for controlling DNA access to nuclear factors is throughchanges in the configuration of chromatin. Classical histology studies defined several forms ofchromatin that vary in their degree ofcompaction and, therefore, accessibility to nuclear factors.34

Heterochromatin is the most compacted form and normally is associated with transcriptionallysilent loci. Euchromatin is the most relaxed form and is usually rich in transcribed genes. In betweenthese two extremes lies facultative heterochromatin, which harbors many molecular signatures ofsilent chromatin, but unlike heterochromatin, is more easily converted to an open state. As such,facultative heterochromatin associates with genes that are under dynamic transcriptional control.The two building blocks of chromatin, nucleosomes and DNA, form a spool-like structure inwhich each nucleosome is wrapped by the DNA helix.35.36 Nucleosomes, in turn, are composedofan octamer ofhistones called H2A, H2B, H3 and H4. Histones H3 and H4 are characterizedby an N-terminal tail that is targeted for a broad panel ofcovalent modifications.

An emerging picture in gene regulation is that covalent modification of histones plays animportant role in determining the accessibility status ofassociated DNA. Indeed, the pattern ofhistone modifications is now thought to constitute a "code" that is recognized by other nuclearfactors to alter local chromatin accessibility.37 For example, most transcriptionally active genesassociate with nucleosomes harboring the following covalent modifications: acetylation ofH3-Lysine 9 (H3K9), acetylation of H4K8 and K12 and methylation of H3K4. Acetylatedhistones attract other nuclear factors containing bromodomains that further augment accessibility, including histone acetyltransferases (HATs) and nucleosome remodeling complexes.38

Transcriptionally inactive genes associate with chromatin lacking the aforementioned modifications, but instead display H3K9 and H3K27 methylation. These methylation marks arerecognized by proteins containing a chromodomain, which facilitate the formation of morehighly compacted chromatin (e.g., histone deacetylase-HDACs}.39 In addition to histones,methylation ofDNA at CpG dinucleotides correlates with levels ofchromatin accessibility andgene activity. In general, CpG dinucleotides are hypermethylated at transcriptionally silent locibut hypomethylated at active genes.4O The modified CpG dinucleotide is targeted by methylCpG-binding proteins (e.g., MeCPI), which have been mechanistically linked to gene repressionvia their association with HDACs and H3K9/K27 methyltransferasesY

Given the connection between transcription and V(D}J recombination, it seemed likely thatchromatin modifications play an important role in recombinase accessibility. In pioneeringstudies,Krangel and colleagues proved this hypothesis to be correct. They showed that H3K9 acetylationassociates with TCRa/o gene segments when they are actively undergoing recombination.42 Sincethis original report, numerous groups have shown that an identical pattern ofhistone and CpGmodifications characterize loci that are transcriptionallyor recombinationallyactive and a separateset ofmodifications decorate inert antigen receptor loci (Fig. 3}.4.43 Importantly, ACE deletions

108

POH DH JH JH JH JH

V(D)jRecombination

EIJ

Me Me Me Me Me

DP

H3K4ac f H3K9ac 'H3K9me t H4acMe CpGMe

I

Figure 3. Epigenetic regulation of recombinase accessibility. Schematic representation ofhistone modifications at the mouse heavy chain locus in pro-B and OP pro-T-cells. RSSs arerepresented as triangles (23 bp spacer black and 12 bp spacer white) flanking each DH and)H gene segment (rectangles). Chromatin associated with the DH)H cluster in pro-B-cells ismarked with an "accessibile" pattern of modifications (H3K4me, H3K9ac and H4ac). Thegermline promoter (diamond) together with acetylated histones recruit the SWI/SNF complex(oval), which remodels chromatin at neighboring gene segments to generate a recombinaseaccessible state in pro-B-cells. In DP thymocytes, the DHJH cluster is decorated with H3K9meand CpG methylation, two chromatin modifications that mediate repression.

from Ig or TCR loci convert histone modifications from an "accessible" to an "inaccessible" pattern.lI•44 In recent studies, we established a cause-effect relationship for one histone modificationin the control ofV(D)J recombination.os Specifically, we showed that recruitment of an H3K9methyltransfcrase inhibited recombination of stably integrated substrates, despite the presenceof requisite ACEs. Thus, H3K9 methylation is sufficient in this setting to repress recombinaseaccessibility. In n important new development, Oettinger and colleagues have shown that theplant homeodomain ofRAG-2 binds specifically to tri-methylated H3K4, a characteristic markoftranscriptionally active chromatin.46,47 Mutations in RAG-2 that disrupt H3K4me3 bindingalso attenuate recombinase activity, indicating that the interaction is functionally significant. Thus,the pattern ofhistone modifications at a given cluster ofgene segments may fine-tune its affinityfor recombinase, providing a level ofcontrol beyond sheer accessibility. Despite these advances, acause-effect relationship has yet to be established for any histone modification and the recombination potential ofgene segments at an endogenous antigen receptor locus.

An added layer of complexity to the "histone code" derives from the large collection ofnuclear factors that can potentially recognize each covalent modification. In this regard, an important role for H3 and H4 acetylation is thought to be the docking ofnucleosome remodelingcomplexes via bromodomains.38 In vitro, these complexes alter the conformation or position ofnUcleosomes on DNA, thereby altering the accessibility ofneighboring sites to nuclear factors.49

Certain complexes, such as SWI/SNF, augment accessibility, whereas other remodeling machinesnormally induce a higher degree ofcompaction.so At the heart ofeach remodeling complex is anATP-dependent motor that provides the energy for reconfiguring nucleosome conformation orsliding the nucleosome to a new position. In the case of SWI/SNF, the major ATP-dependent

DynamicRegulation ofAntigenReceptor Gene Assembly 109

subunits are called Brg1 and Brm. In the context ofantigen receptor gene assembly, early studieson chromatin remodeling were largely restricted to in vitro substrates. In one such trailblazingstudy. Oettinger and colleagues showed that Brg1 could reverse the block to recombinase accessibility when added to nucleosomal substrates." This group went on to show that Brg1 associateswith Ig and TCR loci only when they are poised to undergo rearrangement and are therefore ina recombinase-accessible state.43Most recently. we have triangulated the relationship betweenACE function. SWI/SNF association and recombinase accessibility.Recruitment ofBrgl to theendogenous TCRf} locus requires both the D~1 germline promoter and E~ enhancer elemenrs.VIn chromosomal substrates . ACE function ofthe germline promoter can be replaced completelyby artificial recruitment ofBrgl. Most importantly. RNAi-mediated ablation ofBrgl and Brm inprimary thymocytes abrogates both germline transcription and D~-]~ recombination. Togetherwith prior studies, these findings indicate that ACEs likely function to alter histone modificationpatterns within antigen receptor loci in order to recruit chromatin remodeling complexes thateither impart or impair recombinase accessibility (Fig. 3).

Control ofV(D)J Recombination by Nuclear CompartmentalizationRegulation ofgene accessibilityisa complex process that involvesnot only the covalent modifi

cation ofhistones and DNA, but also the localization oflarge genetic loci to distinct regions withinthe cell nucleus. In general, transcriptionally silent genes are located near the nuclear periphery,whereas expressed genes reside at a more central location within the nucleus.P Although theunderlying mechanisms for this effect remain to be established, recent experiments demonstratethat enforced compartmentalization ofa genetic locus to the inner nuclear membrane repressesits expression.53 With regards to antigen receptor loci, fluorescence in situ hybridization (FISH)analysis has provided insights into how subnuclear relocation may influence their assembly byV(D)] recombination during lymphocyte development. In pro -B-cellspoised to initiate DH -]Hrecombination, both the IgH and IgK alleles are preferentially positioned in the central portionofthe nucleus . In contrast, these loci occupy a perinuclear location in hematopoietic progenitorsand pro -T-cells, which do not rearrange Ig genes." Thus, tissue-specific activation ofIg loci mayinvolve their repositioning from the nuclear periphery in hematopoietic stem cells to a morecentral location upon commitment to the pro-B -cell stage (see below and Fig. 5).

Primary Activation ofAntigen Receptor Locifor D to J Rearrangement

V(D)] recombination at IgH and TCRf}loci progressesin a step-wisefashion, initiatingwith theassembly ofaD] join, followed by rearrangement ofa V gene segment to the existing DJ element.The genetic and epigenetic mechanisms leading to the crucial first step, activating DJ clusters forrearrangement, have been extensivelystudied for the TCRf} locus. Due to its compartmentalizedarchitecture, this antigen receptor locus serves as a tractable model to study the precise role ofACEs in tissue-specific activation of D-J recombination. The mouse TCRf} locus consists oftwo distinct D~J~ clusters, which are both under the control ofa single 3' enhancer element (E~).E~ function is specific for T lineage cellsand this control element becomes activated at the earliest stage of thymocyte development.v In addition. both D~J~ clusters harbor a single germlinepromoter located proximal to each D~ gene segment (PD~l and PD~2) , which are completelydependent on E~ for their function (Fig. 4).56.57Knockout studies in mice have shown thatE~provides a long-range ACE function to direct chromatin modifications, recombination and germlinetranscription at both D~]~ clusters.24.25.44 In contrast. targeted deletion ofthe germline promoterlocated directly upstream of the D~1 gene segment cripples rearrangement and transcription ofthe D~l]~l but not the D~2]~2 cluster, indicating that gerrnline promoters have a more localinfluence on chromatin accessibility."

Although the cis-and trans-elements involved in recombination are unique for individual loci,studies ofTCRf} have provided a mechanistic model for initial activation ofD] clusters for theirtargeting by V(D)J recombinase (Fig. 4) .28 First, thymocyte precursors activate an inherent ACE


PDf3' 013' Jf31(6) C~1 Ef3

PDf3' ° ,Ef3

Jf31(6) C~1

tEf3

PDfl' 03'

Ef3

Jf31(6)

tCf31

Cf31PDf3' Dfl' Jf31(6)

tD~J~ recombination and germline transcription

Figure 4. Ordered activation of DIl-JIl rearrangement . Initially, thymocyte precursors activatethe T-cell specific enhancer, Ell (circle) via the binding of transcription factors (shaded shapes).The ACEfunction of Ell mediates a spread of chromatin accessibility throughout the majority ofthe DIlJIl cluster, with the exception of the Dill-proximal region, which remains refractory tothis opening (gray shaded box and associated nucleosome, gray oval). The germline promoter,PDlll, becomes activated and binds transcr iption factors in an Ell-dependent manner. Onceactivated, these two distal regulato ry elements interact to form a PDIl/EIl holocomplex, whichin turn recruits chromatin remodeling complexes, including SWI/SNF. SWI/SNF remodelsnucleosomes associated with the Dill region, exposing the RSS and TATA box for DIl-JIlrecombination and germline transcription.

function of the Ef3 enhancer,which mediates the spread of partially accessible chromatin overa large region spanningboth Of3Jf3 clusters. Although the mechanisms are currentlyunknown,chromatinassociatedwith the 0f31genesegmentisinitiallyrefractoryto Ep-mediatedopening."However, enhancer-mediatedreorganizationofTCRf3 chromatinpermits binding of additionaltranscriptionfactorsto the OPI germlinepromoter.r' Theloadedpromoter and distalEpelementphysically interact,presumablyviafactorsbound to eachACE,generatingastableholocomplex."Thepromoter/enhancer holocomplexmayserveasastagingplatformto recruit SWI/SNF,whichin turn reorganizes localnucleosomes, especially thoseassociated with the 0131-RSS.52TheSWI/

DynamicRtgulationofAntigm Receptor GmtAsstmbly 111

SNF-mediated remodelinglikely exposes the underlyingTATA boxand RSSto initiate germlinetranscription and D~"""J~ recombination,respectivdy. Future studiesshould be directed at understandingthe generalityof this contingent mechanismfor recombinase accessibility at the IgHand other antigen receptorloci.

Long-Range ControlofV(D)} RecombinationChanges in chromatin accessibility playa key role in controlling recombination of D and J

segments, which areseparatedbyrelatively short distancesin theTC~ and IgH loci«20 kb). Incontrast,theVclusters areseparatedbymuchlongerstretchesofDNA fromtheirpartner D and/orJ segments (>1OOkb) at the IgH,TC~, 19K, Igt..and TCRn loci.Assuch,additionalmechanismsmaybe required to facilitateefficient V......(D)J recombination.Whereasmanyofthe control dements that regulateD......J recombinationhavebeen identified,the ACEs guidingV......DJ remainlargely unknown. Knockout experiments that deleteknown enhancerand gerrulineD promoterdements haveno effecton the transcriptionor accessibility ofdistalVgenesegments at either theIgH orTC~ 10cus.57'59Emergingevidence points to a requirementforV~ or VH promoters inmodulatinglocalchromatinaccessibility to recombinase. Deleeion of theV~13promoter cripplesgerrulinetranscriptionand rearrangement of that genesegment/?However, reporter geneassaysindicatethat transcriptionfromV~ andVH promotersisenhancer-dependent. Thus,theepigeneticregulationof the V geneclusters is under the control of unidentifiedenhancersthat reconfigurechromatin to mediate their recombinase accessibility.

Despite a clear requirement for changes in chromatin accessibility to trigger long-rangeV......(D)J recombination, epigenetic changes alone are not sufficient. Introduction ofthe TCRnenhancerproximalto aV~ genesegmentgreatlyaugmentsitschromatinaccessibility and germlinetranscriptionin DP thymocytes, adevelopmental stagewhereV~ segments arerepressed byallelicexclusion. However,enhancedchromatinaccessibility isinsufficient to driveV~--D~J~ recombination involving the targetedgenesegment," Thus, additional mechanisms are required to mediatethe long-rangeinteractionsbetweendistalV arrays and downstream(D)J regions.

Insights into the mechanisms that control chromosomaldynamics within the TC~ locusderivefrom studiesusingthree-dimensional FISH.Theseanalyses produced astoundinglinksbetween changes in subnuclearlocation, topographyandTC~ generegulation(Fig.5).16 Througha mechanismcalledlocuscontraction, the distalV~ clusterloopsto becomespatiallyproximaltothe D~J~ clusterin DN thymocytes but not inother tissues.62Thechanges inTC~ locustopologywereconfirmedusingmolecular analysis,whichrevealed physical associations not onlybetweenV~and DIU13 clusters but alsoamongstV13 genesegments themselves. A similarcontraction processregulates long-rangeinteractionsofIgH and IgK loci,whichspandistances of2.5 megabases and 3megabases, respectively, Three-dimensional FISH revealed that theVH regionofIgH isjuxtaposedwith the DHJH domainviaaloopingmechanism, whichoccursspecifically at the pro-B-cell stageof development.54.63.64 Likewise, the IgK locusundergoesactive contraction in preB-cells to bringVIC and ]x genesegments into spatialproxlmiry,"Thus, changes in locus topology appear to bea generalmechanismfor long-range control of recombinationbetween distant V and (D)J genesegments.Together, thesefindings suggest that novelcis-elements within both the V and the (D)J clusters mediatetheir physical association, perhapsformingan active hub wheremultipleVgenesegments are in spatialproximity to their partner (D)J segments(Fig. 5).

Allelic ExclusionThe specificity of immune responses is maintained by restrictingeach lymphocyte clone to

express asingleantigenreceptorgenecombination.Followingaproductiverearrangement on oneIgor TCR allele, precursorlymphocytesmust rapidlyinhibit rearrangementat the secondallele,a processcalledallelic exclusion. Indeed, transgenicmiceengineeredto express functionalIgH orTC~ proteinsin precursorlymphocytesfreezeV......DJ rearrangementat their correspondingendogenousalleles.66.67Thesefindings indicatethat thepre-BCR andpre-TCRcomplexes parricipatein a feedback mechanismthat specifically disrupts long-rangeV......(D)J recombination.

112

Common LymphoidProgen itor Pro-B Pro-B

V(D)j Recombination

Pre-B

Figure 5. Regulation of V(D)J recombination by chromosomal dynamics. Cartoon showingchanges in nuclear positioning and IgH locus topology during B-cell development. Prior toB lineage commitment, both IgH alleles are positioned at the nuclear periphery and remaininaccessible to recombinase in common lymphoid progenitors . Upon lineage commitment,both IgH loci migrate to a more central location in the pro-B-cell nucleus and undergo DH-.jHrecombination. IgH loci also contract via a looping mechanism to position the VH clusterinto spatial proximity with the DHjH fusions, permitting efficient VH-.DHjH recombination. If the first IgH allele undergoes a productive rearrangement, preBCR signaling mediatesa repositioning of the second, DHjH rearranged allele to associate with pericentromericheterochromatin (PHC).

Although the mechanisms are poorly understood, allelic exclusion is controlled at multiplelevels including (i) chromatin accessibility, (ii) locus topology and (iii) repositioningofloci topericentromeric heterochromatin. With regards to chromatin accessibility, hyperacetylation ofhistonesassociated with Vpgenesegments islost followingTCR/l expression and the transitionofthymoeytesfrom the DN to DP stageofdevelopment.'" Similarchanges in the pattern of histonemodifications suggest a lossof chromatin accessibility at VB genesegments upon expression ofIgH proteins," However, as stated above, allelic exclusion is still enforcedat a Vp genesegmentdespite the neighboringinsertion of a functional ACE that opens associated chromatin in DPthymocytes."

Similar to activation ofV-.(D)J rearrangement, changes in locus topology may play animportant role in shutting down this stageof antigen receptor gene assembly to enforceallelicexclusion. FISH analyses of thymocyte populations revealed that juxtaposition of the Vp andDP1P regionspersiststhroughout the DN stagebut the locus"decontracts" and becomeslinearat the DP stage,where VP-.DP1P recombination is suppressed by allelicexclusion.62 A similardecontraction process takes place at the IgH locus during the pro-B to preli-cell transition.f'In addition to distancingthe V and (D)J clustersby decontraction, the nonfunctional IgH andTCR/l alleles repositionthemselves to associate with pericentromeric heterochromatin,apotentlyrepressive environment(Fig.5).Recentstudiesindicatethat IgHdecontractionand itsassociationwith heterochromatinaremediatedbythe IgK 3' enhancer,which in somemannerdirectspairingbetween the IgH and IgK loci in pre-Bcells." Thus,in addition to chromatinaccessibility, allelicexclusion ofV-(D)J recombinationislikdy orchestratedbywholescale changes in the structureofthe nonfunctional locus,which (i) separateand therebyisolate the V and (D)J clusters and (ii)move the gene segments into a repressive chromatin environment that is refractory to nuclearfactorssuch as recombinase.

Allelicexclusion at the IgLlociincorporatesmanyof the aforementionedregulatorystrategiesbut alsoinvolves twoadditionallevels ofcontrol BothIgK alleles migratefromthenuclearperipheryto a central location in pro-Bicells.fHowever, upon developmental transition to the preB-cellstagea singleIgK allelebecomesactivatedbasedon chromatinmodifications, itsearlyreplication,germline1" transcription and V,,-1" recombinarion.P The second IgK alleleinitiallyassociateswith pericentromericheterochromatin,which represses its recombinase accessibility. Presumablythe secondalleledissociates from this repressive environmentovertimeifrecombination of the first


allelefails to generatea functionalBCR.65 Thus,unlike IgH or TCRjJloci, allelic exclusion at IgKinvolves sequentialactivation of the twoseparate alleles. Finally. onceafunctional, alloreactive BCRisproduced bya pre-Bvcell, V(D)J recombinationisterminatedviaasignal-dependent repressionof RAG expression. thus providingan irreversible meansto enforceallelic exclusion,"

ConclusionDuring the past five years wehavewitnessed an explosionin our understandingof mechanisms

that controlthe targetingofrecombinase to precise regions withinantigenreceptorlociandtherebyproviderequisitetissue-,stage-and allele-specificity. Wearebeginningto appreciate that, althoughcommon strategies are employedat many loci, the activationor repression of recombination ateach cluster of gene segments likely proceedsvia distinct mechanisms. Indeed, each region appearsto usea unique combinationof crosstalkbetweenACE function. epigeneticmodifications,transcription, chromatin remodeling. nuclear localizationand locus contraction to achieve therequiredlevelofrecombinationefficiency. Additionalmechanisms will likelybe discovered in thenear future. We havebegun to moveforward from correlative studies focusedon the molecularhallmarksof recombinase accessibility to aclearerdefinitionofwhether thesefeatures arecausallylinked. However. similaradvances must be made to establish cause-effect relationships betweenlong-rangechanges in locusconformation(e.g.•contraction)and recombination.Ofequalimportance, a considerable effort should be placedin the discovery of cis-acting elementsthat controllong-rangeinteractions between distant clustersof gene segments. Given what we havealreadylearnedfrom antigenreceptorgeneregulation.wecan anticipatethat manyofthesenewfindingswillhavebroad implications for the control of geneexpression.

AcknowledgementsThe authors acknowledge support received from the National Institutesof Health.

References1. Bassing CH. Swat W; A1t FW The mechanism and regulation of chromosomal V(D)J recombinat ion.

Cell 2002; 109 Suppl:S45-55.2. de Villanay JP, Poinsignon C. de Chasseval Ret at Human and animal models ofV(D)J recombination

deficiency. Curr Opin Immunol 2003; 15(5):592-598.3. Jankovic M, Nussenzweig A. Nussenzweig Me. Antigen receptor diversificarion and chromosome

translocat ions. Nat Immunol 2007; 8(8):801 -808.4. Cobb RM, Oestreich KJ. Osipovich OA er al. Accessibility control of V(D)J recombination. Adv

Immunol 2006 91:45-109.5. Oettinger MA. Schatz DG . Gorka C et al. RAG-1 and RAG-2. adjacent genes that synergistically

activate V(D)J recombination . Science 1990; 248(4962):1517-1523.6. Schatz DG. Oettinger MA. Baltimore D. The V(D)J recombination activating gene, RAG-I. Cell 1989;

59(6):1035-1048.7. Sakano H. Huppi K. Heinrich G et al. Sequences at the somatic recombination sites of immunoglobulin

light-chain genes. Nature 1979; 280(5720):288-294.8. Dudley DD. Chaudhuri J, Bassing CH et al. Mechanism and Control ofV(D)J Recombination versus

Class Switch Recombination : Similarities and Differences. Adv Immunol 2005 ; 86:43-112.9. Yancopoulos GD. Blackwell TK. Suh H et at. Introduced T-cell receptor variable region gene seg

ments recombine in preB cells: evidence that B- and Tvcells use a common recombinase. Cell 1986;44(2 ):251-259.

10. Krangel MS. Gene segment selection in V(D)J recombination: accessibility and beyond. Nat Immunol2003; 4(7):624-630 .

11. Hardy RR. Carmack CEoShinton SA et at. Resolution and characterization of pro-B and prepro-Bscellstages in normal mouse bone marrow. J Exp Med 1991; 173(5) :1213-1225.

12. Melchers F. The pre-Bvcell receptor : selector of fitting immunoglobulin heavy chains for the Bvcellrepertoire . Nat Rev Immunol 2005; 5(7) :578-584.

13. Ehlich A. Schaal S. Gu H et at. Immunoglobulin heavy and light chain genes rearrange independentlyat early stages of Bvcelldevelopment . Cell 1993; 72(5 ):695-704.

14. Krangel MS. Tvcell development : better living through chromatin. Nat Immunol 2oo7; 8(7):687-694.15. Loffen D. Ehlich A. Muller W et at. Surrogate light chain expression is required to establish immuno

globulin heavy chain allelic exclusion during early B-eell development. Immunity 1996; 4(2):133-144.


16. Sen R. Oltz E. Genetic and epigenetic regulation of IgH gene assembly. Curr Opin Immunol 2006;18(3):237-242.

17. Van Ness BG. Weigert M. Coleclough C er al. Transcription of the unrearranged mouse C!C locus:sequence of the initiation region and comparison of activity with a rearranged VIC- CIC gene. Cell 1981;27(3) :593-602.

18. Yancopoulos GD. Alt FW Developmentally controlled and tissue-specificexpression of unrearrangedVH gene segments. Cell 1985; 40(2):271 -281.

19. Sleckman BP, Gorman ]R. Alt FW. Accessibility control of antigen-receptor variable-region geneassembly: role of cis-acting elements. Annu Rev Immuno11996; 14:459-481.

20. Yancopoulos GD, Ale FW. Regulation of the assembly and expression of variable-region genes. AnnuRev Immuno11986; 4:339-368.

21. Stanhope-Baker P, Hudson KM, Shaffer AL er al. Cell type-specific chromatin structure determines thetargeting ofV(D)] recombinase activity in virro. Cell 1996; 85(6) :887-897.

22. Ferrier p, Krippl B. Blackwell TK et al. Separate elements control D] and VOJ rearrangement in atransgenic recombination substrate. EMBO] 1990; 9(1):117-125.

23. Oltz EM, Alt FW; Lin WC ec al. A V(D)J recombinasc-inducible B-cellline: role of rranscriptionalenhancer elements in directing V(D)] recombination . Mol Cell Bioi 1993; 13(10):6223-6230.

24. Bories ]C, Demengeot J, Davidson L et al. Gene-targeted deletion and replacement mutations of theT-cell receptor beta-chain enhancer: the role of enhancer elements in controlling V(D)] recombinationaccessibility. Proc Nad Acad Sci USA 1996; 93(15) :7871-7876.

25. Bouvier G. Warrin F. Naspetti M er al. Deletion of the mouse T-cell receptor beta gene enhancer blocksa~ T'cell development. Proc Nat! Acad Sci USA 1996; 93(15):7877-7881.

26. Chen J. Young F. Bottaro A et al. Mutations of the intronic IgH enhancer and its Ranking sequencesdifferentially affect accessibilityof the]H locus. EMBO J. 1993; 12(12):4635-4645.

27. Serwe M, Sablirzky F.V(D)J recombination in B-cells is impaired but not blocked by targeted deletionof the immunoglobulin heavy chain intron enhancer. EMBO J. 1993; 12(6):2321-2327.

28. Oestreich K]. Cobb RM, Pierce S et al. Regulation ofTC~ gene assembly by a promoter/enhancerholocomplex, Immunity 2006; 24(4) :381-391.

29. SikesML, Meade A. Tripathi Ret al. Regulation ofV(D)J recombination: A dominant role for promoterpositioning in gene segment accessibility. Proc Nat! Acad Sci USA 2002; 99:12309-12314.

30. Sikes ML, Suarez CC . Oltz EM. Regulation ofV(D)] recombination by transcriptional promoters. MolCell Bioi 1999; 19(4):2773-2781.

31. Whitehurst CEoChattopadhyay S. Chen J. Control of V(D)] recombinational accessibility of the D~1gene segment at the TC~ locus by a germline promoter. Immunity 1999; 10(3):313-322.

32. Hawwari A. Bock C. Krangd MS. Regulation of T-cdl receptor alpha gene assembly by a complexhierarchy of germline Ja promoters . Nat Immunol 2005; 6(5):481-489.

33. Abarrategui I. Krangd MS. Regulation of T-cell receptor-alpha gene recombination by transcription.Nat Immunol2oo6; 7(10) ;1109-1115.

34. Fahrner JA, BaylinSB. Heterochromatin: stable and unstable invasions at home and abroad. Genes Dev2003; 17(15):1805-1812.

35. Khorasanizadeh S. The nucleosome: from genomic organization to genomic regulation. Cell 2004;116(2):259-272.

36. Wolffe AP. Guschin D. Review: chromatin structural features and targets that regulate transcription. ]Struct Bioi 2000; 129(2-3):102-122.

37. Jenuwein T, Allis CD . Translating the histone code. Science. 2001; 293(5532):1074-1080.38. Berger SL. Histone modifications in transcriptional regulation. Curr Op in Genet Dev 2002;

12(2):142-148.39. Lachner M. O'Carroll D. Rea S at al. Methylation of histone H3lysine 9 creates a binding site for HPI

proteins . Nature 2001; 410(6824):116-120.40. Bird A. DNA methylation patterns and epigenetic memory. Genes Dev 2002; 16(1):6-21.41. Fuks F. Hurd PJ. Deplus Ret al. The DNA methylrransferases associate with HPI and the SUV39Hl

histone methyltransferase, Nucleic Acids Res 2003; 31(9):2305-2312.42. McMurry MT, Krangel MS. A role for histone acetylation in the developmental regulation of VO]

recombination . Science 2000; 287(5452) :495-498.43. Morshead KB. Ciccone DN. Taverna SD er al. Antigen receptor loci poised for V(D)J rearrangement

are broadly associated with BRG1 and Ranked by peaks of histone H3 dimethylated at lysine 4. ProcNat! Acad Sci USA 2003; 100(20):11577-11582.

44. Spicuglia S. Kumar S. YehJH et al. Promoter activation by enhancer-dependent and -independenr loading of activator and coactivaror complexes. Mol Cell 2002; 10:1479-1487.

45. Osipovich O. Milley R. Meade A et al. Targeted inhibition ofV(D)] recombination by a histone methylrransferase. Nat Immunol 2004; 5(3) :309-316.


46. Liu Y. Subrahmanyam R, Chakraborry T, Sen R. Desiderio S. A PHD-Finger Domain in RAG-2 ThatBinds Hypermethylated Lysine 4 of Histone H3 Is Necessary for Efficient V(D)J Rearrangement. Immuniry 2007; 27(4) :561-571.

47. Matthews AG, Kuo AJ. Ramon-Maiques S, et at RAG2 PHD finger couples histone H3 lysine 4 trimethylation with V(D)J recombination. Nature 2007; 450(7172):1106-1110.

48. Shi X. Hong T. Walter KL er al. ING2 PHD domain links histone H3 lysine 4 methylation to activegene repression. Nature 2006; 442(7098):96-99.

49. Kingston RE, Narlikar GJ. ATP-dependent remodeling and acerylation as regulators of chromatinfluidity. Genes Dev 1999; 13(18):2339-2352.

50. Emerson BM. Specificity of gene regulation. Cell 2002; 109(3):267-270.51. Kwon J. Morshead KB, Guyon JR et al. Histone acetylation and hSWI/SNF remodeling act in concert

to stimulate V(D)J cleavage of nucleosomal DNA. Mol Cell 2000; 6(5):1037-1048.52. Osipovich O. Milley Cobb R. Oestreich KJ et at Essential function for SWI-SNF chromatin-remodel ing

complexes in the promoter-directed assembly ofTcr~ genes. Nat Immuno12007; 8(8) :809-816.53. Reddy KL. Zullo JM. Bertolino E. Singh H. Transcriptional repression mediated by repositioning of

genes to the nuclear lamina. Nature 2008; 452(7184):243-247.54. Kosak ST, Skok JA, Medina KL er al. Subnuclear compartmentalization of immunoglobulin loci during

lymphocyte development . Science 2002; 296(5565):158-162.55. McDougall S, Peterson CL, Calame K. A transcriptional enhancer 3' of C~2 in the T-cell receptor beta

locus. Science 1988; 241(4862):205-208.56. McMillan RE. Sikes ML. Differential Activation of Dual Promoters Alters D~2 Germline Transcription

during Thymocyte Development . J Immunol2008; 180(5) :3218-3228.57. Afshar R, Pierce S, Bolland DJ et al. Regulation of IgH gene assembly: role of the intronic enhancer

and 5'DQ52 region in targeting DHJH recombination. J Immunol 2006; 176(4) :2439-2447 .58. Mathieu N, Hempel WM . Spicuglia Setal. Chromatin remodeling by the T-cell receptor (TCR)-beta

gene enhancer during early T-cell development : Implications for the control of TCR-beta locus recombination . J Exp Med 2000; 192(5):625-636.

59. Perlot T. Ale FW; Bassing CH et al. Elucidation ofIgH intronic enhancer functions via germ-line deletion . Proc Nat! Acad Sci USA 2005; 102(40) :14362-14367.

60. Ryu CJ, Haines BB, Lee HR et al. The T-cell receptor beta variable gene promoter is required forefficient V~ rearrangement but not allelic exclusion. Mol Cell Bioi 2004 ; 24(16) :7015-7023.

61. Jackson A, Kondilis HD, Khor B et al. Regulation of T-cell receptor beta allelic exclusion at a levelbeyond accessibility. Nat Immunol 2005; 6(2) :189-197.

62. Skok JA, Gisler R, Novatchkova M et al. Reversible contraction by looping of the Tern and Tcr~ lociin rearranging thymocytes. Nat Immunol 2007; 8(4) :378-387.

63. Roldan E, Fuxa M, Chong W et al. Locus 'decontracrion' and centromeric recruitment contribute toallelic exclusion of the immunoglobulin heavy-chain gene. Nat Immunol 2005; 6(1) :31-41.

64. Jhunjhunwala S, van Zelm MC . Peak MM , et al. The 3D structure of the immunoglobulin heavy-chainlocus: implications for long-range genomic interactions . Cell 2008; 133(2):265-279.

65. Goldm it M, Ji Y. Skok J et aI. Epigenetic ontogeny of the IS" locus during B-cell development. NatImmunol2005: 6(2):198-203.

66. Gorman JR, Alt FW. Regulation of immunoglobulin light chain isorype expression. Adv Immunol 1998;69:113-181.

67. Manz J, Denis K, Witte 0 et al. Feedback inhibition of immunoglobul in gene rearrangement by membrane mu, but not by secreted mu heavy chains. J Exp Med 1988; 168(4):1363-1381.

68. Tripathi R, Jackson A. Krangel MS. A change in the structure ofV~ chromatin associated with TC~allelic exclusion. J Immunol 2002; 168(5):2316-2324.

69. Chowdhury D. Sen R. Transient IL-7/IL-7R Signaling Provides a Mechanism for Feedback Inhibitionof Immunoglobulin Heavy Chain Gene Rearrangements. Immunity 2003; 18:229-241.

70. MostoslavskyR, Singh N. Tenzen T et al. Asynchronous replication and allelic exclusion in the immunesystem. Nature 2001; 414(6860):221-225.

71. Hardy RR, Hayakawa K. B-ell development pathways. Annu Rev Immunol 2001; 19:595-621.

CHAPTER 10

Molecular Genetics at the T-CellReceptor BLocus:Insights into the Regulation ofV(D)J RecombinationMarie Bonnet, Pierre Ferrier and Salvatore Spicuglia*

Abstract

T he V(D)Jrecombinationmachineryassembles antigenreceptor genes fromgermlineV,Dand] segments duringlymphocytedevelopment. InapT cells, thisleadsto theproductionofthe T-cellreceptor (TCR) a and P chains. Notably, V(D)J recombination at the Tcrb

locus is tightly controlled at various levels, includingcell-type and stagespecificities, intralocusorderingand allelic exclusion. Although manyof thesecontrolsarepartlymediatedat the level ofgenomicaccessibility to the V(D)J recombinase, recentstudieshaveuncovered novelmechanismsthat are also likely to contribute to the developmental regulation of Tcrb gene rearrangementevents. In thischapter,wesummarize our currentknowledge andhighlightunanswered questionsregardingthe regulationofV(D)J recombinationat the Tcrb locus,placingemphasis on mousetransgenesis and gene-targeting approaches.

IntroductionBand T-Iymphocyres, theadaptivearmsoftheimmunesystem invertebrates, cangeneratespecific

responses to a tremendous numberofforeign antigens.' This remarkable propertylargely dependson a series of DNA rearrangements betweengermline V,D and] genesegments that generate variableregiongenes at antigenreceptor-encoding loci.V(D)Jrecombination events areinitiatedin thedeveloping lymphocytes by the lymphoid-specific proteins, RAGI and RAG2 (RAGII2), whichintroduceDNA double-strand breaks precisely at the bordersbetweentwo codingsegments andtheir flanking recombination signal sequences (RSSs). The RSSs are composed of relatively wellconserved heptamerand nonamersequences separated bya less wellconserved spacer of either 12or 23 base-pairs (bp),named 12-and 23-RSSs respectively. V(D)] recombination occurs primarilybetweenone genesegment flanked bya 12-RSS and another bya 23-RSS, a restriction termedthe12/23 rule.Eventually, thebrokenendsarerepaired bythe ubiquitously expressed 'nonhomologousend joining' (NHEJ) machinery to formcodingand RSS(signal) joints.

Thereareseven different antigenreceptorloci.I Theseincludethe immunoglobulin heavy(Igh)andlight (Igk andIgf) chainlociforB-cells and theT-cellreceptor(Tcr) a,b,ganddloci forT-cells.Tcrb and ageneassembly is carriedout at two distinct stages of apT-cell development in the thyrnus.' V(D)J recombinationat the Tcrb locusis initiated within the CD4-CD8- double-negative(DN) compartment and proceedsstepwise with Dp-to-JPoccurringfirst in DN2 cells. prior to

VP-to-DJPjoiningmostlyoccurringinDN3 cells.Expressionofafunctionally rearranged Tcrb geneleadsthymocytes to passthrough p-selection and differentiate into CD4+CD8+ double-positive

'Corresponding Author: Salvatore Spicuglia-Centre d'immunologie de Marseille-Luminy, Case906, Campus de Luminy, Marseille, France. Email: [email protected] r

V(D)JRecombination, edited by PierreFerrier. ©2009 Landes Bioscienceand SpringerScience+Business Media.

Molecular Genetics at the T-CellReceptor PLocus 117

(DP) cells,while instigating Tcra gene rearrangements. Ultimately, TCRa~-expressing cells maybe selected into mature CD4+ or CD8+ single-positive (SP) T-cells.

T-cell development requires temporally-regulated rearrangement and expression ofthe Tcrgenes. The lymphoid-specific expression ofthe Rag1/2 genes restricts V(D)J recombination tothe developing lymphocytes. ' However, Ragl/2 regulated expression alone cannot explain allthe controls ofV(D)J recombination at the Ig and Tcr endogenous loci. In particular, the Tcrblocus is subjected to many levels of regulation which determine a precise developmental orderof rearrangement events and ensure that the Tcrb gene is expressed in an allelically excludedmanner. 2 The present chapter focuses on our current understanding ofhow V(D)J recombination is regulated at the Tcrb locus and places emphasis on mouse transgenesis and gene-targetingapproaches that have revealed essential roles for cis- and trans -regulators of Tcrb gene rearrangements and allelic exclusion .

Overview ofthe Tcrb Genomic Structure and RecombinationProperties

In the mouse germline, the approximately 700-kilobases (kb) Tcrb locus consists of a large(-42S-kb) 5' region containing22 functional V~ genesegments,aswell as 13 additionalV~ pseudogenes and a shorter (-2S-kb) 3' region comprising a duplicated cluster ofD~-J~-C~ gene segments(Fig. lA).31n addition, it contains two groups of trypsinogen genes (not transcribed in T-cells),including one spread over the 2S0-kb region separating the V~ and D~/J~ regions. A distinct V~gene segment (V~14) is situated at the 3' end of the locus, lying in the opposite transcriptionaldirection. Asdetermined by their RSS types and orientations, recombination ofall S'V~, D~ andJ~ gene segments is deletional, whereas that ofV~14 occurs by inversion.' Germline V~ and D~segments are flanked by upstream transcriptional promoters that initiate sterile transcription ina developmentally regulated fashion . A single transcriptional enhancer (E~) lies between C~2

and V~14. The formation of a complete VDJ~ variable region places the promoter of the rearranged V~ segment in the vicinity and under the control of E~, enhancing transcription of thenewly-assembledTC~ unit. Via standard mRNA-maturation processes, the variableVDJ~ andconstant C~ exons are then spliced to produce the full-length TC~ chain. As described furtherbelow, the particular structure and organization ofthe Tcrb locus impose a number ofconstraintsregarding the regulation ofV(D)J recombination. In this context , it is worth noting that defectsin Tcrb locus recombination are suspected to contribute to T-cell pathogenesis (e.g.,oncogenesisand, not as yet disproven, autolmrnuniryj.t"

Tcrb-RSSs and Rearrangement EfficiencyDespite their overall conservation , Tcrb-RSSs exhibit marked sequence variations compared to

the canonical RSSsY·8 Indeed, the frequency ofJ~ gene segment usage at the murine Tcrb locuscorrelates wellwith sequence variations within the correspondingJ~-12RSS residues," In addition,mutagenesis studies have identified various residues in the sequences of the Tcrb-RSSs and theircoding flanks that cooperate to determine the ultimate efficiencyofthe recombination process.'?Thus, theTC~ repertoire naturally reflectsthe subtle interplay between the RSSsand the flankingcoding sequences to direct the activity ofthe V(D)J recombinase. A dramatic, recently uncoveredexample ofthis paradigm is illustrated below.

Beyondthe 12/23 RuleAs mentioned above, Tcrb variable regions are assembled via an ordered two-step process in

which D~-to-J~ rearrangement occurs before the appendage ofa V~ to the rearranged DJ~ genesegment. In this context , the organization ofthis locus, in which V~ and 3'D~ RSSscontain 23-bpspacers and S'D~ andJ~ RSSs contain l2-bp spacers (Fig. lB) , seemed to represent a potentialproblem since direct joiningofV~ segments to nonrearrangedD~ or J~ gene segments would alsobe compatible with the 12123 rule. However, such rearrangements are respectivelyvery infrequentand practically nonexistent within the endogenous Tcrb locus (reviewed by Tillman et al8).These

118 Molecular Genetics at the T-CellReceptor ~ Locus V(D)JRecombination

AVJl cluster

;; ~ ~..

"""2 511 1'2 n n 15 1'& T1 ~S >!:~;; :;-=.- lIIIiI - iii &l I)--(!l II III 1111 I

s ~S S> ~ S~S~ ~

vp cluster

.... S S S~SS S S ~ ~..

5 5 S 'l1 'l1II I I 1111 I I I I I lDlII

"., ..

S~~§ s ..~ ~

..~~ s s s s

3 4 5 6

HJ J

DJCp clustersi i

TtITtI 1'20 Ci,> uS!:; {l ~§ ~a I lIIt~.~D • ~ :

VP-Ukl : >:.. . .. .. .. ... \

.. -.... .

T17D

116III

8 7

, DJ~ k:(J~ ·1 j J

...--_ .... -

TIS..

.. .. ------

T12 113 TH• I II

TIlaD

TlOIB

Clll Op2 Jp2~

HS711ko

CP 2

Epko

Erm.< IdEll/Eprlv kl

~.. kl

8 2 1

VP ~DP~Jp

----.,~-.. <l!IE> e::t[]f-----

b12123

Figure 1. Structural organization and properties of the mouse Tcrb locus . A) The V, D, J andC segments are designed following the conventional ImMunoGeneTics (IMGT) nomenclature(top) and corresponding (published) names (bottom; for Vp segments only) (see: http://imgt.cines.fr/textesllMGTrepertoireJLocusGenes/nomenciaturesimouse/TRB/TRBV/Mu_TRBVnom.htrnl) . Thick black and grey lines represent functional and nonfunctional Vp gene segments,respectively; and grey boxes represent trypsinogen genes. Thin black lines are for D genesegments; white and dashed boxes for J and C gene segments, respectively. An enlargementof the 3' region is shown, where the Epenhancer (grey oval) and transcriptional Dp1 and Vp14promoters (black ovals) are also figured. Published knock-out (ko) and knock-in (ki) alleles areindicated (seeTable I for details and references). B)Schematic representation of Tcrb gene rearrangements. Dp-to-JP joining occurs first (1) followed by Vp-to-DJP assembly (2). The beyond(b)12123 constraint prohibits direct V~-to-J~ joining. 12- and 23-RSSs are indicated.

Molecular Genetics at the T-CellReceptor PLocus 119

paradoxical behaviors were initiallyanalyzedusinga modified Tcrb allele(Jj31"') containing theDj31-Jj31 geneclusteronlyand derivedmutantsproducedbysuccessive gene-targeting(Fig. lA andTable 1).B Assuch,usingallelic mutants (Jj31M3-MS), it wasshownthat the 5'Dj31-12RSS, but nottheJj31-12RSS, efficiently targetsVj3-23RSSs for recombination,aphenomenon termed "beyond12123" (B12/23).11 Moreover, 3'Dj31-23RSS deleted alleles (Jj31M2) readilyproduced Vj3-to-Dj3rearrangements showing,at the veryleast,that theseeventsdo not requireprior DJj3 assembly,"Subsequently, it wasshownthat T-cells from micein which the Vj314-23RSS wasreplacedbythe3'Dj31-23RSSand the Dj3 genesegmentdeleted (Jj31M6allele), exhibited a dramatic increase inVj314 generearrangements, alldirected to theJj3-12RSSs asexpected," However, surprisingly, inthe presenceofan intact Dj31 genesegment(Jj31M7allele), mostwerereadily targetedto DJj3 rearranged intermediates.'! Collectively, these resultsdemonstrate that RSSs can imposesignificantconstraints on Tcrb gene assembly beyond enforcingthe 12123 rule. In particular the Dj3 RSSsmayfunction in anefficient mannerto target the V(D)J recombinase, thus ensuringthe utilizationof a Dj3 genesegmentduringvariable regiongeneassembly and hence the diversityof the TCRj3repertoire.I5However, the impact of this constraintfs) on regulatoryeventssuch as Tcrb orderedrearrangements remainslargely unexplored.

Molecular Mechanism(s) ofB12123Concurrent effortsbyseveral laboratorieshaveprovidedcompellingevidence that B12123 is

establishedlargely by the V(D)J recombinase and Tcrb RSSs and isenforcedat or prior to DNAcleavage.16-19 AdditionalworkusinghybridRSSs has enlightenedthe roleof the j3-12-RSSnonamerand spacersequences in imposingthis constraint.18-20Finally, building on these earlier findings,the Schatzlaboratory recently demonstrated the implementationofthe B12123 constraint viaalowefficiency DNA cleavage reactionat variousdiscretestages dependent on the genesegmentsconsidered(e.g.,initialsinglestrandnickingatJj3 substrates; and,also,synapsis betweenVj3 andJj3substrates)," Stillthequestionremains asto howthe 3'Dj3-23RSSs overcomes apossible 'wedge'inJj3-12RSS cleavage to sustainan efficient Dj3-to-Jj3 recombination.Thehigh conservation(acrossmammalian phylogeny) of Dj3-RSSs compared to Jj3-12RSSs,8 implyinga strong evolutionarypressureon the former, maybe relevant to the maintenanceof this precisefunction,

Cis-Regulatory Elements at the Tcrb LocusComprehensive mappingof DNAse-I hypersensitive (HS) siteswithin the 3' sideof the Tcrb

locushasrevealed 11HS sites(namedHS1to HS11) encompassing a regionfrom 3-kbupstreamofDj31 to 3-kbdownstreamofV1314 (Fig. lA).22.23 Strikingly, HS2 and HS9 respeetivelyoverlapwith Ej3 and thepromoter locatedupstreamofDj31 (pDj31),24.25 both criticalregulatorsof regionaland more localizedrecombinationeventswithin the Dj3-Jj3 clusters (seebelow). Similarly, HS5overlaps with theVj314 promoter.Thisledto theassumption thatother HSsalso represent potentialcis-regulatory elementsfor Tcrb generearrangements and/or transcription.Todate,gene-targetingand the analysis of the resultinglymphoidcellphenotype haveaddressedthe function ofmost oftheseputativerecombinational/transcriptional elements, aswellasthat ofat leastone Vj3 promoter(summarized in Fig. lA and Table1).

The transcriptionalenhancer of the Tcrb locus (Ej3) has been describedas a criticalelementin sustaininghigh levelexpression ofa rearrangedTcrb gene.26-28 InitialcharacterizationofEj3sequenceshasdefinedseventranscription-factor(TF) bindingsites(namedEj31 to Ej3?) .29.30 Effortsto further dissectthe structuralorganizationofEj3and defineminimal (core)enhancersequencesrevealed the importanceofacommoncomposite(ETS/RUNX) TF motif,found within Ej34 andEj36, in mediatingenhanceractivity,3l·32In parallel,byatransgenicapproach,apossible additionaland important role of Ej3 wasshown in the control of Tcrb gene rearrangemcnts.P" Indeed, thegenerationofEj3 knock-outmice(Ej3-I-)demonstrateda requirementof thiselementfor Tcrb generecombinationand, at least in the Dj3-Jj3-Cj3 regions, transcriptionalexpression (refs.35,36;alsoseebelow).Accordingly, the Ej3-I- micedisplayapartial blockof thymocytedifferentiationat theDN cellstageand absence of aj3T-cells in peripherallymphoid organs,37.38 a phenotype similarto


Table 1. Gene-targeting studies at the TCRp locus

Genotype! Description Phenotype2 Refs.3

TCRflko -15-kb deletion, from Jfl1.3 to Blockin aliT-ceildifferentiation 39csz

Efl ko 560-bp deletion of Stronglyreduced D-J and V-DJ 35,36,38EfI-containing sequences rgts; reduced CA at DJCfI; VfI

regions not affected

EfI/EiJj ki EfI-to-EiJjreplacement DfI-JfI and VfI-DJfI rgts restricted 36to T-cells

Efl/Eflrev ki Inversion of Efl No phenotype 73

Cfl2-EfI/EiJj ki l l-kb replacement (from csz to T-cell lineage restriction 44EfI) by EJj disrupted; Jfll, but not Jfl2, rgts

impaired

Efl/Ea ki EfI-to-Ea replacement Rgts still restricted to DN cells; 42CA impaired in DN, but not DPcells; less efficient TCRfI rgts

HSl ko 780-bp deletion of HSl (located No phenotype 23400-bp upstream of EfI)

HS9/10111 ko 3-kb deletion of sequences DecreasedDfll rgts; Dfl2 not 47encompassing HS9, 10and 11 affected; GL transcriptionof the(located upstream of Dfll) Dfll regionabolished

HS9(pDfll) ko 390-bp deletion removingthe Decreased Dfll rgts; Dfl2 not 48pDfll core region (HS9) affected; Reduced CA at the

DfllJfll region

HS7/8 ko Deletion of the intronic No phenotype 46sequences encompassing HS7and HS8

Jfll~ Deletion of the Dfl2-Jfl2 region No phenotype 11

JfllM2 Mutation of the 3'Dfll-RSSin ve-to-net rgtsare readily 12the Jfll~ allele observed

JfllM3 Deletion of Dfl1 and flanking -No VfI rgtsto any of the JfI seg- 11RSS in the jet- allele ments; Developmental block at

the DN cell stage

Jfl1 M4 5'Dfll -RSS to Jfll.2-RSS replace- EfficientDJfI rgts; Reduced levels 11ment in the Jfll~ allele of VDJfI rgts; Developmental

block at the DN cell stage

JfllM5 Jfll.2-RSS to 5'Dfll-RSS replace- VfI segments rearrange exclu- 11ment in the JfllM3 allele sivelyto the Jfll.2-5'Dfll-RSS;

Normal numbers of aflT-celis

JfllM7 Vfl14-RSS to 3'Dfll-RSSreplace- Dramatic increase in Vfl14· 14ment in the Jfl1~ allele T-cells; Vfl14 rgtstargeted to

DJfll, but not to Jfll (B12/23 rulenot broken); Normal numbers ofaliT-celis

continuedon next page

Mo/~cular Genetics at tbe T-C~//Receptor PLocus 121

Table 1. Continued

Genotype' Description Phenotype2 Refs.3

J~lM6 V~14-RSS to 3'D~1-RSS replace- Dramatic increase in V~14' 13ment in the )~lM3 allele T-cells; Direct V~14 to )~1 rgts

Normal numbers of a~T-cells; AE

maintained

vet-est ko (~lD) 475-kb deletion of Tcrbse- Increased rgts using the most 102quences from V~l to 3'C~1 proximal V~ gene (V~10);

Increased CA at V~10; AE main-

tained; Aberrant V~1O-D~2 and

V~10-J~2 rgts are observed

V~8.2/hCD2 ki Insertion of a human CD2 Bi-allelic expression of hCD2 81cDNA downstream of V~8.2 prior to V-D) rgts Distance-

dependent expression of hCD2

in VDJ rearranged alleles

V~12IEa ki Eainsertion upstream of V~12 Increased rgts using V~12: AE 79partially subverted, but feedbackinhibition is maintained

V~13 ki V~13 gene segment and pro- Same utilization frequency of the 53moter inserted 6.8-kb upstream inserted and endogenous V~13

of D~l copies; AE subverted at the level

of rgts, but not at the level oftranscription

pV~13 ko Deletion of a 1.2-kb region con- V~13 rgts inhibited and CA 52taining the V~13 promoter reduced; AE maintained

pV~13/G4SV40 ki 1.2-kb replacement of pV~13 Normal levels of V~13 rgts; 52by the SV40 minimal promoter Cleavage at V~13 not significant-+5 copies of Gal4 binding se- ly affected (but mostly abnormalquences regarding the cleavage site);

TCR~ Tg blocks V~13 rgts, butnot the cleavagesat the aberrantsites

'ko: knock-out mouse; ki: knock-in mouse; rgts: rearrangements; AE: allelic exclusion;CA: chromatin accessibility.

2Additional phenotypes are described in the text.3Additional articles describing the phenotype(s) at the targeted alleles are mentioned in the text.

the one observed in Tcrb knock-out mice." Ofnote, besides these impacts on Tcrb geneexpression,recombination and a~T-cell development, further analysesprovided evidencefor an accwnulationof rare recombination products in E~-ddeted T-cells, including coding joints between the D~l

and D~2 gene segments and intermediate products ofRSS cleavage[so called signal ends (SE)] atV~14and D~ gene segments (in DP Cells).38.40,4( Theimplication(s) ofthese observations regardinga physiological role ofE~ in the correct pairing ofD~-J~ gene segments and/or the processing oftheir recombination products still remains unclear.

Results from transgenic mouse experiments generally argue for a role of lymphoid enhancers in mediating tissue- and stage-specific induction of antigen-receptor gene recombination.'Importantly, this does not necessarily imply that the patterning of endogenous gene rearrangements be simply regulated by enhancer sequences alone. For example, whereas Tcra enhancer


(Ea}-inducedrearrangements within a Tcrb transgenic substrateoccurred efficiently at the DPcellstage, E~-to-Ea replacement within the endogenousTcrb locusstillresultedin DN-restrictedrecombination, although at a low efficiency.33.42Likewise, Tcrb transgenic substrates bearingthe19b intronic enhancer (Ei!L) demonstrated D~-to-J~ rearrangements in both T and B-cells andV~-to-DJ~ rearrangements in T-cells only.43However, on replacing~ for Ei!L at the endogenouslocus,recombinationwas then restrictedto T-cells and D~-to-J~ joining.36 Finally, knock-in replacementofE~ plusC~2-E~ interveningsequences by Ei!L yieldedsignificant levels ofD~-to-J~and V~-to-DJ~ rearrangements in T and Bvcells, unlike the mere E~-to-Ei!L replacement justmentioned.44 Altogether, thesedata stronglysuggest that additionalcis-regulatory elements maycontribute to the controlof lineage- and developmental stage-specificityofV(D}Jrearrangementeventsat the Tcrb locus. In this context, deletion of either the E~-proximal HSI or HS7-HS8sites resulted in no obvious phenotype.23.45.46 Therefore, if functionally relevant, these elementsmaybeexpectedto display high levels ofredundancy. Testing this hypothesis will requirefurthergene-targeting effortsto combinediscretedeletionswithin the sameallele.

Additionalas-elements shownto playarolein regulating recombination includetheD~1-andV~13-associatedpromoters.T-Iymphocytes fromknock-outmiceinwhichpD~ I (eitheraloneortogetherwith the upstreamHS10and HSII sites)hasbeendeleted,exhibitedaspecific blockinrecombinationof the D~ I genesegmentwhereas theD~2 genesegmentrearrangednormally.47.48Asgermlineexpression ofthe D~2 segmentlikely dependson an asyet uncharacterized regionalpromoter, common opinion statesthat D~2 recombinationalsorelies on such a promoter-mediated, localized control, though this has yet to be formally demonstrated.Such reasoningwasextendedto the control of rearrangementofindividualV~ genesegments, aseachisprecededbya transcriptionalpromoter,onlyafewofwhichhavebeenthoroughlycharacterized.49·51 Indeed,a1.2-kbdeletionofsequences upstreamofV~13resultedin the inhibition of transcriptionofthecorrespondingV-gene segmentand decreased RAG1I2-mediated rearrangement.P Moreover,when inserted -6-kb upstreamofD~ I, a supplementalV~13genesegment(and associated promoter) displayed the samefrequencyof rearrangementas the V~13 endogenoushomologue.53Thesestudies support the hypothesis that V~ promoters maygenerally be sufficient to inducerecombinationof their associated V-gene segments. Nonetheless, the presence ofaglobalregulator within the Tcrb locusand moreparticularlythe 5'V/3 region,notablycontrollinglong-rangeas-interactions betweenthe V~ and D~-J/3-C/3 clusters(seebelow)cannot be ruled out.

Trans-Regulators ofTcrb Locus Expression/RecombinationThecritical roleoftranscriptionalcis-regulatoryelements inorchestratingV(D)Jrearrangements

maybe clearly linked to (at leastsome of) the TFs they normallyrecruit. Initially, TF bindingmotifs (E~I-E~7) ofE/3 wereanalyzed usingdedicated bandshifi: and/or in vitro footprintingaswellasgene-reporterassays.2.3Q The identifiedmotifs includedthose for TFs belongingto theGATA,ATF/CREB, bHLH,ETSandRUNX families. Subsequently, mostofthesesitesappearedreadilyoccupied in developing T-cells (according to in vivofootprinting assays; ref.32 and ourunpublisheddata);andchromatinimmunoprecipitation (ChIP)experiments furtherdemonstratedthe specific binding ofTFs RUNXI and ETSI to E~ overlapping sequences.S4.55 Concordantly,overexpression of Rum} in an ex-vivo T-celldifferentiation model system resultedin increasedlevels of Tcrb expression.56 Likewise, the combinationof in vivogenomicfootprintingand ChIPassays hasimpliedtheloadingofabatteryofTFs topD~ l-overlappingsites, includingIkaros, ETSI,RUNXI, ATF/CREB, GATA, SPI and KLFS.54,s7,ss Although not yet completely characterized,promoter regions 5'ofV~ genes generallyappearedto bealsoenrichedwithavarietyom-bindingsites,includingtheETS/RUNX and ATF/CREB compositemotifs.3,sO,51

Due to the pleiotropiceffects that thesefactorsgenerallyexertthroughout embryonicand/orT-celldevelopment(reviewed byRothenberget al59), adefinitive demonstrationof their genuinerole in the control of genetranscriptionand/or rearrangement at the Tcrb locusrepresents a difficult challenge. To date, five nuclearfactors/signaling pathways havebeen reponed as directlyinterferingwith Tcrb generearrangements.i TheseincludeTFs HEB,60 c_MYB,61 and BCL1lb;62

Molecular Genetics at the T-C~"ReceptorPLocus 123

the scaffold/matrix-associated region 1 (SMAR1)protein;63and, probably, transcriptionaleffectors downstreamof the Notch1signaling cascade.64 Remarkably, disruption of most of the correspondinggenes(and overexpression ofSMAR1), resultedin aselective impairmentofV~-to-DJ~

recombination,60.61.63.64 underliningthe uniqueness of this particularstepof the Tcrb variable regiongeneassembly. Although not a trivialtask, the complementaryapproachof mutating the cognatebinding motif{s) within TO'b endogenouscis-regulatory elementsmayin the end be required tofirmly ascertainthe direct contribution of each and everyfactor to this process.

Chromatin AccessibilityLongbelievedto be solely confinedto a staticroleasa DNA packagingenvelope, chromatin is

nowalso recognized asamasterplayer in thedynamicregulationofallgenomicDNA transactions.Indeed,at the commandofdevelopmental/signalinginputs,chromatinevolves fromahermeticallypacked(heterochromatin) shellto amorerelaxed, 'metabolism-amenable' (euchromatin) structure.In thiscontext,fromthewidespread correlationbetweenongoingrecombination and transcriptionthrough nonrearrangedgenesegments/regions that prevails at antigen receptorloci, to the demonstrationofacriticalimpactofboth Ig/TO' transcriptionalcis- and trans-regulators on the formeractivity, allargumentsconverged to the modelwherelineage- and/or temporal-specific actionsofacommonV(D)J recombinase areprimarilyregulatedat the levelof chromatinpermissivity.65 Theformaldemonstration that thesecontrols indeed depend on the regulatedchangesin chromatinaccessibility camefrom experiments usingisolatednucleiand purifiedRAG1/2 factors.66Withina givenlocus, chromatin, associated with V(D)J rearranginggene segments, generally adopts a'less-compacted' configuration compared to that spreading over recombination inert regions(for review seerefs.67,68).Asfor the Tcrb locus,several molecularparameterssynonymous witheuchromatin (germlinetranscription; lackof CpG methylation; enrichment in histone H3/H4acetylation and H3K4 methylation; accessibility to restriction enzymes) have been correlatedwith stage-specific (DN2IDN3) V(D)J recombinationevents.23•38,41.42,48.57.69.71Conversely, fromthe DN3-to-DP thymic-celltransition [Le, ~-selection) onwards,heterochromatinfeatures weregenerally found alongthe chromosomalregionscomprisedof nonrearrangedV~ genes.23.41.70.72

Chromatin Remodeling byEnhancer-Promoter InteractionThestringentT-cellphenotypeobservedin the E~-I- mousemadeit an excellent modelsystem

to investigate the roleoftranscriptional enhancers in regulatingrecombinational accessibility.Thesestudiesweresignificantly helpedbythe possibilityof analyzingchromatin-associated parametersin EJ3-deleted,early-developingT-cells unableto proceedwith V(D}Jrecombinationdue to RAGdeficiency (E~-I- x Rag-I. mice). Indeed, detailed analysis ofDN T-cellsfrom these animalsandcomparison with those from Rag-I. controls, provided compellingevidence that E~ mediateschromatin remodelingwithin the proximalD~-J~-C~ domainsand, likely, activates the germlinepromoters flankingthe D~ genesegments (Fig.2).S4·57.69 In sharp contrast, the unaffectedappearanceof the chromatinstructureat E~-deleted alleles on the distal S'V~ and neighboring3'V~14regions suggested no impingingeffectofE~ at theseparticular locations. Despitesucha polarizedenhanceractivity, however, inversion ofE~ sequence orientationdid not alter Tcrb generegulation(in termsoftranscription andrecombination), thussupportingaDNA-loopingratherthan linking!trackingmechanism for thisenhancer'srolein the cis-activation of targetpromoter/recombinationsequences." In parallel, experimental approaches analyzingthe chromatinremodelingfunction ofthe D~l and V~13 promoters, showedthat both act to reduce localchromatin accessibility, i.e.,without affecting the neighboringD~2-J~2 or V~ genesegments, respectively (Fig. 2).48.52.54.74

Thorough molecularanalysis of Ep-deleted thymoeytes subsequently demonstrated that thiselement contributes to the assemblyof a functional nucleoprotein complexat pD~l, includingthe loadingof discreteTFs and basaltranscriptionalmachlnery," Such a dedicated process likelyinvolves a physical interaction between the two cis-regulatoryelements, possibly contributing to

the formationofastableholocomplex(Fig.3).S4 In thiscontext, the chromatinaccessibilityofJ~1genesegments is relatively unaffected at pDJ3l-deleted alleles, implyingthat E~ may exert both


V~ (- 35) D~1 Jj}1 Cj}1 D~2 Jp2 Cj}2 E~ VJH4

Figure 2. Effects on chromatin accessibility of mouse knock-out deletion of Tcrb cis-regulatoryelements. The upper line shows a schematic representation of the Tcrb locus in wild-typemice. The lower (2nd, 3rd and 4th) lines summarize the effects on chromatin accessibility ofthe Ep, pDPl and pVp targeted deletions, respectively. Shadowed areas symbol ize relaxed,accessible chromatin; arrows indicate germline transcription.

pD~I-dependentand -independent chromatinremodelingfunctions." Overall. theseresultssupport a modelwhereby the two elements act in coordination to regulatechromatinaccessibility atthe proximalD~-)~-C~ regions.

Chromatin RemodelingEnzymes andthe ControlofTcrb Locus Expression/Recombination

Chromatinstructurecanbealteredinanumberofdifferent ways. includingcovalentmodifications(e.g.•acetylation. methylation. phosphorylation and ubiquitinylation) ofhistonetails. replacementby distinct histone variants and changes in nucleosome positioningvia the action of specializedenzymatic and/or chromatin remodeling activities." A synopsis of these activities and dedicatedfactors potentially involved in regulating V(D)J recombinational accessibility-extendingbeyondthe scopeof this chapter-has recently been proffered.68 Regarding Tcrb D-) genesegments furinstance. the likelihood that suchfactors directly interfere with theirchromosomal accessibility wasbrought to light in experiments wheretargetedrecruitmentto pD~I of the H3K9 specific histonemethyltransferase G9a(usingproteinfusion to theDNA-bindingdomainofGal4andamodified Tcrbminilocus transgene) was shownto induceanextensive changein localchromatin environment (Le.,towards anheterochromatin-like layout) andasignificant decrease intheonsetofE~-mediated D~-)~transgenic expression and recombtnadon.f Thusfar. evidence indeedexists for a feweuchromatininducers having an impacton the onsetofV(D)) recombination at the Tcrb endogenous locus.67.68

UsingChIP assays. weand othershave shownthat BRGI (asubunitof the nucleosome-disruptingcomplexSWI/SNF) and thehistoneacerylases CBP/p300and PCAF,areassociated withE~- and/orpD~ I-overlappingsequences inprimaryDN thymoeytes and/or pre-T-celilines.S4•57,71 Moreover.Gal4-mediated targeting ofBRGI to a Tcrb transgenic substrate completely substitutes the pD~1functionininducinglocal recombinationalaccessibility; andknock-down byRNAinterferenceoftwoessentialSWI/SNF components (BRGI andBRM)results indecreased accessibility ofendogenousD~-RSSS .77 Assuggestedbytheseauthors. formation ofanEf3/pD~ l-basedholocomplex maygenerateanewinteractionsurface forthestable association ofSWI/SNF components.whichwouldthencontributeto remodel or displace a neighboring nucleosome(s),thusexposing sites requiredfor theinitiationofgermline transcription and/or recombination.Consistentwiththishypothesis. D~-RSSsand immediatesurroundingsequences display highnucleosome densities"andmaycomprise Strongnucleosome-positioningsites.78Whetheradditional chromatin regulators playasimilar.complementary, or differing rolefs) in controlling Tcrb locusaccessibility remains to bedetermined.

Molecular Genetics at the T-Cell Receptor PLocus 125

DN

co'.0=c:seo

~

tll',re"".IIo"1Allele 1

DP

DJlll

ce'~

.5

.ceoC.Jeo

:z:

Figure 3. Chromosomal cont raction/expansion events at the Tcrb locus during early T-cell development and prospective impact on the control of V(D)j recombination and allelic exclusion.Legends for the gene segments, regulatory elements and locus organization are as in Figure 2.In DN thymocytes, chromosomal contraction and DNA looping bring the various cis-regulatoryelements within the Tcrb locus closer, enabling potential interaction(s) and the onset of V(D)jcis-recomb ination . Differentiation to the DP cell stage results in locus decontraction and heterochromatinization of most unrearranged V~ gene segments, thus preventing further V~-to-Dj~rearrangement at these sites. At this stage,the WDj~ rearranged variable gene region, under thecontrol of E~, is highly transcribed (Allele 1).However, unrearranged V~ gene segments locatedimmediately upstream of the V~Dj~ rearranged region (Allele 1) and the V~14 gene segment(Alleles 1 and 2), remain accessible, yet do not rearrange. For didactic purposes, the allelicconformations represented here were chosen to result from (i) a productive rearrangement onthe first Vs-to-Dls joining attempt; (ii) a V~Dj~ rearranged variable gene region made of as',middle-located V~ gene segment joined to D~lIj~l gene segments.

126 V(D)JR~comb;nation

Beyond Chromatin AccessibilityThe B12123constraint discussed abovealone provided evidence that the regulationof Tcrb

V(D)J recombinationgoesfar beyond the alreadysophisticatedprocessofan appropriatespatialand temporal tuning of chromosomalaccess to the particular RSSs. In addition however. studiessurrounding the inhibition of Tcrb gene rearrangements once the developingthymoeytes havepassedthrough ~-selection and reachedthe DP thymiccellstage(whererearrangedTcrb geneexpressionismaintainedand V(D)J recombinationtargetedtowardsthe Tcr-Ja locus).haverecentlyrevealed puzzlingresultson this matter.Firstly. the fewE~-I- thymocytes whichdifferentiate intoDP cells (viatransientexpression ofaybTCR-see ref 37).partiallyrecover chromatinaccessibility overtheir D~-J~-C~ (andV~14) domains.indicatingno further requirementofE~ to unwrapheterochromatin over these domains at this stage.38.41 YetV~I4-to-DJ~ rearrangements are stillnot observedat thesesites.Secondly. in micein whichE~ wasreplacedbyEa. or in whichEnwasinserted downstreamofa V~ gene segment in the 5' V~ domain. Tcrb gene rearrangement wasstillinhibited in DP cellsdespitesignsofchromosomalaccessibilitywithin the regionadjacenttothe replaced/newlyinsertedenhancer.42.79Thirdly.V~ genesegments located 5' (up to 150-kb)ofa rearrangedVDJ~ unit are maintained in a relaxed chromosomalform. yet remainrefractorytoV(D)J recornbination.P!' It appears that once a functional variable gene region has been made.further Tcrb gene rearrangement is inhibited via an (epigenetic?) processtes) acting relativelyindependentlyof merechromatin accessibility.

What are the molecularmechanism(s) that enforce the inhibition of Tcrb gene recombination in DP thymocytes? Pioneeringstudiesusingfluorescence in situ hybridization(FISH) haverevealed allelesubnuclear(re)positioningand large-scale locuscontraction/ chromosomal loopingas novelprocesses that mayalsobe involved in the developmental regulationofgeneexpressionand recombination at immune loci.67.82Recently. usingboth FISH and chromosomeconformation capture (3C) assays. the Tcrb and Tcra lociwereindeed shown to undergo long-rangeDNAcontraction in DN and DP rhymocytes, respectlvely.v The foldingof the Tcrb locus is reversedat the DP stage. raisingthe intriguing possibilitythat this locus contraction/expansion processcontributes to regulatingboth V~-to-DJ~ recombination in earlyDN cellsand its inhibition inmoredeveloped DP cells. respectively (Fig.3).HowIg/Tcr locuscontraction/expansion isachievedis still unclear. Deciphering the underlyingmolecular mechanism(s) and factors involvedwillsignificantly improveour understanding of long-rangesynapsis eventsin V(D)J recombinationand their regulation.

Allelic Exclusion at the Tcrb LocusIndividual lymphocytes generally synthesize antigen receptors of a unique specificity.

Accordingly. the vast majorityof mature a~T-cells bear a singleTC~ chain (out of a possibletwo. one encoded at each Tcrb allele) .84-86 Thisis in part achieved bya phenomenon referredto asallelicexclusion." In fact.similarto the situation firstdescribedat the 19b and Igk loci in B-cells,88approximately 60%of a~T-cells harbor asingleproductive(in framewith C~) VDJ~ rearrangedvariable regiongene, whereas the remaining40% carry two VDJ~s rearrangedin nonproductiveand productiveconfigurations. respectively. The 60/40 ratio isconsistentwith one productiveineverythree rearrangements at individualV-to-DJ joints and a feedbacksignalwherebyaTCRlIgproduct fromone functionallyrearrangedalleleleadsto the inhibition offurther V-to-DJ joiningon the opposite allele. This regulatedmodel of allelic exclusion clearly impliesan initiation stepwherebyhomologousalleles are sequentially targeted for recombination.87

Initiation ofAllelicExclusion at the Tcrb LocusStudiesmainlycarriedout atIgh andIgk locihaveledto twodistinct types ofmechanistic model

explaininghow V(D)J recombination maybe initiated in an allelic-specific manner.82,87 On theone hand. instructivemodelscallupon epigenetics [i,e.•DNA (de)methylation. histone tagging,nuclear (pericentromeric) positioning and/or other less-well definedepigeneticmark(s) shownby asynchronous DNA replication] to differentially label the two alleles such that only one will

MolecularGenetics at the T-eellReceptor fJLocus 127

be prone to rearrange. On the other hand, stochastic models evoke inter-allelic competition (andgenerally a low probability in allelic activation) as a means to dissociate allele rearrangements. Inthis context. analysesat the Tcrb locus brought contrastingresults.2.89 Indeed, severalfindings arguedagainst an intrinsically favored recombinational accessibility towards one allele only at this locus.When investigated, DNA demethylation at the D~-J~-C~ regions and germline transcription ofa V~-containing region appeared, at least initially, biallelic (refs. 23.69,81 and our unpublishedresults). In addition, V~-to-DJ~ recombination has been shown to initiate on one allele beforecompletion of all possible VDJ~ rearrangements on the opposite allele." Nonetheless, like Igalleles, Tcrb alleles seem to replicate asynchronously in developing thymocytes, with one oftenbeing recruited to pericentromeric heterochromatin.P'" Additional efforts willthus be requiredto reconcile these apparently conflicting observations.

Feedback Inhibition ofTcrb RecombinationIn late DN3 thymocyres, the assembly ofthe newly formed TCR~ chain with the invariant

preTa chain and CD3 complex subunits forms the pre-TCR (reviewed by von Boehmer et al92).Pre-TCR-mediated signaling-which in addition to the pre-TCR components involves manydownstream kinases and adaptator molecules such as e.g., p5Scltand SLP-76-ensures ~-selection

and a number ofcritical outcomes for cells bypassing this checkpoint [i .e., the maintenance ofcell survival, induction ofcell proliferation and differentiation into DP thymocytes and allelicexclusion) . As a result. disruption of pre-TCR signaling by gene inactivation ofpTa, cd3 orslp-76.blocks a~T-cell development at the DN thymic celIstage and impairs allelic exclusion.93'9S

Conversely, enforced expression of a Tcrb transgene in DN thymocytes inhibits endogenousV~-to-DJ~ rearrangements and promotes developmental progression into DP celIs.96,97 A varietyofbasic processes have been suggested to account for the suppression of Tcrb gene rearrangements by pre-TCR signaling, including cell-cycle-dependent degradation of RAG2 and V~

gene silencing." In addition, as described below, a dissection ofpre-TCR downstream signalingdemonstrated that immature T-cells utilize distinct pathways to achieve allelic exclusion versuscell expansion and differentiation.2.89

The pre-TCR promotes activation ofmultiple signaling pathways including Ca2+ flux, proteinkinase C (PKC) and RAS-RAF-MAP kinase (MAPK) signaling pathways. Activation ofpre-TCRproximal p56H or more distal PKC kinasesissufficient to induce all aspects of~-selection, includingallelicexclusion.88,98Strikinglyhowever, in DN thymocytes, smallGTPases RAS-or RAF-mediatedactivation of the MAPK pathway induces T-cell expansion and cellular differentiation but doesnot block Tcrb gene rearrangement, implying a partition ofSignaltransduction for the feedbackinhibition ofV~-to-DJ~ joining and for cellular expansion/differentiation somewhere betweenthe PKC and RAS/RAF activation nodes.88.99 However, a normal heterochromatin layout is observed along the V~ locus in DP thymocytes generated via MAPKactivation,100 indicatingdistinctrequirements in DP and DN cells to sustain inhibitory features. Reduced accessibility ofV~ genesegments in DP thymocytes likely contributes to lock out allelic exclusion,41,7o.72 relying on thesetting ofan appropriate developmental program via the induction ofdiscrete TFs (reviewed byRothenberg et aI59). Concordantly, notably, Etsl-deficient thymocytes were shown to display adisruption ofallelic exclusion and impaired DN3-to-DP cell transition.'?' An ultimate goal willbe to link pre-TCRsignal transduction cascadeswith all the nuclear functions involved in securingallelic exclusion. including those particular factors that possibly reduce chromatin accessibility atV~ promoters thereby repressing transcription and recombination.

Allelic exclusion likely involves multiple layers of control in addition to the mere changesin chromatin accessibility," Indeed, inhibition ofV~ gene rearrangement in D P thymocytesis preserved when V~ gene segments are maintained in a transcriptionally active (accessible)configuration by insertional knock-in ofEaY·79 In this context, as mentioned, 'taking away' V~genes by pericentromeric allele recruitment or Tcrb locus decontraction could playa significantrole." In support ofthis latter possibility, distinct gene knock-in experiments, in which aV~ genewas introduced in proxim ity (5') to the D~1 or D~2 region. demonstrated an escape from allelic


exclusion (at the level ofV~-to-DJ~ or V~-to-D~ recombination, respectively).53.J02.I03 However,these mechanisms still cannot explain the maintenance ofan inhibition for recombination oftheunrearranged (and apparently accessible)V~ segments located S'proximal (in cis) ofa rearrangedV~DJ~ variable gene region.80,81 Hence, supplementary constraints may act in an as-yet undefinedway to eventually enforce allelic exclusion.

Conclusion and Future DirectionBecause the immune system is not absolutely required for survival in a pathogen-free environ

ment, antigen receptor geneshave served as tractable models to study the regulation ofgeneexpression at complex genomic loci usingmost notably gene rargetingtechnologies. Ashas been illustratedhere . the Tcrb locus is one of the most genetically modified in the mouse so far. During the pastdecade. these studies have led to a better understanding ofthe function ofdistinct cis-regulatoryelements and their hierarchical impact in the control oflocus expression and recombination interms ofchromatin accessibility (Fig. 2 and 3). Moreover, previously unsuspected processes havebeen unraveled, including subnuclear localization and locus contraction/expansion, not to mention the RSS biased usage. It remains to be resolved which processes are of general use during(lymphoid-) cell differentiation programs and which are more specific to the control ofV(D)Jrecombination, or even Tcrb locus expression.

Additional findings will undoubtedly complete our knowledge in the near future. In particular.progresses should include a better characterization ofthe nucleoprotein complexes recruited toTcrb-regulatory elements. the resulting epigenetic features orchestratingaccessibility/heterochromatinization at this locus and plausible interplays with DNA-modifying machineries [includingthe RAG2 and/or additional component of the V(D)J recombinase; refs. 104-106]. The recentdevelopment oflarge-scale.genome-wide ChIP-on-chip methodologies looks particularlypromising in this regard. In parallel. it will be important to establish whether intergenic and/or antisensetranscripts (see chapters by M. Krangel and A. Corcoran) are also produced at the Tcrb locus and.ifso, investigating what impact they have on accessibilityto the V(D)J recombinase. Likewise.does'transcriptional interference' playa role at this locus? Finally.a better understandingofthe precisemechanisms leading to the establishment and enforcement ofallelic exclusion isstill expected and,further still are insights into the potential risks for the immune system/organism caused by relaxedaccessibility/allelic exclusion during Tcrb locus recombination.

AcknowledgementsWe thank members of the PF lab for critical reading of this manuscript. Work in this

lab is supported by institutional grants from Inserm and the CNRS and by specific grantsfrom the 'Fondation Princesse Grace de Monaco', the 'Association pour la Recherche sur leCancer' (ARC 327S & 3136). the 'Agence Nationale de la Recherche' (IMPBlOO-04S632 &ANR-BLAN-0396-03/BYOS-0006-01) and the Commission of the European Communities(MRTN-CT-03S733). The E~-targeted mice studied in the PF lab were generated with the helpof the functional genomic platform of the Marseille -Nice Genopole",

References1. Jung D. Alt FW. Unraveling V(D)J recombination; insights into gene regulation. Cell 2004;

116:299-311.2. Jackson AM. Krangel MS. Turning T-cdl receptor ~ recombination on and off: more questions than

answers. Immunol Rev 2006; 209:129-141.3. Glusman G. Rowen L. Lee I ec al, Comparative genomics of the human and mouse T-cell receptor loci.

Immunity 2001; 15:337-349.4. Cauwelier B. Dastugue N. Cools J et aI. Molecular cytogenetic study of 126 unselected T-ALL cases

reveals high incidence of TC~ locus rearrangements and putative new Tscell oncogenes. Leukemia2006; 20:1238-1244.

5. Schlissel MS. Kaffer CR. Curry JD. Leukemia and lymphoma: a COSt of doing business for adaptiveimmunity. Genes Dev 2006; 20:1539-1544.

6. Nemazee D. Receptor selection in Band T-lymphocytes. Annu Rev Immuno12000: 18:19-51.

Molecular Geneticsat the T-CellReceptorPLocus 129

7. Ramsden DA. Baetz K. Wu GE. Conservation of sequence in recombination signal sequence spacers.Nucleic Acids Res 1994; 22:1785-1796.

8. Tillman RE. Wooley AL. Hughes MM et al. Regulation of T'cell receptor p-chain gene assembly byrecombination signals: the beyond 12/23 restriction . Immunol Rev 2004; 200:36-43.

9. Livak F,Burtrum DB. Rowen L et al. Genetic modulation ofT-cell receptor gene segment usage duringsomatic recombination.] Exp Med 2000; 192:1191-1196.

10. Olaru A. Patterson DN. Cai H et al. Recombination signal sequence variations and the mechanism ofpatterned T-cell receptor-fllocus rearrangement. Mol Immunol 2004; 4O:1189-1201.

II . Bassing CH. Alt FW; Hughes MM et al. Recombination signal sequences restrict chromosomal V(D)]recombination beyond the 12/23 rule. Nature 2000; 405:583-586.

12. Slecianan BP. BassingCH. Hughes MM et al. Mechanisms that direct ordered assemblyofT-cell receptor p locus V. D and] gene segments. Proc Natl Acad Sci USA 2000; 97:7975-7980.

13. Wu C. Ranganath S. Gleason M er al. Restriction of endogenous T-cell antigen receptor fl rearrangements to Vp14 through selective recombination signal sequence modifications. Proc Natl Acad Sci USA2007; 104:4002-4007.

14. Wu C. Bassing CH. lung D et al. Dramatically increased rearrangement and peripheral representationofVP14 driven by the 3'Dfll recombination signal sequence. Immunity 2003; 18:75-85.

15. Hughes MM. Yassai M. Sedy ]R et al. T-cell receptor CDR3 loop length repertoire is determinedprimarily by features of the V(D)] recombination reaction. Eur] Immuno12003; 33:1568-1575.

16. Tillman RE. Wooley AL. Hughes MM et al. Restrictions limiting the generation of DNA double strandbreaks during chromosomal V(D)J recombinarion.} Exp Med 2002; 195:309-316.

17. Olaru A. Patterson DN. Villey I et al. DNA-Rag protein interactions in the control of selective D geneutilization in the TCRfllocus. J Immunol 2003; 171:3605-3611.

18. lung D. Bassing CH. Fugmann SD et al. Extrachromosomal recombination substrates recapitulatebeyond 12/23 restricted V(D)J recombination in nonlymphoid cells. Immunity 2003 ; 18:65-74.

19. Tillman RE. Wooley AL. Khor B et al. Targeting ofVfl to Dp rearrangement by RSSs can be mediatedby the V(D)J recombinase in the absence of additional lymphoid-specific factors. ] Immunol 2003;170:5-9.

20. Hughes MM. Tillman RE. Wehrly TD et al. The Bl2/23 restriction is Criticallydependent on recombination signal nonamer and spacer sequences. ] Irnmunol 2003; 171:6604-66 10.

21. Drejer Teel AH. Fugmann SD. Schatz DG. The beyond 12/23 restriction is imposed at the nicking andpairing steps of DNA cleavage during V(D)] recombination. Mol Cell Bioi 2007; 27:6288-6299.

22. Hashimoto Y. T-cell receptor fl gene has two downstream DNase I hypersensitive regions. J Exp Med1989; 169:2097-2107.

23. Chattopadhyay S. Whitehurst CEo Schwenk F et al. Biochemical and functional analyses of chromatin changes at the TCR-fl gene locus during CD4-CD8- to CD4'CD8' thymocyte differentiation.J Imrnunol 1998; 160:1256-1267.

24. Sikes ML. Gomez RJ. Song] et al. A developmental stage-specificpromoter directs germline transcription of Dfl]fl gene segments in precursor T-Iymphocytes.] Immuno11998; 161:1399-1405 .

25. Dory RT. Xia D. Nguyen SP er al. Promoter element for transcription of unrearranged T-ccIl receptorfl-chain gene in pro -T-cells. Blood 1999; 93:3017-3025 .

26. Ktimpenfort P. De long R. Uematsu Y et al, Transcription of T-cell receptor fl-chain genes is controlledby a downstream regulatory element. EMBO] 1988; 7:745-750.

27. McDougall S. Peterson CL. Calame K. A transcription enhancer 3' of csz in the T-cell receptor fllocus. Science 1988; 241:205-208.

28. Busse CEo Ktotkova A. Eichmann K. The TCRfl enhancer is dispensable for the expression of rearranged TCRfl genes in thymic DN2/DN3 populations but not at later stages. ] Immunol 2005;175:3067-3074.

29. Takeda]. Cheng A. Mauxion F et al. Functional analysis of the murine T-cell receptor fl enhancer andcharacteristics of its DNA-binding proteins. Mol Cell Bioi 1990; 10:5027-5035.

30. Leiden ]M, Thompson CB. Transcriptional regulation ofT-cell genes during Tvcell development. CurOp Immunol 1994; 6:231-237.

31. Carvajal 1M. Sen R. Functional analysis of the murine TCR fl-chain gene enhancer. ] Immunol 2000;164:6332-6339.

32. Tripathi RK. Mathieu N. Spicuglia S et al. Definition of a T-cell receptor fl gene core enhancer ofV(D)J recombination by transgenic mapping. Mol Cell Bioi 2000; 20:42-53.

33. Capone M. Watrin F. Fernex C et al. TCRfl and TCRn gene enhancers confer tissue- and stage-speci6city on V(D)] recombination events. EMBO] 1993; 12:4335-4346.

34. Okada A. Mendelsohn M. Alr FW: Differential activation of transcription versus recombination oftransgenic T-cell receptor fl variable region gene segments in B and T lineage cells.] Exp Med 1994;180:261-272.


35. Bouvier G, Watrin F, Naspetti M er al. Deletion of the mouse T-cdl receptor ~ gene enhancer blocksa~ T-cdl development, Proc Nat! Acad Sci USA 1996; 93:7877-7881.

36. Bories JC , Demengeot J, Davidson L et al. Gene-targeted deletion and replacement mutations of theT-cdl receptor ~-ehain enhancer: the role of enhancer dements in controlling V(D)J recombinationalaccessibility. Proc Nat! Acad Sci USA 1996; 93:7871-7876.

37. Leduc I. Hempel WM, Mathieu N ec al. T-cdl development in TCRfl enhancer-deleted mice: implications for a~ 'I-cell lineage commitment and differentiation. J Immunol20oo; 165:1364-1373.

38. Ryu CJ. Haines BB. Draganov DD et al. The T-cdl receptor ~ enhancer promotes access and pairing of D~ and J~ gene segments during V(D)J recombination. Proc Nat! Acad Sci USA 2003 ;100:13465-13470.

39. Mombaerts P, Clarke AR, Hooper M ee al. Creation of a large genomic deletion at the T-cdl antigenreceptor ~-susunit locus in mouse embryonic stem cdls by gene targeting. Proc Nat! Acad Sci USA1991; 88:3084-3087.

40. Hempel WM. Stanhope-Baker P, Mathieu N et al. Enhancer control of V(D)J recombination at theTCRfllocus: differential effects on DNA cleavage and joining. Genes Dev 1998; 12:2305-2317.

41. Mathieu N, SpicugliaS. Gorbarch S et al. Assessing the role of the T'cell receptor ~ gene enhancer in regulating coding joint formation during V(D)J recombination . J BioI Chern 2003; 278:18101-18109.

42. Senoo M. Mochida N. Wang L et al. Limited effect of chromatin remodeling on D~-to-J~ recombinationin CD4+CD8+ thymocyte: Implications for a new aspect in the regulation ofTCR ~ gene recombination. Int Immuno12001; 13:1405-1414.

43. Ferrier P, Krippl B, Blackwell TK et al. Separate dements control DJ and VOJ rearrangement in atransgenic recombination substrate. EMBO J 1990; 9:117-125.

44. Eyquem S. Lagresle C. Fasseu M et al. Disruption of the lineage restriction of TCRfl gene rearrangements. Eur J Immunol 2002; 32:3256-3266.

45. Chattopadhyay S, Whitehurst CE, Chen J. A nuclear matrix attachment region upstream of the T-cellreceptor ~ gene enhancer binds Cux/CDP and SATBI and modulates enhancer-dependent reportergene expression but not endogenous gene expression. J BioI Chern 1998; 273:29838-29846.

46. Whitehurst CE, Hu H, Ryu CJ et al. Normal TCRfl transcription and recombination in the absenceof the J~2-C~2 intronic cis dement. Mol Immunol 2001; 38:55-63.

47. Wh itehurst CE, Chattopadhyay S. Chen J. Control ofV(D)J recombinational accessibilityof the D~1

gene segment at the TCRfllocus by a germline promoter. Immunity 1999; 10:313-322.48. Whitehurst CE, Schlissd MS. Chen J. Ddetion of germline promoter PD~1 from the TCRfl locus

causes hypermethylation that impairs D~I recombination by multiple mechanisms. Immunity 2000;13:703-714.

49. AlvarezJD . Anderson SJ, Loh DY. V(D)J recombination and alldic exclusion of a TCR ~-chain minilocus occurs in the absence of a functional promoter. J Immuno11995; 155:1191-1202.

50. Halle JP, Haus-Seuffcrt p, Woltering C et al. A conserved tissue-specific structure at a human T-cellreceptor ~-chain core promoter. Mol Cell Bioi 1997; 17:4220-4229.

51. Chen F, Rowen L, Hood L et al. Differential transcriptional regulation of individual TCR V~ segmentsbefore gene rearrangement. J Immunol 2001; 166:1771-1780.

52. Ryu CJ. Haines BB. Lee HR er al. The T-cell receptor ~ variable gene promoter is required for efficientV~ rearrangement but not allelic exclusion. Mol Cell BioI 2004; 24:7015-7023.

53. Sieh P. Chen J. Distinct control of the frequency and alldic exclusion of the V~ gene rearrangement atthe TCRfllocus. J Immunol 2001; 167:2121-2129.

54. Oestreich KJ, Cobb RM, Pierce S et al. Regulation ofTCRfl gene assembly by a promoter/enhancerholocomplex. Immunity 2006; 24:381-391.

55. Hollenhorst PC , Shah AA. Hopkins C et al. Genome-wide analysesreveal properties of redundant andspecific promoter occupancy within the ETS gene famlly. Genes Dev 2007; 21:1882-1894.

56. Kawazu M. Asai T, Ichikawa M et al. Functional domains of runxl arc differentially required for CD4repression. TCRfl expression and CD4/8 double-negative to CD4/8 double-positive transition in thymocyte development, J Immuno12005; 174:3526-3533.

57. Spicuglia S. Kumar S. YehJH et al. Promoter activation by enhancer-dependent and -independent loading of activator and coactivator complexes. Mol Cell 2002; 10:1479-1487.

58. Yang XO. Doty RT. Hicks JS et al. Regulation of T-cell receptor D~1 promoter by KLF5 throughreiterated GC -rich motifs. Blood 2003; 101:4492-4499.

59. Rothenberg EV, Taghon T. Molecular genetics of T-cell development. Annu Rev Immunol 2005;23:601-649.

60. Barndt R. Dai MF, Zhuang Y. A novd role for HEB downstream or paralld to the pre-TCR signalingpathway during a~ thymopoiesis. J Immuno11999; 163:3331-3343.

61. Bender TP. Kremer CS, Kraus M et al. Critical functions for c-Myb at three checkpoints during thymocyte development, Nat Immunol 2004; 5:721-729.

Molecular Genetics at tbt T-CtllReceptor PLocus 131

62. Wakabayashi Y, Watanabe H. Inoue J et al. Bcl11b is required for differentiation and survival of aflT-Iymphocyres. Nat Immunol 2003; 4:533-539.

63. Kaul-Ghanekar R. Majumdar S. Jalota A er al. Abnormal V(D)J recombination of T-cell receptor fllocus in SMARI transgenic mice. J Bioi Chern 2005; 280 :9450-9459.

64. Wolfer A. Wilson A. Nernir M et al. Inactivation of Notch 1 impairs VDJfl rearrangement and allowspre-TCR-independent survival of early afllineage thymocytes. Immunity 2002; 16:869-879.

65. Yancopoulos GD. Ale FW. Developmentally controlled and tissue-specific expression of unrearrangedVH gene segments. Cell 1985; 40:271-281.

66. Stanhope-Baker P. Hudson KM. Shaffer AL er al. Cell type-specific chromatin structure determines thetargeting ofV(D)J recombinase activity in vitro. Cell 1996; 85:887-897.

67. Spicuglia S. Franchini DM. Ferrier P. Regulation ofV(D)J recombination. Curr Opin Immunol2006;18:158-163.

68. Cobb RM. Oestreich KJ. Osipovich OA ec al. Accessibility control of V(D)J recombination. AdvImmunol 2006; 91:45-109 .

69. Mathieu N. Hempel WM. Spicuglia S ee al. Chromatin remodeling by the T-cell receptor (TCR)-fl geneenhancer during early T-cell development : implications for the control of TCR-fllocus recombination.J Exp Med 2000; 192:625-636.

70. Tripath i R. Jackson A. Krangel MS. A change in the structure of Vfl chromatin associated with TCRfl allelic exclusion. J Immunol2002; 168:2316-2324.

71. Morshead KB. Ciccone DN. Taverna SD er al. Antigen receptor loci poised for V(D)J rearrangementare broadly associated with BRG1 and flanked by peaks of histone H3 dlmethylared at lysine 4. ProcNatl Acad Sci USA 2003; 100:11577-11582.

72. Senoo M. Shinkai Y. Regulation of Vfl germline transcription in RAG-deficient mice by theCD3r-mediated signals: implications of Vfl transcriptional regulation in TCR fl allelic exclusion. InrImmuno11998; 10:553-560.

73. Huang F, Cabaud O. Verrhuy C et al afl T-cell development is not affected by inversion of TCR flgene enhancer sequences: polar enhancement of gene expression regardless of enhancer orientation.Immunology 2003; 109:510-514.

74. SikesML. Meade A. Tripathi Ret al. Regulation ofV(D)J recombination: A dominant role for promoterposition ing in gene segment accessibility. Proc Natl Acad Sci USA 2002; 99:12309-12314.

75. Krangel MS. T-cell development : better living through chromatin. Nat Immunol2007; 8:687-694.76. Osipovich O. Milley R. Meade A et al. Targeted inhibition ofV(D)J recombination by a histone merh

yltransferase. Nat ImmunoI2004; 5:309-316.77. Osipovich O. Milley CR. Oestreich KJ et al. Essential function for SWl-SNF chromatin-remodeling

complexes in the promoter-directed assembly of Tcrfl genes. Nat Immuno12007; 8:809-816.78. Baumann M. Mamais A. McBlane F er al. Regulation ofV(D)J recombination by nucleosome positioning

at recombination signal sequences. EMBO J 2003; 22:5197 -5207.79. Jackson A. Kondilis HD. Khor B et al. Regulation ofT-cell receptor fl allelic exclusion at a level beyond

accessibility. Nat Immunol 2005; 6:189-197.80. Jackson AM. Krangel MS. Allele-specificregulation ofTCRfl variablegene segment chromatin structure .

J Immuno12005; 175:5186-5191.81. Jia J. Kondo M. Zhuang Y. Germline transcription from T-cell receptor Vfl gene is uncoupled from

allelic exclusion. EMBO J 2007; 26:2387-2399 .82. Corcoran AE. Immunoglobulin locus silencing and allelic exclusion. Semin Immunol 2005 ;

17:141-154.83. Skok JA. Gisler R. Novatchkova M et al. Reversible contraction by looping of the Tcra and Tcrfl loci

in rearranging thymocytes. Nat Immunol 2007 ; 8:378-387 .84. Balomenos D. Balderas RS. Mulvany KP et al. Incomplete T-cell receptor Vfl allelic exclusion and dual

Vfl-expressingcells. J Immuno11995; 155:3308-3312.85. Davodeau F. Peyrat MA. Romagne F et al. Dual T-cell receptor fl chain expression on human

T-Iymphocytes. J Exp Med 1995; 181:1391-1398.86. Padovan E. Giachino C. Cella M et al. Normal T-Iymphocytes can express two different T-cell receptor

fl chains: implications for the mechanism of allelic exclusion. J Exp Med 1995; 181:1587-1591.87. Moscoslavsky R. Alt FW. Rajewsky K. The lingering enigma of the allelic exclusion mechanism. Cell

2004 ; 118:539-544.88. Gartner F. Ale FW. Monroe R er al. Immature thyrnocytes employ distinct signaling pathways for allelic

exclusion versus differentiation and expansion. Immunity 1999; 10:537-546.89. Khor B. Sleckman BP. Allelic exclusion at the TCRfllocus. Curr Opin Immunol 2002; 14:230-234.90. Khor B. Sleckman BP. Intra- and inter-allelic ordering of T-cell receptor fl chain gene assembly. Eur J

Immunol 2005; 35:964-970.


91. Mostoslavsky R, Singh N, Tenzen T et al. Asynchronous replication and allelic exclusion in the immunesystem. Nature 2001; 414:221-225.

92. von Boehmer H. Aifantis I, Azogui 0 ec al. Crucial function of the pre-T'cell receptor (TCR) inTCRjl selection. TCRjl allelic exclusion and afl versus ylliineage commitment. Immunol Rev 1998;165:111-119.

93. Aifantis I. Buer ]. von Boehmer H er al. Essential role of the preT-cell receptor in allelic exclusion ofthe Tvcell receptor fllocus . Immunity 1997; 7:601-607 .

94. Ardouin L. Ismaili ]. Malissen B er al. The CD3-yllt and CD3-l;/lJ modules arc each essential forallelic exclusion at the T-cell receptor fllocus but arc both dispensable for the initiation of V to (D)Jrecombination at the Tscell recepror-B, -y and -II loci. ] Exp Med 1998; 187:105-116.

95. Aifantis I, Pivniouk VI, Gartner F et al, Allelic exclusion of the T-cell receptor fllacus requires the SH2domain-containing leukocyte protein (SLP)-76 adaptor protein.] Exp Med 1999; 190:1093-1102 .

96. Uernatsu Yo Ruser S. Dernbic Z et al. In transgenic mice the introduced functional T-cell receptor flgene prevents expression of endogenous fl genes. Cell 1988; 52:831-841.

97. Pircher H, Ohashi P. Miescher G et al. T-cell receptor (TcR) fl chain transgenic mice: studies on allelicexclusion and on the TcR+ alII populanon.] Immuno11990; 20:417-424.

98. Michie AM, Soh ]W, Hawley RG et al, Allelic exclusion and differentiation by protein kinaseC-mediated signals in immature thymocyres . Proc Natl Acad Sci USA 2001; 98:609-614.

99. Iritani BM. Alberola-Ila ], Forbush KA et al. Distinct signals mediate maturation and allelic exclusionin lymphocyte progen itors. Immunity 1999; 10:713-722.

100. Jackson AM, Krangel MS. A role for MAPK in feedback inhibition of tcrb recombination] Immunol2006 ; 176:6824-6830.

101. Eyquem S, Chemin K, FasscuM et aJ.The Ets-I transcription factor is required for complete pre-T-cellreceptor function and allelic exclusion at the T-cell receptor fllocus. Proc Natl Acad Sci USA 2004;101:15712-15717.

102. Senoo M. Wang L, Suzuki D et al. Increase ofTCR Vfl accessibility within Efl regulatory region influences its recombination frequency but not allelic exclusion.] Immunol 2003; 171:829-835.

103. Suzuki D, Wang L. Senoo M et al. The positional effect of Efl on Vfl genes of TCRjl chain in theordered rearrangement and allelic exclusion. Int Immuno12005; 17:553-1560.

104. Liu Yo Subrahmanyam R. Chakraborty T et al. A plant homcodomain in Rag-2 that binds hypermethylated lysine 4 of histone H3 is necessary for efficient antigen-receptor-gene rearrangement. Immunity2007; 27:561-571.

105. Matthews AG. Kuo A], Ramon -Maiques S et al. RAG2 PHD finger couples histone H3 lysine4 trimerhylatlon with V(D)J recombination. Nature advance online publication. 2007 (DOl 10.10381nature06431 ).

106. Ramon-Maiques S, Kuo A], Carney D er al. The plant homeodomain finger of RAG2 recognizes histoneH3 methylated at both lysine-4 and arginine-2. Proc Nad Acad Sci USA 2007; 104:18993-18998.

CHAPTER 11

Molecular Pathways and MechanismsRegulating the RecombinationofImmunoglobulin Genesduring B-Lymphocyte DevelopmentKristen Johnson, Karen L. Reddy and Harinder Singh"

Abstract

T he hallmark ofB-cell development is the ordered recombination ofimmunoglobulin (Ig)genes. Recently, considerable progress has been achieved in assembling gene regulatorynetworks comprised of signaling components and transcription factors that regulate

B-cell development. In this chapter we synthesize experimental evidence to explain how suchsignaling pathways and transcription factors can orchestrate the ordered recombination ofimmunoglobulin (Ig) genes. Recombination of antigen-receptor loci is regulated both by thedevelopmentally controlled expression of the Rag] and Rag2 genes and the accessibility ofparticular loci and their gene segments to recombination. A new framework has emerged thatinvokes nuclear compartmentalization, large-scale chromatin dynamics and localized changes inchromatin structure in regulating the accessibilityofIg loci at specificstagesofB -celldevelopment.We reviewthis emergent framework and discuss new experimental approaches that will be neededto explore the underlying molecular mechanisms.

IntroductionB-cell development is orchestrated by the coordinated action of signaling pathways and

transcription factors that promote survival, proliferation and differentiation of hierarchicallyordered progenitors (Fig. 1). B-lineage cells are directly derived from hematopoietic progenitorsin the bone marrow termed early lymphoid progenitors (ELP) or common lymphoid progenitors(CLP) that alsohave the potential to differentiate into T, NK and to a lesserdegree myeloid cells.1.2

B-cell fate specification and commitment occur at the pro-B-cell stage. The rearrangement ofimmunoglobulin D-to-]H gene segments is activated within the ELP/C LP pool ofdevelopmentalintermediates and is completed at the pro-B-cell stage.3.4 Since ELPs and CLPs can differentiateinto alternate cell types, DlH rearrangements are not a defining molecular feature of B-lineagecells. However, the joining of V-to-D] H segments occurs exclusively within pro -B-cells andis a hallmark of commitment to the B-lineage. This step of rearrangement is highly regulatedrequiring multiple signaling and transcription factor inputs, thereby ensuring developmentaltiming and lineagespecificity(Fig. 1). Productive rearrangement ofan Igheavychain allelegenerates

·Corresponding Author : Harinder Singh-Howard Hughes Medical Institute, Department ofMolecular Genetics and Cell Biology, The University of Chicago, Chicago, IL, USA 60637,USA. Email: hsinghwuchlcago.edu

V(D)JRecombination, edited by Pierre Ferrier. ©2009 Landes Bioscienceand Springer Science+ Business Media.

134 V(D}]Recombination

()fe-OCR OCR: :

~ IIF!:~

7 ~ r u,

"E .l vw.,.. I .. Pu5 .Em2. .. IRF" .I E2A IRF".1 ~ ..E2AEBF YVt , _ Rag v"L~ ~ exe... v.J.

eXCL'1",pp ELPIClP ..0-8 ...·8 .....8 ammatureB

IIo<geqdmgl (sma ll rnt:ng~

Figure 1. Overview of signaling pathways and transcription factors controlling B-cell development and V(D)J recombination. The init ial and final configurations of the immunoglobulin loci are denoted in MPP and immature B-cells, respectively. Specific recombinationevents that occur in developmental intermediates are shown along with their regulators. Inmultipotent progenitors (MPP) that give rise to all hematopoietic lineages, the Ig loci arein their germline configuration (not recombined) and are transcriptionally inactive. Earlylymphoid progenitors (ELP) or common lymphoid progen itors (CLP) express low levels of thetranscr iption factors EBF and E2A that regulate the initial expression of the Rag genes andD-to-JH recombination. IL-7 signaling has been implicated in regulating expression of EBF.B-cell fate specification is directed by the upregulation of EBF that induces Pax5 and resultsin the generation of pro-B -cells. IL-7signaling via the transcr ipt ional activator Stat-5 regulatesdistal VHgene transcription and accessibility. In addition Pax-5, Ezh2 and YY1 are requiredfor distal V-to-DJH recombination . After successful V-to -DJH recomb ination , the pre-BCR isdisplayed on the cell surface. Constitutive signaling through this receptor upregulates IRF4/8expression while continued signaling through the IL-7receptor initially inhibits V-to-JKrecombination by blocking E2A accessibility to the intronic IgK enhancer and repressing Rag geneexpression. IRF4,8 directly bind to and activate the kappa 3/ enhancer. IRF4also induces thechemokine receptor CXCR4 that is proposed to move pre-B-cells away from IL-7 producingstroma in the bone marrow resulting in loss of IL-7 signaling. Consequently, E2A access tothe IgK intronic enhancer (E2A*) is enabled, the Rag genes are highly expressed and efficientV-to -JK recombination ensues. Productive rearrangement of an Ig light chain allele results inassembly and expression of the B-cell receptor (BCR).

pre-B-cells in which the heavy-chain protein pairs with the surrogate light-chains , AS and Vpre-B,to form the pre-BCR.5Pre-B-cells undergo a self-limitingproliferative expansion mediated by thepre-BCR and the IL-7 receptor, thereby amplifying clones with successful IgH rearrangements.During this cycling pre-B-cell phase, the Raggenes are downregulatedP Upon cell cycleexit, Raggene expression is re-induced and the cellsactivate rearrangement oftheir Iglight-chain loci. Bothheavy and light chain rearrangements are subject to allelic exclusion, a process that ensures thatonly a single productively rearranged allele isgenerated and expressed thereby ensuring that a givenB-cell expresses a unique antigen receptor. Inthis chapter we synthesize experimental evidence tounderstand how signaling pathways and transcription factors can orchestrate the developmentallyordered recombination ofimmunoglobulin (Ig) genes and also enforce allelic exclusion.

Recombination ofantigen-receptor loci is regulated both by the developmentally controlledexpression of the Rag] and Rag2 genes that encode the V(D)J recombinase as well as by theaccessibility of particular loci and their gene segments to recombination. Recently, a newframework has emerged that invokes nuclear compartmentalization and large-scale chromatindynamics in addition to localized changes in chromatin structure in regulating the accessibilityoflgloci at specificstages ofB-celldevelopment. It has been established that the fundamental structuralunit of chromatin, the nucleosome, can inhibit recombination when positioned directly over arecombination signal sequence (RSS).8.9 Thus it is widely accepted that accessibility to

recombination ofindividual gene segments must involvelocalized changes in nucleosome structureand positioning (Fig.2).10 Such changesare brought about by chromatin modifying complexes thatare recruited to specific nucleosomal regions by transcription factors. Considerable progress hasbeen achieved in elucidating distinct molecular mechanisms by which transcription factors and

Molecular Pathways and Mechanisms RegulatingtheRecombination ofImmunoglobulin Genes 135

A

HJK9-JM"H3K9-3Me/

B

Chrcmabn directed

C

Scqucnco o.,ccled

Figure 2. Molecular mechanisms regulating chromatin accessibility and recruitment of theRAG complex to recombination signal sequences (RSS). A) In the inactive state, Ig gene segments contain repressive H3K9 tri-methylation (H3K9-3Me) nucleosome modifications. Thismodification is proposed to inhibit recombinase accessibility to the RSS (red triangle). Inthis scenario, both the RSS and the V gene promoter (green DNA segment) are inaccessible.Activation of germline transcription (green arrow) is accompanied by nucleosome modifications that include histone H4 acetylation (H4-Ac) , histone H3K9 acetylation (H3K9-Ac)and histone H3K4 tri -methylation (H3K4-3Me). This open chromat in structure is proposedto increase accessibility of RAG complex to the RSS. B) Chromatin directed RAG complexrecruitment. This mechanism involves RAG-2 interaction with an RSS-proximal nucleosomecontaining H3K4-3Me. C) Sequence directed RAG complex recruitment. In this scenario theRAG complex is recruited by direct interactions w ith the transcription factor Pax-S bound toa site adjacent to an RSS. Pax-S interacts with both RAG-1 and RAG-2 proteins.

chromatin modifying complexes can locally regulate accessibility ofantigen receptor gene segmentsto the V(D)] recombinase. Chromatin modifying complexes catalyze posttranslational modifications ofhistone tails that can serve to recruit nucleosome remodeling complexes that in turn alterthe positioningofnucleosomes thereby directly changing the accessibility ofrecombination signalsequences (RSSs) to the V(D)] recombinase (Fig. 2A) . In addition, the histone modifications canfunction as molecular scaffolds for more favorable binding of the RAG 1,2 complex (Fig. 2B).Interestingly, a direct interaction between RAG2 and K4rrirnethylated-histone H3 has recentlybeen dernonstrated.U'P'This int eraction appears to promote recombination in vivo. Transcriptionfactors can also directly interact with the RAG 1,2 proteins and recruit them to a nearby RSS (Fig.2B). In this regard, the transcription factor PaxS has been shown to bind to the RAG 1 and RAG2proteins and facilitate their recruitment to an RSS in vitro and promote recombination ofa VH

gene substrate.13It is now recognized that developmentally regulated changes in nuclear compartmentalization ofIg loci also impact their accessibility to recombination (Fig. 3). Ig loci have beenshown to associate in a regulated manner with two distinct repressive nuclear compartments, theinner nuclear membrane-nuclear lamina and pericentrorneric heterochromatin, each ofwhich appear to impair accessibility ofthese loci to recornblnarion.v"!" Given the fact that recombinationofV with D or] gene segments often involve molecular synapses ofRSSs separated by distancesas large as I -2Mb, an important issue is how such long-range recombination events are facilitated.Recent evidence suggests that Igloci undergo compaction or contraction.14,1 7,181his phenomenon

136

MPP

Nuclear Envelope

V(D)j R~,ombination

- -Emem _ _ ONM

- _ IN'"lamlna

(Lamin NC . B)

..... IfIH petmlme or non-pmdudiWtly f8arranpecl _

• ItJHprodllClw./y f8arranpecl allele

.... 1fI~ permline or non-pmduetivoly fllarranpecl a_

t contrlldod ItJH or 1fI~ allele

Figure 3.Regulation oflggene recombination vianuclearcompartmentalization and DNA looping.Igloci undergodevelopmentally regulated changes in nuclearcompartmentalization that havebeen proposed to regulate recombination. These includeassociations with repressive compartmentssuchas the innernuclearmembrane-lamina and pericentromeric heterochromatin. Ig locialsoundergolarge-scalechangesinchromatin configuration, termedcompactionorcontractionthat are thought to represent DNA loops which facilitate long range DNA recombination. Therecombination status and nuclear disposition of the Ig heavy and kappa light chain alleles isdepicted during variousstages of B-cell development. The gray outline indicates the nuclearenvelope that comprises the nuclear membrane and lamina. In multipotential progenitor cellsor non-B-Iineage cells (light green cell), both IgH and IgK alleles are positioned at the nuclearperiphery. At this stage, the IgH loci are decontracted and the distal VHgenes (red oval) arepositionedcloserto the peripherythan the DH, JH or CHregions(greenoval). D-to-JH recombination can occur in thisstate. Shown beloware the proposed interactions of the IgH loci with theinner nuclearmembrane (INM). ONM indicates outer nuclearmembrane. Thedistal VHgenes,but not the CHregion, are in molecularcontact with componentsof the INM-Iamina includingemerin and lamin B. As the cellstransition to the pro-B-cell stage (yellow cells), the Igloci arepositioned awayfrom the nuclearperipherybyan unknownfactor(s) X. Inaddition,at thisstage,the IgH loci undergocompaction (contraction) thereby bringing the distal VHgene segmentsinclose proximity to the DH-JH region by loopingout the intervening DNA. This contraction hasbeen shown to be dependent on Pax-S, a positiveregulator ofV-to-DIHrecombination and YYlthat binds to the Ell enhancer in the CHregion. Finally, at the pre-B-cell stage (pink cell), one ofthe IgK alleles undergoescontraction, while the other remains decontracted and is associatedwith pericentromeric heterochromatin. Moreover, the decontracted IgK allele is often foundto be associated with a decontracted IgH allele at the same pericentromeric heterochromatincluster. Such association with pericentromeric heterochromatin is proposed to contribute toallelic exclusion. IRF4,B, are related transcription factors thatare required for IgK recombination.Theyregulate positioning of an IgK alleleawayfrom pericentromericheterochromatinand maypromote contractionor DNA looping.

appears to reflect higherorderchromosomal DNA loopsthat helpto bringwidely separated genesegments in closer proximity forDNA recombination.Wereview thenewframework foranalyzingIgrecombination accessibility atvarious levels,includingnuclear compartmentalization,chromosomeand chromatinstructure.

Molecular PathwaysandMechanismsRegulating theRecombination ofImmunoglobulin Genes 137

B-Cell Fate Specification andtheJoining ofD-to-JH SegmentsD-to-]H rearrangements are used to define hematopoietic progenitors in which lymphoid po

tential has been induced. Such recombination events are found at low levels in ELPs and CLPS.I,2Pro-Bvcells invariably display DJH rearrangements at both IgH alleles. Genetic experiments haverevealed a requirement for the cytokine receptors Flk2/Flt3 and IL-7R, as well as the transcription factors PUl , Ikaros, E2A and EBF in the generation of pro-Bvcells." These signaling andtranscription components have been assembled into contingent gene regulatory networks thatinitiate B-cell development. In this section we discuss the known functions of these regulatorsin the activation of DlH rearrangement and the molecular mechanisms underlying this earliestrecombination event .

E2A and EBF are key regulators ofB-cell fate specification and loss ofeither factor results in anearly and profound block to B-cell development, in vivo, that appears to be at the stage involvinginitiation ofD-to-]H rearrangemenes.P'" Interestingly, ectopic expression ofeither transcriptionfactor in conjunction with the RAG proteins in a nonlymphoid cell line is sufficient to activateD-to-]H recombination." Consistent with the view that either transcription factor can induceD-tO-]Hrecombination, it has been shown that neither E2A nor EBF are absolutely required forthis rearrangement event during B-cell development. EBF can bypass the requirement for E2A inearly B-cell development and induce D-]H rearrangements in E2A mutant cells when expressedat sufficient levels." Conversely, EBF mutant progenitors when propagated in the presence ofFLand IL-7 express E2A and display Dxo-]Hrearrangements." The molecular mechanisms by whichthese factors are able to activate recombination ofthe D]Hsegments remain to be elucidated. Anattractive possibility is that they bind to sites within and near the intronic IgH enhancer andlocally remodel chromatin structure enabling accessibility ofthe nearby RSSs. In suppon ofthispossibility, the transactivatingdomain ofE2A that interacts with chromatin modifying complexesis required to promote ectopic IgH recombination,"

Localized histone modifications have been implicated in the onset ofD-to-]Hrecombinationon the basis oftheir selective appearance at the D-]H locus as it is poised to undergo recombination (Fig.4). In CD 19+ pro-B-cells isolated from Ragdeficient mice, the D-] Hcluster is associatedwith hyperacerylated histones, suggesting a role for increased histone acetylation in creating alocal region ofaccessibility that can be targeted by the recombinase machinery." The concept ofregion-specific chromatin alterations as a means ofeffecting developmentally ordered changes inrecombination accessibility has recently been strengthened by gene targeting studies that place aVHsegment in close proximity to the DHelements. This resulted in a lossofordered rearrangementfor the targeted VHgene segment. " Discrete chromatin domains within the IgH locus such as theone exemplified by the D-]Hcluster suggest the existence ofboundary elements that may functionto restrict recombination to gene segments within the domain.

Another process that is manifested at the D]H locus prior to recombination is antisensetranscripdon.v-" Antisense intergenic transcription through this region is dependent on the EIJ.enhancer, an element that has been shown to be required for D-to-]Hrecornbination.Y" Basedon these results , it has been suggested that processive antisense transcription through the D]Hregion (60 kb) disrupts repressive chromatin structure thereby facilitating recombinacion.P-"Intriguingly, whereas the 5 .- and 3 ' - D Hsegments are associated with active H3-K9 acetylationmarks the intervening DHsegments, that are infrequently recombined, appear to undergo active histone deacerylation." It has been proposed that the intervening D Hgenes are transcribedin both the antisense and sense orientations leading to the generation of low levels of dsRNAthat promotes repeat induced epigenetic silencing. It remains to be determined ifboth of thefore-mentioned mechanisms are utilized in shaping the repertoire of Dvto-]H recombinationevents in pro-B-cells.

Despite the obvious requirement for D-tO-]Hrearrangement duringB-celldevelopment, asnotedabove this step islessstringently regulated than V-to-D]H recombination. An intriguingexplanationfor this difference in regulated accessibilityissuggested by the topology ofthe IgH locus in lymphoidprogenitors. Specifically, the heavy-chain locus appears to be anchored at the nuclear lamina through

138

Oct·1+

Stat5 ~04------

V(D)j Recombination

E2A+

---+. EBF- - -+. Rag1,2

p

Pax5 +-- - - - - - - - - - - -1Rag recruitment

DNAlooping

CJl Co Cy3

Figure 4. Signaling pathways and transcription factors regulating IgH recombination. Thenetwork depicts signaling pathways and transcriptional regulators that are required for heavychain gene recombination in pro-B-cells. Gray boxes represent the indicated gene segments(not to scale). Early in pro-B-cell development, the CHand DWJH region adopts an open chromatin structure accompanied by localized histone acetylation (pink bars). Recombination ofthe D-to-JH segments is proposed to influence the local chromatin structure and accessibilityof proximal VHgenes. After D-to-JH recombination, the VHdomain becomes differentiallyacetylated, with the more distal gene segments displaying relatively higher levels histoneH4 acetylation. IL-7 signaling via Stat-S and its interaction with Oct-l regulates acetylationof the distal VHgenes. Pax-S is dispensable for the acetylation and transcription of the distalVHgene segments, but is requ ired for their compaction and recombination. Importantly, H4acetylation is localized to nucleosomes pos itioned near VHgene promoters (dark green) andRSS (red triangle) but not in the intergenic regions .

thedistalVH domain.14•16.17.32Asthelocusisinanextended confonnation at thisstageofdevelopment.theVH genesegments aremoreclosely associated witharepressive compartmentthantheD]H cluster(Fig. 3). Consequently, the D]Hgenes segments maybemore accessible to the V(D)J recombinasethan the VH segments (seebelow). Though the roleof nucleartopologyof the IgH locusin differentially regulating itsaccessibility remains to berigorously established. theseanalyses encompassingtranscription factors. chromatinstructure. antisense transcription andnuclear organization highlightboth localand globalmechanisms that likely regulate recombination.

B-Cell Fate Commitment and V-to-D}H RearrangementThetranscriptionfactorEBFinducesB-cell fatespecification and alsoinitiatesB-cellfatecom

mitment by restrictingalternate myeloid lineageoptions.25•33 IL-7R signaling is required for thedevelopmental induction of the EBFgenein lymphoidprogenitors(Fig. 1).34 EBFin turn induces

Molecular Pathways andMechanisms RegulatingtheRecombination ofImmunoglobulin Genes /39

the expression of'Pax-S,a transcriptionfactor that isrequiredforB-cellfatecommitmenr.v" EBFand PaxS areessential for the generationof committed pro-B-cells in whichV-to-DlH rearrangements areactivated. It is important to note that theseB-lineage specific recombinationeventsatthe IgH locusnot only coincidewith B-cellfate commitment but they are directlyregulatedbysignalingpathways (IL-7R)and transcriptionfactors(EBFand PaxS) that areneededto establishthe committed state.

The VHdomain of the IgH locus spansapproximately 2 Mb and includesapproximately 1SOfunctionalgenesegments. eachofwhichhasitsownpromoter and RSSelement.36.37Thegenesegments aregrouped into families basedon sequencehomology. Distinct regulatorypathways andmechanisms areinvolved in controllingthe recombinationaccessibility ofDHproximalversus DHdistalVHgenes. During B-celldevelopmentin the fetal liver. the DHproximalVHgenesegmentsare preferentially recombined.P'" Thisselective rearrangementof VHgenesis consideredto be aconsequence of their closerproximityto the D]Hsegments. Intriguingly, it hasbeendemonstratedthat the DHproximalportion of theVHdomain becomesassociatedwith hyperacetylated histonesfollowing successful D]Hrearrangement. Theseresultssuggest an attractivemechanisminvolvingthe limited spreadingofactivatinghistone marksto accountfor the preferentialrearrangement ofproximalVHgenesseenearlyin development(Fig.4).261hey alsoprovidea meansfor sequentiallyorderingthe recombination ofD-to-]Hsegments andproximalV-to-D]Hgenesegments within theIgH locus. Recently. it hasbeen shownthat the transcription factorEBFis requiredfor V-to-D]Hrecombination." It will be important to determineifEBF targetsproximalVHgenesand regulatestheir chromatinstructure therebypromoting their recombination.

Considerable progress hasbeenachieved inanalyzingthe recombinationof the distalVHgenes.The IL-7 signalingpathwayspecifically regulates recombinationof distal VHgene segments.42.43Locally restricted histone acetylationassociated with the individualdistal VHgene segments isdependent on IL-7 signaling.26•44 The transcription factor STATS. a downstreamsignaling component of the IL-7 signalingpathwayhas been shown to be required for efficient distal VHgenerearrangement. thus establishing a molecularlink between the IL-7 signalingpathwayand IgHrecombination." StatSis recruited to VHgenepromoters via the transcription factor Oct-I thatbinds to the conservedoctamer element.StatS functions as a co-activator. stimulatinggermlinetranscription.histoneacetylation and recombinationof the distalVHgenesegments." It shouldbenoted that StatSisnot requiredfor the repositioningofIgH alleles away from the nuclearlaminaor for their compaction.both of thesehigher-orderstepsare alsoimplicatedin regulatingdistalVHgene recombination (seebelow).Therefore, IL-7 signalingvia the transcription factor StatSappears to function specifically in regulatinglocalizedchanges in accessibility of distalVHgenesegments through histone modifications and possibly nucleosome remodeling.

A second key regulator of distal VHgene recombination is the transcription factor PaxS(BSAP). Importantly, PaxS regulates the recombinationofdistalVHgenesegments viaamolecularmechanismthat isdistinct from the one detailedabovefor StatS.43In the absence ofPaxS. B-celldevelopmentis arrestedat the pro-B-cellstage." In contrast to the block seen in StatSdeficientcells. the distal VHgene segments are associated with highly acetylatedhistones and undergonormal levels of germline transcription in the absence of PaxS.46 Although, the IgH alleles arecentrallypositioned in the nucleiofPaxS mutant pro-Bvcells, they do not undergo compaction.also termed contraction (Fig. 3),17 Importantly. restoration of PaxS expression in PaxS. Rag2mutant cellsinducescontractionofIgH alleles.Theseresultsdemonstratethat PaxS canpromoteIgH locus contraction thereby increasingthe spatialproximity of distal VHgene segments andthe D]Hcluster, in the absence of recombinase activity. Intriguingly, PaxS is also implicated inthe lossof H3-K9 methylation in pro-Bvcells by promoting exchange with the histone variantH3.3 (seebelow)." We note that PaxS has been shown to bind to multipleVHgenesegments inB-lineage cells."Moreover. asnoted above. PaxS interactswith the RAG1,2proteins." ThusPaxSappearsto regulatedistal VHgenerecombinationviamultiple mechanisms that includeremovalofrepressive histone modifications, promotinglocuscontraction and directlyrecruitingRAG1.2complexes (Figs. 2.3).


IgH locuscompactionor contraction isa manifestation of higher-order chromosomal DNAloops that juxtaposedistal VHgene segments with the DJH cluster and has been proposed topromote distalVHgenerecombinarion.F'" Interestingly. ectopicexpression ofPax5 in T-lineagecells inducesV-to-DJH recombination but paradoxically resultsin a similarIgH recombinationphenotype as seen in Pax5 mutant pro-B-cells. Pax5 mis-expressing T-lineagecells rearrangeproximalbut not distal VHgene segments despite the fact that these genesegments are highlytranscribed and the IgH loci are centrallylocated.17•48 It should be noted that Pu5 expressingT-lineage cells canalsoactivate the EBFgeneand the latter factormayaccountfor their abilitytoundergoproximal V-to-DJHrecombination.Theseresults have ledto the suggestion that paxs actsin conjunctionwith another B-cell specific factor to induce IgH locuscontractionand facilitatedistalV-to-DJH rearrangement.

Thezincfinger transcription factorITI has alsorecently beenshownto beinvolvedin IgHlocuscontraction." B-lineage specific deletionof the ITI generesultsin a blockto development at thepro-B-cellstageand impairedV-to-DJHrecombinationthat ismostseverely manifested fordistalVHgenesegments,"ITI bindsto theheavy-chain intronicenhancerand has beenproposedto playadirect rolein locuscontractioni.e.•DNA loopingbypromotingenhancer-promoter interactions.Importantly, Pax5expression is not altered in ITI mutant pro-Bvcells, Thus Pax5 and ITI areindependently requiredforIgHlocuscontraction.Locus contractionislikely to facilitate molecularsynapsis oftwowidely separated andcompatible RSSs bythe RAGproteincomplexes.Oncewidelyseparatedgenesegments are brought into proximitywith one another. RAG proteins can thenachieve molecularsynapsis. RAGproteinshavebeen inferredto preferentially associate with RSSelementscontaininga 12 bp spacerrather than with an RSS containinga 23 bp spacerin vivo.50

Thesedata support the "capture" model.which posits that oligomeric RAG complexes initiallybind to an RSScontaininggenesegmentand then capture the complementary RSScontainingsegmentleadingto molecularsynapsis and the initiation of recombination viaDNA cleavage.

In addition to the fore-mentioned transcriptionfactors. the polycombgroupprotein Ezh2 isalsorequiredfor rearrangement of the distalVHgenesegments." Strikingly. the molecularphenotype ofEzh2 mutant pro-Bvcells isverysimilarto that ofPu5 mutant cells in that distalVHgenesegments arehighlytranscribedand associated with hyperacetylated histonesdespitethe blocktotheir recornbinarion.t-" Ezh2isa histonemerhyltransferase, that methylates histone H3 at K27.Lossof Ezh2 resultsin reduced H3 methylationat distal VHgene segments." It remains to bedetermined ifEzh2 asis the case for Pu5 and ITI isrequiredfor IgH locuscontraction.

Thereisan additionaldevelopmentally regulated chromatinmodification. H3-K9methylation.which appears to regulateheavy-chain recomblnarion ." H3-K9 methylationis associated withthe VHlocus in non-B-lineage cells but is removed in pro-B-cells. H3-K9 methylationhasbeendemonstratedto severely inhibit recombination uponitstargeting to anengineeredRSScontainingsubstratein a Bvcell line", Interestingly, asis case for the DJH region.antisense transcriptionalsooccursat the VHlocusand maybe involved in promotingexchange of repressive histones.?

Theabovestudiesenableusto proposeamodelforthedevelopmental controlofimmunoglobulinheavy chainrecombination in developingB-cells. In non-B-cells andhematopoieticprogenitorsthe germlineheavy-chain alleles are associated with the inner nuclearmembrane-nuclear laminacompartment and assembled in a repressive chromatin structure involving H3-K9 methylation(Fig.2).As the VHgenesegments aremoreclosely interactingwith the INM-laminathan the DJHcluster. the latter segments mayundergorecombinationwhilepositionedat the nuclearperiphery(Fig. 3). B-cell fate specification requires the transcription factors E2A and EBE In lymphoidprogenitors.thesetwo factorsappearto regulatethe initial low-level expression ofthe Raggenesaswellasthe accessibility of the DJHclusterto recombination. likely bypromotingantisense transcription and chromatin modifications. Increased expression of EBFas a consequence ofIL-7Rsignalingpromotes B-cell fatespecification and the generationof committedpro-B-cells via theinduction ofPax5.DuringB-cellfatespecification the Igheavy-chain alleles arerepositioned awayfromthe nuclearlaminathroughasyetunidentifiedfactorsand mechanisms,"EBFisan attractivecandidate regulator for inducing repositioning of IgH alleles. In EBF-I- lymphoidprogenitors

Molecular Pathways and Mechanisms Regulating theRecombination ofImmunoglobulin Genes 141

the D-] recombined IgH alleles are positioned at the nuclear lamina (I.DeMarco and H . Singh,unpublished results) . In pro-Bvcells, IL-7 signaling via Stat5, induces localized chromatin altera tions in distal VH gene segments and activates their transcription." Pax5 along with YYI promotesIgH locus contraction facilitating the recombination ofdistal VH gene segments. F'" A recent studycompared the distance distributions ofFISH signalsfrom multiple smallprobes (1Okb) that hybridize to the IgH locus and used computer modeling and triangulation to demonstrate that the locusisorganized into compartments containingclusters ofloops separated by linkers. 54 Importantly, inpro-Bvcells, the entire 2Mbp region containing the VH genes appears to be juxtaposed to the DH

elements , thus facilitating long-range genomic interactions.54It will be important to determine howthe transcription factors Pax5 and IT1 that appear to impact distinct domains of the IgH locuscontribute to its structural reconfiguration in pro- B-cells.The molecular functions ofthe transcrip tion factors EBF,Pax5 and IT1 in regulating antisense VH transcripts remain to be explored ." Wenote that proximal VH gene rearrangement requires EBF but not Pax5 or ITL25.46,4~ Thus regulatedchromatin alterations, interactions with the INM-Iamina compartment that are domain specificand locus reconfiguration accompanied by compaction appear to promote the accessibility ofthelarge repertoire ofVH gene segments to recombination in developing B-cells.

The Pre-B-Cell Checkpoint and the Induction ofLight-ChainRecombination

Following productive heavy-chain rearrangement B-cells progress through a critical developmental checkpoint (Fig. 1). This process consists ofa self-limiting clonal expansion culminatingin cell cycleexit and initiation oflight -chain rearrangement. Successful light-chain rearrangementresults in the generation ofimmature IgM+ B-cells.The Raggenes are down regulated during thecycling pre-B-cell phase and re-induced upon cell cycle exit. Signaling through the pre-BCR andthe IL-7R regulates the pre-B to B-cell transition. IL-7 signaling is active during the large cyclingpre-B-cells stage. However, the pre -BCR reduces the dependence ofpre-B-cells on IL-7 and thisis also correlated with a change in the anatomic distribution ofpro-B and pre-B-cells in the bonemarrow. The former are associated with IL-7 expressing stroma whereas the latter are positionedawayfrom such stromal cells.55.56One oftwo light-chain loci, IgK or IgA, undergo productive rearrangement at the pre-B-cell stage. Their genomic structures are depicted in Figure 5. In mice, theIgK locus is more frequently rearranged, at a ratio of20:1 and consequently recombination ofthislocus has been more intensively studied.57

Signaling through both the pre-BCR and the IL-7R drives the limited clonal expansion ofpre-Bvcells.Raggene expression is down regulated during this phase (Fig. 1)?22 Therefore, theproliferative burst separating the IgH and IgL recombination event s during B-cell developmentprovides pre-B-cells with the opportunity to pause recombination in the absence of an activerecombinase and redirect chromatin accessibility from the heavy-chain locus to the light-chainloci. Until recently, it was considered that cell-cycleexit, may be sufficient to initiate Ig light-chainrecombination. However, a combination of loss-of-function and gain-of-function experimentsinvolving key cell cycle regulators, have demonstrated that exit ofpre-B-cells from the cell-cycleis not a sufficient condition for the activation of recomblnation." Instead. acquired pre-BCRsignaling followed by attenuated IL-7R signaling results in alteration of chromatin accessibilityof Ig light chain loci and cell cycle exit. Attenuation of IL-7R signaling also contributes to theoptimal expression of the Raggenes and high recombinase activity.

An area ofintense investigation concerning the regulation ofIglight-chain recombination hasinvolved the analysis of transcription factors that bind to and activate transcriptional enhancerswithin the IgK locus. The simplest explanation for the restriction oflight-chain recombination tothe pre-Bvcell stage would be developmentally appropriate expression of Igx-specific transcription factors. As detailed below, the molecular mechanism is not quite that simple. Nevertheless.recent insight suggests an exquisitely regulated process that integrates the developmental signalingprograms found in pre-B-cells to the activities ofkey transcription factors ultimately leading tostage-specific IgK recombination.


• Proliferation

Rag1.2 CXCR4 4 ~rJE2A Il~ II

Igo: -D-O-D-I/ IIIII • • 'I-Q- ~ Ig).i

iE" 3'E" E)'1-3

Figure 5. Signaling pathways and transcription factors regulating IglC and Ig>.. recombination .The network depicts signaling pathways and transcriptional regulators that are required forlight-chain recombination at the pre-B-cell stage. Arrows represent positive regulat ion andbarred lines represent repression. As indicated, IRF-4plays a central role in inducing light-chainrecombination downstream of the pre-BCR by directly engaging the 3'ElC and x light-chainenhancers. IRF-4,B are also suggested to induce cell cycle arrest and modulate IL-7 signalingthereby resulting in robust induction of Rag gene expression and E2A binding to iElC. IRF4,Binduce chemokine receptors that are proposed to induce migration of pre-B-cells away fromIL-7 producing stroma, leading to attenuation of IL-7 signaling and activation of the IglC enhancer (through E2A binding) as well as increased Rag gene expression.

Genetic analyseshave demonstrated that the transcription factors, E2A, Pax5 and the relatedfamily members IRF-4 and IRF-8 are required for light-chain recombination (Fig, 5).59-61 Thesefactors have known binding sites within the Igkappa enhancers and in vivo DNA footprintinganalysishasshown that their sitesareoccupied in primary pre-Bvcells.S Interestingly, footprintinganalysiscomparing the binding ofthese keytranscription factorsduring the transition from pro-Bto pre-B-cellsdemonstrates no changewith the exception of the composite site for PU-l /IRF-4.62Interestingly, IRF-4 expression increases at the pre-B-cell stage.58.63.64 Loss of IRF-4 along withIRF-8, results in a complete block to B-celldevelopment at the large cyclingpre-B-cellstagewitha failure to undergo IgK or Igf.recombination/" A detailed analysisof the molecular mechanismsby which IRF-4 and IRF-8 activate recombination of Ig light-chain loci by is provided below.Unlike IRF-4 and IRF-8, the transcription factors PaxS and E2A also function earlier in B-celldevelopment at the pro-Bscell stage,where they are required for Ig heavy-chain recombination .UsingPax5deficient or E2A deficientpre-Bvcells, it has been shown that both factors additionallyregulate IgK germline transcription and recomblnanon.P-"

Signalingby the pre-BCR has been widely considered to activate light-chain recombination.Expressionof a transgene encoding the 19J.L heavy-chainprotein increasesIgK locusaccessibilityinRagdeficientpro_B_cells.65.67Additionally,the enforced expressionofactivatedRas, adownstreamsignalingcomponent ofthe pre-BCR,promotes Iglight-chain recombination in the absenceofanIgheavy-chain." Conversely,lossofsignalingmolecules including BLNK, Btk and PCLy, that liedownstream of the pre-BCR, results in fewercellsthat have rearranged their Igkappa loci.69.70Asnoted above, the transcription factor IRF-4 is induced by pre-BCR signaling and Iglight-chainrecombination is blocked in Irf4,B-i-pre-B-cellsdespite the high expression of the pre-BCR.60Restoring either IRF-4 or IRF-8 expression rescuesdevelopmental progression and activates Iglight-chain rearrangernent.W" IRF-4promotes histone acetylation at criticalenhancers within IgKand Igf.lociand induces their germline transcription (Fig.5).58 Intriguingly,IRF-4 alsocounteractsassociation of an IgK allele with pericentromeric heterochromatin, an interaction that has been

MolecularPathwaysand MechanismsRegulatingthe Recombinationoflmmunoglobulin Genes 143

proposed to inhibit recombination(Fig. 3).58 Thesedata delineatea molecularpathwaybywhichpre-BCR signalingregulates both 19K and Ig1.. recombinationand alsoprovide insightsinto theunderlyingmolecularmechanisms.

Several studieshaveimplicated IL-7 signalingin the negative regulationofIglight-chain recombination. Withdrawal ofIL-7 in pro-Bvcell culturesappearsto induce Iglight-chainrecombination.? However, Ig light-chain recombinationcan occur in the presenceof high concentrationsofIL-7 and it hasbeen argued that IL-7 withdrawalmerelyselects for cellsthat haveundergoneproductivelight-chain recombination," Until recentlythe preciseroleofIL-7 signalingin regulating Ig light-chain recombination had remained unclear.55•73 UsingIrf4,s-t-pre-Bvcells, it hasbeen demonstrated that IL-7 signalingcan regulateIg light chain recombinationIndependentlyof pre-BCRsignalingand IRF-4.Attenuating IL-7 signalingin Irf4,B-i- pre-Bvcells activates 19Kbut not Igk recombination." Recombination is accompaniedby the induction of 19K germlinetranscriptsand substantialupregulationof Ragtranscripts. Intriguingly, binding of E2A to theintronic 19K enhancer and localized histone acetylationincreases within 24 hours of attenuatedIL-7signaling.ThusIL-7 signalingmodulatesIgK rearrangementin pre-B-cells bycontrollingtheactivityof the intronic 19K enhanceraswellasoptimalexpression of the Raggenes.AsIL-7 signaling is activein pro-Bvcells it would inhibit Ig light-chain recombinationat this stage. As notedabove, IL-7 signalingpromotes Ig heavychain rearrangement in pro-B-cells and this pathwayisdependenton StatS.It remains to bedetermined ifinhibition oflgKrearrangement byIL-7signaling isalsodependent on StatSand if so what is the nature of the molecularmechanismbywhichStatS regulates accessibility ofE2A at the intronic 19K enhancer.

Despitethe fact that pre-BCRand IL-7 signalingpathways canfunction independendyofoneanother in promoting 19K recombination, it is highly likelythat their activities are coordinatedduring B-celldevelopment. Consistent with this view, the two pathways function synergisticallyto induce IgK recombinationand the generationofIgM expressing B-cells.58The molecularbasisof synergy in promoting IgK recombination appears to be manifestedat two steps. Firstly, eachpathwaytargetsadistinct IgK enhancerand synergy islikely aconsequence of simultaneously activatingboth enhancers. Secondly, IRF-4preferentially inducesIgK germlinetranscriptionwhereasattenuationofIL-7 signalingmorehighlyinducesRaggeneexpression therebyoptimizingchangesin accessibility with expression of the recombinase.

An intriguing model has been proposed for the regulation of Ig light chain recombinationvia the coordination of pre-BCR and IL-7 signalingpathways in vivo. Genome-wide expressionanalysis using Irf4,B-I- pre-B-cells revealed a number of genes involved in cell migration andadhesion that are regulatedby IRF-4.58 Ofparticular interest was the geneencoding CXCR4, achemokinereceptorthat promotesmigrationin response to CXCLl2. IRF-4dependentupregulation ofCXCR4 wasshownto resultin achangein the chemotacticpropertiesof pre-Bvcells, SinceCXCLl2 expressing stromalcells arespatially separatedfrom IL-7expressing stromalcells in thebone marrow, it hasbeenproposedthat IRF-4 regulatedchemotaxis towardsCXCLl2 expressingstomalcellsresultsin repositioningof pre-B-cells away from the IL-7 expressing stroma.56.581his

movementwould result in attenuation of IL-7 signalingand promote the synergistic inductionofIg light-chainrecombinationby the two molecularpathways detailed above.

Allelic ExclusionAllelicexclusion of both IgH and IgL loci ensures the generation of B-cells that express a

singletype of antigen receptor. For eachlocusproductiverearrangementofone allele culminatesin feed back inhibition of further rearrangementof the other allele, We will initiallydiscuss themolecularmechanisms that havebeen suggested to regulate allelic exclusion of the 19K locus,asit has been more Intensively studied. Allelicexclusion at the 19K locus is initiated by a singlealIde beingchosento undergo recombinationat the pre-Bvcell stage.Two fundamentallydifferentmechanisms, stochasticversus directed, have been proposed to explain this phenomenon. Thestochasticmechanisminvokes limiting amounts of either a transcription factor(s) that regulateslocus accessibility or limiting expression of the recombination machinery. Either condition is


proposed to leadto inefficient recombinationtherebydecreasing the probabilitythat both allelesundergorecombinationsirnultaneously.Y" Data insupport ofthis mechanismhas beenobtainedbymonitoring GFP expression in a knock-inmousethat expresses a GFP eDNA from an unrearranged lC allele." Only a smallpercentageof pre-Bvcells wereseen to express GFP and such 19Kgermlinetranscriptionwasmonoallelic. Thisdata hasbeen interpreted to suggest that a limitingtranscriptionfactorthat activates IgK germlinetranscriptionin pre-B-cdlsalso restricts recombination to the smallfractionofactivatedalleles.An alternativeexplanationfor allelic exclusion of theIgK locusproposesa series of directed epigeneticchangesthat occur differentially on individualkappa alleles. In agreement with this hypothesis, tight correlations have been found betweenmonoallelicDNA demethylationoflgK alleles and their replication timing?6-~ More detailedanalyses haverevealed that at the pre-Bvcell stagethe earlyreplicating19K alleleis assembled intoan activechromatinstructureandpreferentially undergoes DNA demethylation therebyincreasingits accessibility to recombinarion," In contrast,the late replicatingalleleisassembled into inactivechromatincomprisinghypoacetylatedhistonesandmethylatedH3-K9. This allele isalso associatedwith pericentromericheterochromatinand suggested to be a poorer substratefor recombination.Intriguingly, a cis-element, termed Sis, hasbeen discovered in the Y-JK interveningsequenceandthis element targetsan 19K transgeneto pericentromeric heterochromatin." Usingyeastartificialchromosome-based single copytransgenicmicethe Siselementwasshownto negatively regulateIgK recombination." Moreover, this elementwasshown to interact with the zinc fingerproteinIkaros,a transcriptionfactor that has beenshownto be associated with transcriptionally inactivegenes, includingx allelethat areassociated with pericentromeric hererochromarin.P'" Thesedatahave led to the suggestion that Ikaros-Sis complexes actively participate in the processof allelicexclusion bypromotingsilencingofasingleIgK alleleviainteractionwith pericentromeric heterochromatin. Thesedistinct setsof observations concerningmonoallelic activationofthe 19K locushaveutilizeddifferentmethodologies and cannot be easily reconciled. It ispossible that adirectedmechanismisusedto distinguishthe two alleles and alimitingtranscriptionfactorfurther restrictsthe activationof the more accessible alleleto a smallpercentageof pre-B-cells.

Allelicexclusion at the heavy-chain locusinvolves feedbackinhibition by the product of theproductivelyrearranged allele (assembled into the pre-BCR) and attenuation ofIL-7 receptorsignaling.5•82 It has beenshownthat the nonproductivelyrearrangedheavy-chain alleleisrecruitedto pericentromeric heterochromatin and undergoes locus decontraction followingsuccessfulrearrangementof the other allele." Recently, an intriguing mechanisminvolving specific interchromosomal interactions between the heavy-chain and light-chain loci has been proposed tolink allelicexclusion at both 10ci.84 Using3D FISH, IgH and IgK alleles werefound to colocalizewith pericentromericheterochromatin in pre-B-cells. This inter-chromosomal interaction wasdependent on the Ig 3'lC enhancer. Deletion of this cis-regulatory element resulted in not onlylossof the association between IgH and IgK alleles but prevented IgH locusdecontraction. Thiswassuggested to promote continued accessibility of the Igheavy-chainlocusto recombinationinpre-B-cells and a breakdownof allelic exclusion.

PerspectivesThe analysis of transcription factors and signalingpathways that regulate immunoglobulin

gene recombinationduring B-lymphoeytedevelopmenthas resultedin considerable progress. Aplausible developmental schemecannowbeformulatedfor theorderedrecombinationofIg heavyand light chainloci.The transcriptionfactorsnot onlyappear to regulateIglocusaccessibilityvialocalized changes in chromatinstructurebut alsolikely modulaterecombination byalteringnuclearcompartmentalization ofIgalleles and their large-scale chromatindynamics. Futureresearch shoulduncovernovelmolecularcomponentsthat mediatethe interactionsofIg lociwith the INM-laminacompartment or pericentromeric heterochromatin and test if they regulate recombination.Furthermore, the molecularmechanisms underlyinglarge-scale DNA loops at Ig loci remain tobe elucidated. Formationof these intrachromosomalloopsis likely to be requiredfor long-rangeY(D)J recombinationand the generationof a diverse repertoireof antigen receptors.

MolecularPathways and Mechanisms Regulatingthe Recombinationoflmmunog/obulin Genes 145

References1. Igarashi H , Gregory SC, YokotaT et aI. Transcription from the RAG1 locus marks the earliest lymphocyte

progenitors in bone marrow. Immunity 2002; 17(2):117-130.2. Allman D. Miller JP. Common lymphoid progenitors , early B-lineage precursors and IL-7: character

izing the trophic and instructive signals underlying early B-cell development. Immunol Res 2003; 27(2-3) :131-140.

3. Li YS, Hayakawa K, Hardy RR. The regulated expression of B-lineage associated genes during B-celldifferentiation in bone marrow and fetal liver. J Exp Med 1993; 178(3):951-960.

4. Bassing CH, Swat W; Ale FW The mechanism and regulation of chromosomal V(D)J recombination.Cell 2002; 109 Suppl:S45-55.

5. Geier JK, Schlissel MS. Pre-BCR signals and the control of Ig gene rearrangements. Semin Immunol2006; 18(1):31-39.

6. Lin H, Grosschedl R. Failure of Bvcell differentiation in mice lacking the transcription factor EBF.Nature 1995; 376(6537):263-267.

7. Li Z, Dordai DI, Lee J er al. A conserved degradation signal regulates RAG-2 accumulation during celldivision and links V(D)J recombination to the cell cycle. Immunity 1996; 5(6) :575-589 .

8. Golding A, Chandler S, Ballestar E et aI. Nucleosome structure completely inhibits in vitro cleavagebythe V(D)J recombinase. EMBO J 1999; 18(13) :3712-3723.

9. Kwon J, Imbalzano AN, Matthews A et al. Accessibility of nucleosomal DNA to V(D)J cleavage ismodulated by RSS position ing and HMG1. Mol Cell 1998; 2(6):829-839.

10. Krangel MS. Gene segment selection in V(D)J recombination: accessibility and beyond. Nat Immunol2003; 4(7) :624-630.

11. Matthews AG, Kuo AJ, Ramon-Maiques S er al. RAG2 PHD linger couples histone H3 lysine 4 trimethylation with V(D)J recombination . Nature 2007; 450(7172) :1106-1110.

12. Liu Y, Subrahmanyam R. Chakraborty T er al, A plant homeodomain in RAG-2 that binds Hypermerhylated lysine 4 of histone H3 is necessaryfor efficient antigen-receptor-gene rearrangement. Immunity2007; 27(4) :561-571.

13. Zhang Z, Espinoza CR, Yu Z et al. Transcription factor Pax5 (BSAP) transacrivates the RAG-mediatedV(H)-to-DJ(H) rearrangement of immunoglobulin genes. Nat Immunol 2006; 7(6) :616-624.

14. Kosak ST. Skok JA, Medina KL et aI. Subnuclear compartmentalization of immunoglobulin loci duringlymphocyte development. Science 2002; 296(5565):158-162.

15. Skok JA, Brown KE, Azuara V et aI. Nonequivalent nuclear location of immunoglobulin alleles inB-Iymphocytes. Nat Immunol 2001; 2(9) :848.

16. Reddy KL, Zullo JM, Bertolino E er al. Transcriptional repression mediated by repositioning of genesto the nuclear lamina. Nature 2008; 452(7184):243-247.

17. Fuxa M, Skok J, Souabni A et al. Pax5 induces V-to-DJ rearrangements and locus contraction of theimmunoglobulin heavy-chain gene. Genes Dev 2004; 18(4):411-422.

18. Sayegh C, Jhunjhunwala S, Riblet Ret aI. Visualization of looping involving the immunoglobulinheavy-chain locus in developing B-cells. Genes Dev 2005 ; 19(3):322-327.

19. Medina KL, Singh H. Gene regulatory networks orchestrating B-cell fate specilication, commitmentand differentiation. Curr Top Microbiol Immunol2oo5; 290:1-14.

20. Bain G, Maandag EC, lzon DJ er aI. E2A proteins are required for proper Bvcell development andinitiation of immunoglobulin gene rearrangements. Cell 1994; 79(5) :885-892.

21. Zhuang Y, Soriano P, Weintraub H. The helix-loop-helix gene E2A is required for B-cell formation .Cell 1994; 79(5) :875-884.

22. Lin WC, Desiderio S. V(D)J recombination and the cell cycle. Immunol Today 1995; 16(6):279-289.23. Romanow WJ, Langerak AW;Goebel Pet al. E2A and EBF act in synergy with the V(D)J recombinase

to generate a diverse immunoglobulin repertoire in nonlymphoid cells. Mol Cell 2000; 5(2):343-353.24. Seer CS, Brumbaugh RL, Kee BL. Early B-cell factor promotes B-Iymphopoiesiswith reduced interleukin

7 responsiveness in the absence of E2A. J Exp Med 2004; 199(12):1689-1700 .25. Pongubala JM, Northrup DL , Lancki DW et al. Transcription factor EBF restricts alternative lin

eage options and promotes Bvcell fate commitment independently of Pax5. Nat Immunol 2008 ;9(2) :203-215.

26. Chowdhury D, Sen R. Stepwise activation of the immunoglobulin mu heavy chain gene locus. EMBOJ 2001; 20(22):6394-6403.

27. Bates JG . Cado D, Nolla H er aI. Chromosomal position of a VH gene segment determines its activation and inactivation as a substrate for V(D)J recombination. J Exp Med 2007; 204(13):3247-3256.

28. Chakraborty T, Chowdhury D, Keyes A er aI. Repeat organization and epigenetic regulation of theDH-Cmu domain of the immunoglobulin heavy-chain gene locus. Mol Cell 2007; 27(5) :842-850.

29. Bolland DJ. Wood AL, Afshar R et aI. Antisense intergenic transcription precedes 19h Dvro-] recombination and is controlled by the intron ic enhancer Emu. Mol Cell BioI 2007; 27(15) :5523-5533.


30. Afshar R. Pierce S. Bolland DJ et al. Regulation of IgH gene assembly: role of the intronic enhancerand 5 ' DQ52 region in targeting DHJH recombination. J ImmunoI2006: 176(4):2439-2447.

31. Chakalova L. Debrand E. Mitchell JA ee al. Replication and transcription: shaping the landscape of thegenome. Nat Rev Genet 2005; 6(9) :669-677.

32. Yang Q, Riblet R. Schildkraut CL. Sites that direct nuclear compartmentalization are near the 5' endof the mouse immunoglobulin heavy-chain locus. Mol Cell BioI 2005: 25(14):6021-6030.

33. Kikuchi K. Lai AY. Hsu CL et aI. lL-7 receptor signaling is necessary for stage transition in adult B-celldevelopment through up-regulation ofEBEJ Exp Med 2005: 201(8):1197-1203.

34. Dias S. Silva H Jr. Cumano A et al. Inrerleukln-? is necessary to maintain the B-eell potential in common lymphoid progenitors. J Exp Med 2005; 201(6) :971-979.

35. Busslinger M. Transcriptional control of early B-cell development. Annu Rev Immunol 2004;22:55-79.

36. Chevillard C. Ozaki J. Herring CD et al. A three-megabaseyeast artificial chromosome contig spanningthe C57BL mouse Igh locus. J Immunol 2002: 168(11):5659-5666.

37. Johnston CM. Wood AL. Bolland DJ et at. Complete sequence assembly and characterization of theC57BL/6 mouse Ig heavy chain V region. J ImmunoI2006: 176(7):4221-4234.

38. Yancopoulos GD. Desiderio SV. Paskind M et al, Preferential utilization of the most JH-pro:rimal VHgene segments in pre-B-celllines . Nature 1984; 311(5988) :727-733.

39. Jeong HD. TealeJM. VH gene family repertoire of resting B-cells. Preferential use ofD-proximal familiesearly in development may be due to distinct B-cell subsets. J Immuno11989; 143(8):2752-2760.

40. Malynn BA. Yancopoulos GD. Barth JE et at. Biased expression of JH-prox imal VH genes occurs inthe newly generated repertoire of neonatal and adult mice. J Exp Med 1990; 171(3):843-859.

41. Ten Boekel E. Melchers F, Rolink AG. Changes in the V(H) gene repertoire of developing precursorB-Iymphocytes in mouse bone marrow mediated by the pre-Bvcell receptor. Immunity 1997;7(3):357-368.

42. Corcoran AE. Riddell A, Krooshoop D et al, Impaired immunoglobulin gene rearrangement in micelacking the IL-7 receptor. Nature 1998: 391(6670) :904-907.

43. Bertolino E. Reddy K. Medina KL et al. Regulation of interleukin 7-dependent immunoglobulin heavy-chain variable gene rearrangements by transcription factor STAT5. Nat Immunol 2005;6(8) :836-843.

44. Johnson K. Angelin-Duclos C. Park S et aI. Changes in histone acetylation are associatedwith differencesin accessibilityofV(H) gene segments to V-DJ recombination during B-cellontogeny and development.Mol Cell BioI 2003; 23(7):2438-2450.

45. Nun SL, Heavey B. Rolink AG et at. Commitment to the B-Iymphoid lineage depends on the transcription factor Pax5. Nature 1999: 401(6753) :556-562.

46. Hesslein DG, PRugh DL, Chowdhury D et at. Pax5 is required for recombination of transcribed,acetylared, 5' IgH V gene segments. Genes Dev 2003; 17(1):37-42.

47. Johnson K, PRugh DL. Yu D et at. B-cell-specific loss of histone 3 lysine 9 methylation in the V(H)locus depends on Pax5. Nat Immunol 2004: 5(8):853-861.

48. Hsu LY, Liang HE, Johnson K et at. Pax5 activates immunoglobulin heavychain V to DJ rearrangementin transgenic thymocytes. J Exp Med 2004: 199(6):825-830 .

49. Liu H. Schmidr-Supprian M, Shi Y et at. Yin Yang1 is a critical regulator ofB-eell development. GenesDev 2007: 21(10) :1179-1189.

50. Curry JD , Geier JK. SchlisselMS. Single-strand recombination signal sequence nicks in vivo: evidencefor a capture model of synapsis. Nat Immuno12005; 6(12) :1272-1279.

51. Su IH, Basavaraj A, Krutch insky AN et at. Ezh2 controls B-cell development through histone H3methylation and Igh rearrangement. Nat Immunol 2003: 4(2) :124-131.

52. Osipovich O. Milley R. Meade A et at. Targeted inhibition of V(D)J recombination by a histonemethyltransferase. Nat Immunol2004: 5(3):309-316.

53. Bolland DJ, Wood AL, Johnston CM et at. Ant isense intergenic transcription in V(D)J recombination.Nat Immunol 2004: 5(6):630-637.

54. Jhunjhunwala S. van Zelm MC, Peak MM et at. The 3D structure of the immunoglobulin heavy-chainlocus: implications for long-range genomic interactions. Cell 2008; 133(2):265-279.

55. Rolink AG. Winkler T. Melchers F et at. Precursor Bvcell receptor-dependent B-eell proliferat ionand differentiation does not require the bone marrow or fetal liver environment. J Exp Med 2000:191(1):23-32.

56. Tokoyoda K, Egawa T, Sugiyama T et at. Cellular niches controlling B-Iymphocyre behavior withinbone marrow during development. Immunity 2004; 20(6):707 -718.

57. McGuire KL. Vitetta ES. kappa/lambda Shifts do not occur during maturation of murine B-cells.J Immuno11981 ; 127(4):1670-1673.

Molecular Pathways andMechanismsRegulatingtheRecombination ofImmunoglobulin Genes 147

58. Johnson K, Hashimshony T, Sawai CM ee al. Regulation of Immunoglobulin Lighr-Chain Recombination by the Transcription Factor IRF-4 and the Attenuation of Inrerleukin-? Signaling. Immuniry 2008;28(3):335-345.

59. Lazorchak AS, Schlissel MS, Zhuang Y. E2A and IRF-4/Pip promote chromatin modification andtranscription of the immunoglobulin kappa locus in pre-Bvcells. Mol Cell Bioi 2006; 26(3):810-821.

60. Lu R, Medina KL, Lancki DWet al. IRF-4,8 orchestrate the pre-Bxo-B transition in lymphocytedevelopment, Genes Dev 2003; 17(14):1703-1708.

61. Sato H. Saito-Ohara F, InazawaJet al. Pax-5 is essential for kappa sterile transcription during Ig kappachain gene rearrangemem. J Immunol 2004; 172(8):4858-4865 .

62. Shaffer AL. Peng A. Schlissel MS. In vivo occupancy of the kappa light chain enhancers in primarypro- and pre-Bvcells: a model for kappa locus activation. Immunity 1997; 6(2):131 -143.

63. Muljo SA. Schlissel MS. A small molecule Abl kinase inhibitor induces differemiation of Abelsonvirus-transformed pre-Bvcell lines. Nat Immunol 2003; 4(1) :31-37.

64. Thompson EC. Cobb BS. Sabbatdni P et al. Ikaros DNA-binding proteins as integral components ofBvcell developmental-scage-specific regulatory circuits. Immunity 2007; 26(3) :335-344.

65. Spanopoulou E. Roman CA. Corcoran LM et al. Functional immunoglobulin rransgenes guide orderedBvccll dlfferentiatton in Rag-l-deficienr mice. Genes Dev 1994; 8(9) :1030-1042 .

66. Young F, Ardman B. Shinkai Y et al. Influence of immunoglobulin heavy- and light-chain expressionon Bvcell differendadon. Genes Dev 1994; 8(9) :1043-1057 .

67. Stanhope-Baker P. Hudson KM, Shaffer AL et al. Cell type-specific chromatin structure determines thetargeting ofV(D)J recombinase activity in vitro. Cell 1996; 85(6):887-897.

68. Shaw AC, Swat W. Davidson L et al. Induction of Ig light chain gene rearrangemem in heavy chaindeficient Bvcells by activated Ras. Proc Natl Acad Sci USA 1999; 96(5):2239-2243.

69. Flemming A, Brummer T. Reth M et al. The adaptor prote in SLP-65 aces as a rumor suppressor thatlimirs pre-Bvcell expansion. Nat Immunol 2003; 4(1) :38-43.

70. Xu S, Lee KG. Huo J et al. Combined deficiencies in Bruton tyrosine kinase and phospholipaseCgamma2 arrest Bvcell development at a pre-B'CR» stage. Blood 2007; 109(8) :3377-3384.

71. Ma S. Turetsky A, Trinh L ec al. IFN regulatory factor 4 and 8 promote Ig light chain kappa locusactivation in pre-Bvcell development. J Immunol 2006; 177(11) :7898-7904.

72. Grawunder U. Haasner D, Melchers F et al. Rearrangement and expression of kappa light chain genescan occur without mu heavy chain expression during differemiation of prc-Bscells. Int Immuno11993;5(12) :1609-1618 .

73. Milne CD. Fleming HE, Paige C]. IL-7 does not prevem pro-Bzpre-Bscell maturation to the immature/slgM(+) srage. Eur J Immunol 2004; 34(10) :2647-2655.

74. Coleclough C, Perry RP. Karjalainen K et al. Aberrant rearrangemems contribute significantly to theallelic exclusion of immunoglobulin gene expression. Nature 1981; 290(5805):372-378.

75. Schlissel MS. Regulation of activation and recombination of the murine Igkappa locus. Immunol Rev2004; 200:215-223 .

76. Liang HE. Hsu LY. Cado D et al. Variegated transcriptional activat ion of the immunoglobulinkappa locus in pre-Bvcells contributes to the allelic exclusion of light-cha in expression. Cell 2004;118(1) :19-29 .

77. Mostoslavsky R. Singh N, Tenzen T et al. Asynchronous replication and allelic exclusion in the immunesystem. Nature 2001; 414(6860):221-225.

78. Mostoslavsky R, Kirillov A. Ji YH et al. Demethylation and the establishment of kappa allelic exclusion.Cold Spring Harb Symp Quant Bioi 1999; 64:197-206.

79. Goldmit M. Schlissel M, Cedar H et al. Differencial accessibility at the kappa chain locus plays a rolein allelic exclusion. EMBO J 2002; 21(19) :5255-5261.

80. Goldmit M, Ji Y, Skok J et al. Epigenetic ontogeny of the Igk locus during B-cell development , NatImmunol2005; 6(2) :198-203.

81. Liu Z, Widlak P, Zou Y et al. A recombination silencer that specifies heterochromatin positioning andikaros association in the immunoglobulin kappa locus. Immunity 2006; 24(4) :405-415.

82. Chowdhury D. Sen R. Transient IL-7/IL-7R signaling provides a mechanism for feedback inhibitionof immunoglobulin heavy chain gene rearrangemems. Immunity 2003; 18(2):229-241.

83. Roldan E, Fuxa M. Chong W er al. Locus 'deconrractlon' and centromeric recruitment contribute toallelic exclusion of the immunoglobulin heavy-chain gene. Nat Immunol 2005; 6(1) :31-41.

84. Hewirr SL. Farmer D. Marszalek K et al. Association between the Igk and Igh immunoglobulin locimediated by the 3' Igk enhancer induces 'decontraction' of the Igh locus in pre-Bvcells. Nat Immunol2008; 9(4):396-404.

CHAPTER 12

Regulation ofV(D)J Recombinationby E-Protein Transcription FactorsMary Elizabeth]ones and YuanZhuang*

Abstract

ExtensivestudyoftheE-proteinsE2AandHEBduringlymphoeyte development has revealedvarious functionsforthesebHLH transcriptionfactorsin regulatingV(D)Jrecombinationin both B-andT-cells.The studyofE-proteinsin mammals beganwith the identification of

E2Abyits abilityto bind immunoglobulinheavyand light chainenhancers. Subsequentanalysishasidentifiednumerousrolesfor E2Aand HEB at the immunoglobulin and T-cellreceptorloci.E-protein targetsalsoincludethe rag genes and other factorscriticalfor recombination and forregulationof the developmental windowswhen cells undergorecombination. E-proteinsappearto bemasterregulators that coordinateantigenreceptorgenerearrangement andexpression.Thischapter focuses on how E-proteins regulateV(D)J recombination by activating transcription,initiatingrearrangement and drivingdifferentiation duringB-and T-celldevelopment.

IntroductionE2A, the foundingmemberof the E-proteinfamily of transcriptionfactorsin mammals, was

originallyidentifiedby its ability to bind enhancerregionsof the immunoglobulin heavy chain(IgH) andlightchain (IgL) genes.Earlyanalysis ofthe IgH andIgL enhancers identified aconservedsequence that serves asatissue-specificproteinbindingsitein B-cells.1,2The twoalternatively splicedproductsof the e2agene,E47and E12,werelaterisolatedasthe proteinsbindingto thisconservedsequence, which isdefinedas an E-boxsiteY A much broaderrolefor EM in development wasimmediately predicted due to its structuralhomologyto the Drosophilagenedaughterless (da),involved in celldeterminationand differentiation.3.5 Following their identification, E2Aand theadditional membersof the mammalian E-protein family, HEB and E2-2, havebeen extensivelystudied for their criticalrolesduringlymphocytedevelopment,"

E-proteinsarebasichelix-loop-helix(bHLH) transcriptionfactorsthat function asdimerstobind DNA and regulategeneexpression.TheHLH regionmediates protein dimerizationand thebasic regionmediates DNA binding.E-proteindimersbind to E-boxsites,definedbythe consensussequence CANNTG. E2Ahomodimersand E2A/HEB heterodimersare the primaryE-proteindimersfunctioningin B-and T-cells, respectively. TheDNA bindingactivity ofE-protein dimersisnegatively regulatedbythe four membersofthe Id (inhibitor ofdifferentiation) protein family,Idl-Id4. Idproteinscontainan HLH motiffordimerizationbut lackaDNA bindingbasicregion,thusallowingcompetitive dimerization to inhibit E-proteinactivity. ThebalanceofE-protein andId expression is tightlyregulatedthroughout B- and T-celldevelopment.

Association ofE2A with the Igenhancers stronglysuggests a rolefor E-proteinsin regulatingV(D)Jrecombination.E2Abindsdirectlyto E-boxsiteswithin theIgH EJ.l enhancerandIgL kappa

·Corresponding Author: YuanZhuang-Department of Immunology, Duke University MedicalCenter, Box 3010, Durham, NC 27710, U.S.A. Email: vzhuangesduke.edu

V(D)JRecombination, edited by PierreFerrier. ©2009 LandesBioscienceand SpringerScience+Business Media.

Regulation ofV(D)jRecombination byE-Protein Transcription Factors 149

(Igk) intronic and 3' enhancers.r'" Additional regions within the Ig and T-cell receptor (TCR)loci also contain E-box sites. For example, putative E-box sites have been identified downstreamof the recombination signal sequence (RSS) within most Igk V gene families" and the TCRnenhancer contains an E-box site with sequence similarity to the Ig enhancer site." E-box sites arealso located within the TCRfJ enhancer.P" Two E-box motifs are located in the core ~ enhancerregion responsible for enhancer-dependent recombination activity, and nuclear factor bindinghasbeen suggested at one of these sites by DNA footprinting analysis." In addition to sequenceanalysis of Ig and TCR regulatory regions, gene knockout and over-expression models havefurther suggested roles for E-proteins in V(D)J recombination during lymphocyte development.Accumulating evidence indicates multiple ways through which E-proteins directly or indirectlyimpact V(D)J recombination in both B- and Tcells.E-proteins can regulate V(D)J recombinationat various levels, including the transcriptional control of Ig and TCR associated genes, initiationofgene rearrangement and regulation ofdifferentiation through the developmental stages whenIg and TCR loci recombine.

Transcriptional Control ofIg and TCRAntigen Receptor and TheirAssociated Genes

E-proteins activate transcription ofmultiple factors essential for V(D)J recombination, including the Ig and TCR genes themselves. Sterile germline transcripts through Ig and TCR loci havebeen hypothesized to playa role in increasing chromatin accessibility prior to recombination."An example of this role for transcription has recently been shown at the TCRn locus.'? Whentranscription is blocked within the In locus , both rearrangement and chromatin remodeling aresuppressed. It is therefore possible that E-proteins may be impacting chromatin accessibility andrecombination through activation ofgermline transcription. There have been various examples ofE2A inducing transcription within the Igand TCRloci, mostly through in vitro studies in cell lines.Over-expression ofE2A in nonB-celllines is sufficient to induce ectopic expression ofgermlinetranscripts from the IgH and Igk loci. Forced expression ofE47has been shown to induce IgHtranscription in pre-T and fibroblast cell lines.P'" E12 has been shown to induce Igk transcriptionin a mitogen stimulated macrophage cell line'? and E12 or E47 can also activate Igk transcriptionin a kidney cell line. 21 Consistent with these results. Id over-expression in Bvcell Iines inhibits theactivity of both IgH and Igk enhancers to induce transcription, indicating the role for E2A inactivating enhancer-dependent transcription at these loci." In addition, loss ofE2A in pre -B-celllines results in a loss ofIgk rranscription." E2A may not only regulate Igk transcription throughinteraction with the intronic and 3 ' enhancers. but may also function at the Igk promoters, whereconserved E-box sites can also be found."

A similar role for E-protein mediated transcriptional activation has been suggested for theTCR loci as well. Over-expression ofE2A and/or HEB in a kidney cell line activates Vy and Vbgermline transcription." In this study, E2A and HEB activate only a specific subset ofVy andVb genes and upon cotransfection with Ragl and Rag2, rearrangements utilizing these specificV segments are induced. This correlation suggests E-protein activation of transcription is linkedto recombination at these loci . Putative E-box sites have been described within the V~ promoterregions,26but whether or not E-proteins playa similar role in activating germline transcription atthe TCRfJ loci is still under investigation.

E-protein downstream targets relative to V(D)J recombination also include genes encodingthe recombinase machinery and several receptor components that pair with the functionallyrearranged Ig and TCR chains. Two ofthese targets most essential to V(D)J recombination arethe recombination activating genes. rag] and rag2. Rag] expression is induced upon over-expression ofE12 in a macrophage cell line and Rag] and Rag2 expression levels increase uponover-expression ofE47 in a pre-Tscell line.v-" E2A has also been implicated in regulating Ragexpression by interactingwith the Erag enhancer, critical for Ragexpression in Bvcells." Forcedexpression ofId3 in T -cell progenitors inhibits the up-regulation ofRag] and Rag2. furtherdemonstrating a role for E-proteins in initiat ion of rag gene expression." Another E2A target


criticalduring V(D)] recombination is the geneencoding terminal deoxynucleotidetransferase(TdT). E2A binding has been observed at the 5' region of the tdt locus and E47 can activateTdT expressionin a nonlymphoid celllineY8

Finally, E-proteins regulate components of both the pre-Bvcell receptor (pre-BCR) andpre-TCR. E-proteins activatetranscription of surrogate light chain genes (;S and Vpre B) andpre-Ta,which are required to pair with IgH and TCRfJ,respectively.7·10.20.29-331his pairingallowsdevelopingB-cells to express a pre-BCR and developing a~ T-cellsto express a pre-TCR. E2Aalsoregulates expression of mb-l and possiblyB29, additionalcomponentsofthe pre-BCR.7.34·35Surface expression of a pre-BCR or pre-TCR triggers entry to the next stage of developmentwhere the cells will then undergo rearrangement ofIgL and TCRa genes, respectively. This rolefor E-proteinsin regulatingdifferentiationthrough the stageswhen recombinationoccurswill befurther discussed in a later sectionof this chapter.

Induction ofIg and TCR Gene RearrangementEctopicexpression ofE-proteinsin nonlymphoidcellsnot onlyactivates transcription,but also

inducesrearrangementeventsin the Igand TCR lociupon co-expression with Ragl and Rag2.Asmentioned above, introduction of E2A and/or HEB with the Ragproteins in a kidneyce1llineinducesrearrangements within the TCRy and TCRb loci.25.36In separatestudies,transfcctionofE2A and Ragwasshown to induce IgH D-] rearrangementin a pre-T-celllineand IgHD-] andIgkVkl-] rearrangements in akidneycellline.19.21J7In eachof thesecases, E2Agenerates adiverserepertoire,yet only certainsubsetsof genesegments are targeted for recombination.Themechanism by which E-proteinsmediate recombinationis not entirelyunderstood. One possibilityisthat E-proteinscreatelocalizedaccessibility for recombinationand thereforemayinfluencetherelative rearrangement efficiency ofspecific genesubsets."

The physiological role of E2A in V(D)] recombination has been further defined by in vivoand in vitro studies of Igk rearrangement in B-cells. Targeted mutation of the two functionalE-boxsiteswithin the Igkintronic enhancer resultsin a severe reduction in Igkrearrangement indevelopingB-cells38and deletion ofE2A in pre-Bscell linesblocksIgk rearrangement." In addition, re-introduction ofE47 to theseE2Adeficientpre-B-celllinesrescues Igkrecombinadon."Thesestudiessuggest that E-proteinsregulateinitiation ofV(D)J recombinationat least in panbydirectlybinding to cis-regulatoryelementswithin the recombiningloci.

E-proteins have also been proposed to regulate secondary IgL rearrangement in immatureB-cells.39E2Awild-type miceexpressinganauto-reactive BCR transgene displayasignificantpopulationofperipheralB-cells that haveundergoneasecondaryrearrangement of the endogenousIgLto replace the auto-reactive BeR. However, E2Aheterozygous miceexpressing the auto-reactiveBCR transgenecontainveryfewmatureB-cells. Thissuggests that E2AdosageiscriticalforB-cellsto undergo receptorediting, allowingreplacement of an auto-reactive receptor.

Regulation ofthe Developmental Window for V(D)J RecombinationIn addition to directlyactivating transcriptionand initiatingrearrangement asdescribedabove,

E-proteins also indirectly regulate V(D)J recombination by controlling differentiation duringB- and T-celldevelopment. SinceE-proteinsareexpressed in both B- and Tvcells, there areobviouslyadditional factorsdeterminingthe lineageand stagespecific recombinationeventsat the Igand TCR loci.Failureof cells to enter the stagewhen thesefactorsarefunctioningwouldpreventinitiation of rearrangement events. Defects in Ig or TCR recombination in E-protein deficientmodelsmayoftenresultfroma blockin developmentprior to the stagewhencellswouldundergorearrangement. For example, E2A deficientmiceexhibit a block in Bvcell developmentprior to

IgHrearrangemene.P'" E2AdeficientB-cells are blockedat the prepro-B-cell stage,a stagepriorto the pro-B-cellstagewhere IgH intronic enhancer deficientmice demonstratea block." Thissuggests that eventhough E2Ahasbeenshownto playa rolein activatingthe IgHenhancer,E2Ahas additional rolesprior to this role that contribute to the block in IgH recombinationin E2Adeficientmice. Eventhough many of the E2A targets at this earlystageof B-celldevelopment

Regulation ofV(D)]Recombination byE-Protein Transcription Factors 151

remain unidentified, potential targets have been revealed through microarrayanalysis of E2Adeficientcellsand upon overexpression ofE2A in these cells.10,43,44 The remainingchallenge is toidentifywhich of thesetargetsarecriticalfor E2Amediateddevelopmentto the pro-Bvcell stagefor subsequent rearrangement ofIgH genes.

Once developing B-cells have undergone IgH rearrangement, E2A remains critical for theexpression of the surrogate light chain components.?·10.20.3Q.32 Vpre-B- and AS are required forsurface pre-BCRexpression and proper differentiationto the prc-Bvcell stagewhere the cells willundergo IgL recombination." Although E2A is alsocriticalduring IgL rearrangement, E2Afirstregulates differentiation to the pre-Bvcell stage. IfE2A is required throughout developmentofpro and pre-Bvcells, how doesit regulateIgH and IgLin stagespecific manners?Tissueand stagespecific expression of factors that cooperatewith E2Acan result in activationof differentsetsofgenes. Forexample, E2Acooperates with the B-cell specific transcriptionfactorsearlyB-cellfactor(EBF) and PaxS to regulateexpression of mb-I in pre-Bvcells.v The differential transcriptionalnetworksestablishedbyE2Aat the pro-Bvspre-B-cellstages couldcontribute to the stagespecificeffects ofE2A at the IgH and IgL loci.Other potential mechanisms responsible for E-proteinstageand lineagespecific regulationof receptorgene lociwill be discussed further in the finalsectionof this chapter.

A similarrolealsoexists for E2Aand HEB duringdifferentiationofdevelopingT-cells. SinceT-celldevelopment isregulated bythecombineddosage ofE2A andHEB,single knockouts exhibitonly parrialblocksin T-celldeveloprnenr.w" To inhibit total E-protein activity, miceexpressinga dominant negative form ofHEB weregenerated.v The dominant negative HEB protein isableto form nonfunctional heterodimerswith E2A to inhibit both E2Aand HEB activity, thereforeresultingin a more severe phenotype than that seenin the singleknockout mice.Dominant negative HEB miceexhibit a block in T-celldevelopmentat the CD4-CD8- double negative (DN)stageand adefectinTCRfl V(D)Jrecombination.Introduction ofa functionala~TCRtransgeneisunable to rescue this developmental block, indicatingthat the rearrangementdefect is not theonlycausefor the blockat DN stage.Theseresultsdemonstratethat E-proteinshavemultiplerolesduring this windowof development. Sincethese rolesinclude regulationof differentiation, Ragexpression and perhapsTCRfl expression and rearrangement, it is likelythat multipleE-proteintargets are responsible for coordinatingV(D)J recombination at this stage. E-proteinsare thenalsorequiredforprogression fromDN to DP,partlythrough the inductionofpre-Ta expression,"E-proteinsthereforeregulatethe entryandprogression through stages criticalfor both TCRfl andTCRa recombination.

Proper regulationof genesegmentusageduring V(D)] recombinationwithin the TCRy ando loci is also dependent on E-proteins. There is a differential usageofVy and Vo genesduringrearrangement in fetal vs, adult thymocyte development.50 Adult E2A deficientmice display adefect in usage of adult predominant Vy2and VoS geneswhereas rearrangements utilizing thefetalspecific Vy3and Vol genesegments persist." Theseresultsindicate that E2Apositively andnegatively regulates specific V genesduring the windowof adult T-celldevelopment.Thisstudyalsodemonstratesa requirement for E2Aduring fetal thymocytedevelopmentfor usageof a fewV genesegments. but Vy3and VOlfetalusageappearscomparable to wild-type. The mechanismby which E2A activityresultsin the increasedusageofsomeV genesand repression of others isnot wellunderstood. The mechanismby which E2A promotes usage of genesegments in adultbut not fetaldevelopmentis suspectedto result from differentdosages ofE2A activityY·52Eventhough eM is expressed at comparablelevels in both adult and fetal thymus.Jd2 expression ishigher in fetal thymus.which would be expectedto result in reducedE2A activityin fetal compared to adult thymus.f'

Finally, accumulatingdata indicates that E-proteins can also influence the duration of therecombiningwindowof development. An example of this role is seenat the TeRa locusduringthe CD4+CD8+ double positive (DP) stage. The transcription factor RORyt, an isoformof theorphan nuclearreceptor RORy, is required in DP thymocytesto regulatethe survival windowatthis stagebyinducingBcl-XL expression.P'"This DP survival window is criticalfor establishing

Bee

llT

ee

ll... ~

88'Ig

Hge

rmlin

etr

ans

crip

tion

and

-..re

com

bin

atio

n

IgH

en

ha

nce

r'r

eg

ula

tion

ofch

rom

atin

acce

ssib

ility

?

88-..

'Rag

and

TdT

ellp

r&s

sIo

n

Era

g,T

dT

S'

reg

ion

88

'8l1

pr&s

slo

nof

surr

ogat

elig

htch

ain

and

ad

diti

ona

l-..

pre-

BC

Rco

mpo

nent

s

'req

uire

dfo

rdev

elo

pm

en

tV

pre

6,"S

,m

b-1

pro

mo

ters

IIIIg

Lre

com

bin

ing

sla

ge

pr<

>8

62

9S

're

gio

n

--

pnH

lO

P

ON

'pre

-To

ellp

r&ss

lon

're

qu

lred

for d

ew

lop

me

nt

IIIT

CR

are

com

bin

ing

stag

e

·tra

ns

crip

tion

and

r&co

mb

lna

tion

ofa

sub

set

ofVI

'end

\Ill

gena

-..

seg

me

nts

,rol

aln

adul

tvs.

lela

Iusa

ge

oreg

ulat

ion

ofT

CR

Ptr

ans

crip

tion,

reco

mbi

nat

ion

,or

chro

mat

inac

cess

ibili

ty?

'Rag

and

TdT

ellp

r&ss

lon

asIn

Bce

lls?

-..

·RO

R.,t

ellp

ress

ion

-..

ge

ne

rate

sD

Psu

r.iw

lw

ind

owfo

rpr

oper

TC

Ra

reco

mb

ina

tion

'reg

ula

tion

olT

CR

a-.

.tr

ans

crip

tion

,re

com

bin

atio

n,o

rch

rom

atin

acce

ssib

ility

?

8 8R

OR

ytp

rom

ote

r

TC

Ra

en

ha

nce

r

pre

-Ta

en

ha

nce

r

TC

R-,

an

dsl

oci?

TC

R/3

en

ha

nce

r/p

rom

ofe

rs

8--..

Rag

locu

s,

Tcf

T5

'reg

ion

'Ig<

germ

line

lI'an

scrip

tion

and

reco

mb

inat

ion

-..

'aca

l)la

tion

'IgL

seco

nd

ary

rea

rra

ng

em

en

ts?

IgK

en

ha

nce

r

'RF~188

Fig

ure

1.E

-pro

tein

med

iate

dre

gula

tion

ofV

(DlJ

reco

mbi

natio

ndu

ring

B-

and

T-c

ell

de

velo

pm

en

t.E

2Ah

om

od

ime

rsan

dE

2A/H

EB

hete

rodi

mer

sar

edi

spla

yed

for

thei

rro

les

rela

tive

toV

(D)J

reco

mbi

natio

nin

B-

vsT

-cel

ls,r

espe

ctiv

ely.

E-p

rote

ins

regu

late

Igan

dT

CR

loci

,ge

nes

requ

ired

for

rear

rang

em

ent

and

com

pone

nts

ofth

epr

e-B

CR

and

pre-

TC

Rdu

ring

proB

and

DN

stag

esof

B-

and

T-c

ell

deve

lopm

ent,

resp

ectiv

ely.

For

B-c

ell

and~T

-cel

l

deve

lopm

ent,

each

ofth

ese

even

tsis

criti

calf

orce

llsto

diffe

rent

iate

top

re-B

-an

dD

Pst

ages

toth

enun

derg

oIg

Lan

dT

CR

are

com

bin

atio

n.

E-p

rote

ins

are

not

only

requ

ired

for

diff

ere

ntia

tion

toth

ese

stag

es,

but

are

then

furt

he

rre

quire

dfo

rp

rop

er

IgL

and

TC

Ra

reco

mb

ina

tion

.E

-pro

tein

sar

eal

socr

iti

calf

orpr

oper

Vge

neus

age

du

rin

gfe

tal

and

adul

tyc

'lT-c

ell

deve

lopm

ent.

De

pic

ted

exam

ples

here

are

limite

dto

kno

wn

orp

red

icte

dsi

tes

ofE

-pro

tein

DN

A-b

ind

ing

wit

hin

targ

etge

nes.

Pot

entia

lE

-pro

tein

bin

din

gsi

tes

wh

ich

have

not

been

dem

onst

rate

das

dir

ect

bin

din

gsi

tes

inth

esp

ecifi

edce

llty

peby

Chi

Pan

alys

isar

ein

italic

s.

E-p

rote

ins

may

also

serv

efu

nctio

nsin

de

pe

nd

en

tof

DN

A-b

ind

ing

,w

hic

har

eno

tin

clu

de

din

this

figur

e.

~ ~ '-::; ~ <l li! s ~. i· :s

Regulation ofV(n)jRecombinationby£ -Protein Transcription Factors 153

a diverserepertoire ofTCRa rearrangements. Since rearrangementsthrough theJa locusduringTCRa recombination occur in a proximal to distal manner, S' to 3 ', the lifespanofDP cellscaninfluence the repertoire.56-59 RORyt deficientmice, exhibiting a shorter DP lifespan, alsoexhibita defect in usageof3 ' Ja gene segmenes."E2A has been shown to activateexpressionofRORytin thymoeytesby binding to criticalE-boxsiteswithin the promoter region.53Inagreement withthese findings,a recent study demonstrates a similar S' skewingofJa usagewhen both E2A andHEB aredeleted at the DP stage(MEJones and Y Zhuang data to be published). Likelythroughregulation of RORyt expression, E2A and HEB indirectly influence the TCRa repertoire byensuringa sufficientwindow for rearrangement.

ConclusionE-proteins demonstrate considerableinvolvementin variousaspectsofV(D)J recombination,

a fewofwhich aredepicted in Figure1.Twomain questions remain.First,what are the underlyingmechanismsguiding E-protein mediated transcriptional regulation in a lineageand stagespecificfashion?Second,what rolesare E-proteinsplayingin addition to actingastranscriptional regulators? As mentioned earlier, E-proteins are suspected to generate localized accessibility aroundspecificgene segmentswithin various receptor loci," E-proteins havealso been suggestedby additional studies to playa role in chromatin modification. Cooperative efforts ofE2A, EBF andPaxShavebeen shown to regulate CpG demethylation and nudeosome remodeling at the mb-Iprornorer" More relative to V(D)J recombination, E2A has also been shown to playa role inIgk enhancer acetyladon," If E-proteins can induce chromatin accessibility for recombinationin certain localizedregionswithin both the Ig and TCR loci,how do E-proteins regulate Ig andTCR receptors specifically in B- and T-cells, respectively? Even though expressionof E2A innonlymphoid cellscan induce rearrangements,it is important to remember that these are mostlyover-expression studiesand E2Amaybe inducingexpression ofadditionalfactorsthat arerepressedin B- or T-cells.Theseresultssuggestthat overallE-protein dosagemayplaya role in differentialgene activation. For example,some targets may require a certain threshold of E-protein activityto be activated.This threshold would be expected to be exceededin over-expression studies,butmaybe differentiallyregulatedin B-and T-cells. Also,limited access to E-boxsitesin B-vsT-cellscould potentially contribute to E-proteins' B-vsT-cell specific effects.

Another wayE-proteins could be exhibiting lineageand stagespecific affectsis through regulated interactions with different binding factors. So faronly a few co-activatorsinteracting withE-proteins in lymphocytes havebeen identified. One group of factors that have been shown toassociatewith E-proteins are the histone acetyltransferases (HATs) p300, CBP and PCAF.6<J.63One study showsthese interactions existingin B-cells and demonstrates that HATscan enhanceE2A transcriptional activity.62 However,which E2A target genes are dependent on E2A-HATinteractions haveyet to be determined.The corepressorETO hasalsobeen shown to interact withE-proteins and in doing so, blocksthe recruitment ofHATs.64 ETO isalsoable to bind to histonedeacetylases (HDACs).64.65 The ability of E-proteins to recruit either HATs or HDACs couldpotentially contribute to the lineageand stagespecificeffectsofE-proteins at the Igand TCRloci.Another wayE-proteins could havelineageand stagespecific functions is through recruitment ofE-proteinsbyfactorswith more restrictedexpressionpatterns. Forexample, IRF-4hasbeen shownto promote E2A recruitment at the Igk3 ' enhancer in pre-B-cells.23 Future studieswill likelyshedmore light on how the somewhat ubiquitous, yet tightly regulated, expressionofE-proteins canresult in lineageand stagespecific regulation of the Igand TCR genes.

References1. Ephrussi A, Church GM, Toncgawa S ec al. B lineage-specific interactions of an immunoglobulin

enhancer with cellular factors in vivo. Science 1985: 227(4683):134-140.2. Church GM, Ephrussi A, Gilbert W et al. Cell-type-specific contacts to immunoglobulin enhancers in

nuclei. Nature 1985; 313(6005) :798-801.3. Murre C, McCaw PS, Yaessin H et al. Interactions between heterologous helix-loop-helix proteins

generate complexes that bind specificallyto a common DNA sequence. Cell 1989; 58(3):537-544.


4. Murre C, McCaw PS, Baltimore D. A new DNA binding and dimerization motif in immunoglobulinenhancer binding, daughterless, MyoD and myc proteins. Cell 1989; 56(5):777-783.

5. Jan YN,]an LY. HLH proteins, fly neurogenesisand vertebrate myogenesis. Cell 1993; 75(5):827-830.6. Murre C. Helix-loop-helix proteins and lymphocytedevelopment,Nat Immunol2005; 6(11):1079-1086.7. Greenbaum S, Zhuang Y. Identification of E2A target genes in B-lymphocyte development by

using a gene tagging-based chromatin immunoprecipitation system. Proc Nat! Acad Sci USA 2002;99(23): 15030-15035.

8. Murre C, Voronova A, Baltimore D. Bvcell- and myocyte-specific E2-box-binding factors contain EI2/E47-like subunits. Mol Cell BioI 1991; 11(2):1156-1160.

9. Henthorn P, Kiledjian M, Kadesch T. Two distinct transcription factors that bind the immunoglobulinenhancer microE5/kappa 2 motif. Science 1990; 247(4941) :467-470.

10. Greenbaum S, Lazorchak AS, Zhuang Y.Differential functions for the transcription factor E2A in positive and negative gene regulation in pre-B.lymphocytes.] BioI Chern 2004; 279(43) :45028-45035.

11. Brekke KM, Garrard WT. Assembly and analysisof the mouse immunoglobulin kappa gene sequence.Immunogenetics 2004; 56(7):490-505.

12. Ho IC , Yang LH, Merle G et aI. A 'f-cell-specific transcriptional enhancer dement 3 ' of C alpha inthe human Tvcell receptor alpha locus. Proc Nat! Acad Sci USA 1989; 86(17) :6714-6718.

13. Takeda], Cheng A, Mauxion F er aI. Functional analysis of the murine T'cell receptor beta enhancerand characteristics of its DNA-binding proteins. Mol Cell BioI 1990; 10(10):5027-5035.

14. Gottschalk LR, Leiden ]M. Identification and functional characterization of the human T-cell receptorbeta gene transcriptional enhancer: common nuclear proteins interact with the transcriptional regulatorydements of the T-cell receptor alpha and beta genes. Mol Cell BioI 1990; 10(10):5486-5495.

15. Tripathi RK, Mathieu N, Spicuglia S er aI. Definition of a Tvccll receptor beta gene core enhancer ofV(D)] recombination by transgenic mapping. Mol Cell BioI 2000; 20(1):42-53.

16. Krangd MS. Gene segment selection in V(D)] recombination : accessibiliryand beyond. Nat Immunol2003; 4(7):624-630.

17. Abarrategui I, Krangd MS. Regulation of T-cell receptor-alpha gene recombination by transcription .Nat Immuno12006; 7(10) :1109-1115.

18. Choi ]K, Sheri CP, Radomska HS et aI. E47 activates the Ig-heavychain and TdT loci in non-Bvcells,EMBO] 1996; 15(18):5014-5021.

19. Schlissel M, Voronova A, Baltimore D. Helix-loop-helix transcription factor E47 activates germ-lineimmunoglobulin heavy-chain gene transcription and rearrangement in a pre-Tvcell line. Genes Dev1991; 5(8):1367-1376.

20. Kee BL, Murre C. Induction of early Bvcell factor (EBF) and multiple B lineage genes by the basichelix-loop-helix transcription factor E12. J Exp Med 1998; 188(4):699-713.

21. Romanow W], Langerak AW,Goebel P ee al. E2A and EBF act in synergy with the V(D)] recombinaseto generate a diverse immunoglobulin repertoire in nonlymphoid cells. Mol Cell 2000; 5(2):343-353.

22. Wilson RB, Kiledjian M, Shen CP et al. Repression of immunoglobulin enhancers by the helix-loop-heltrprotein Id: implications for B-lymphoid-cell development. Mol Cell BioI 1991; 11(12):6185-6191.

23. Lazorchak AS, Schlissel MS, Zhuang Y. E2A and IRF-4/Pip promote chromatin modification andtranscription of the immunoglobulin kappa locus in pre-Bvcells. Mol Cell BioI 2006; 26(3) :810-821.

24. Bemark M, Liberg D, Leanderson T. Conserved sequence elements in K promoters from mice andhumans: implications for transcriptional regulation and repertoire expression. Immunogenetics 1998;47(3):183-195.

25. Ghosh ]K, Romanow W), Murre C. Induction of a diverse Tvcell receptor gamma/delta repertoire bythe helix-loop-helix proteins E2A and HEB in nonlymphoid cells] ExpMed 2001; 193(6):769-776.

26. Chen F, Rowen L, Hood L er aI. Differential transcriptional regulation of individual TCR V betasegments before gene rearrangement.] Immunol2oo1; 166(3):1771-1780.

27. Hsu LY, Lauring], Liang HE et aI. A conserved transcriptional enhancer regulates RAG gene expressionin developing B-cells. Immunity 2003; 19(1):105-117.

28. Blom B, Heemskerk MH, VerschurenMC er aI. Disruption of alpha beta but not of gamma delta T-celldevelopment by overexpression of the helix-loop-helix protein Id3 in committed Tvcell progenitors.EMBO] 1999; 18(10):2793-2802.

29. Herblot S, Steff AM, Hugo P et aI. SCL and LMOI alter thymocyte differentiation: inhibition ofE2A-HEB function and pre-T alpha chain expression. Nat Immunol2000; 1(2):138-144.

30. lkawa T, Kawamoto H, Wright LY et aI. Long-term cultured E2A-deficient hematopoietic progenitorcells are pluripotent. Immunity 2004; 20(3) :349-360.

31. Sigvardsson M. Overlapping expression of early Bvcell factor and basic helix-loop-helix proteins asa mechanism to dictate Bvlineage-specific activity of the lambda5 promoter. Mol Cell BioI 2000;20(10) :3640-3654.

Regulation ofV(D)JRecombination byE-Prottin Transcription Factors 155

32. Sigvardsson M. O'Riordan M, Grosschedl R. EBF and E47 collaborate to induce expression of theendogenous immunoglobulin surrogate light chain genes. Immunity 1997; 7(1):25-36.

33. Reizis B. Leder P. Expression of the mouse pre-Tvcell receptor alpha gene is controlled by an upstreamregion containing a transcriptional enhancer. J Exp Med 1999; 189(10):1669-1678.

34. Maier H. Ostraat R, Gao H et al. Early B-cell factor cooperates with Runxl and mediates epigeneticchanges associated with rnb-I transcription . Nat ImmunoI2004; 5(10):1069-1077.

35. Sigvardsson M, Clark DR, Fitzsimmons D er al, Early Bvcell factor, E2A and Pax-5 cooperate to activatethe early Bvcell-specific mb-I promoter. Mol Cell Bioi 2002; 22(24):8539-8551.

36. Langerak AW,Wolvers-Tettero IL. van Gastel-MolEJ et aI. Basichelix-loop-helixproteins E2A and HEBinduce immature T-cell receptor rearrangements in nonlymphoid cells. Blood 2001; 98(8):2456-2465.

37. Goebel P. Janney N. ValenzuelaJR et al. Localized gene-specific induction of accessibility to V(D)J recombination induced by E2A and early Bvcell factor in nonlymphoid cells. J Exp Med 2001;194(5):645-656.

38. Inlay MA, T ian H. Lin T et al. Important roles for E protein binding sites within the immunoglobulin kappa chain intronic enhancer in activating Vkappa ]kappa rearrangement. J Exp Med 2004;200(9) :1205-1211.

39. Quong MW, Martensson A. Langerak AW et al. Receptor editing and marginal zone B-cell developmentarc regulated by the helix-loop-helix protein, E2A. J Exp Med 2004; 199(8):1101-1112.

40. Bain G. Maandag EC, lzon DJ et al. E2A proteins are required for proper Bvcell development andinit iation of immunoglobulin gene rearrangements. Cell 1994; 79(5) :885-892.

41. Zhuang Y, Soriano P, Weintraub H. The helix-loop-helix gene E2A is required for Bvcell formation.Cell 1994; 79(5) :875-884.

42. Perlot T. Alt FW, BassingCH et al. Elucidation of IgH intronic enhancer functions via germ-line deletion. Proc Nat! Acad Sci USA 2005; 102(40):14362-14367.

43. Schwartz R. Engel I, Fallahi-Sichani M er al. Gene expression patterns define novel roles for E47 in cellcycle progression. cyrokine-mediared signaling and T lineage development. Proc Nat! Acad Sci USA2006; 103(26):9976-9981.

44. IkawaT, Kawamoto H. Goldrath AWet al. E proteins and Notch signaling cooperate to promote T-celllineage specification and commitment. J Exp Med 2006; 203(5):1329-1342.

45. Shimizu T, Mundt C. Licence S et al. Vpre-Bl/Vpre-Bz/lambda 5 triple-deficient mice show impaired B-cell development but functional allelic exclusion of the IgH locus. J Immunol 2002;168(12):6286-6293.

46. Bain G. Engel I. Robanus Maandag EC et al. E2A deficiency leads to abnormalities in alphabeta T-celldevelopment and to rapid development ofT-cell lymphomas. Mol Cell Bioi 1997; 17(8):4782-4791.

47. Barndt R. Dai ME Zhuang Y. A novel role for HEB downstream or parallel to the pre-TCR signalingpathway during alpha beta thymopoiesis, J Immunol 1999; 163(6):3331-3343.

48. Barndt RJ. Dai M, Zhuang Y. Functions of E2A-HEB heterodimers in Tcell development revealed bya dominant negative mutation ofHEB. Mol Cell Bioi 2000; 20(18) :6677-6685.

49. Wojciechowski]. Lai A. Kondo M et al. E2A and HEB Arc Required to Block Thymocyte ProliferationPrior to Pre-TCR Expression.] ImmunoI2007; 178(9):5717-5726.

50. Raulet DH. The structure, function and molecular genetics of the gamma/delta T-cell receptor. AnnuRev Immuno11989; 7:175-207.

51. Bain G. Romanow WJ. Albers K er al. Positive and negative regulation ofV(D)J recombination by theE2A proteins. J Exp Med 1999; 189(2):289-300.

52. David-Fung ES. Yui MA. Morales M ec al. Progression of regulatory gene expression states in fetal andadult pro-T-cell development. Immunol Rev 2006; 209:212-236.

53. Xi H. Schwartz R, Engel I et al. Interplay between RORgammat. Egr3 and E proteins controls proliferation in response to pre-TCR signals. Immunity 2006; 24(6) :813-826.

54. Sun Z. Unutmaz D. Zou YR et al. Requirement for RORgamma in thymocyte survival and lymphoidorgan development. Science 2000; 288(5475) :2369-2373.

55. Kurebayashi S. Ueda E, Sakaue M er al. Retinoid-related orphan receptor gamma (RORgamma) isessential for lymphoid organogenesis and controls apoptosis during thymopoiesis. Proc Nat! Acad SciUSA 2000; 97(18) :10132-10137.

56. Wang E Huang CY, Kanagawa O. Rapid deletion of rearranged T-cell antigen receptor (TCR)Valpha-Jalpha segment by secondary rearrangement in the thymus: role of continuous rearrangementof TCR alpha chain gene and positive selection in the T-cell repertoire formation. Proc Nat! Acad SciUSA 1998; 95(20) :11834-11839.

57. Thompson SD, Pelkonen J. Hurwitz JL. First T-cell receptor alpha gene rearrangements during T-cellontogeny skew to the 5 ' region of the J alpha locus. J Immuno11990; 145(7):2347-2352.

58. Petrie HT. Livak E Burtrum D et al. T-cell receptor gene recombination patterns and mechanisms: celldeath, rescue and T-cell production. J Exp Med 1995; 182(1):121-127.

156 V(D)JR~combjnatjon

59. Guo J. HawwariA. Li H er al. Regulation of the TCRaipha repertoire by the survivalwindow of CD4( +)CD8(+) thyrnocytes. Nat Immuno12002; 3(5):469-476.

60. Eckner R, YaoTP, Oldread E et aJ. Interaction and functional collaboration of p300/CBP and bHLHproteins in muscle and B-cell differentiation. Genes Dev 1996; 10(19):2478-2490.

61. Qiu Y, Sharma A, Stein R. p300 mediates transcriptional stimulation by the basic helix-loop-helixactivators of the insulin gene. Mol Cell Bioi 1998; 18(5):2957-2964.

62. Bradney C, Hjelmeland M, Komatsu Y et aJ. Regulation of E2A activities by histone acetyltransferasesin B-lymphocyte development. JBioi Chern 2003; 278(4) :2370.2376.

63. Massari ME. Grant PA, Pray-Grant MG et aJ. A conserved motif present in a class of helix-loop-helixproteins activates transcription by direct recruitment of the SAGA complex . Mol Cell 1999;4(1):63-73.

64. Zhang J, Kalkum M, Yamamura S er aJ. E protein silencing by the leukemogenic AMLl-ETO fusionprotein. Science 2004; 305(5688): 1286-1289.

65. Hug BA. Lazar MA. ETO interacting proteins. Oncogene 2004; 23(24) :4270-4274.

CHAPTER 13

Temporal and Spatial RegulationofV(D)J Recombination:Interactions ofExtrinsic Factorswiththe RAG ComplexYun Liu,LiZhangandStephen Desiderio"

Abstract

I n the course oflymphoid development, V(D}J recombination is subject to stringent locusspecific and temporal regulation. These constraints are ultimately responsible for severalfeatures peculiar to lymphoid development, including the lineage specificity of antigen

receptor assembly, allelic exclusion and receptor editing. Inaddition, cell cycle phase-dependentregulation ofV(D}J recombinase activity ensures that DNA rearrangement is completed bythe appropriate mechanism of DNA repair. Regulation ofV(D}J recombination involves interactions between the V(D}J recombinase-a heteromeric complex consisting ofRAG -I andRAG -2 subunits-and macromolecular assemblies extrinsic to the recombinase. This chapterwill focus on those features ofthe recombinase itself-and in particular the RAG-2 subunitthat interact with extrinsic factors to establish patterns oftemporal control and locus specificityin developing lymphocytes.

Functional Organization ofRAG-l and RAG-2RAG -l and RAG -2 are 1040 and 527 amino acid residues long, respectively. Residues 384

through 1008 ofRAG-l constitute the core fragment, which contains the catalytic site for DNAcleavage,"? mediates binding to recombination signal sequences (RSSs}4-6 and makes contactswith the coding flanks.?.8The core RAG-2 fragment (Fig. I), consisting of residues 1 through387, extends interactions of RAG-l with the RSS and is essential for helical distortion near thescissilebond, a possible prerequisite for transesteriticarion.v-?Accordingly, mutations that impairrecombinase-rnediated cleavageand joining have been identified in core RAG-2.w

Residues 387 through 527 ofRAG-2 comprise the non-core region (Fig. I) and are dispensable for DNA cleavageby the RAG proteins in vitro. Nonetheless, removal ofthis region reducesthe efficiency of extrachromosomal recornbinadon.!':" increases production of hybrid joints,"impedes endogenous Vwto-DJH joiningI2.18.19 and promotes aberrant recornbination.P Themechanisms underlying these effects may be complex, as the non-core region includes multiplefunctional domains (Fig. lB) .

*Corresponding Author: Stephen Desiderio-Department of Molecular Biology and Geneticsand Institutefor Cell Engineering,TheJohns Hopkins UniversitySchool of Medicine,Baltimore, Maryland 21205, USA.Email: sdesider@jhmLedu

V(D)j Recombination, edited by Pierre Ferrier. ©2009 Landes Bioscienceand Springer Science+ Business Media.

158 V(D)j Rtcombi1l4tion

Temporal Regulation ofV(D)J Recombination through Interactionswith the RAG-2 Non-Core Region

Thenon-coreregionofRAG-2 supportsthe periodicdestructionofRAG-2 protein. RAG-2accumulates in quiescent cells and in dividingcells during the GI phase; rapid degradation ofRAG-2 begins at the GI -to-S transition and continuesuntil the following entry into GUI•23

Consequently, the appearance of recombination signal end intermediates24.2S and RAG-signalend complexes" is restrictedto GO/G1.Destruction ofRAG-2 is triggered byphosphorylationof threonine490,whichlieswithin aphylogenetically conserved cyclin-dependent kinase (Cdk)targetsiteand isalsodependenton a lysine-rich intervalspanningaminoacidresidues 499-508.21

Overlapping the RAG-2 degradation domain (Fig. IB) is a noncanonical nuclearlocalizationsequence that supportsbinding of importin 5 and nuclearimport of RAG-2P At the Gl-ee-Stransition,phosphorylation ofRAG-2 bycyclinA/Cdk2permitsassociation ofRAG-2 with theSkp2-SCF ubiquitinligase.Thisphosphorylation-dependentinteractionismediatedbytheF-boxprotein Skp-2and itsassociated protein Cksl .Uponpolyubiquitylation ofRAG-2bySkp2-SCF,RAG-2issubjectedto proteasomal degradation."

The cellcycle dependence ofV(D)] recombination mayplaya role in the couplingof DNAcleavage bythe RAGcomplex to DNA repair.V{D)J recombination isnormally completedbyaform of DNA repairtermed nonhomologous end joining(NHEJ). NHE] is active throughoutthe cellcycle, but an alternative mechanism for double-strandDNA repair, homologous recombination (HR), isnearlyinactive duringG I. 29ln thymocytes of miceexpressing RAG-2{T490A),aberrantrecombinants resembling productsof abortivehomologous recombination areobservedto accumulate.P These observations suggest that restriction ofRAG-2 accumulation to the GOand GI cellcycle phases promotes the correct repair ofV(D)J recombination intermediates byNHE], perhapsbytemporalsequestration of RAGactivityfrom HR.

A core non-core

B

Y402 Y415N4030406E407

M443W453 T490 K503

Figure 1. Regulatory domains of RAG-2. A) Schematic representation of mouse RAG-2. Coreand non-core regions are designated; amino acid residues are numbered below. KL, Kelch-likepropeller domains; L, linker domain; PHD, plant homeodomain finger; D, doma in governingprogrammed degradation and nuclear import of RAG-2. B) Detailed representation of thenon-core region. Amino acid residues at domain boundaries are numbered above. L (blackrectangle), PHD (gray rectangle) and D (hatched rectangle) as defined in (A). The hatchedinterval denotes the extent of the domain governing cell cycle-dependent degradation ofRAG-2; the shaded region within this interval marks the nuclear import signal that resideswithin the degradation domain. Shadedarrowheads, sitesof mutations in the linker doma in thatimpair V(D)J recombination. Open arrowheads, targets of mutations in the PHD domain thatabolish H3K4me3 binding and impair V(D)J recombination. Black arrowhead, cyclinNCDK2phosphorylation site, essential for programmed degradation of RAG-2 at the G1-S transition.Shaded diamond, target of mutation that selectively impairs nuclear import of RAG-2.

Temporal andSpatial Regulation ojV(DJjRecombination 159

Locus Specificity: General RemarksThe V(D)J recombinase is directed toward particular sets of gene segments, depending on

lymphoid lineageand developmental stage.Recentwork has begun to provide a frameworkforunderstandinghow this targetingisachieved. At the levelof unchromatinizedDNA, the V(D)Jrecombinase is targeted to antigen receptorgenesegmentsby meansof specific interactionswithflanking RSSs and this recognitiondoes not require the non-core regions of RAG-I or RAG-2.Not all RSSs support recombination with the sameefficiency, because RSSs exhibit considerablesequence variation.Although sequence variationamong RSSs can indeedaffectgenesegmentusage,30 thesedifferences cannot accountfor the dynamicshiftsin locusspecificity that accompanycommitmentto distinct lymphoidlineages and developmental transitionswithin lineages. Rather,ordered rearrangement of antigen receptor gene segments is associated with the imposition orreliefof epigenetic marks. Specific chromatin modifications in the vicinityof RSSs are stronglyassociated with the presence or absence of ongoingrearrangement. The propensityof a particularlocus to undergo rearrangement has been thought to be determined by accessibility to the RAGcomplex, aviewthat ascribes a passive roleto the recombinase. Recentfindings, however, indicatethat the recombinase-through direct bindingto modifiedchromatin-is an active partner in theepigenetic regulationof rearrangement. Wediscuss belowhow epigenetic marksinteractwith theV(D)J recombinase to promote locus-specific rearrangement.

Epigenetic Modifications ofPossible Relevance to V(D)JRecombination

An alteration in gene function is termed epigenetic if it is maintained through cell divisionand does not involve a changein the DNA sequence. One extensively studied epigenetic mark isDNA methylationon cytosine, whichin mammals occursat most CpG dinucleotides. A farmorecomplex setofepigenetic marksareassociatedwith theproteincomponentsofchromatin.Thebasicunit of eukaryoticchromatin is the nucleosome. This consists of a histone core-two moleculeseachof the histonesH2A, H2B, H3 and H4-around whicharewrappedabout 146basepairsofDNA. Histones are subject to a varietyof posttranslationalmodifications includingacetylation,methylation, phosphorylation, ubiquirylation and sumoylation. Differences in the degreeand stereospecificity of modification contributesubstantially to the complexity of thesemarks. Lysine. forexample,canbe rnono-,di- or trimethylated, whileargininecanbe dimethylatedsymmetrically orasymmetrically. Inadditionto chemicalmodification, the register inwhichDNA iswrappedaroundthe histone core-termed nucleosome phasing-can haveprofound effects on the accessibility ofspecific sequences to interaetingfactors. Observations relatingthesemodesofepigenetic regulationto the activation or suppression ofV(D)J recombinationaresummarized in turn below.

DNA MethylationMethylation ofCpG dinucleotides isnormallyassociated with the suppression of transcription.

Consistentwith ageneralcorrelationof recombination with transcription,CpG methylationoverantigen-receptor-gene segments is also associated with suppression ofV(D)J recombination."Deletion ofPD~I, a promoter locatedS' to the D~1 genesegmentor E~ an enhancerlocated 3'to the TC~ locus, is accompanied by increased CpG methylation in the D~I-]~1 region anddefectsinTC~1rearrangement .Pv' Conversely, demethylationof DNA has beenassociatedwithactivation of rearrangement. In developing Bvcells, for example, the IgK allelethat is firstactivatedfor rearrangement isdemethylated overthejx-CKregion,whilethe oppositeallele remains hypermethylatedand is recruited to heterochromatin.3s.36

Nucleosome PhasingTogether, thecoreRAG-l and RAG-2fragments catalyze RSS-specific nickingandtransesterifi

cationof DNA substrates invitro.Efficient cleavage isnot observed.however, whenchromatinizednuclearsubstrates are used." RAG-mediatedDNA cleavage in vitro is impededwhen the targetRSSis incorporated into a nucleosome;38-lO the degreeofinhibition hasbeenvariously proposed


to be dependent" or independenc" ofnucleosome phasing relative to the RSS. The resistance ofmononucleosomal substrates to cleavage may result from inaccessibility of histone-associatedDNA to the RAG complex as well as from helical distortion induced by wrapping of the DNAaround the histone core.40.41 The impediment to RAG-mediated DNA cleavage observed withmononucleosomal substrates in vitro can be relieved synergistically by histone acetylation andSWI/SNF-dependent remodeling, possibly as a result ofalterations in chromatin structure thatenhance accessibility ofthe RSS to the RAG complex.40•41

Histone AcetylationAcetylation ofhistones H3 and H4 is associated with active chromatin. A positive correlation

between histone acetylation and active antigen receptor gene rearrangement has been widelydocumented. Decreased acetylation ofH3 and H4 is associated with diminished germline transcription at unrearranged antigen receptor loci and is important for allelicexclusion.36.4z.45 DuringB-cell development, diminished IL-? signaling is associated with decreased histone acetylationand reduced accessibility to nucleasesover distal VH segments," A similar relationship is observedoverV~ segments during the transition ofintrathymic T-cell progenitors from the CD4-CD8- tothe CD4+CD8+ stage.46Thus, decreases in histone acetylation are associated with diminished rearrangement. Consistent with this relationship, IgKallelesat which recombination is active exhibitincreased acetylation ofhistone H3.36

Histone H3 K9 MethylationDimethylation ofhistone H3 at lysine 9 (H3K9me2), which is associated with silent chro

matin, is positively correlated with inhibition ofV{D)J recombinadon.f'" Dimethyl marksat H3K9 are removed over VH segments at the pro-B to pre-B-cell transition, at which stageVwto-DJ H joining occurs ; H3K9 demethylation is dependent on expression ofthe transcriptionfactor PaxS in pro-Bvcells." A role for H3K9me2 in the control ofV{D)] recombination wassuggested in an experiment that targeted G9a, a histone H3K9 methyltransferase, to aTC~minilocus. In th is setting. directed H3K9 methylation was found to inhibit both germlinetranscription and V{D)J recombination, overriding the presence of cis-acting accessibilitycontrol elements." An interpretation of these findings is complicated. because ablation of theG9a methyltransferase in mice had no significant effects on lymphoid development or stagespecificity ofV(D)] recombination. despite suppress ive effects on Alight chain usage. B-cellproliferation and plasma cell differenriaeion."

Histone H3 K4 MethylationMethylation ofhistone H3lysine4 (H3K4) isaphylogeneticallyconserved modification thathas

been linked to transcriptional activation in yeastand metazoans," The relationship between histoneH3K4 methylation and V{D)J recombination has been the subject ofmuch recent smdy.36.48.52.53Dimethylated histone H3K4 {H3K4me2)48.53 and trimethylated H3K4 {H3K4me3)S4·55 exhibitdistinct patterns ofenhancement within the D-] H cluster in pro-B-cells poised to undergo D-to-]H

rearrangement. Moreover. the recombinationally active Ig K allele in pre-B-cells is marked by hypermethylation ofH3K4.36

Monoubiquitylation ofhistone H2B at lysine 123 (ubH2B) promotes histone H3K4 methylation in yeast.56-58 UbH2B isassociated with transcriptionally activechromatin both in yeast59-63 andin mammalian cells.60.64 Patterns ofubH2B deposition have yet to be extensivelymapped. AsH2Bubiquitylation appears to be a prerequisite for H3K4 hypermethylation, it will be of interest toknow whether the density ofubH2B is enhanced at sites ofactive V{D)J recombination. possiblyextending the chain ofcausation one step upstream.

Direct Recognition ofModified Histone H3 by the V(D)j RecombinaseThe observations outlined above, while essential to an understanding ofepigenetic control,

do not in themselves provide mechanistic insight into how histone modification is linked

Temporal andSpatial Regulation ojV(D)j Recombination 161

mechanistically to V(D)J recombination.Buildingon recent progressin the understanding ofhowhistonemethylationpatternsareread,several studieshavecombinedbiochemical, structuraland geneticapproachesto outline how one such linkage isestablished.

Avariety ofprotein domainsarecapable ofbindingthe N-terminalregionofhistoneH3 whenthis is hypermethylated at lysine 4. These includethe chromodomains of CHD 1,6S,66 the doublerudor domain of]MJD2A67and the plant homeodomain(PHD) fingers ofING2,68'71 BPTF68.71and Yngl." Crystallographic analysis reveals that the PHD fingers ofING2.69BPTF68 andYngl72

all contain an aromatic cage that mediates binding to methyl-lysine. a feature shared by othermerhyl-lysme-binding domains." Thestructuralbasis ofH3K4me2 or H3K4me3 bindingbythePHD finger isofparticularlybroadsignificance, because thisrecognition domainispresentinmanychromatin-associated proteinsthat carryout histone modification.Y"

The abilityof the PHD finger to mediate binding to H3K4me2 and H3K4me3 led severalgroupsto examine the functionofasimilardomainthat earlierhadbeenidentifiedwithin residues419 through 481 of the non-core regionofRAG-V6This noncanonicalPHD finger(Fig. IB)wasshown to mediatedirect bindingofRAG-2 to histone H3 di- or trimethylared at K4,witha preference for H3K4me3.S4.SSMutations that abolish binding of the RAG-2 PHD finger toH3K4me3 (Fig. IB) werefound to impairV(D)J recombinationboth within extrachomosomalsubstrates and at endogenous 10ci.S4•ssMoreover. the association ofthe RAG-2 PHD finger withchromatinacross the immunoglobulin heavy chainlocusispositively correlated with thedensityofH3K4me3.S4Mutationsthat disruptH3K4me3bindingor Zn++ coordinationbythe RAG-2PHDfingerhad been associated earlierwith combinedhereditaryimmunodeficiencies in humans.77•

81

underscoring the physiologic importanceof theseinteractions.The crystalstructuresof complexes betweenthe RAG-2 PHD finger and modifiedH3 pep

tides haveshown that this domain. while functionally related to its canonicalcousins. exhibitsthe unusualabilityto integrateepigenetic marks.821n the complex with apeptidebearingK4me3.the rrimethyl ammonium group of K4 is buried in an "aromatic cage" similar to that of othermethyl-lysine-binding domains. An important difference between the PHD finger of RAG-2and other H3K4me3-bindingdomains. however. wasobserved: an enhancedaffinityforadoublymodifiedhistone-namely, H3 bearingboth K4Me3and asymmetricallydimethylated arginineatposition2 (R2Me2s).Thisispossible because the RAG-2PHD finger lacks asidechaincarboxylatethat in homologous domainsformssalt bridgeswith unmodifiedR2.1n RAG-2 this is replacedby tyrosine, which mediates interactionswith H3R2me2s.82An important consequence is thatbinding ofRAG-2 to an H3 peptide bearingK4me3is enhancedby the presence ofR2Me2s.82

While the differential affinities of RAG-2 for singly and doublymodifiedhistone H3 could inprinciplecontributeto locusdiscrimination bytheV(D)Jrecornbinase, thephysiological relevanceof this property remains unclear, because symmetric methylationof histone H3 R2hasasyet notbeen detected in vivo.

Evidence for Allosteric Regulation ofV(D)J RecombinaseActivity by Histone H3 Trimethylated at Lysine 4

The engagement of histone H3K4me3 by the RAG-2 PHD fingerprovides a bridgebetweenone chemicalmark of active chromatin and the V(D)J recombinase machinery. Paradoxically.whileV(D)J recombination isprofoundlyimpairedbyapoint mutation that abolishes H3K4me3bindingbythe RAG-2 PHD finger. completeremovalofthe non-coreregion.includingtheentirePHD finger. hasonlya modestdebilitatingeffeet.S4•ssTo reconcile theseobservations it hasbeenproposedthat an inhibitory domain resides within the non-core regionofRAG-2 and that suppressionof recombinase activityby this domain is relieved upon engagement of the PHD fingerby H3K4me3 (Fig. 2). Consistent with this proposalis a crystalstructure in which the RAG-2PHD finger-in the absence ofan H3K4me3ligand-is occupiedbyan amino-terminalpeptideencoded by the expression construct.F It maybe that hypermethylated H3K4 does not simplyact asa dockingsitefor the recombinase but rather plays a moreactive roleasan allosteric triggerof RAG catalysis.

Constitutive


Active

. ~~InactiVe /

~

Figure 2. A model for allosteric activat ion of the RAG complex by modified histone H3.White figures represent RAG-2; C and NC denote core and non-core regions. respectively.Shaded object represents trimethylated lysine 4 of histone N3 (H3K4me3). In a hypotheticalinactive conformation (left), the aromatic channel of the RAG-2 PHD finger is occup ied byan inhibitory domain residing elsewhere in the non-core region. In the hypothetical activeconformation (upper right), the PHD finger is bound by histone H3K4me3 and the putativeinhibitory domain is released. The RAG-2 core fragment (lower right) lacks both the PHDfinger and the putative inhibitory domain. In this configuration RAG-2 is proposed to assumean active configuration constitutively. For further discussion, see text.

Future Directions: Deposition and Integration ofEpigeneticSignals Controlling V(D}) Recombination

The link between transcriptional activation and locus-specificity ofV(D)J recombination haslong suggestedthat transcription and V(D)J recombination are controlled by shared epigeneticmechanisms.Progressin understanding these mechanismshas awaitedthe chemicalcharacterization ofepigeneticmarks and the development ofmethods bywhich the genomicdistribution ofthesemarkscouldbe mapped.Theseapproacheshavebegun to provideadetailedviewofepigeneticchange at antigen receptor genesas a function of development. Several important questions willcontinue to dominate the field.

The first is to defineprecisely the structural features that confer locus specificity to the V(D)Jrecombinase. While recognition of histone H3K4me3 by RAG-2 providesa link between activechromatinandV(D)J recombination, it isobviousthat H3K4me3-ageneralmarkoftranscriptionallyactivechromatin-is too broadlydistributedto actaloneindirectingthe recombinase to specificsitesofaction.Clearlyother modesofregulationmust contributeto locusspecificityofrecombinaseactivity. While it seemslikely that this will involve a combinatorialsummationof chromatinmodificationsand DNA sequenceelements, the answerisfar from clear. A relatedquestionconcernsthedirect roleofmodifiedchromatin in regulatingRAG activity. The proposalthat the recombinase isallostericallyactivatedupon bindingofthe RAG-2PHD fingerto modifiedchromatinwillneed to betestedand the relative contributionsofmodifications at H3K4, H3R2 and elsewhere will need to bedefined.Regionsof the RAG-2other thanthe PHD fingermayalso mediatefunctionalinteractionswith chromatin.The RAG-2linker region,whichliesat the amino-terminalsideofthe PHD finger(Fig. lB), has been reported to bind corehistonesand mutationswithin this regionwerefound toimpairVH-to-DJHjoining;83the basis for this apparentlyselective effectisunclear.A third questionconcernshowdevelopmental signals,suchasthosethat emanatefromthepreBCR,govern depositionand removalofepigeneticmarksat antigen receptor loci.A resolution ofthese outstanding issueswill provide a starting point from which to addressthe largerproblem ofallelicexclusion.

Temporal andSpatial Regulation ofV(D)j Recombination 163

References1. Fugmann SD. Villey I]. Ptaszek LM ecaI. Identification of two catalytic residues in RAG1 that define a

single active site within the RAGl/RAG2 protein complex. Mol Cell 2000; 5:97-107.2. Kim DR. Dai Y,MundyCL et aI. Mutationsof acidic residuesin RAGI define the active siteof the V(D)

] recombinase. GenesDev 1999; 13:3070-3080.3. Landree MA. Wibbenmeyer ]A. Roth DB. Mutational analysis of RAGI and RAG2 identifies three

catalytic amino acids in RAGI critical for both cleavage stepsofV(D)J recombination. GenesDev 1999;13:3059-3069.

4. Akamatsu Y, Oettinger MA. Distinct roles of RAGI and RAG2 in binding the V(D)] recombinationsignal sequences. Mol Cell Bioi 1998; 18:4670-4678.

5. Difilippanronio M]. McMahan C]. Eastman QM et aI. RAG1 mediates signal sequence recognition andrecruitmentofRAG2 in V(D)] recombination. Ccll1996; 87:253-262.

6. Swanson PC. Desiderio S. V(D)J recombination signal recognition: distinct. overlapping DNA-proteincontactsin complexes containing RAGI with and without RAG2. Immunity1998; 9:115-125.

7. Eastman QM. Villey I]. Schatz DG. Detection of RAG protein-V(D)J recombination signal interactionsnear the site of DNA cleavage by UV cross-linking. Mol CelI Bioi 1999; 19:3788-3797.

8. Swanson PC. Desiderio S. RAG-2promoteshepramer occupancy by RAG-l in the assembly of a V(D)Jinitiationcomplex. Mol Cell Bioi 1999; 19:3674-3683.

9. Hiom K. Gellert M. A stable RAGl-RAG2-DNA complex that is active in V(D)J cleavage. Cell 1997;88:65-72.

10. Qiu ]X. KaleSB. Yarnell SchultzH er aI. Separation-of-function mutants reveal critical roles for RAG2in both the cleavage and joiningsteps ofV(D)] recombination. Mol Cell 2001; 7:77-87.

11. Cuomo CA. OettingerMA. Analysis of regions of RAG-2 important for V(D)] recombination. NucleicAcids Res 1994; 22:1810-1814.

12. Kirch SA.RathbunGA. OettingerMA. Dual roleofRAG2 in V(D)Jrecombination: catalysis and regulation of orderedIg geneassembly. EMBO] 1998; 17:4881-4886.

13. McMahan C]. Difilippantonio M]. RaoN et aI. A basic motif in the N-terminal region of RAG1 enhancesV(D)J recombination activity. Mol CelI Bioi 1997; 17:4544-4552.

14. Sadofsky M]. Hesse ]E. Gellert M. Definition of a core region of RAG-2 that is functional in V(D)Jrecombination. Nucleic Acids Res1994; 22:1805-1809.

15. Sadofsky M]. Hesse ]E. McBlane ]F et aI. Expression andV(D)J recombination activity of mutatedRAG-lproteins. Nucleic Acids Res 1993; 21:5644-5650.

16. SteenSB.Han jO, MundyC et aI. Roles of the "dispensable" portions of RAG-l and RAG-2 in V(D)]recombination. Mol Cell Bioi1999; 19:3010-3017.

17. Sekiguchi ]A. Whitlow S. Alt FW Increased accumulation of hybrid V(D)] joins in cells expressingtruncatedversus full-length RAGs. Mol Ce1I2001; 8:1383-1390.

18. Akamatsu Y. MonroeR. DudleyDD et aI.Deletionof the RAG2C terminus leads to impaired lymphoiddevelopment in mice. Proc Natl AcadSci USA 2003; 100:1209-1214.

19. LiangHE. Hsu LY, Cado D ct aI.The "dispensable" portion of RAG2 is necessary for efficient V-to-D]rearrangement duringB- and T-celldevelopment. Immunity2002; 17:639-651.

20. Talukder SR. DudleyDD. Alt FWet aI. Increased frequency of aberrant V(D)] recombination productsin core RAG-expressing mice. Nucleic Acids Res2004; 32:4539-4549.

21. Li Z. Dordai DI. Lee] et aI.A conserved degradation signal regulates RAG-2 accumulation duringcelIdivision and linksV(D)] recombination to the cellcycle. Immunity 1996; 5:575-589.

22. Lin WC. Desiderio S. Regulation ofV(D)] recombination activator protein RAG-2by phosphorylation.Science 1993;260:953-959.

23. Lin WC. Desiderio S. CelIcycle regulation ofV(D)] recombination-activating protein RAG-2. Proc NatlAcadSciUSA 1994;91:2733-2737.

24. Desiderio S. Lin WC. Li Z. ThecelI cycle and V(D)] recombination. Curr Top Microbiol Immuno11996;217:45-59.

25. Schlissel M. Constantinescu A. Morrow T et al. Double-strand signal sequence breaks in V(D)] recombination are blunt. S'-phosphorylared, RAG-dependent and cell cycle regulated. Genes Dev 1993;7:2520-2532.

26. Jiang H. Ross AE. Desiderio S. Cell cycle-dependent accumulation in vivo of transposition-competentcomplexes between recombination signal ends and full-length RAG proteins. ] BioI Chern 2004;279:8478-8486.

27. Ross AE.Vuica M. Desiderio S.Overlapping signals for proteindegradation and nuclear localization definea role for intrinsicRAG-2 nuclear uptakein dividing cells. Mol Cell BioI2003; 23:5308-5319.

28. JiangH. ChangFC. Ross AE er aI. Ubiquitylation ofRAG-2 bySkp2-SCF linksdestruction of the V(D)] recombinase to the cellcycle. Mol Cell 2005; 18:699-709.

164 V(D)jRt:combination

29. Takata M, Sasaki MS, Sonoda E et al. Homologous recombination and nonhomologous end-joiningpathways of DNA double-strandbreakrepair have overlapping roles in the maintenance of chromosomalintegrity in vertebrate cells. EMBOJ 1998; 17:5497-5508.

30. Feeney AJ.Factors that influence formation of B-cell repertoire. Immunol Res2000; 21:195-202.31. Nakase H, Takahama Y,Akamatsu Y. Effect ofCpG methylation on RAGI/RAG2 reactiviry: implications

of direct and indirectmechanisms for controlling V(D)J cleavage. EMBO Rep2003;4:774-780.32. Whitehurst CE, Chattopadhyay 5, ChenJ. Control ofV(D)J recombinational accessibiliry of the D beta

1 genesegment at the TCR beta locus by a germline promoter. Immunity 1999;10:313-322.33. Mathieu N, Hempel WM, Spicuglia 5 et al.Chromatin remodeling bythe T-cell receptor (TCR)-beta gene

enhancer dotingearly T-celldevelopment: Implicationsfor the controlofTCR-betalocus recombination.J Exp Med 2000; 192:625-636.

34. WhitehurstCE, Schlissel MS,ChenJ. Deletion of gennline promoterPD beta 1 fromthe TCR betalocuscauses hypermethylation that impairs D beta 1 recombination by multiple mechanisms. Immuniry 2000;13:703-714.

35. Mostoslavsky R, Singh N, Kirillov A et al. Kappa chain monoallelic demethylation and the establishmentof allelic exclusion. Genes Dev 1998;12:1801-1811.

36. Goldmit M, Ji Y, SkokJ er aL Epigenetic ontogeny of the Igk locus doting B-cell development. Nat Immunol2005; 6:198-203.

37. Stanhope-Baker P, Hudson KM, Shaffer Ai et al. Cell type-specific chromatin structuredetermines thetargeting of V(D)J recombinase activity in vitro. Cell 1996; 85:887-897.

38. Kwon J, Imbalzano AN, Matthews A er al. Accessibiliry of nucleosomal DNA to V(D)Jcleavage is modulated by RSS positioning and HMG1. Mol Cell 1998; 2:829-839.

39. GoldingA, Chandler5, Ballestar E et al. Nucleosome structurecompletely inhibits in vitro cleavage bythe V(D)J recombinase. EMBOJ 1999;18:3712-3723.

40. Kwon J, Morshead KB, Guyon JR ecal. Histone acerylation and hSWI/SNF remodeling act in concertto stimulate V(D)J cleavage of nucleosomal DNA. Mol Cell 2000;6:1037-1048.

41. Baumann M, Mamais A, McBlane F et aL Regulation ofV(D)J recombination by nucleosome positioningat recombination signal sequences. EMBOJ 2003;22:5197-5207.

42. Chowdhury D, Sen R. Stepwise activation of the immunoglobulin mu heavy chaingenelocus. EMBOJ2001; 20:6394-6403.

43. Chowdhury D, Sen R. Transient IL-7/IL-7Rsignaling provides a mechanism for feedback inhibition ofimmunoglobulin heavy chaingenerearrangements. Immunity 2003; 18:229-241.

44. Johnson K, Calame K. Transcription factors controlling the beginning and end of B-cell differentiation.CUrtOpin Genet Dev 2003; 13:522-528.

45. McMurry MT, Krangel MS. A role for histone acetylation in the developmental regulation of VOJ recombination. Science 2000;287:495-498.

46. Tripathi R, Jackson A, Krangel MS.A change in the structureof Vberachromatin associated with TCRbeta allelic exclusion. J Immunol 2002; 168:2316-2324.

47. JohnsonK, PAugh DL, Yu D er aL Bvcell-spccific loss of histone3 lysine 9 methylation in the V(H) locusdepends on Pax5. Nat Immunol 2004;5:853-861.

48. Morshead KB, Ciccone DN, Taverna SD er al. Antigen receptor locipoisedfor V(D)J rearrangement arebroadly associated with BRG1 and flanked by peaks of histone H3 dimethylated at lysine 4. Proc NadAcadSci USA2003; 100:11577-11582.

49. Osipovich 0 , Milley R, Meade A et al. Targeted inhibitionofV(D)J recombination by a histone rnethyltransferase. Nat Immunol2004; 5:309-316.

50. Thomas LR,Miyashita H, Cobb RM et al.Functional analysis of histonemethyltransferase g9ain B- andT-Iymphocytes. J Immunol2008; 181:485-493.

51. Sims R] 3rd, Reinberg D. Histone H3 Lys 4 methylation: caught in a bind? Genes Dev 2006;20:2779-2786.

52. Perkins E], KeeBL, Ramsden DA. Histone 3 lysine 4 methylation doting the pre-B to immature B-celltransition. Nucleic Acids Res 2004;32:1942-1947.

53. Chakraborry T, Chowdhury D, Keyes A et al. Repeat organization and epigenetic regulation of theDH-Cmu domain of the immunoglobulin heavy-chaingenelocus. Mol Cell 2007; 27:842-850.

54. Liu Y, Subrahmanyam R, Chakraborry T et al. A plant homeodomain in RAG-2 that binds Hypermethylated lysine 4 of histoneH3 is necessary for efficient antigen-receptor-gene rearrangement. Immunity2007;27:561-571.

55. Matthews AG, Kuo A], Ramon-Maiques 5 et al. RAG2 PHD lingercouples histone H3 lysine 4 trimethylation with V(D)J recombination. Nature 2007;450:1106-1110.

56. Dover ], Schneider ], Tawiah-Boateng MA et al. Methylation of histone H3 by COMPASS requiresubiquitination of histoneH2B by Rad6. J BioI Chern2002;277:28368-28371.

Temporaland SpatialRegulation ofV(D)jRecombination 165

57. Sun Zw, Allis CD. Ubiquitination of histone H2B regulates H3 methylation and gene silencing in yeast.Nature 2002; 418:104-108.

58. LeeIS. Shukla A. Schneider] et al. Histone crosstalk between H2B monoubiquitination and H3 methylation mediated by COMPASS. Cell 2007; 131:1084-1096.

59. Henry KW; Wyce A. Lo WS ec al. Transcriptional activation via sequential histone H2B ubiquitylationand deubiquirylanon, mediated by SAGA-associated Ubp8. Genes Dev 2003; 17:2648-2663.

60. Zhang Y. Transcriptional regulation by histone ubiquitination and deubiquitinarion . Genes Dev 2003;17:2733-2740.

61. Kao CF. Hillyer C. Tsukuda T et al. Rad6 plays a role in transcriptional activation through ubiquitylationof histone H2B. Genes Dev 2004; 18:184-195.

62. Xiao T. Kao CF, Krogan N] et at Histone H2B ubiquitylation is associated with elongating RNA polymerase 11. Mol Cell Bioi 2005; 25:637-651.

63. Minsky N. Shema E. Field Y er al. Monoubiqu itinated H2B is associated with the transcribed region ofhighly expressedgenes in human cells. Nat Cell Bioi 2008; 10:483-488.

64. Kirmizis A. Santos-Rosa H. Penkert C] et al. Arginine methylation at histone H3R2 controls depositionofH3K4 rrimethylation. Nature 2007; 449:928-932.

65. Flanagan ]F. Mi LZ. Chruszcz M et al. Double chromodomains cooperate to recognize the methylatedhistone H3 tail. Nature 2005; 438:1181-1185.

66. Sims RJ 3rd. Chen CF. Santos-Rosa H et al. Human but not yeast CHDI binds directly and selectively to histone H3 methylated at lysine 4 via its tandem chromodomains. ] Bioi Chern 2005;280:41789-41792.

67. Huang Y. Fang]. Bedford MT et al. Recognition of histone H3lysine-4 methylation by the double tudordomain of]M]D2A. Science 2006; 312:748-751.

68. Li H. Din S. Wang W et al. Molecular basis for site-specific read-out of histone H3K4me3 by the BPTFPHD finger of NURF. Nature 2006; 442:91-95.

69. Pena PV; Davrazou F, Shi X et al. Molecular mechanism of histone H3K4me3 recognition by planthomeodomain ofING2. Nature 2006; 442:100-103.

70. Shi X. Hong T. Walter KL et al. ING2 PHD domain links histone H3 lysine 4 methylation to activegene repression. Nature 2006; 442:96-99.

71. Wysocka]. Swigut T. Xiao H et al. A PHD finger ofNURF couples histone H31ysine 4 trimethylationwith chromatin remodelling. Nature 2006; 442:86-90.

72. Taverna SD. Ilin S. Rogers RS et al. Yngl PHD finger binding to H3 trimethylated at K4 promotesNuA3 HAT activity at K14 of H3 and transcription at a subset of targeted ORFs. Mol Cell 2006;24:785-796.

73. Ruthenburg A]. Allis CD. Wysocka J. Methylation of lysine 4 on histone H3 : intricacy of writing andreading a single epigenetic mark. Mol Cell 2007; 25:15-30.

74. Bienz M. The PHD finger. a nuclear protein-interaction domain. Trends Biochem Sci 2006; 31:35-40.75. Mellor J. It takes a PHD to read the histone code. Cell 2006; 126:22-24.76. Callebaut I. Momon JP. The V(D)J recombination activatingprotein RAG2 consistsofa six-bladed propel

ler and a PHD fingerlike domain. as revealed by sequence analysis. Cell Mol Life Sci 1998; 54:880-891.77. Gomez CA. Ptaszek LM. Villa A et al. Mutations in conserved regions of the predicted RAG2 kelch

repeats block initiation ofV(D)J recombination and result in primary immunodeficiencies. Mol Cell Bioi2000; 20:5653-5664.

78. Noordzij JG. de Bruin-Versteeg S. Verkaik NS ec at The immunophenorypic and immunogenotypic B-celldifferentiation arrest in bone marrow of RAG-deficient SCID patients corresponds to residual recombination activities of mutated RAG proteins. Blood 2002; 100:2145-2152.

79. Schwarz K. Gauss GH . Ludwig L et al. RAG mutations in human Bvcell-negarive SCID. Science 1996;274:97-99.

80. Villa A. Sobacchi C. Notarangelo LD er at V(D)J recombination defects in lymphocytes due to RAGmutations : severe immunodeficiency with a spectrum of clinical presentations. Blood 2001; 97:81-88.

81. Villa A. Santagata S. Bozzi F et al. Partial V(D)] recombination activity leads to Omenn syndrome. Cell1998; 93:885-896.

82. Ramon-Maiques S. Kuo A]. Carney D et al. The plant homeodomain finger ofRAG2 recognizes histoneH3 methylated at both lysine-4 and arginine-2. Proc Natl Acad Sci USA 2007; 104:18993-18998.

83. West KL. Singha NC. De Ioannes P et al. A direct interaction between the RAG2 C terminus and thecore histones is required for efficient V(D)] recombination. Immunity 2005; 23:203-212.

CHAPTER 14

V(D)J Recombination:OfMice and SharksEllen Hsu*

Abstract

T he adaptiveimmune systemofjawedvertebratesisbasedon avast,anticipatory repertoireof specific antigen receptors, immunoglobulins(Ig) in B-Iymphoeytes and T-cellreceptors(TCR) in T-Iymphoeytes.The Igand TCRdiversityisgeneratedbyaprocesscalledV(D)J

recombination, which is initiated by the RAG recombinase,Although RAG activityisverywellconserved, the regulated accessibility ofthe antigen receptor genesto RAG has evolvedwith thespecies' organizational structure, which differs most significantly between fishes and tetrapods.V(D)J recombination wasprimarilycharacterizedin developinglymphocytesofmiceand humanbeingsand isoftendescribedasanordered, two-stageprogram.Studiesin rabbit,chickenand sharkshowthat this processdoes not haveto be ordered, nor does it need to take placein two stagestogenerate a diverserepertoire and enable the expressionofa singlespeciesofantigen receptor percell, a restriction calledallelicexclusion.

Introduction

Origins oftheAdaptive Immune SystemV(D)J recombination is the process by which antigen receptors, immunoglobulin (Ig) and

T-cellreceptor (TCR), areassembled for expressionduring developmentofthe respective B- andT-Iymphoeytes. Somaticrearrangementofthe V (variable), D (diversity)andJ (joining) genesegments ' is initiated bythe recornbinaseRAG (recombination-activatinggene)2.3 in a cut-and-pasteprocess that entails joining of these separate gene components to encode the V region, theN-terminus ofthe receptor polypeptide.The V region is 100-120 amino acid long and forms theligand-binding site in heterodimers ofheavy (H) and light (L) chains of Ig, the alpha and betachainsofTCRa~ and the gammaand delta chains ofTCRyl'>.

RAG and lymphocytes expressingIg and TCR are present in all jawed vertebrates (Fig. 1),from cartilaginous fishes to mammals.Neither RAG nor the rearranging receptors are found inprotochordares or lamprey and hagfish, which suggests that the present RAG function becameestablished in a vertebrate ancestor sometime in the 80 million yearsbetween the divergenceofjawless fishes and cartilaginousfishes.4•5 Extensive duplication events,either two whole-genomeduplications or one genome-wideduplication and multiple segmentalduplications occurred before and afterdivergenceofjawlessfishes.6-9The incipienceand evolutionofthe adaptiveimmunesystemtook placeduring this period ofextensivegenomic restructuring,'?

The origin of the rearranginggenes was first suggestedby Sakano and coworkers," who remarked that the recognition motifs (recombination signal sequences, RSS) adjacent the V(D)Jgenesegmentswerereminiscentofsignals found at the termini ofintegratedtransposableelements.

*Ellen Hsu-Department of Physiology and Pharmacology, State University of New York,Health Science Center at Brooklyn, Brooklyn , New York, USA. Email: [email protected]

V(D)J Recombination, edited by Pierre Ferrier.©2009 Landes Bioscienceand Springer Science-Business Media.

V(D)jRecombination: O/MiceandSharks 167

Figure 1. Evolution of the chordates. The phylogenetic relationships among chordates isshown (boxes) with notations of the major animal models in each taxon beneath the boxes.The adaptive immune system, defined by RAG-mediated rearranging antigen receptor genesof the Ig superfamily and by the major histocompatibility complex, has been found only inthe jawed vertebrates (gnathostomes, beige boxes). Protochordates include Cephalochordates(Amphioxus) and Urochordates (Ciona/tunicates). Numbers denote when the taxa emerged inevolution (millions of years ago). Reprinted with some alterations; Hsu E, Pulham N, RumfeltLL et al. The plasticity of immunoglobulin gene systems in evolution. Immunol Rev 2006;210:8-26. Copyright Blackwell Munksgaard 2006 .

Subsequently the chemistryof the RAG-mediatedpathwaywasfound to resemble thosedescribedfor transpositional recombination bymobileelements," With the growingavailability of genomeinformationfromdifferent species, it becamefeasible toattemptdelving into theoriginsofRAGandits recognitionsequences. Terminalinvertedrepeats with motifsand spacerintervalsimilarto RSSwereobservedin the Transib transposonin nematodes, insectsand seaurchin.13RAGiscomposedof two componentsand both RAG1- and RAG2-like sequences weredetected in the seaurchingenome, although their function is not yet dear.14Thesediscoveries, togetherwith demonstrationof latent transposase activityin RAG,15.16 arguefor RAG havingbeen part of a DNA transposonthat wasintroducedearlyinto the vertebratelineage, evolving to its roleofV(D)J recombinase byretaining the excision component of transposase activityP-19The presence of RAG sequences inechinoderms couldindicateentryof the transposonat afarearliertimeandlost incertainphylaandclasses (proeochordates, jawless fishes) but retainedin others,or else a separatehorizontal transferin jawedvertebrates.

It ishypothesized that in the ancestralvertebratethe RAGtransposonbecameintegratedintoa V-likegene,splittingit into two components that can rejoin afterRAG-induceddouble-strandbreakage and removal of the interveningDNAY Thecleavage occursin the sameplacedue to RSSrecognition, but because of the nucleotidelossand/or gainarisingfrom the repairprocess, the new

168 V(D)JRt!combination

joints would bevariedin sequence[next section). Breakage and repairofDNA inducedby RAGthus generates molecularheterogeneity, which mayarguablyhavebeen the selecting factor iftheoriginalV genehad an immune function that wasenhanced bydiversified sequences.

V(D)Jrecombination becameestablishedinearlyvertebratesabout500million years agoandistheprocess that assembles Igand TCRgenes inallspecies and the species-specific receptorgenes liketheIgNAR(newantigenreceptor)andNAR-TCRinsharIcsW-21 andTCRJ1in marsupials.PThis chapterdealsmainlywithcomparative studieson theIggenesystem.Thereisconsiderablymoreinformationon antibodyin earlyvertebrates, due to the much longerhistoryof studiesofIg protein and to therelativeeaseofdetectingVB sequences across specieswithheterologousprobes.TheIgMmolecule isverywellconserved fromsharksto mammals23

,24 in overallsequence and structure, beingthe antigenreceptoron naive B-cells that in plasmacells issecreted asapolymeric antibody, usually a pentamer.TCR eDNA sequences characterized in all animals show that they are cellsurface receptors only.TCRa~ and TCRybhavebeen clonedfrom allclasses of jawedvertebrates,25-29 includingall threemajorgroupsof mammals (marsupials suchasopossum." monotremes suchasduckbillplarypus"and placentals of mostorders, includingrodents,rabbits, ruminantsand primates).

There are two evolutionarily conservedfeatures ofV(D)J recombination: the mechanismofRAG action and the regulationof this processto ensureone end result-that only one kind ofantigen receptor isexpressed per cell(for a review, seeref 32).This restrictioniscalledallelic exclusion.Although the recombinationpathwaymediatedbyRAG iswellconserved, the regulatedaccessibility of the antigenreceptorgenesto RAG hasevolved with the organizationalstructure,which differsmost significantly between cartilaginous fishes and tetrapods (Fig. 2).

V(D)J Rearrangement

RAG Recognition andJointResolutionThe rearranging elements-the gene segments V, D and J with their adjacent RSS-are

present in all classes of jawedvertebrates,as are the keyenzymes involvedin DNA nickingandmodification, RAGlIRAG2 and terminal deoxynucleotidyl transferase (TdT). Although theselymphocyte-specific enzymes havebeenstudiedalmostentirelyin mouseor invitrosystems,33 theirhighly conservedmode of action in other animalsmay be deduced. First,pairwiserecognitionof the RSSis required and the RSSpair to be recombinedmust consistof one RSScontaininga12-bpspacerand the other a 23-bp spacer("12123rule").' In all vertebrates where the genomicorganizationof the genesegmentshasbeen determined, the RSSthat flankpotentiallyrecombinogenicgenesegmentsreflectthis pairingrelationship.

RAGinitiates thepathwaythatleadstodouble-strand breaks ateithergenesegment andthecodingendsbeingsubsequentlyjoinedbythecell's DNA repairprocesses (Fig. 3).Double-strandbreakage isobtainedinatransesterification reactionthat results inacovalentlyclosedhairpinon thecodingendand a freeblunt RSSat the other.The hairpinisopenedasymmetrically, creating an overhang withinvertedrepeat,someofwhichisoccasionally retained(P region) aspart of the ligatedjoint.

The presenceof P region is thus indicativeof a hairpin intermediate createdduring the double-strandbreakand joiningprocess. Examinationof the VD and DJ junctionsin IgH chainsandTCR ~ and bchains,or VJjunctionsin IgL chainsand TCR a and y chains,the portion of the VsequencecalledCDR3 (complementarity-determiningregion3),showsgermlinecontribution (Vand ] genesegmentflanks, portions ofD genesequence) and occasional P region in all animals,suggesting that V(D)J recombination at different loci and in variousspecies undergo the sameunique processinvolving hairpinned codingends.

SelectionforJunctionalDiversificationA secondcategoryofsomatically-generated additions at the junction is N region,which con

sistsof nonrernplared, mostlyGC-richsequences catalyzedbyTdTJ4.3S that, togetherwith codingend-processingmediatedbyexonucleases, arethe maincontributorsto generatingthediversificationat the junctions ofthe rejoinedgenesegments.

V(D)j Recombination: O/MiceandSharks 169

OH Ol .IH (II other (

;--f... f-•....

'---'lOObp

OJ ,.........After 0 to J r.arrangem.nt

Germline mouse 19 heavy chain locusVHI- 2OOI

Roar ra"lled VDJ: VOJ

Germline shark 19 heavy chain lociIgH G2A

O. D1 JH ( II

'}--tI~~rl-lr-. ;-fL........J t.......J L.......J '---J

416 bp J76bp399bp 6JI~b 120 0b

IgHG5

(II

Rearranged VOJ:

Figure 2. Comparison of Ig H chain genes in mouse and shark. Germline mouse Ig H chainlocus: the mammalian H chain locus consists of a series of tandemly duplicated VH, 0 and JHgene segments that rearrange during B-cell development. The recombined VOJis transcribedwith one of the downstream constant (C) region genes, here simplified as single units (bluebox is CI1). The VHis represented by olive boxes, preceded by the leader sequence in darkgreen and flanked by the recombination signal sequence (RSS, white triangle) at the 3' end,that consists of heptamer and nonamer motifs separated by a 23 bp spacer sequence. Asindicated, the distance between the 3'-most VHand the first functional D is 90 kb. The Dgene segments in red, flanked on both sides by RSS (black triangles) containing 12 bp spacersand the JH gene segments (orange) with 23 bp spacer RSS. After D to J rearrangement: thefirst stage of rearrangement involves recombination between D and JH, with the interveningDNA excised. The DJ product is depicted as a fusion of the red and orange boxes, withthe RSS flanking its 5' end. Rearranging V to DJ: locus contraction and looping of the DNAallows linearly distant VHgene segments to recombine with the DJ. The final VOJ productis shown as Rearranged VOJ. Germline shark Ig H chain loci: the IgM H chain genes insharks and skates (cartilaginous fishes) are multiple miniloci each consisting of VH, two D,one JH and one CI1 gene (blue box). The gene segments in any nurse shark IgH gene arelocated about 400 bp apart as shown but are distant (e.g., 6.3-6.8 kb) from the CI11 exon.The physical relationships among the loci are not clear except for one instance, where theywere located 120 kb apart ." Rearranged VOJ: the four gene segments rearrange within theminilocus to VOOJ (called VO)). Whereas in mouse IgH gene rearrangement takes placein a strict order (0 to JH before VHto 0)), the rearrangement of the four gene segments inthe shark takes place at once and without any strict order. Reprinted with permission fromMalecek K, Lee V, Feng W et al. Immunoglobulin heavy chain exclusion in the shark. PLoSBioi 2008; 6:e157. Copyright 2008 Malecek et al. A color version of this image is availableat www.landesbioscience.com/curie.

170 V(D)j&combination

hepwmer 12 t>p none.....

c odln g"• • _~G - - ---

TlllCAC - - ---

~codlnnaa==~::::::

'*'-, 23 DP n.~,

~ RAGlIRAG2 nicking

OIl

=====~1c.v:.\GTO - N:NNNI:. -- OTlllCAC -- TlJTTTTTTO ---

---GTTTTTTGT ~~ - - - -

---CAMMN:.A aroN;N;ItO," - - - • ~

p

-----... 'CACAOTO-- N:NNNI:. --_____~) ,GTGTCN;-- TIJTTTTTTO ---

mmI,

:::::;=:---= .::::\

p

--........------------- - .........--

Double·.tranded breakHairp in lormallon

Signa. end Joi ning

G A}

p" -----A' - - - - --A

Ha irp in opening

-----GATC T - - - - -

TAA -----~

-----GATCggT- - - - - -

-----CTAGocA- - - - - -

Trimm ing , retenllon 01P nucleolld.. ITCI, N addIlion (ggl

and coding Joint lormallon

Figure 3. RAG-mediated recombination. Detailsof V(D)J recombination aredescribedin otherchapters of thisbook.Theflanksof genesegments beforerearrangement areshownwith the RSSenclosed by triangles, to correlatewith the symbols in Figures 2 and 4. The RSS pair is boundby RAG, which introducesnicks.The nicking occurs5' of the 7-merend of the R5S on the topstrandof each of the two Ig genesegments, producing a 3'-hydroxy on the coding end of theIg genesegment and a 5' phosphoryl on the RSS (signal) end. The result is a duplex nicked ateitherRSS. Thesecond stepinvolvesintramoleculartransesterfication reactions wherethe3'-OHattack the opposing phosphodiester bonds, causing the coding ends to become a covalentlyclosed hairpinsand freeing the blunt signal ends. Joining of the endsis carried out by the nonhomologous end joining repair pathway. The hairpin coding endsare opened asymmetricallyby the nuclease Artemisand the resultant single-stranded overhang consists of a portion of thecoding end and its complementarysequence. Sometimes the overhang could be included aspart of the final joined product and is observed as inverted repeat sequence (P region). TheDNA endsaretrimmed;TdTmayinsertnontemplated nucleotides(lowercase letters). Reprintedwith some alterations; Hsu E. Immunoglobulin recombination signal sequences: somatic andevolutionary functions. In: Caporale L, ed. The Implicit Genome, New York: Oxford UniversityPress, 2005, Chapter9. Copyright2006 by Oxford University Press, Inc.

V(D)JRecombination: O/MiceandSharks 171

The other members of the mammalian Family X DNA polymerases, to which TdT belongs,arenot restrictedto precursorlymphocytes, but two of them, Polu and PolA., arealsoinvolved atdifferentpoints during codingend-processing and appear to modulate the extent ofcodingendnucleolytic processing." TdT and Pol u areveryclosely related" and their presence in fishes," incontrast to the single-copy ancestral form in the urochordate Ciona [nmicates, Fig. 1), suggeststhat the lymphocyte-specific TdT evolved to its current role in the immunesystem by the timeof divergence of cartilaginous fishes. The earlyinvolvement ofTdT in the evolution ofV(D)Jrecombinationreflects its importancein amplifying the selectedattributesconferredbygenerearrangement: sequence and sequence length variationasa resultof RAG-inducedbreakage.

Diversification mechanisms like mutation or gene conversion may exist in invertebraees"and predate the rearranging Iggenesystem, but theseprocesses do not generatesequence lengthdiversity repeatedly and reliably in one location that will tolerate a loop sizespectrum of 2-23amino acidsin the human H chain" or 9-13 in the shark L chain." The greatestcontribution tothe combiningsite topology is thus madeby the variable CDR3. Fromcrystallographic studiesof antigen-antibody complexes H chainCDR3 appears to playthe mostsignificant role,not onlyin the number of contactswith antigen but alsoin its potential for conformationalchanges for"inducedfit" upon bindingligand.42.43

NovelRearranging Genes in Sharks andMarsupialsSome species requireantigen receptor diversity additional to that provided by heterodimer

specificities of the IgITCR repertoire. Theyare (1) shark IgNAR20: neither Ig nor TCR but anearlydivergent gene,(2) sharkNAR-TCR21: TCR isoforms, produced bygraftingan additionalV region onto an existing TCRb rearrangement by splicing, forming two successive V regionsand (3) marsupial TC~22 : a hybrid of Ig and TCR components whose product also containstwo joinedV regions. IgNARisa secretedserumprotein and the other two presumably areactivein cell-mediated processes.

IgNAR is a dimer but the V regionsare not paired; the ligand-binding site is thus a singleV region. The IgNAR V region is generated by four rearrangements-c-V, three D and J genesegments-providinghighlyvariable andexceptionallyplasticCDR3that arepostulatedto adoptmultipleconformations forinduced-fit binding.44IgH chaindimerswithsingle-domain Vregions(VHH) arealsoexpressed in camels." although thesegenesegments arepart of the IgH locus."SharkIgNARand the camelVHH arethe resultofconvergent evolution, asaresharkNAR-TCRand opossumTC~. NAR-TCR ispart of the sharkTCRb locus, whereasTC~ in opossumareencodedbyindependentgeneclusters ."TC~, likeIgNAR,involves rearrangement of2-3 D elements. Because TCRIl and IgNARare

encodedbya fewrniniloci, their repertoireisbasedsolely on CDR3junctionaldiversity.Theuseof a single V domain, somewith longerCDR3, in cartilaginous fishes and in mammals suggeststhat thereexists somecategory of antigensthat requireligand-bindingsitesperhapsmoreflexiblethan providedbythe classical Igor TCR heterodimer.

One ReceptorPer CellBoththestrengthand theweakness ofV(D)J recombination isitsrandomnature.An immensely

diverse,anticipatory repertoire isgeneratedconcomitantwithcelland resource wastage.Theprocesscannotensure that theVbecomes joinedin-frame with respectto theJ (andC region)sequence, sothat leasttwo ofthreerearrangements attemptsarenonfunctional. Moreover, randomly-generatedspecificities alsoincludethose that recognize selfcomponentsand theseareeliminatedat the immaturelymphocyte stage whentriggered byaselfligand.Selection forself-tolerance or mountinganimmuneresponse ismostefficaciously (i.e.,specifically) mediatedwhenonlyonespeciesofreceptor isexpressedpercell," Thelastphenomenon, knowngenerallyasallelic exclusion, results fromregulatedRAGaccess to the recombinogenic elements.V(D)Jrecombination islineage- andcellstage-specific,meaningthat DNA from nonlymphoidcells or fromcells of the incorrectdevelopmental stagearenot actedupon by RAG.48 Thereare,however, someinterestingexceptions in cartilaginous fishesand thesearedescribed in a latersection("Rearrangement ofIg genes in non-Bscells").


V(D)J Rearrangement PatternsIn the mouseand human systems the rearrangementofIg Hand L chainsis oftendescribedas

an ordered, regulatedprogram.49Thewell-studiedstepsinvolve formationofthe DJbeforerecombinationofVH to the DJin pro B-cells,followed bycelldivisionand subsequentrearrangement ofthe Lchaingenesin pre Bvcells, wherethe kappaL chain (1gK) locusisactivatedbeforethe secondL chain isorype, lambda (Ig"-). The regulated accessibility of different genesand genesegmentsto RAG enableone H chain alleleto be expressed (allelic exclusion, H chain exclusion) with onealleleof either 1C or 'A. L chain (allelic and isotypicexclusion); hence, one kind of antigenreceptorper lymphocyte. Outside of the mousemodel there is currentlylittle information on TCR or Igchromatin and DNA modification,bur V(D)] rearrangementpatterns reflectthe order of geneaccessibility to RAG and thesearecomparedamongmouse,rabbit, chickenand shark.

MouseIn the mouseIgH generearrangementtakesplacein asetorder and in step-wise fashion.59.51At

thepro B-cell stagethechromatindomainencompassingthe D,]H and Cf.l genesbecomeactivated,probablythrough the intronic enhancerand allowD to]H recombinationon both chromosomes.Thisisfollowed byactivation ofthe chromatindomaincontainingtheupstreamVH genes. Becauseof the verylargedistancebetweenthe VH genesegments and the D], locuscontractionand loopingofthe DNA52.54 arerequiredto bring them into closeproximiryfor rearrangement (Fig. 2).Hchain exclusion is the outcomeof the staggeringof the V to D] step betweenthe two alleles. IfthefirstVD] isnot viable, rearrangement continueson the homologouschromosome.

The initiation and maintenanceofallelicexclusion involves relocationof the genesin nuclearcomparrments.t-" In pro B-cells IgH repositions awayfrom the nuclear periphery and thismay have to do with its activation; in pre B-cells the nonrearranged allele is recruited to thepericentrometric heterochromatin, an interaction thought to be repressive forrecombination.How rearrangement beginsat one allelebefore the other is not clearand the basismaydiffer atthe IgH, TCR~ and IgK genes. An explanation for asynchronous rearrangement at the TCR~locus has been recentlyproposed after finding that in rearrangingT-cellsboth alleles ofTCR~interacted with repressive nuclearcompartmentsat equaland high frequency.561his observationsuggests a limited window ofopportunity to achievethe V to DJ step and that any rearrangement is consequently a very low frequency event. Two simultaneous rearrangements in a cellare thus unlikelyto occur and allelicinclusion is avoided.

RabbitMouseIgH configurations in hybridoma celllines? reflect the frequency of the recombination

events, which render 51% of them VD]ID], 44%VD]/VD]- and 5% VD]lgermline (VD] is theexpressed rearrangement, VD]- isnonfunctional). In contrast,the IgH configurations in rabbitcelllineswere:40%VD]ID], 10%VD]/VD]- and50%VD]/germline,"Since theD to]H stepoccurredon both alleles in 95%ofmouseB-cells, the findingthat it has not done soin 50%of rabbitB-cellssuggests that asynchrony betweenIgH alleles can existto a greaterextent in rabbit. Lanningandcoworkers'? hypothesized that the D to]H rearrangement in rabbit involves slower kinetics and isthe rate-determining step;onceD] isachieved on one chromosome thereis rapidrecombination toVD]. Because ofthe overall inefficiency of the D to]H step,the relatively fewnumbersof cells withVD]/VD]- reflect a restricted timeopportunity for the laggard allele to achieve VD].

Although some infrequent VH to D rearrangement was observed in rabbit splenocytes,"its significance is unknown, sincethe recombined D] is the primary intermediate isolatedfrompro-Bvcells. Cloned fetal rearrangements carried the two D-proximalVH genes, VHI and theneighboringpseudogeneVH2,60 despite>100 available VH upstream; this earlyrearrangementbias together with clonal expansionofB-cells with VHI-expressingVD] causes such H chainsto be 70-90%of expressed Igmolecules," Usage of the D-proximalVH 1in rabbit canbe likenedto the preferential rearrangement of the D-proximalVH genesin fetal mouse liver,62.63 but themolecularbasisof either remainsbe to elucidated.

V(D)JRecombination: O/Miceand Sharks 173

ChickenTheearliestrecombinedcells arein the yolksacat day5and 6ofincubationand carryD] only;

VD] isfound on day9.64Rearrangement occursexclusively to the D-proximalVH gene,the onlyfunctional geneout of multiple VH elements; the other VH act as donor templatesduring thegeneconversion processin the bursa.Theprimary Igrepertoirein chicken,asin rabbits,isgenerated by posrrearrangement gene conversion.65.62 There is a distinct D]/D] step that is B-lineagespecific in chickenand this is followed by simultaneousV rearrangement at the H and L chain10ci.64.66 Clonescarryingonlythe V] or onlythe VD] couldbeobserved,66 showingthat there isnoorderedH and L chain rearrangement, as there exists in mouseand rabble." Thus,in the chicken,L chain rearrangement isnot dependent on the success ofH chain rearrangementand there is nopre-B-cell stageas in mammals.

More than 90% of bursalfollicles contain the VD]/D] configurationand none carriedVD]rearrangements on both chromosomes.Similarly, onlyone alleleof the L chainrecombined. Itwassuggested that the V rearrangementoccursafter removalof repression from one allelerandomlyand that this is an event of such lowefficiency that there is little probabilityof its occurrenceonboth alleles.67

Multiple IgH Loci in Other Vertebrate SpeciesThe contrastingexamples ofmouse,rabbit and chickenshow that the V(D)] recombination

program isadapted fur eachspecies.Thereisat leastone step that is limited by RAG accessibilityand/or time constraints'? and the factorsthat determinetheseparametersremainto be elucidated.Thesethree systems all involve a choiceof two H chain alleles, but when one recombinationsteptends to be limitingor occurringat verylowfrequency, then the presenceof additional allelesone or two,equallysubjectedto the constraints-would not greatly increase the chancesfur allelicinclusion. Model systems genetically manipulated to carry multiple H chain genes (interspecieshybrid tetraploid and triploid Xenopusf and mice triallelicfor IgH69) do exhibit monoallelicH chain expression and thus the samewould be expectedfur those animalswith more thanonenaturally-occurring IgH locus. Polyploid Xenopusspecies carry multipleactiveIgH genes." Bonyfish,aloneofallvertebrateclasses, underwent an additionalgenome-wide duplication" and somespecies support more thanone IgH locusalthough in most only one remains.

Ig Rearrangement in the SharkThe IgH minilocus organization in cartilaginous fishes, representatives of the earliest ver

tebrates, is considered primitive and ancestral to the classical IgH locus in other vertebrates.Sharks, rays and skatescarry 15-200 miniloci ("clusters") each consistingof a fewgenesegments(VH-D 1-D2-]H-CJl)4.23 asshownin Figure2. In mostspecies the rearrangingelementsarelocatedwithin a total spanof2 kb.Theclusters themselves arelocatedfarapart fromeachother," >120kband canbe situatedon differentchromosomes." V(D)] recombinationtakesplaceamongthe fourgenesegments of the mlnilocus: there isno evidencefur interclusterrearrangement in B-cells andhence no need for locuscontraction in such a system. The closeproximityof the genesegments(400 bp apart) also makes unlikelyany separatelyactivatedchromatin domainswithin a cluster.In fact, there isno strict order of rearrangement of the VH, D I, D2 and]H. Once an IgH geneisactivatedin a precursorBvcell, its genesegments recombineall at once and to completion."

In singleBvcell studies,fewIg transcripts" and fewgenomicrearrangements" wereobservedper lymphocyte. Inthe nursesharkthereare9-12functional IgH genesand in anyB-cellthere are1-3VD] genomicrearrangements of which only one appeared to encode a viable receptor. Lessthan 10%of the cells carriedanypartiallyrearrangedgenesand the rest of the IgH geneswereingermline configuration.Thissuggests that onceinitiated,recombination occursefficientlybetweenthefourgenesegments.These datashowthat H chain exclusion exists in theshark,despiteitsuniqueIgH organization. As in higher vertebrates, H chainexclusion in sharksisbasedon limitation ofrearrangement, but the mechanismofrepression (or activation)must accommodate the largeandvariednumbersofIgH loci in differentcartilaginous fishspecies.


The process producing monoallelic Ig H chain expression at the murine IgH locus evolvedwith and is a consequence ofthe complex gene organization. whose multiple gene segments arescattered over 2 Mh. If you take away the locus contraction and the separatdy activated domains ,what shark and mouse have in common is that initiation of rearrangement is an inefficient, lowfrequency event. Whether there are regulatory features in common between shark and tetrapodIgHgene systems remains to be established. However. a few conclusions can be extracted. Becauseof the large number of IgH loci and their dispersed locations, it is unlikely that H chain exclusion in the shark is based on any mechanism that prederermines'Y" rearrangement preference athomologous chromosomes. In nurse shark at least two IgH genes are adjacent" and the moddfor kappa L chain exclusion based on rearrangement preference evinced by the earlier replicatingchromosome will not distinguish multiple, linked genes. It is not clear whether the 1-3 rearrangements in a B-cell occurred simultaneously or sequentially and we suggest that their activation wasprobably stochastic . Ifit happened that one rearrangement at a shark IgH gene is nonfunctionalit seems unlikdy that its allde is more apt to be the one next (or simultaneously) targeted forrecombination than an adjacent or any other IgH in the genome.

Rearrangement ofIg Genes in Non-BrcellsThere exist pre-rearranged Ig genes in the germline of cartilaginous fishes. catfish and

chicken .78•80 In sharks. skates and rayssome IgH clusters carry partially or fully recombined VD-]or VD] and the IgL clusters joined V] .4 Examination ofnurse shark L chain junctions in somegermline-joined V] showed P region sequence that may indicate a one-time hairpin formation.This evidence and the fact that the "12/23 rule" is always obeyed, suggest that there was RAGactivity in germ cells of some animals. 81•19 It was hypothesized that RAG-mediated changesin germline Ig genes produced the VD templates used in chicken H chain gene conversion orperhaps generated D elements during antigen receptor gene evolution.

The function ofrecombined genes in the shark antibody repertoire is not known; it appearsthat many are pseudogenes. In a species with many pre-rearranged VD] there would be a stronglikelihood for allelic inclusion ifmore than one IgH is activated at a time, but at the momentthe germline genes in these animals have not been fully characterized. Nurse shark is an instancewhere all its IgM clusters have been characterized and none are pre-rearranged, showing thatgermline-joined genes are particular to the species."

Once initiated, somatic rearrangement in B-cellsleads to VDJ. Partially rearranged IgH on theother hand have been observed in abundance in nurse shark thymoeytes and 3-7 can be isolatedper celF4Thymic H chain transcripts could not be detected, implying that availability ofDNAto RAG does not require transcription. That many thymocyte rearrangements are incomplete asVD-D-], V-DD], etc., suggests that transcription may be part of the process that recruies" RAGto its target for efficient recombination. This IgH rearrangement-permissive state in thymocytesmay have characteristics in common with that in germ cells enabling RAG, when present, toeffect recombination. However, the state of the IgH chromatin in either cell type has yet to becharacterized.

AboutL ChainIn the course ofevolution, whole-locus duplications produced the multiple cluster organi

zation of cartilaginous fish IgH and IgL, whereas successive tandem duplications of the genesegments V, (D) and] generated the "translocon" organization that exists in tetrapods. WhileH chain genes are organized either as translocon or multiple clusters. the evolution ofL chaingenes" is more complex.

The number of L chain isotypes varies among vertebrates. In chicken there is only the onelocus, Igt.; in mammals there are two, IgA. and 19K. In Xenopus there are three : Igo (sigma) and thehomologs of'Igx (Igp, called rho) and Igi-. (called Type III). In shark there are four: cartilaginousfish-specific"Igo-cart" (called Type IINSS) and the homologs ofsigma.Jgx (called Type III/NS4)and 19i-. (called Type II/NS3}.1gK is thus present in all animals except birds and its organizationvaries considerably. In tetrapods IgK is one locus. In nurse shark the IgK homolog exists as >60

V(D)]Recombination: O/MiceandSharks 175

rniniloci, with one V,oneJ and one C exonand tend to be separatedbysomedistance.Howeverin a bony fishlikezebrafish, IgK genes{Type 1/3)84 arearrangedclosely in serialarrays (examplesin Fig. 4) and on at leastfour differentchromosomes.'"

It isnot clearhow L chain expression is regulatedin zebrafish (or anybonyfish). In cod it wasshownthat multipleenhancers existed in theserial dustersbut not everyIgL C regionwasassociatedwith downstreamenhanceractivity.86 It cannot be anticipatedfrom meredistancehow regulatorycontrol is exercised. Because there can be additional possibilities for intracluster rearrangementfollowing an initial V toJattempt (Fig. 4), wehavesuggested that the bony fishorganizationallowsfor correctionnot onlyof nonproductiveVJ but also in-frameVJ that contribute to forminga self-reactive specificity.84 In other words,there existsa potential for receptorediting" in fishes,sincethe organizationalset-upappearsto allowfor secondaryrearrangements.

In zebrafish, the IgH organization is translocon like teerapods'" so that both types of arrangement exist for its Iggenes. It is clear that H and L chain gene organizations do not haveto co-evolve-as they did not in bony fish89-and information from this and the other modelsystemssuggestthey can be regulated independently.L chain exclusionis not asstringent as H

Zebrafish Light Chain Type 1 19kb

Figure 4. Organizationof representative genes encodingzebrafish Lchains. Some L chainType1 clusters on chromosome 24 are represented on top line; the names of segments aresomeofthose identified in reference 84; their updated linkage, polarity and distances were obtainedfrom theZv7zebrafish genomeassembly (www.ensembl.org) andreference 85.V (yellowboxes)and J (blue)genesegments are flanked by RSS (white triangle is RSS with 12 bp spacer, blacktriangle is RSS with 23 bp spacer) and C exons aredepicted by black boxes. Thetranscriptionalpolaritiesareindicatedby overhead arrows. A hypotheticalseries of rearrangements isdepicted.Inversion recombination 1: rearrangement betweenllc andV1 i immediatelyupstream to form VJ(indicatedasfusedrectangles) and blunt-end joined RSS (fused triangles). Deletion recombination 2: The RSS-23 of the fusedsignal joint recombines with downstreamV genesegment anddeletes interveningDNA. Inversion recombination 3: the remainingJ rearranges to upstream V,forming again VJ and blunt-end joined RSS. This VJ can be excised by deletion recombination4 and replaced by rearrangement at anothercluster, inversionrecombination 5. A color versionof this imageis available at www.landesbioscience.com/curie


chain, but the mechanismfor restricting their expression in zebrafish must managea largearrayof clusters, manyofwhich carrymultiple recombinogenicelementson either sideofthe Cexon. How V(D)J recombination issorted out in zebrafishwill elucidatethose aspectsofRAGaccessibility that evolve with individual species' immune systemrequirements.

ConclusionV(D)J rearrangement wasestablished in an ancestral jawedvertebrateabout SOO millionyears

ago.Fromsharksto mammals two features areevolutionarily conserved-the mechanism ofRAGrecombinase actionandaprocess forlimitingrearrangement activity inordertoproducemonospecificlymphocytes. Theregulatedaccessibility ofantigenreceptorgenes to RAGwas characterizedin precursorlymphocytes of miceand human beings, whereit is usually described asan ordered.two-stage program.However. acomparison ofIg rearrangement patternsfromrabbit, chicken andshark showsthat this process neither has to be strictlyordered nor must takeplacein two stagesto generatea diverse repertoireand bringabout allelic exclusion.

AcknowledgementsI wishto thankLouisDu PasquierandMartin Flajnikforreadingthe manuscript andKarolina

Malecekfor her help with the figures. This work was supported by grants from the NationalInstitutesofHealth and the NationalScience Foundation.

References1. Tonegawa S. Somatic generation of antibody diversity. Nature 1983; 302:575-581.2. Schatz DG. Oettinger MA. Baltimore D. The Y{D)] recombination activating gene. RAG-I. Cell 1989;

59:1035-1048.3. Oettinger MA. Schatz DG. Gorka C et at RAG-l and RAG-2. adjacent genes that synergistically

activate Y{D)] recombination. Science 1990; 248:1517-1523.4. Rast ]P. Litman Gw. Towards understanding the evolutionary origins and early diversification of rear

ranging antigen receptors. Immunol Rev 1998; 166:79-86.5. Flajnik MF. Du Pasquier 1. Evolution of innate and adaptive immunity : can we draw a line? Trends

Immunol 2004; 25:640-644.6. Spring]. Genome duplicat ion strikes back. Nature Genetics 2002; 31:128-129.7. Furlong RF. Holland PWH. Were vertebrates ocroplotd r Philos Trans R Soc Lond B Bioi Sci 2002;

357:531-544.8. Holland PWH. Garcia-Fernandez], Williams NA er al. Gene duplications and the origins ofvertebrate

development . Development Supp11994; 125-133.9. Escriva H. Manzon L. Youson] ec al. Analysis of lamprey and hagfish genes reveals a complex history

of gene duplications during early vertebrate evolution. Mol bioi Evol 2002; 19:1440-14506.10. Kasahara M. Suzuki T. Du Pasquier 1. On the origins of the adaptive immune system: novel insights

from invertebrates and cold-blooded vertebrates. Trends lmmunol2004; 25:105-111.11. Sakano H. Hupp i K. Heinrich G et aI. Sequences at the somatic recombination sites of immunoglobulin

light-chain genes. Nature 1979; 280:288-294.12. van Gent DC. Mizuuchi K. Gellert M. Similarities between initiation of Y{D)] recombination and

retroviral integrat ion. Science 1996; 271:1592-1594.13. Kapitonov vv,jurka ], RAG1 core and Y{D)] recombination signal sequences were derived from Transib

transposons . PLoS Bioi 2005; 3:el81.14. Fugmann SD. Messier C. Novack LA er al. An ancient evolutionary origin of the Ragl12 gene locus.

Proc Narl Acad Sci USA 2006; 103:3728-3733.15. Agrawal A. Eastman QM . Schatz DG. Transposition mediated by RAG I and RAG2 and its implications

for the evolution of the immune system. Nature 1998; 394:744-751.16. Hiom K. Melek, M. Gellert M. DNA transposition by the RAGI and RAG2 proteins: a possible source

of oncogenic translocations . Cell 1998; 94: 463-470.17. Thompson CB. New insights into Y{D)] recombination and its role in the evolution of the immune

system. Immunity 1995; 3:531-539 .18. Fugmann SD. Lee AI. Shockett PE et al. The RAG proteins and Y{D)] recombination : complexes, ends

and transposition. Annu Rev lmmunol20oo; 18:495-527.19. Lewis SM. Wu GE. The old and the restless, ] Exp Med 2000; 191:1631-1636.20. Greenberg AS. Avila D. Hughes M et al. A new antigen receptor gene family that undergoes rearrange

ment and extensive somatic diversification in sharks. Nature 1995; 374:168-173.

V(D)jRecombination: O/MiceandSharks 177

21. Criscitidlo MF. Saltis M. Flajnik ME An evolutionarily mobile antigen receptor variable region gene:doubly rearranging NAR-TcR genes in sharks. Proc Natl Acad Sci USA 2006; 103:5036-5041.

22. Parra ZE. Baker ML. Schwarz RS ec al. A unique T-cell receptor discovered in marsupials. Proc NarlAcad Sci USA 2007; 104:9776-9781.

23. Flajnik ME Comparative analyses of immunoglobulin genes: surprises and portents. Nat Rev Immunol2002; 2:688-698.

24. Litman GW; Anderson MK. Rast JP. Evolution of antigen binding receptors. Annu Rev Immuno11999;17:109-147.

25. Rast JP. Anderson MK. Strong SJ er al. a. ~. y and b T-cell antigen receptor genes arose early in vertebrate phylogeny. Immunity 1997; 6:1-11.

26. Wilson MR. Zhou H. Bengcen E et al. T-cell receptors in channel catfish: structure and expression ofTCR alpha and beta genes. Mol Immuno11998; 35:545-557.

27. Haire RN, Kitzan Haindfield MK et al. Structure and diversity ofT-lymphocyte antigen receptors alphaand gamma in Xenopus. Immunogenetics 2002; 54:431-438.

28. Andre S, Kerfourn F. Affaticati P et al. Highly restricted diversity ofTCR delta chains of the amphibianMexican axolotl (Ambystoma mexicanum) in peripheral tissues. Eur J Immunol2oo7; 37:1621-1633.

29. Kubota T. WangJ, Gobel TW et al. Characterization of an avian (Gallus gallus domesticus) TCR alphadelta gene locus. J Immuno11999; 163:3858-3866.

30. Parra ZE, Baker ML. Hathaway J et al. Comparative genomic analysis and evolution of the T-cell receptor loci in the opossum Monodelphis domestica. BMC Genomics 2008; 9:111.

31. Parra ZE, Arnold T. Nowak MA et al. TCR gamma chain diversity in the spleen of the duckbill platypus(Ornithorhynchus anarinus). Dev Comp ImmunoI2006; 30:699-710.

32. Krangel MS. Gene segment selection in V(D)J recombination: accessibility and beyond. Nat Immunol2003; 4:624-630.

33. Gellert M. V(D)J recombination: RAG proteins. repair factors and regulation. Annu Rev Biochemistry2002; 71:101-132.

34. Komori T. Okada A. Stewart V et al. Lack of N regions in antigen receptor variable region genes ofTdT-deficient lymphocytes. Science 1993; 261:1171-1175 .

35. Gilfillan S. Dicrich A, Lemeur M et al. Mice lacking TdT: mature animals with an immature lymphocyterepertoire. Science 1993; 261:1175-1178.

36. Bertocci B. De Smet A. WeillJC et al. Nonoverlapping functions of DNA polymerases rnu, lambda andterminal deoxynucleotidyltransferase during immunoglobulin V(D)J recombinat ion in vivo. Immunity2006; 25:31-41.

37. Dominguez 0 , Ruiz JF. Lain de Lera T et al. DNA polymerase mu (Pol mu), homologous to TdT,could act as a DNA mutator in eukaryotic cells, EMBO J 2000; 19:1731-1742.

38. Beerz S, Diekhoff D. Steiner LA. Characterization of terminal deoxynucleotidyl transferase and polymerase mu in zebrafish. Immunogenetics 2007; 59:735-744.

39. Zhang SM. Adema CM , Kepler TB et al. Diversification of Ig superfamily genes in an invertebrate.Science 2004 ; 305:251-254.

40. Wu IT. Johnson G. Kabat EA. Length distribution of CDRH3 in antibodie s. Proteins 1993; 16:1-7.41. Fleurant M. Changchien L. Chen CT et al. Shark Ig light chain junctions are as diverse as in heavy

chains. J Immunol 2004; 173:5574-5582.42. Wedemayer GJ. Patten PA. Wang LH et al. Structural insights into the evolution of an antibody com

bining site. Science 1997; 276:1665-1669.43. Wilson lAoStanfield RL. Antibody-antigen interactions: new structures and new conformational changes.

Curr Opin Struct BioI 1994; 4:857-867.44. Stanfield RL. Dooley H. Verdino P et al. Maturation of shark single-domain (lgNAR) antibodies:

evidence for induced-fit binding . J Mol Bioi 2007; 367:358-372.45. Nguyen VK, Hamers R, Wyns L et aI. Camel heavy-chain antibodies : diverse germline V(H)H and

specific mechanisms enlarge the antigen-binding repertoire . EMBO J 2000; 19:921-930.46. Achour I. Cavdier P. T ichit M et al. Tetrameric and homodimeric camelid IgGs originate from the

same IgH locus. J Immuno12008; 181:2001-2009 .47. Burnet FM. The clonal selection theory of acquired immunity. Cambridge : Cambridge University Press.

1959.48. Stanhope-Baker P, Hudson KM. Shaffer AL et al. Cell type-specific chromatin structure determines the

targeting ofV(D)J recombinase activity in vitro. Cell 1996; 85:887-897.49. Yancopoulos GO. Ale FW; Regulation of the assembly and expression of variable-region genes. Annu

Rev Immuno11986; 4:339-368.50. Alt FW; Yancopoulos GO , BlackwellTK et al. Ordered rearrangement of immunoglobulin heavy chain

variable region segments. EMBO J 1984; 3:1209-1219.


51. Chowdhury D. Sen R. Regulation of immunoglobulin heavy-chain gene rearrangements. Immunol Rev2004; 200:182-196.

52. Fuxa M. Skok J. Souabni A et aI. Pax5 induces V-to-DJ rearrangements and locus contraction of theimmunoglobulin heavy-chain gene. Genes Dcv 2004; 18:411-422.

53. Kosak ST. SkokJA, Medina KL et al. Subnuclear compartmentalization of immunoglobulin loci duringlymphocyte development. Science 2002; 296:158-162.

54. Jhunjhunwala S. van Zelm MC. Peak MM et aI. The 3D structure of the immunoglobulin heavy-chainlocus: implications for long-range genomic interactions. Cell 2008; 133:265-279.

55. Roldan E. Fuxa M. Chong W et aI. Locus "decontraction" and centromic recruitment contribute toallelic exclusion of the immunoglobulin heavy-chain gene. Nat Immunol 2005; 6:31-41.

56. Schlimgen RJ. Reddy KL. Singh H ee aI. Initiation of allelic exclusion by stochastic interaction of Tcrballeles with repressive nuclear compartments. Nat Immunol 2008; 9:802-809.

57. Perlot T. Alr FW; Bassing CH er aI. Elucidation of IgH intronic enhancer functions via germ-line deletion . Proc Nat! Acad Sci USA 2005; 102:14362-4367.

58. Tunyaplin C. Knight KL. IgH gene rearrangements on the unexpressed allele in rabbit B-cells. J Immuno11997; 158:4805-4811.

59. Lanning D. Jasper P, Knight K. IgH haplotype exclusion in rabbits. Semin Immunol 2002;14:163-168.

60. Tunyaplin C. Knight KL. Fetal VOJ gene repertoire in rabbit: evidence for preferential rearrangementofYH1. Eur J lmmuno11995; 25:2583-2587.

61. Mage RG. Lanning D. Knight KL. Bvcell and antibody reperto ire development in rabbits: the requirement of gut-associated lymphoid tissues. Dev Comp Immunol 2006; 30:137-153.

62. Yancopoulos GD. Desiderio Sv, Paskind M et aI. Preferential utilization of the most JH-proximal YHgene segments in prell-cell lincs. Nature 1984; 311:727-733.

63. Knight KL. Becker RS. Molecular basis of the alIelicinheritance of rabbit immunoglobulin YH alIorypes:implications for the generation of antibody diversity. Cell 1990; 60:963-970.

64. Reynaud CA. Imhof BA. Anquez V er aI. Emergence of committed B-Iymphoid progenitors in thedeveloping chicken embryo. EMBO J 1992; 11:4349-4358 .

65. WeillJC. Reynaud CA. The chicken B-celI compartment. Science 1987; 238:1094-1098.66. Benatar T. Tkalec L. Ratcliffe MJ. Stochastic rearrangement of immunoglobulin variable-region genes

in chicken B-celI development. Proc Nat! Acad Sci USA 1992; 89:7615-7619.67. Weill JC. Cocea L. Reynaud CA. Allelic exclusion: lesson from GALT species. Semin Immunol 2002;

14:213-215.68. Du Pasquier L. Hsu E. Immunoglobulin expression in diploid and polyploid interspecies hybrid of

Xenopus : evidence for allelic exclusion. Eur J Immunol 1983; 13:585-590.69. Barreto V. Meo T, Cumano A. Mice triaUclic for the Ig heavy chain locus: implications for YHDJH

recombination. J Immunol 2001; 166:5638-5645 .70. Du Pasquier L. Blomberg B. The expression of antibody diversity in natural and laboratory-made poly

ploid individuals of the clawed toad Xenopus. Immunogenerics 1982; 15:251-60.71. Hoegg S. Brinkmann H. Taylor JS er al. Phylogenetic timing of the fish-specific genome duplicat ion

correlates with the diversification of teleost fish. J Mol Evol 2004; 59:190-203.72. Lee V. Huang jl., Lui MF er aI. The evolution of multiple isotypic IgM heavy chain genes in the shark.

J Immunol 2008; 180:7461-7470.73. Anderson M. Amemiya C. Luer C ec aI. Complete genomic sequence and patterns of transcription of

a member of an unusual family of closely related. chromosomalIy dispersed Ig gene clusters in Raja. InrImmuno11994; 6:1661-1670.

74. Malecek K. Lee V, Feng W er aI. Immunoglobulin heavy chain exclusion in the shark. PLoS Bioi 2008;6:cl57.

75. Eason DD. Litman RT. Luer CA et al. Expression of individual immunoglobulin genes occurs in anunusual system consisting of multiple independent loci. Eur J Immunol2004; 34:2551-2558.

76. MostoslavskyR. Singh N. Kirillov A et aI. Kappa chain monoalIelic demethylation and the establishmentof allelic exclusion. Genes Dev 1998; 12:1801-1811.

77. Mostoslavsky R. Singh N. Tenzen T et aI. Asynchronous replication and allelic exclusion in the immunesystem. Nature 2001; 414:221-225.

78. Kokubo F, Litman R. Shamblott MJ et al. Diverse organization of immunoglobulin YH gene loci in aprimitive vertebrate. EMBO J 1988; 7:3413-3422.

79. Ghaffari SH. Lobb CJ. Structure and genomic organization of a second cluster of immunoglobulinheavy chain gene segments in the channel catfish. J Immuno11999; 162:1519-1529.

80. Reynaud CA. Dahan A. Anquez V et al. Somatic hyperconversion diversifies the single YH gene of thechicken with a high incidence on the D region. Cell 1989; 40: 283-291.

V(D)JRecombinatio«: O/Miceand Sharks 179

81. Lee 55. Fitch D. Flajnik MF er al. Rearrangement of immunoglobulin genes in shark germ cells. J ExpMed 2000; 191:1637-1648 .

82. de Villartay J-P. Passera ou ne passera pas-accessibility is key. Nature ImmunoI2006; 7:1019-1021.83. Criscitiello MF, Flajnik ME Four primordial immunoglobulin light chain isorypes, including lambda and

kappa. identified in the most primitive living jawed vertebrates . Eur J Immunol 2007; 37:2683-2694.84. Hsu E, Criscitiello ME Diverse immunoglobulin light chain organizations in fish retain potential to

revise B-cell receptor specificities. J Immunol 2006; 177:2452-2462.85. Zimmerman AM, Yeo G. Howe K et al. Immunoglobulin light chain (IgL) genes in zebrafish: Genomic

configurations and inversional rearrangements between (VL-JL-CL) gene clusters. Dev Comp Immunol2008; 32: 421-434 .

86. Bengren E, Stromberg S, Daggfeldt A et al. Transcriptional enhancers of immunoglobulin light chaingenes in Atlantic cod (Gadus morhua) . Immunogenetics 2000 ; 51:647-658.

87. Nemazee D. Receptor editing in lymphocyte development and central tolerance. Nat Rev Immunol2006; 6:728-740.

88. Danilova N, Bussmann J. jekosch K et al. The immunoglobulin heavy-chain locus in zebrafish: identification and expression of a previously unknown isotype, immunoglobulin Z. Nat Immunol 2005;6:295-302.

89. Daggfeldt A, Bengren E, Pilsrrom L A cluster type organization of the loci of the immunoglobulin lightchain in Atlantic cod (Gadus morhua L) and rainbow trout (Oncorhynchus mykissWalbaum) indicatedby nucleotide sequences of cDNAs and hybridization analysis. Immunogenetics 1993; 38:199-209.

CHAPTER 15

Normal and Pathological V(D)JRecombination:Contribution to the Understanding ofHumanLymphoid MalignanciesSaidaDadi, SandrineLe Noir, VahidAsna6., Kheira Beldjordand ElizabethA. Macintyre*

Abstract

T he majority ofhaematological cancers involve the lymphoid system. They include acutelymphoblastic leukemias (ALL), which are arrested at variable stagesofdevelopment andpresent with blood and bone marrow involvement and chronic leukemias , lymphomas

and myelomas, which present with infiltration ofa large variety ofhematopoietic and non hematopoietic tissuesby mature lymphoid cellswhich expressasurface antigen receptor.The majorityinvolve the B-celliineage and the vast majority have undergone clonal rearrangement of theirIg and/or TCR rearrangements. Analysis ofIglTCR genomic V(D)J repertoires by PCR basedlymphoid clonaliry analysis within a diagnostic setting allows distinction ofclonal from reactivelymphoproliferative disorde rs, clonal tracking for evidence oftumor dissemination and follow-up,identification ofa lymphoid origin in undiagnosed tumors and evaluation ofclonal evolution. Ig/TCRVDJerrors are also at the origin ofrecombinase mediated deregulated expression ofavarietyofproto-oncogenes in ALL , whereas in lymphoma it is increasingly clear that IgH containingtranslocations result from abnormalities other than VD] errors (somatic hypermutation and/orisotype switching). In addition to th is mechanistic contribution to lymphoid oncogenesis, it ispossible that failure to successfully complete expression ofan appropriate Ig or TCR may leadto maturation arrest in a lymphoid precursor, which may in itself contribute to altered tissuehomeostasis, particularly if the arrest occurs at a stage ofcellular expansion.

IntroductionApproximately 5% ofhuman cancers overall and over 70% ofhaematological cancers involve

the lymphoid system, with the majority involving the B-celllineage. Lymphoid cancers includeimmature, "blastic" lymphoid proliferations which involve essentially the blood and/or bonemarrow (Acute Lymphoblastic Leukemia or ALL), mature lympho-proliferations involvingpredominantly secondary lymphoid organs (non-Hodgkin's Lymphomas or NHL) or blood andbone marrow (chronic lymphocytic leukemias or CLL) and expansions ofplasmocytes, with predominant bone marrow and tissue involvement (multiple myeloma or MM). Dysirnrnune statessuch as Hodgkin's disease or Angioirnrnunoblastic lymphadenopathy (AlLD), at the interface

·Corresponding Author: Elizabeth A. Maclntyre-Hopital Necker Enfants Malades, Laboratoired'hematologie, bat. Pasteur, 149 rue de Sevres 75015 Paris, France.Email: e lizabeth [email protected]

V(D)] Recombination, edited by Pierre Ferrier. ©2009 Landes Bioscienceand Springer Science+Business Media.

Normaland Pathological V(D)JRecombination 181

between reactive immune disorders and lymphoid malignancies, also exist. Whether these areclassified as lymphoid malignancies or not often depend on the techniques available for theircharacterization, notably analysis of the V(D)J status of their immunoglobulin (Ig) and T-cellReceptor (TCR) loci by techniqueswhich will be collectively referredto as lymphoid clonalityanalysis here. Lymphoid malignancies are also frequendy characterized by V(D)J recombinaseerrors which lead to transcriptional deregulationof lymphoid "oncogenes" by junapositioningto, most commonly, Igor TCR regulatorysequences. This represents a lymphoidspecific formof"physiological genetic instability"which includesV(D)J recombinase errors and abnormalitiesofisotype switchingand/or somaticmutation. Only the formerwill be consideredhere; they arecollectively, if imprecisely, referred to as V(D)J translocations. Sucherrors can be consideredtobe, at a minimum, mechanisticelementsinvolved in lymphoid oncogenesis. It is, however, possiblechatfailureto successfully completefabricationand expression of an appropriateIgor TCRmayin itselfrepresentan oncogenicevent within the multistageprocesschat is now recognizedto preceedclinicalpresentationof the majorityof human cancers.

Since lymphoid cancers represent homogeneouspopulations arrested at different stages ofdevelopment, theyprovideinvaluable modelsfor the studyofmolecularand cellulareventsleadingto interruption oflymphoiddevelopment. Within thiscontext,"readingthelanguage" ofIg/TCRrearrangements can provideusefulinformation regardingthe type oflymphoid (sub)populationinvolved, the stageof maturation arrest and the chromatin accessibility of the different Ig/TCRloci. It shouldhoweverbeemphasised that thephenol genotypeof thebulk lymphoidcancerisnotnecessarily synonymous, but mostprobablydownstream,to the lymphoidcancerstemcell.Sinceany detectableclonalV(D)J rearrangement or translocationsuggests at a minimum that the Ig/TCR lociwereaccessible duringpreceedingstages oflymphoid oncogenesis, suchrearrangementsrepresentusefulfingerprints of upstream oncogenicevents.We haveundertaken to review thesedifferent, but interlinked,applicationsof the analysis of normaland abnormalV(D)J codingjointrepertoires appliedto understandingoflymphoidmalignancies and theirdysimmune close relatives.Suchan approach isby definition nonexhaustive and we apologies to allindividualcontributorswhichwehaveonlyreferenced indirectly, in the interestsofbrevity.Wewill not discuss therapeuticaspectsofV(D)J manipulation, nor analysis oftranscribed,functionalV(D)J repertoires and willonly briefly touch on detection of signaljunction rearrangements.

Diagnostic Clonality AnalysisMolecular analysis of Ig/TCR genomic repertoires in diagnosticevaluation of (suspected)

human lymphoidmalignancies wasinitiallyperformedby Southern blot analysis,"?but wasprogressively replacedfrom the 1980sonwardsbyPCR analysis from DNA.4-9 Both arebasedon theprincipal chat reactive lymphoproliferations are associated with polyclonalIgITCR repertoireswhereas the majorityof lymphoidcancersdemonstrate clonal, homogeneousrearrangements ofIgand/or TCR 10ci,IO with the pattern of clonal rearrangements reflecring the lymphoid lineageinvolved and its stageof maturation arrest.lO•n

TechnicalandPracticalAspectsSouthern blotting predominantly reflected homogeneous V and J segment usagewhereas

PCR V(D)J amplification also exploitsheterogeneityofVDJ junctional sequences at the thirdcomplementarity determiningregion(CDR3).Thelongerthe CD R3,the easierthe distinctionofclonalandpolyclonal rearrangements.I4 Detection ofVDJ, DJ,VD DD and DJ rearrangements arepossible ifappropriateprimersareused.ISThemajorityofdiagnosticsystems useconsensus primersdirected to relatively conservedframework regions, ofien in a multiplexformat.IS Predictably, therisk of false negative results is dependent on the complexity of the repertoire (Table1) and thedegreeof homologybetween the V,D and] primersand their target sequences. Theother mainfactorcontributingto false negativityissomaticmutation involving PCR primer targetsequencesbut others include: presenceof inhibitors; analysis of uninvolvedtissueand DNA degradationof fixed tissues.


Table 1. Human IgflCR repertoires, combinatorial complexity and chromosomallocalisation

Number of GermlineEncoded Segments

v oApproximate COR3Length(bp)

NumberofNRegions

ChromosomalLocalization

IgH 46-52 27 6 50 1-2 14q32.3

IgK 31-36 0 5 10 1 2pll .2

Igi.. 30-33 0 4 10 1 22qll .2

TCRII 7 3 4 5-50 1-4 14q11.2

TCRa 45-47 0 50 10 14q11.2

TCRy 9 0 5 10 7q14

TCR~ 39-47 2 13 10 1-2 7q34

ThenumberofVsegments varies. Certain Va/II segments can rearrange to bothTCRII and TCRa loci.Number of N region varies with incomplete VD, DDor DJ rearrangements.

Distinctionofclonal. oligoclonal and polyclonal PCR productsisbasedon eithernondenaruringpolyacrylamide gelelectrophoresis (PAGE). usuallyunderconditionsencouragingheteroduplexformation.or "genescan" sizing of fluorescent PCR products.The fOrmer has the advantage of optimisingdistinctionof clonalhomoduplexes from polyclonal heteroduplexes but requires optimalPAGEconditions.Genescan sizingallows precise informationregarding clonalproduct size. usefulformolecular follow-up andcomparison ofdifferent samples fromagiven tumorandcanallow identification ofV and] segment usage ifdifferently labelled primersareused(Fig. 1).Underqualitativeconditions.both havean approximate sensitivity of 1-5%. although this dependson the positionof clonaland polyclonal populations. sincea clonalpopulationwhichissituatedat the peak of theGaussian distributionof polyclonal PCR productswill be detectedwith lowersensitivity than onewhichiseitherlargeror smaller than thesefragments (Fig. 1).Quantitationofclonalrearrangementsby real-time PCR ispossible usingCRD3 specific probes.or moreusually primers(Fig. 2) (ref 16and references therein) Thisrequires sequencing ofdiagnostic material and has beendeveloped essentially for follow-up ofpatientswithALL.In general. diagnostic strategies aim onlyto distinguishclonalfrom polyclonal populationsand do not attempt to identifysegmentusage. Judicious useofappropriately situated.variably labelled fluorescent primersallows identification ofY, D and] segments from a limited number of multiplex PCR. basedon PCR product sizeand fluorescenceY"Readingthe language" ofIgITCR rearrangements in thiswaycan contributeto identification ofthe stageof maturationarrestand lineage affiliation. Suchanalyses do not allowdetermination offunctional. in-frame rearrangement. unless combinedwith sequence analysis.

Diagnostic PCRhavebeendevelopedforalllociother than TCRn.The mostwidely usedlocifordiagnostic clonality analysis areIgHVD] and TCRyV].sinceboth rearrange relatively earlyduringnormalBandT-lymphoiddevelopment respectively. includinginallsubsets ofeachlineage.Backuploci for the B-celllineage include IgK and IgHD] rearrangements. whereas Igt.. clonality analysiswithin adiagnostic settingiscomplex and rarely addsadditionalinformation. Forsuspected T-cellmalignancies.TCRycanbecomplementedbyTCRj3 VDJanalysis.whichisamoreappropriate targetthan TCRb; due to the deletionof thislocusduringTCRn rearrangement and the consequent riskofpseudo-clonaliryfrorn rareresidualTCRI)rearrangements. UseofTCRI)isessentially restricted toclonaliryanalysis inALLandraresuspectedTCRyl) lymphoproliferativedisorders.Detailsregardingthe incidence andpatternsofIglTCR rearrangements in the main categories oflymphoproliferativedisorders (LPD) canbefoundinTable2.15,18.24 Succinctly. matureBlineage LPD rearrange IgHand

~ l ~ "'-~ So ~ s <>:

i. ;: .... ~ ,2 <;;;:;~ ~ ~ ..... ~

.

~.

FR

t

no!

c

1'4

0.e

o,e

o-.-

--

.-_

211'10

IAIX

tYdc

rItIIB

aii3

lgeJ

i;.

.'0

140

Z10

----

-20

Vt1

JQU

AFR

2.I

lGF

Rt

JHl.2

.4$

JH3

Fig

ure

1.ly

mp

ho

idcl

on

alit

yan

alys

is.G

enes

can

and

hete

rodu

plex

imag

esof

TC

Ran

dIg

Hg

en

om

icre

pert

oire

s.1)

-Ana

lysi

so

fmu

ltip

lex

TC

Ryr

earr

ange

men

tby

PC

Rfr

omD

NA

usin

gflu

ores

cent

prim

ers

and

gene

scan

(GS

)eva

luat

ion

(left)

com

pa

red

ton

on

de

na

turi

ng

PA

GE

anal

ysis

ofh

ete

rod

up

lex

(HD

)P

CR

pro

du

cts

(rig

ht):

A=

po

lycl

on

al

rear

rang

emen

ts.

Ba

nd

C-

Clo

nal

bi-

alle

licV

fl-Jy

ll2re

arra

ngem

ents

ina

B-c

ell

pre

curs

orA

ll,

anal

ysed

byG

S(B

)or

HD

(C).

Slo

wm

igra

ting

clon

alhe

tero

dupl

exes

are

seen

whe

nbo

thal

lele

su

nd

erg

ore

arra

ngem

ent

wit

hth

esa

me

Vy

and

Jyse

gmen

ts,

bu

tdiff

ere

ntC

DR

3.H

om

od

up

lexe

sar

ein

dic

ate

dby

arro

ws.

The

use

ofd

iffe

ren

tlyla

belle

dpr

imer

sa

llow

sid

en

tifi

cati

on

ofth

eV

and

Jse

gmen

tsus

edon

the

basi

sof

PC

Rp

rod

uct

size

and

colo

ur,

assh

own.

Siz

em

arke

rsar

ein

red.

D=

po

lycl

on

alT

CR

Vy9

-Jg

rear

rang

emen

t(g

reen

)and

can

on

ica

lVy9

JP(b

lack

).T

hela

tte

rre

pres

ents

ase

lect

ed,

fun

ctio

na

lre

pert

oire

,as

evi

de

nce

dby

the

3bp

spac

ing

betw

een

peak

s,as

seen

wit

hIg

H,

but

not

wit

ho

the

rT

CR

yre

arra

ngem

ents

.2)

-A

naly

sis

ofm

ult

iple

xm

ult

iflu

o

resc

ent

IgH

rear

rang

emen

tby

PC

Rfr

omD

NA

usin

gflu

ores

cent

prim

ers

and

gene

scan

eval

uatio

n.P

olyc

lona

l(A

)an

dcl

on

al

(Ban

dC

)re

arra

ngem

ents

are

show

n.T

heus

eof

3d

iffe

ren

tlyla

belle

d[H

prim

ers

(JH

1/2,

4,5

inbl

ue,

JH6

ingr

een

and

jH3

inbl

ack)

give

sim

pro

ved

reso

lutio

nin

po

lycl

on

al

rear

rang

emen

ts,

the

ma

jori

tyof

wh

ich

use

JH4

-6se

gmen

ts.

IgH

rear

rang

emen

tsar

ea

mp

lifie

dw

ith

bo

thFR

1an

dF

R2

cons

ensu

sp

rim

ers

inse

para

tere

actio

nsbu

tan

alys

edto

geth

er,

inor

der

toin

crea

secl

onal

info

rma

tivity

inca

ses

havi

ngu

nd

eg

on

eso

ma

ticm

uta

tion

.F

orex

ampl

e,of

the

2B

-cel

lp

recu

rso

rA

lls

sho

wn

,sa

mpl

eB

dem

onst

rate

sa

maj

orJH

6an

da

min

orJ

H3

rear

rang

emen

twh

ich

are

info

rma

tive

wit

hb

oth

FR1

(rig

ht)

and

FR

2(le

ft)V

Hpr

imer

s.F

orsa

mpl

eC

,in

cont

rast

,th

em

ajo

rJH

1,2,

4,5

rear

rang

emen

tis

am

plif

ied

effic

ient

lyfr

omth

eFR

1b

ut

po

orl

yfr

omth

eF

R2

pri

me

ran

dth

em

ino

rjH

6re

arra

ngem

ent

iso

nly

seen

wit

hFR

1.A

colo

rve

rsio

nof

this

imag

eis

avai

labl

eat

ww

w.la

nd

esb

iosc

ien

ce.c

om

/cu

rie

.T

heu

nif

orm

inte

nsity

,re

dpe

aks

coo

resp

on

dto

size

mar

kers

...... ~


JUNCTIONAL REGIONS (CDR3region)ARE USED AS TARGETS FOR MRD·PCR

•J

•v'-- -l_'-- -'

• II~··---.

~II

ASOP'rEr. aIeIe $peciIe~ (bwa'tl f1 retlrSe)

• • ADesarl pOOe (asTBqlm pOOe)

Figure 2. Real time quantitative Ig/TCR CDR3 specific strategies. Quantification by RQ-PCRof the tumor load or minimal residual disease (MRD). For each Ig/TCR rearrangement, thejunctional region is amplified, sequenced and several "clone-specific" primers or allele specific oligonucleotides (ASO) are designed. Specific CDR3 specific primers are then used forclone specific amplification of follow-up material using CDR3 and v, D or J primers and Vor JTaqman probes. Quantification is performed using a standard curve constructed fromthe RQ-PCR assay by serial dilutions of patient's blasts in a peripheral blood mononuclearcell pool (10-1 to 10-5).

IgK (VJ or Kappadeletingelement-KDE) in the vastmajorityof cases. with extensive repertoireswhichleadto littlerisk.offalsepositive results andariskoffalsenegative results whichisproportionalto thedegreeofsomatic mutation.MatureT lineageLPDrearrangeTCRyandTCRf3 andoccasionallyTCRb.The restricted repertoireofTCRy VJ reartangements leadsto ariskoffalsepositive detectionof pseudoclonaliry, particularly ifPAGEconditionsaresuboptimal.IS PAGEheteroduplex analysisispreferable to fluorescent genescan analysis in adiagnostic setting.sincethereisalowerriskoffalsepositives. The presence of canonical "invariant" rearrangements. suchasVy9-]P rearrangements incirculatingTCRyblymphocytes.canalso beerroneously interpretedasindicatingdonalexpansion byinexperienced operatorsand for this reasonnot alldiagnostic multiplex strategies includea]P (alsoreferred to as]y1.2)specific primer.IS Twoclassifications forhumanTCRyVand] segments exist."Thepresence of minor normalclonal/invariant populationsiswellrecognised in circulating CD8+T-Iymphoeytes fromolderindividualsandin reactivedisorders suchaslymphomatoidpapulosis.Theriskoffalsepositiveresults canbeminimisedbysimultaneous useofTCRf3analysis" andrestrictionof theseanalyses to high throughputlaboratories. inorder to maximise experience. Interpretationoflymphoidclonaliryprofiles shouldbe undertakenin close interactionwith theprescribingphysicianor pathologistand with knowledge of the clinical context.

Cross lineagerearrangements. alsoreferredto as"illegitimate rearrangements" (Igrearrangements in aT LPD or viceversa) arerarein matureLPD.Theyarecommonin acutelymphoblasticleukemias. with the majorityofB lineage ALLdemonstratingTCRyrearrangement and/or TCRbor.morerarelyTCRf3 rearrangements. Igrearrangements inT-ALLareless commonand arepreferentiallyfound in the TCRyb lineage.2S•27 illegitimaterearrangements reflectthe fact that ALLsremain recombinase competent and consequently rearrangeall loci in an accessible chromatinconfiguration. Rearrangement patterns differwith oncogenicsubtype.with stageof maturationarrest and with patient age.funongst B-cellprecursorALLs. for example. relatively maturecaseswhich express Igeyt /.I. rarelydemonstrateTCRy rearrangements. whereas the majorityof CD 1O'cyt /.1.- ETV6-RUNXl or BCR-ABLcases do SO.28,29DetailsofIglTCR reartangementprofiles inALLcanbe foundinTable2.Detectionoflymphoiddonaliry israrelyrequiredto makeadiagnosisof ALL and isessentially usedfor molecularfollow-up (seebelow)." Extensive sequenceanalysisof these clonal rearrangements has,however. allowedaccumulationof a largedatabank allowinganalysis of V;D and] segmentusage and CDR3 diversity. whichmayeventually leadto improvedunderstandingofthe pathogenicstages leadingto ALL development.

Normaland Pathological V(D)JRecombination

Table 2. Approximate incidenceof clonallgnCR rearrangement in lymphoidmalignancies, asdetected by PCR from DNA. Only diagnostic PCR targetsarecited

185

IgH IgK Igl.. TCRli TCRy TCRp

B-Cell Proliferation

BCP-All 90 30 20 30 60 30

Cll 100 100 30 10 20 25

Non-Hodgkin's Lymphoma

Fl 90 85 20 5 5 5

MCl 100 100 45 5 10 10

BlBCl 85 80 30 15 15 20

MZl 95 80 30 10 15 20

T-Cell Proliferation

T-All 5 0 0 50 90 90

T-lGl 0 5 5 30 95 95

AllT 30 30 5 35 90 90

Abbreviations: BCP-All: B-cell precursor Acute Lymphoblastic Leukemia; Cl.L: ChronicLymphocytic leukemia; FL:Follicular Lymphoma; MCL: Mantle Cell Lymphoma; BLBCL: DiffuseLarge B-Cell Lymphoma; MZL : Marginal Zone Lymphoma; T-ALL: T-cell acute lymphoblas-tic leukaemia ; T-LGL: T-Large Granular lymphocytic leukaemia; AILT: AngioimmunoblasticT-Cell Lymphoma.

ClinicalApplicationsDiagnosticclonaliry analysis ismainlyusedto distinguishreactive.polyclonalLPD fromclonal.

probablybut not necessarily. malignantLPD. Once a clonalpopulation has been identified. it ispossible to trackthis clonein differenttissuesamples. in order to assess dissemination at diagnosis.or to determine clonal identity at relapse. Clonal trackinghas also been used within a minimalresidual disease setting in ALL and certain NHL. once apparent complete remission has beenobtained. to stratify individualpatient management.basedon the cineticsof response to remission induction at diagnosis. Succinctly. clonaltrackingwith CDR3 specific probes. usedwithin astrictlystandardised. quantitativesetting, allowthe detection of minor clonalpoulations with areproduciblesensitivity ofat least 10-4 (1 malignantcellamongst 10000normal cells). It has alsobeenusedto "back-track"preclinical development ofALL.in conjuctionwith molecular oncogenicmarkers. allowingthe identificationof leukemicclonesmanyyears beforeclinicalpresentation.including in postnatal samples prior to developmentof pediatricALL,30,31

Recombinase Mediated OncogenesisAnalysis of structural chromosomalabnormalitiesby classical morphologicalkaryotypingin

lymphoid malignancies allowed the identification of recurrent translocations involvingthe Igloci in B lymphoid malignancies and TCR loci in T-cellmalignancies. The adventofmoleculartechniquesled to identificationof the IgITCR partner genesand the demonstration that karyotypicanalysis largelyunderestimatedthe incidenceand complexityof these rearrangements. Thelargenumber of partner genesidentifiedhas allowednumerous insightsinto normal and pathologicallymphoid developmentand function. but their verynumber precludes their descriptionhere and readersare invited to consult the followingreviews on the subjeet.32-36 Only generalaspects relevant to V(D)J rearrangementwill be detailed here.Within the context oflymphoid


•+

-fifio~iiiiii:1-- 1::::;1":1

RAG tiil'ili£ii;:BijEBawl

• ~ ?

H--.... I~·r.:~·I!1.~·•••••• ~~ .~~~:~~ : I,tetbtQ:O;-:iiWtii'iiriil;1-- 1 _

•I>--

.................

t RAG!

-iP1om;;&jEBii~ [)

Figure3. Type 1 and 2 Ig/TCR rearrangements. Recognitionsignalsequences (RSS) are representedby trianglesat V, D and Jsegmentextremities. Gray trianglesrepresent cryptic RSS. InTypeI rearrangements, RAGtargetsboth thebona-fideandthecryptic RSS.ln Type2 junctions,the break in the proto-oncogene is targetedby unknown mechanisms. In both cases, there isexcision of intervening DNA, in the signal joint shown at the bottom right hand corner.

malignancies, the term"illegitimate" rearrangement isusually reserved for cross-lineage intralocusrearrangements, such as the TCR rearrangements identifiedin B lineageALL describedabove."Trans-rearrangement" refers to rearrangement betweendistinct Igand TCR loci,abnonnalitieswhich havebeen principallydescribedin patients with AtaxiaTelangiectasia." V(D)J translocations usually impliesstructuralkaryotypicabnormalities involving aproto-oncogeneand an IgorTCRlocus(Fig.3 and Table3).Theincreasing recognitionofrecombinase mediatedderegulationofgeneswith no involvement of an IgITCR locus,including those resultingfrom microscopic,intragenicrearrangements not associated withevidentkarytoypic abnormalities, justifies useof themoregeneralterm "recomblnase mediatedoncogenesis". Comparativegenomichybridizationhasdemonstratedthat in pediatricBlineageALL,manyofthesedeletions involve genes whichregulateB-celldevelopment,includingTCF3 (alsoknown asE2A),EBFI, LEFI, IKZFI (IKAROS) andIKZF3 (AIOLOS).38 At leasta proportion of these aremediated by the recombinase.

Recombinase mediated eventscan occur at the site of any RSS-like sequence which is in anaccessible chromatinconfigurationduringrecombinase activity. One ofthe beststudiedexamplesoutside the lymphoid oncogenesis context is deletionsof the HPRT locus.39-42These havebeenusedasameasureofgenomicinstability,someofwhicharemediatedbythe recombinase complex.Within the present context,onlyV(D)J recombinase mediatedeventswith oncogenicpotentialwill be detailed.The role of recombinase abnormalities in IgiTCR rearrangements is illustratedby their high incidencein patientswith AtaxiaTelangiectasia and similardisorders."

VD]Errors in LymphoidMalignanciesDuring lymphoid development, recombinase activity targeted to recombination signal se

quences(RSS)would ideally be restrictedto legitimatetargetswithin IgITCR loci and allgenescontrolling tissuehomeostasis would be protected from this lymphoid specific fonn of "physiologicalgenomic instability". The existence, however, of a large number of RSS-like sequencesthroughout the genome (10 million or I cryptic RSSevery 1-2 kb on average) meansthat nonspecific targeting of RAGI can induce double stranded breaksoutside IgITCR loci, leadingto

NormalandPathologicalV(D)j Recombination 187

Table 3. Deregulation of lymphoid oncogenesby IglTCR juxtapositioningin ALL

Oncogene (Ig1tcr TranslocationsPartner Genes) Protein Family Group Involved References

B-ALL Translocation InvolvingIg Genes

104 Inhibitor of DNA t(6; 14)(p21;q32) 61

binding(lD)HLH

LHX4 L1M-homeodomain t(l; 14)(q25; q32) 60

BCL9 Not identified t(l ; 14)(q21 ; q32) 59

ILJ 4H Cytokine t(5; 14)(q32; q32) 62

c-Myc bHLH-Zip t(8; 14)(q24; q32) 58,55

t(2;8)(p1 2; q24) 56

t(8; 22)(q24; qll) 57

CEBP bZIP t(1 4; 19)(q32; q13); 63

t(8; 14)(q11 ; q32);

Inv(14)(ql1 ; q32)!

t(14; 14)(qll; q32)

t(14; 20)(q32; q13)

T-ALL Translocation Involving TCR Genes

HOXA cluster Class I homeodomaincontaining Inv(7)(p15q34)t(7; 7) 90,91

TLXI(HOXII) Class" homeodomaincontaining t(7; 1O)(q34; q24) 65,66

t(lO; 14)(q24; q11)

TLX3* (HOXIIL2j Class " homeodomaincontaining t(5; 14)(q35; q32) 86,87

LM01 LIM-only domain till ; 14)(p15; qll ) 68

LM02 LIM-only domain t(ll ; 14)p13; ql l), 69

t(7; 11)(q35; q13) 71

TALI b HLH Type " t(l; 14)(p32; qll ), 100

t(l ; 7)(p32; q34) 91

TAL2 b HLH Type" t(7; 9)(q34; q32) 75

LCK SRC family of tyrosine kinase t(l; 7)(p34; q34) 95,101

BHLHB1 b HLH Type" t(14; 21)(qll .2; q22) 76

LYLI b HLH Type" t(7; 19)(q34; p13) 74

CCN02 D-type cyclin t(7; 12)(q34; p13) 102

t(l2; 14)(p13; q11) 103

NOTCH1 Notch receptor family t(7; 9)(q34; q34.3) 94

*TLX3 is included desp ite the fact that the predominant tiS; 14) involves BeU1B, not IgH,since these BCL11B-TLX3 translocations are med iated by the recombinase and since raretranslocations involving TLX3 and TCRalfJ are descr ibed .

188 V(DJ]Recombination

intergenic rearrangements and deregulationof genesby junapositioning to IgITCR regulatorysequences (promoters or enhancersj." This can lead to increasedexpression or nonextinction ofthe juxtaposed"proto-oncogene"bypromoter/enhancer substitution or byseparationof codingsequences fromnegative regulatorydements.Onlythoserearrangements whichleadto deregulatedtissuehomeostasiswill beassociated with lymphoidmalignancies.Ifthe deregulatedgenes inducea survival or proliferative advantage or a block to maturation, the clonebearingthe translocationwill be transformed,or at least immortalised.Basedon these considerations, V(D)J errors willonly occur in cells which are recombinase competent and will target proto-oncogenes which areaccessible during this phaseofrecombinase activity.

It isincreasinglyrecognised that the transcriptional and phenotypicprofileobservedinacancerat diagnosis isnot necessarily identical,but isprobablymoremature,whencomparedto the cancerinitiating or stem cell.Genetic modifications which occur in this cancerstem cellare, however,transmitted to allclonaldescendants. Within thiscontext,both bona-fideIgITCR rearrangementsand recombinase mediatedoncogenicrearrangements detectedindiagnostic material canrepresentgenetic fingerprints of earliereventswhich haveoccurred in lymphoid cancer stem cells, or inintermediate malignantprecursorpopulations.Ifsuchmarkers arepresent in the majorityof thetumor at diagnosis, it is likely that they reflectan upstreameventduringoncogenicdevelopment,wherasthosepresentin minor subclones aremorelikdy to representdownstreameventsoccurringin tumor subclones. The capacityto accurately evaluatethe proportion of cells demonstratingagivenmarkerdependson the techniquesused.Briefly, molecularPCRand CGH basedtechniquesusingextractedDNA arepoorlyadapted to precisequantificationand cytogeneticanalysis of mitotic materialisbiasedbypotential nonrepresentativity of the cells undergoingmitosisunder theculture conditions used. FISH analysis of interface nucleihas the advantage of beingcellbased,but is only applicable to certain oncogenicmarkers, not to V(D)J rearrangements and is heavilydependent on the qualityof materialanalysed (barenucleivs.tissuesections, for example). Giventhese reserves, detection of an Igor TCR rearrangementin an apparentlynonlymphoid cancer,impliesprior exposure of malignantprecursorsto recombinase activity. Identification ofIglTCRrearrangements in Acute Mydoid leukaemia, for example, is preferentially found in cases withMLL generearrangement, with the MLL fusion transcriptparmers beingassociated with different Ig/TeR profiles." Similarly, detection of a recombinase mediatedoncogenicmarkerimplieschromatin accessibility of the parmer geneduring a phase of recombinase competenceprior totumor development,What levd ofqualitativeand/or quantitativerecombinase competenceand/or RAG1/2 activityisrequiredfor theserecombinase errorsisnot clear. Rearrangement ofTCRband TCRy can occur in the presenceof much lowerlevels of RAG1 activitythan that requiredforTC~ rearrangement46 and it is possible to induce TCRb rearrangement in kidney cellsinthe presenceofE2A and HEB.47.48

Categories ofRecombinase ErrorsTwo categories of recombinase errorsarerecognised:44•49.50 TypeI rearrangements demonstrate

breaksat RSSat both loci,oneofwhichisusually an Igor TCR; in TypeII rearrangements, onlytheIgITCR breakismediatedbyRAGand the mechanisms targetingthedoublestrandedbreakon theparmer geneare incompletely understood(Fig. 3). Once generated, thisDNA fragment becomesincluded in the recombinase complex, with the translocation resulting from a DNA repairerror,rather thanmistargetting of the recombinase. A recombinase mediatederror is characterized by i)involvement ofan IgITCRlocus;ii)recurrentgenomic breakpoints; iii) identification ofa bona-fideRSS-likesequence at thebreakpoint on theparmerchromosomeiv)additionofnongermline encodednucleotides at the translocation breakpointand v) generation of a signal joint. Recombinase mediated translocations werefirstidentified in Blymphoidnon-Hodgkins lymphoma (NHL) with thet(14; 18) translocation involving IgH and BCU,5l-53 Translocations involving Iglocipreferentiallyinvolve the IgH locusand arefound in relativdy mature,sIg+lymphomas.Thesetranslocations areessentially TypeII and primarily involve abnormalities of class switchand somatic hypermutation;54 assuch, theyarebeyondthe scopeof thisarticle,whichisrestricted to V(D)J recombinase errors

Normaland Pathological V(D)j Recombination 189

in immature lymphoproliferative disorders, essentiallyALL.Aproportionoftheseabnormalities arealso found in certainlymphomas, notablythoseinvolvingMYC in Burkitt's lymphoma and thoseinvolvingHOXII/TLXI in T-Iymphoblastic lymphoma.

VD]Deregulation with Oncogenic PotentiatIg translocations are found in approximately I% of B lineage ALL, when they are virtually

restrictedto mature,slg+ cases. Partnergenes includeMYC,55.58 BCL-9,59 LHX4/,() ID4,61 IL362

or thedifferentmembers of theCEBPfamily63 (Table3). In contrast, chromosomal abnormalitiesinvolving theTCR lociareamongthosemostfrequently encounteredinT-ALL.Mostinvolve theTCRa/b locuson chromosome 14qII or morerarely, TCRf3 on chromosome 7q34;35.64 rearrangementsinvolvingTCRyareexceptional.ThefirstTCRtranslocations to bedescribed inT-ALLwerethoseinvolvingHOXI1/TLXI atchromosome IOq2465-67andLMO1/2 on chromosome IIp.68-71Theincidence ofTCR translocations byclassical, morphological karyotypingwas underestimatedand it wasonly with the adventof screening by FISH that the true incidencewasappreciated.ScreeningforTCR translocations demonstratedthat approximately45%ofT-ALLsdemonstratetranslocations, includinga minoritywith asyetunidentifiedpartners.PredominantknownTCRpartner genes can bedividedinto thoseof the bHLH, LMO and HOX/TLX families.

ThemostcommonlyencounteredbHLH partner isTALI/SCL,whichwasinitiallydescribedin the rare t(l; 14)(p32;ql I)." Muchmorefrequent is the SIL-TALlrecombinase mediatedintrachromosomal deletion,whichplaces theentireTALI codingsequence undercontrolof the SILpromoter;" SIL-TALldeletions arefound in20%ofpediatricand 5-10%ofadultT-ALLs. OtherbHLH translocations includethe raret(7; 19)(q34;p13),74t(7; 9)(q34:q32)75 and t(14; 21)(qll;q22)76 involvingLYL-I, TAL2and bHLHB I respectively. Thefrequentinvolvement ofmembersof the bHLH family oftranscriptional regulators iscoherentwith the fundamentalroleofbHLHproteins in regulationofT and Blymphoidlineagedevelopment. Thisis further emphasised bythe fact that the LMO proteinsderegulatedbyTCRjuxrapositioning in translocations involvingLMOI (llpI5)68 or LM02 (llp13)69.71 form part of a complex which also includes TALI andits bHLH partner, E2A.77.81

Deregulationofhomeoboxgeneexpression is increasingly recognised in T-ALL.Theorphanhomeoboxgene,HOXI1/TLXI, ispredominantlyinvolved in the t(10;14)(q24;qII) and morerarelythe t(7; 10)(q34;q24).65-67Forcedexpression ofTLXI in murinebone marrowgives risetoT-ALL-like malignancies with longlatency, suggesting that other eventsare necessary to induceleukemia; but with TLXI expression representing an earlyevent.82.83 TLXI regulates the G1/ScheckpointofT-ALLviaitsbidingcapability to the protein serine/threoninephosphatases PP2Aand PP1.84.85 Chromosomaltranslocations t(lO; 14)(q24;ql l ) involvingTLXI are amongsttheclearest example of recombinase involvement in T-ALL.Deregulatedexpression ofHOXIIL2ITLX3 isfrequently found inpediatricT-ALL,due in mostcases to at(5; 14)involving the TLX3locusat Sq3S and CTIP2I BCLllB at 14q32,7000 kb proximalto the IgH locus.86.87Despitethe absence of IgITCR involvement, this translocation is mediatedby the recombinase and raretranslocations involving TLX3 and TCRa/b havebeendescribed."TLX3 hasveryclose homologyto TLXI, asevidenced bymicroarray studiesshowingthat TLXI and TLX3 T-ALLs clustertogether.89.9OTCR translocations involving the HOXA clusteron chromosome 7 predominantlyinvolve theTCR~ locus,leadingto a crypticintrachromosornal inverslon.Y" Another commonabnormalityin T-ALLisdeletionof the p16/INK4/Cdk2 gene;92.93this isrecombinase mediatedin at leasta proportion of cases. Other rarerecombinase mediatedabnormalities includetranslocationsinvolving TCRf3 and Notchl in the t(7; 9)(q34; q34)94 and t(l; 7)(q34; q34) involvingLCK and TCRf3.95.%

In general, these recombinase mediated errors are restricted to T-ALLs of the TCRa~lineage,which express RAGI and haveundergone extensive TCR rearrangement. The H OX/TLX cases are arrested prior to TCRa rearrangement, in contrast to SIL-TALI cases, whichhave undergone TCRa rearrangement on at least one allele. They are rarelyfound in TCRybexpressingT-ALLs,with the exceptionofTLX3 expressing cases, whichfrequentlyexpress both


TCRyband cytoplasmic TCRfl.97 RecombinaseV(D)Jerrorsareclearlyan important mechanismin the development ofimmature T-cell malignancies. Attempts to recreate these malignanciesin murine models havefrequentlydemonstrated long latency and/or a low proportion of micedevelopingleukaemia/lymphoma, in keepingwith multistage oncogenesis.9B•99Inkeepingwiththis, low levels oftranslocations involvingLM02 havebeen identified in normal thymus."

ConclusionTheaforementioned abnormalities represent amechanistic rolefor the recombinase indevelop

ment oflymphoid malignancies. It is howeveralsopossible that failureto completeproductionofa mature, functional Ig or TCR may favourmalignant expansion, particularlyif the cellsarearrested at a stagewhen the pre B or TCR is expressed and capable ofmediatingligand drivencellularexpansion. The majorityof acute leukaemias do not express a surface Ig/TCR, despitehavingundergoneextensive Ig/TCR rearrangement. This failure to completesuccessful Ig/TCRrearrangementislikely to be at leastpartiallyat the originof the recombinase competenceand themaintenanceofRAGI expression. It is thereforeat least theoreticallypossible that abrogationofthe factorsblockingcompletionofIg or TCR assembly couldleadto expression ofthe appropriateIg/TCR at the surface, downregulationof RAGexpression and possibly evenleukemiccelldeathbydifferentiation. As mentioned above, a significant proportion ofHOXllL2/TLX3+ T-ALLsexpress unusual TCRyb receptorsand cytoplasmic TCRfl. TheseT-ALLsmaintain high levelsof RAG1 transcripts, despite the expression of a surfaceTCR, suggesting that expression of an"inappropriate,default"TCRyb in cells havingundergone beta selectionis insufficient to allowextinction of the recombinase. Explorationof the mechanisms underlyingthe failure to rearrangeTCRa mayfurther our understandingofT-ALL oncogenesis.

Inconclusion, understandingand exploitingnormaland abnormalrecombinase activitycanbeused both in individualpatient managementand in understandinglymphoidoncogenesis.

AcknowledgementsWethank all the techniciansand studentsfrom the immunogenetics sectionof the diagnostic

haematologyplatform and INSERM EMIU 0210 at Necker-EnfantsMalades who havecontributed overthe years to accumulationofour experience oflymphoid clonalityanalysis. Thanksalsoto ClaudineSchiff, BertrandNadeland PierreFerrierat Centre d'ImmunologieMarseille Luminyfor fruitful discussions and collaborations and to Jacquesvan Dongen et al for the coordinationofthe Biomed2programswhich allowedstandardisationand optimisedexploitationof clonalityanalysis in lymphoproliferative disorders. Wealsothank all Frenchadult and pediatric clinicianswho manage patients with lymphoid malignancies, without whose collaboration it would beimpossible to further our understandingof thesedisorders.

NoteSaidaDadi and SandrineLe Noir havecontributed equallyto this work.

References1. BeishuizenA, VerhoevenMA, Mol EJ et al.Detection of immunoglobulin heavy-chaingene rearrangements

by Southern blot analysis: recommendations for optimal results. Leukemia 1993; 7(12):2045 -2053.2. Langerak AW, Wolvers-Tettero IL, van Dongen J], Detection ofT-cell receptor beta (TCRB) gene rear

rangement patterns in T-cell malignancies by Southern blot analysis. Leukemia 1999; 13(6) :965-974.3. van Krieken JH, Elwood L, Andrade RE et al. Rearrangement of the T-cell receptor delta chain gene

in Tvcell lymphomas with a mature phenotype. Am J Patho11991; 139(1) :161-168.4. Langerak AW, SzczepanskiT, van der Burg M et al.Heteroduplex PCR analysisof rearranged T-cell recep

tor genes for clonality assessment in suspect T-cell proliferations. Leukemia 1997: 11(12):2192-2199.5. Gonzalez M, Gonzalez D, Lopez-Perez Ret al. Heteroduplex analysis ofVO] amplified segments from

rearranged IgH genes for clonality assessments in Bvcell non-Hodgkins lymphoma. A comparisonbetween different strategies. Haematologica 1999; 84(9) :779-784.

6. Derksen pw, Langerak AW, KerkhofE er al. Comparison of different polymerase chain reaction-basedapproaches for clonality assessment of immunoglobulin heavy-chain gene rearrangements in Bvcellneoplasia . Mod Pathol 1999; 12(8) :794-805.

Normaland Pathological V(D)JRecombination 191

7. KrafIi: AE, Taubenberger]K . Sheng ZM et aI. Enhanced sensitivitywith a novelTCRgarnrna PCR assay forclonality studies in 569 formalin-fixed, paraflin-embedded (FFPE) cases, Mol Diagn 1999; 4(2):119-133.

8. Assaf C. Hummel M. Dippel E ec al. High detection rate of Tscell receptor beta chain rearrangementsin Tvcell Iymphoproliferations by family specific polymerase chain reaction in combination with theGeneScan technique and DNA sequencing. Blood 2000; 96(2) :640-646.

9. Theriault C . Galoin S. Valmary S et al. PCR analysis of immunoglobulin heavy chain (lgH ) andTcR-gamma chain gene rearrangements in the diagnosis of lymphoproliferarive disorders: results of astudy of 525 cases. Mod Pathol 2000; 13(12):1269-1279 .

10. van Dongen ]J. Analysis of immunoglobulin genes and Tvcell receptor genes as a diagnostic tool for thedetection of lymphoid malignancies. Neth ] Med 1987; 31(5 -6):201-209 .

11. Gleissner B. Maurer ]. Thiel E. Detection of immunoglobulin heavychain genes rearrangements in Bvcellleukemias, lymphomas. multiple myelomas,monoclonal and polyclonal gammopathies, LeukLymphoma2000; 39(1-2):151-155.

12. Sen F. Vega F. Medeiros LJ. Molecular genetic methods in the diagnosis of hematologic neoplasms.Sernin Diagn PathoI2002; 19(2):72-93.

13. LangerakAW,van Krieken]H. Wolvers-Tettero lLet aI. The role of molecular analysis of immunoglobulinand T'cell receptor gene rearrangements in the diagnosis of lymphoproliferarive disorders. ] Clin Pathol2001; 54(7):565-567.

14. Miqueu P. Guillet M. Degauque N et al. Statistical analysis of CDRJ length dist ributions for theassessment ofT and Bvcell repertoire biases. Mol ImmunoI2007; 44(6) :1057-1064.

15. van Dongen JJ. Langerak AW. Bruggemann M et al. Design and standardization of PCR primers andprotocols for detection of clonal immunoglobulin and Tvcell receptor gene recombinations in suspectIymphoproliferations: report of the BIOMED-2 Concerted Action BMH4-CT98-3936. Leukemia 2003;17(12) :2257-2317.

16. van der Velden YH. Cazzaniga G. Schrauder A et al. Analysis of minimal residual disease by IgiTCRgene rearrangements: guidelines for interpretation of real-time quantitative PCR data. Leukemia 2007;21(4):604-611.

17. Delabesse E. Burtin ML. Millien C et al. Rapid. rnuldfluorescenr TCRG Vgamma and]gamma typing:application to Tscell acute lymphoblastic leukemia and to the detect ion of minor clonal populations .Leukemia 2000; 14(6):1143-1152.

18. van Krieken ]H. Langerak AW. San Miguel]F er al. Clonal ity analysis for antigen receptor genes: preliminary results from the Biorned-Zconcerted action PL 96-3936. Hum PathoI2003; 34(4):359-361.

19. Droese], Langerak AW, Groenen P] et al. Validation ofBIOMED-2 multiplex PCR tubes for detectionofTCRB gene rearrangements in T-cell malignancies. Leukemia 2004; 18(9):1531 -1538.

20. EvansPA. Port C. Groenen P] et al. Signi6cantly improved PCR-based clonaliry testing in Bvcell malignancies by use of multiple immunoglobulin gene targets. Report of the BIOMED-2 Concerted ActionBHM4-CT98-3936. Leukemia 2007; 21(2):207-214.

21. van Krieken ]H. Langerak AW, Macintyre EA et al. Improved reliability of lymphoma diagnostics viaPCR-based clonaliry testing: report of the BIOMED-2 Concerted Action BHM4-CT98-3936. Leukemia2007; 21(2):201-206.

22. Bruggernann M. White H. Gaulard P et al. Powerful strategy for polymerase chain reaction -basedclonality assessment in Tvcell malignancies Report of the BIOMED-2 Concerted Action BHM4CT98-3936. Leukemia 2007; 21(2) :215-221.

23. Langerak AW, Molina 1']. Lavender FL et al. Polymerase chain reaction-based clonality testing in tissuesamples with reactive lymphoproliferanons: usefulness and pitfalls. A report of the BI0MED-2 Concerted Action BMH4-CT98-3936. Leukemia 2007; 21(2) :222-229.

24. Liu H. Bench A]. Bacon CM et al, A practical strategy for the routine use ofBIOMED-2 PCR assays fordetection of B- and Tvcellclonality in diagnostic haematopathology. Br] Haematol 2007; 138(1):31-43.

25. Asna6 V. Beldjord K, Garand R et al. IgH D] rearrangements within T-ALL correlate with cCD79aexpression. an imrnature/Tf.Rgammadelta phenotype and absence of IL7Ralpha/CD127 expression.Leukemia 2004; 18(12):1997-2001.

26. Kitch ingman GR. Rovigatti U. Mauer AM et al. Rearrangement of immunoglobulin heavy chain genesin Tvcell acute lymphoblastic leukemia. Blood 1985; 65(3) :725-729.

27. Szczepanski T. Pongers-Willernse M]. LangerakAW er at Igheavychain gene rearrangements in T-cell acutelymphoblastic leukemia exhibit predominant DH6-19 and DH7-27 gene usage, can result in complete V-Do]rearrangementsand arc rare in T-cdl receptor alpha beta lineage. Blood 1999; 93(12):4079-4085.

28. Chen Z. Le Pasher D. Dausser ] et al. Human Tvcell gamma genes are frequently rearranged inBvlineage acute lymphoblastic leukemias but not in chronic B-cell proliferations. ] Exp Med 1987;165(4):1000-1015.

29. Brumpt C. DelabesseE. Bddjord K et al, The incidence ofclonal T'cell receptor rearrangements in Bvcellprecursor acute lymphoblastic leukemia varies with age and genotype. Blood 2000; 96(6):2254-2261.


30. Greaves M. Molecular generics. natural history and the demise of childhood leukaemia. Eur J Cancer1999: 35(14) :1941-1953.

31. Greaves M. Childhood leukaemia. BMJ 2002: 324(7332) :283-287.32. Graux C, Cools J, Michaux L et al. Cytogenetics and molecular genetics of T-cell acute lymphoblastic

leukemia: from thymocyte to lymphoblast. Leukemia 2006: 20(9) :1496-1510.33. O'Neil J, Look AT. Mechanisms of transcription factor deregulation in lymphoid cell transformation .

Oncogene 2007: 26(47) :6838-6849.34. Argiropoulos B. Humphries RK. Hox genes in hematopoiesis and leukemogenesis. Oncogene 2007:

26(47) :6766-6776.35. Armstrong SA. Look AT. Molecular genetics of acute lymphoblastic leukemia. J Clin Oncol 2005:

23(26) :6306-6315.36. Willis TG. Dyer MJ. The role of immunoglobulin translocations in the parhogenesis of B-cell

malignancies. Blood 2000: 96(3) :808-822.37. Stern MH. Lipkowitz S, Aurias A et al. Inversion of chromosome 7 in ataxia telangiectasia is gener

ated by a rearrangement between T-cell receptor beta and T-cell receptor gamma genes. Blood 1989:74(6):2076-2080.

38. Mullighan CG. Goorha S, Radtke I er al. Genome-wide analysisof generic alterations in acute lymphoblastic leukaemia. Nature 2007: 446(7137) :758-764.

39. FuscoeJC , Zimmerman LJ, Lippert MJ et al. V(D)J recombinase-like activity mediates hprt gene deletion in human feral T-lymphocytes. Cancer Res 1991: 51(21):6001-6005.

40. Fuscoe JC, Zimmerman LJ. Harringron-Brock K et al. Deletion mutations in the hprt gene ofT-lymphocytes as a biomarker for genomic rearrangements important in human cancers. Carcinogenesis1994: 15(7):1463-1466.

41. Fuscoe JC, Zimmerman LJ, Harrington-Brock K et al. Large deletions are tolerated at the hprt locusof in vivo derived human T-lymphocytes. Murat Res 1992: 283(4) :255-262.

42. Finerre BA. Kendall H. VacekPM. Mutational spectral analysis ar the HPRT locus in healthy children.Murat Res 2002: 505{1-2):27-41.

43. Digweed M. Human genetic instability syndromes: single gene defecrs wirh increased risk of cancer.Toxicol Lett 1993: 67(1-3):259-281.

44. Marculescu R. Vanura K. Montpellier B et al. Recombinase, chromosomal translocations and lymphoidneoplasia: rargering mistakes and repair failures. DNA Repair (Arnsr) 2006: 5(9-10):1246-1258.

45. Dupret C. Asna/i V,Leboeuf D et al. IgH/TCR rearrangements are common in MLL translocated adultAML and suggesr an early T/myeloid or B/myeloid maturation arrest, which correlates with the MLLpartner, Leukemia 2005: 19(12):2337-2338.

46. AsnafiV. Beldjord K, Boulanger E et aI. AnalysisofTCR. pT alpha and RAG-l in T-acure lymphoblasricleukemias improves undemanding of early human T-lymphoid lineage commlrmenr, Blood 2003:101(7) :2693-2703.

47. LangerakAW,Wolvers-Tertero IL. van Gasrel-MolEJ er al. Basichelix-loop-helixproteins E2A and HEBinduce immature T-cell receptor rearrangements in nonlymphoid cells. Blood 2001: 98(8) :2456-2465.

48. Ghosh JK, Romanow WJ, Murre C. Induction of a diverse T-cell receptor gamma/delta repertoire bythe helix-loop-helix proteins E2A and HEB in nonlymphoid cells.J Exp Med 2001: 193(6):769-776.

49. Marculescu R, Le T, Simon P et al. V(D)]-mediated translocations in lymphoid neoplasms: a functionalassessment of genomic insrabiliry by cryptic sites. J Exp Med 2002: 195(1):85-98.

50. Lieber MR. Yu K, Raghavan SC. Roles of nonhomologous DNA end joining, V(D)] recombination and class swirch recombination in chromosomal rranslocations. DNA Repair (Amsr) 2006:5(9-10) :1234-1245.

51. Tsujimoto Y, Gorham J. Cossman Jet al. The r(14: 18) chromosome translocations involved in Bvcellneoplasms result from mistakes in VOJ joining. Science 1985: 229(4720):1390-1393.

52. Cleary ML, Smith SD. SldarJ. Cloning and structural analysis of cDNAs for bcl-2 and a hybrid bcl-2/immunoglobulin transcript resulting from the r(14: 18) translocation. Cell 1986: 47(1):19-28 .

53. Tsujimoro Y, Croce CM. Analysis of the structure , transcripts and protein products of bcl-Z, the geneinvolved in human follicular lymphoma. Proc Nad Acad Sci USA 1986: 83(14):5214-5218.

54. Schlissel MS. Kalfer CR. Curry JD. Leukemia and lymphoma: a cost of doing business for adaptiveimmunity. Genes Dev 2006: 20(12) :1539-1544.

55. Dalla-Favera R. Martinotti S. Gallo RC et al. Translocation and rearrangements of the c-myc oncogenelocus in human undifferennared B-celllymphomas. Science 1983: 219(4587) :963-967.

56. Davis M. Malcolm S. Rabbitts TH. Chromosome translocation can occur on either side of the c-myconcogene in Burkitt lymphoma cells. Nature 1984: 308(5956):286-288.

57. Hollis GF, Mitchell KF, BatteyJ et al. A variant translocation places the lambda immunoglobulin genes3' to the c-myc oncogene in Burkitt's lymphoma. Nature 1984: 307(5953): 752-755.

Normaland Pathological V(D)j Recombination 193

58. Manolov G, Manolova Y. Marker band in one chromosome 14 from Burkitt lymphomas. Nature 1972;237(5349) :33-34.

59. Willis TG, Zalcberg IR, Coignet L] et at. Molecular cloning of translocation t(l ; 14)(q21; q32) definesa novel gene (BCL9) at chromosome lq21. Blood 1998; 91(6) :1873-1881.

60. Kawamata N, Sakajiri S, Sugimoto K] er al. A novel chromosomal translocation t(l ; 14)(q25; q32)in preB acute lymphoblastic leukemia involves the LIM homeodomain protein gene, Lhx4. Oncogene2002; 21(32) :4983-4991.

61. Bellido M, Aventin A, Lasa A et at. Id4 is deregulated by a t(6; 14)(p22; q32) chromosomal translocation in a B-celliincage acute lymphoblastic leukemia. Haemarologica 2003; 88(9):994-1001.

62. Meeker TC, Hardy 0, Willman C et al. Activation of the interleukin-3 gene by chromosome translocation in acute lymphocytic leukemia with eosinophilia. Blood 1990; 76(2):285-289.

63. AkasakaT, BalasasT, Russell L] et al. Five membersof the CEBP transcription factor family are targetedby recurrent IGH translocations in B-cell precursor acute lymphoblastic leukemia (BCP-ALL). Blood2007; 109(8):3451-3461.

64. Raimondi SC, Behm FG, Roberson PK et at. Cytogenetics of childhood T-cell leukemia. Blood 1988;72(5) :1560-1566.

65. Hatano M, Roberts CW, Minden M er al. Deregulation of a homeobox gene, HOXll, by the t{10; 14)in T-cell leukemia. Science 1991; 253(5015):79-82.

66. Dube 10, Kamel-Reid S, Yuan CC et at. A novel human homeobox gene lies at the chromosome10 breakpo int in lymphoid neoplasias with chromosomal translocation t(10; 14). Blood 1991;78(11):2996-3003.

67. Kennedy MA, Gonzalez-Sarmiento R, Kees UR et at. HOXll, a homeobox-containing T-cell oncogeneon human chromosome 10q24. Proc Natl Acad Sci USA 1991; 88(20) :8900-8904.

68. McGuire EA, Hockett RD, Pollock KM et at. The t(l1 ; 14)(p15; q11) in a T-cell acute lymphoblasticleukemia cell tine activates multiple transcripts, including Tcg-I, a gene encoding a potential zinc lingerprotein. Mol Cell Bioi 1989; 9(5):2124-2132.

69. Royer-Pokora B, Loos U, Ludwig WO. TTG-2, a new gene encoding a cysteine-rich protein with theLIM motif, is overexpressed in acute T-cell leukaemia with the t(l1; 14)(p13 ; q11). Oncogene 1991;6(10):1887-1893.

70. VanVllerberghe P, van Gretel M, BeverlooHB et at. The cryptic chromosomal deletion del(II)(pI2p13)as a new activation mechanism ofLM02 in pediatric T-cell acute lymphoblastic leukemia. Blood 2006;108(10):3520-3529.

71. Boehm T, Foroni L, Kaneko Y et at. The rhombotin family of cysteine-rich LIM-domain oncogenes:distinct members arc involved in T-cell translocations to human chromosomes 11p15 and 11p13. ProcNat! Acad Sci USA 1991; 88(10) :4367-4371.

72. Bernard 0, Barin C, Charrin C et at. Characterization of translocation t(l; 14)(p32 ; q11) in a T andin a B acute leukemia. Leukemia 1993; 7(10) :1509-1513.

73. Aplan PO, Lombardi DP, Ginsberg AM et al, Disruption of the human SCL locus by "illegitimate"V-(D)-] recombinase activity. Science 1990; 250(4986) :1426-1429.

74. Mellentin ]0, Smith SD, Cleary ML. lyl-I, a novel gene altered by chromosomal translocation in T-cdlleukemia, codes for a protein with a helix-loop-helix DNA binding motif Cell 1989; 58(I):77-83.

75. Xia Y,Brown L, Yang CY et at. TAL2, a helix-loop-helix gene activated by the (7; 9)(q34; q32) translocation in human T-cell leukemia. Proc Natl Acad Sci USA 1991; 88(24) :11416-11420.

76. Wang], ]ani-Sait SN, Escalon EA ee at. The t( 14; 21)(q11.2; q22) chromosomal translocation associatedwith T-cell acute lymphoblastic leukemia activates the BHLHBI gene. Proc Natl Acad Sci USA 2000;97(7):3497-3502.

77. Baer R. TAll, TAL2 and LYLl: a family of basic helix-loop-helix proteins implicated in T-cell acuteleukaemia. Sernin Cancer BioI 1993; 4(6) :341-347.

78. O'Neil], Billa M, Oikcmus S et at. The DNA binding activity of TAL-l is not required to induceleukemia/lymphoma in mice. Oncogene 2001; 20(29) :3897-3905.

79. Bain G, Engel I, Robanus Maandag EC er at. E2A deficiency leads to abnormalit ies in alphabeta T-celldevelopment and to rapid development ofT-cell lymphomas. Mol Cell BioI 1997; 17(8):4782-4791.

80. Yan W; Young AZ, Soares VC et at. High incidence ofT-cell tumors in E2A·null mice and E2A/Idldouble-knockout mice. Mol Cell BioI 1997; 17(12):7317-7327.

81. O'Neil], Shank], Cusson N et at. TALl /SCL induces leukemia by inhibiting the transcriptional activityofE47/HEB. Cancer Cell 2004; 5(6):587-596.

82. Hawley RG, Fong AZ, Lu M er at. The HOXll homeobox-containing gene of human leukemiaimmorralizes murine hematopoietic precursors. Oncogene 1994; 9(1):1-12.

83. Hawley RG, Fong AZ, Reis MD et at. Transforming function of the HOXll/TCL3 homeobox gene.Cancer Res 1997; 57(2):337-345.


84. Kawabe T, MuslinAJ. Korsmeyer SJ.HOX11 Interacts with protein phosphaeases PP2A and PPI anddisrupts a G2/M cell-cycle checkpoint. Nature 1997; 385(6615):454-458.

85. RizI. Hawley RG. Gl/S transcriptional networks modulatedbythe HOX11/TLXI oncogene ofT-cellacute lymphoblastic leukemia. Oncogene2005; 24(36):5561-5575.

86. BernardOA. Busson-LeConiat M, Ballerini P et aI.A new recurrentand specific cryptic translocation.t(5; 14)(q35; q32). isassociated with expression of the Hox11L2 genein T acutelymphoblastic leukemia.Leukemia 2001; 15(10):1495-1504.

87. MacLeod RA, NagelS. Kaufmann M ecaI. Activation ofHOX11L2 byjuxtaposition with 3'-BCLlIBin an acutelymphoblastic leukemia cellline (HPB.ALL) with t(5; 14)(q35; q32.2). Genes ChromosomesCancer 2003; 37(1):84-91.

88. Hansen-Hagge TE, Schafer M. Kiyoi H et aI.Disruption of the RanBPI7/HoxIIL2 regionby recombination with the TCRdelta locus in acutelymphoblastic leukemias with t(5; 14)(q34; qll). Leukemia2002; 16(11):2205-2212.

89. Ferrando AA.Neuberg DS. Staunron J er al, Geneexpression signatures define novel oncogenic pathwaysin T-cell acute lymphoblastic leukemia. Cancer Cell 2002; 1(1):75-87.

90. SoulierJ, ClappierE, Cayuela JM et al. HOXA genes are included in geneticand biologic networksdefininghuman acuteT-cell leukemia (T-ALL). Blood 2005; 106(1):274-286.

91. Speleman F,Cauwelier B.Dasrugue N er aI. A new recurrentinversion. inv(7)(pI5q34), leads to transcriptionalactivation of HOXAIO and HOXA11 in a subset ofT-cell acute lymphoblastic leukemias.Leukemia 2005; 19(3):358-366.

92. OhnishiH. Kawamura M, IdaK et al,Homozygous deletions of pI6/MTSI genearefrequent but mutations are infrequent in childhoodT-cell acute lymphoblastic leukemia. Blood 1995;86(4):1269-1275.

93. Cayuela JM. Madani A. Sanhes L et aI. Multiple tumor·suppressor gene I inactivation is the mostfrequent geneticalteration in T-cell acute lymphoblastic leukemia. Blood 1996; 87(6):2180-2186.

94. Ellisen LW; BirdJ, WestDC et aI.TAN-I, the humanhomologof the Drosophila notch gene. is brokenby chromosomal translocations in T lymphoblastic neoplasms. Cell 1991;66(4):649-661.

95. Burnett RC. Thirman MJ. Rowley JD et aJ. Molecular analysis of the T-cell acute lymphoblasticleukemia-associated t(l; 7)(p34; q34) that fuses LCK and TCRB. Blood 1994; 84(4):1232-1236.

%. BurnettRC, David JC. HardenAM ecat TheLCK gene is involved in the t(l; 7)(p34; q34) in the T'cellacute lymphoblastic leukemia derived cellline, HSB-2. Genes Chromosomes Cancer 1991; 3(6):461-467.

97. Asnafi V, Beldjord K. LiburaM et aI. Age-related phenotypicand oncogenic differences in T'cell acutelymphoblastic leukemias mayreflecr thymic atrophy. Blood 2004; 104(13):4173-4180.

98. Larson RC. OsadaH, Larson TA ecal, Theoncogenic LIM protein Rbtn2 causes thymicdevelopmentalaberrations that precedemalignancy in transgenic mice. Oncogene1995; 11(5):853-862.

99. Aplan PD, Jones CA, Chervinsky DS er aI. An sci gene product lacking the transactivation domaininduces bony abnormalities and cooperates with LMOI to generate T-cell malignancies in transgenicmice. EMBOJ 1997; 16(9):2408-2419.

100. Carroll AJ. Crist WM, Link MP et aJ. The t(l; 14)(p34; q11) is nonrandom and restricted to T-cellacute lymphoblastic leukemia: a Pediatric Oncology Group study. Blood 1990; 76(6):1220-1224.

101. Tycko B, Smith SD.SklarJ. Chromosomal translocations joiningLCK and TCRB loci in humanT-cellleukemia. J Exp Med 1991; 174(4):867-873.

102. Le Coniar M. DellaValle V,Marynen P er aI.A new breakpoint. telomeric to TEL/ETV6, on the shortarm of chromosome 12 in T'-cell acute lymphoblastic leukemia. Leukemia 1997; 11(8):1360-1363.

103. Karrman K, Andersson A. Bjorgvinsdottir H et aI. Deregulation of cyclin D2 by juxtaposition withT-cell receptor alpha/delta locus in t(12; 14)(p13; q11)-positive childhood T-cellacute lymphoblasticleukemia. EurJ Haematol2006; 77(1):27-34.

INDEX

A

Accessibility 4.33-35.59.60,62.63,65,68.73.75-77,82.84.90,93-99, 103.105-113.116.121 .123.124.126-128.133-144.149.150,153.159,160,166.168.172.173.176.181 ,188

Accessibility to recombination 76.135,144Acute lymphoblastic leukemia(ALL) 180.

182-187,189,190Allele 34,38,40.53.60.63.64,66-69.74.

75,94,104.105,109,111-113,118-123,125-127,133.134.136.137.139-144.159,160.172-174.183,184.189

Allelic exclusion (AE) 34.68, 104,105. 111113.116,117,121,125-128.134,136,143.144.157,160,162.166.168.171 .172,176

Alternative NHE] see Nonhomologousendjoining

Antibody 60,68.73-76.79.83, 168. 171,174

Antigen receptorgeneevolution 174Antisense transcription 64.66.69,137.138 ,

140Artemis 2,3 , 17. 18.26.27,36.37.46.48.

49,51.52.170ATM 17.34.36,39.40,52Atypical scm 49

B

Base-flipping 6-9,11Basic helix-loop-helix (bHLH) 148.187,

189B-cell 24.26 .33-35,40.46-50.54,60.62

64,66-69.7~77,79.94.95, 104.105.108,109.111 -113,116,122.126.133.134,136-144.148-153.159 .160,168 ,169,171-174,180,182-186

development 48.60,77, 105. 112, 133,134.136-139.141 -143.150 .160.169, 186

fate commitment 138, 139fate specification 133.134.137.138, 140

cCancer 11.49.53.54.91,128.180,181.

188Cell cycle 36,38,40,52.127.134, 141,142,

157, 158Cernunnos 2.3.17.37.46.49.51 ,53.54Cernunnos/XLF 53, 54c-Fos 10Chemokine 134,142,143Chromatin 33-36.38 ,40,52.59.62,63,

65-69,73.75,76.78-80,90,93-99,103.106-113.121-124.126-128.133-141.144.149.153,159-162,172-174.181.184.186.188

accessibility 63.68.90,93.106,107,109-112,121,123,124,126-128.135,141 .149.153.181,188

immunoprecipitation (ChIP) 73.76.78 .79,122,124,128.152

remodeling 59.62.63,65.69.78.80,95.99.109.110,113,123,124,149

structure 35.40. 62, 65. 67.75 .76. 78,79, 93. 95-98, 123. 124. 133-140.144.160

Chromosomaltranslocation 16.19,21-23.32,54.103 .189

Chromosometerritory 38-40, 63cis-Regulatory element 119,122-125.128,

144.150Cryptic RSS ~36, 186CXCR4 134.143CyclinA/Cdk2 158

D

Disintegration 19.21.25DNA 1-6.8-11.16-27,32-40.46,47.49-54.

59.60,62.68.69, 74.75.77.79.80,82,~86.88.89,95.96,99.103.104,

107-109.111.116.119 .123 -128,135.136.140.142,144,148.149.152.157 160,162.167-172.174.175,181.183 .185-188

coding 18,24.26damagesensing 34. 39

196

footprinting 142,149ligase IV 2,3, 18,27,36,37looping 68,69,125,136,140methylation 159repair 19,32,34,37,39,46,47,50-53,

104,157,158,168,188DNA-PK 36,38 ,51 ,52,54DNA-PKc 2,3,36,37,38,51,52D-region 66, 67

E

E2A 77-79,94,134,137,140,142,143,148-153,186,188,189

E12 148,149E47 148-150Ea 94,97EBF 77,78,134,137-141,151,153E-boxsite 148-150,153Electrophoreticmobilityshift assay (EMSA)

3Enhancer 22,35,62,63,66,68,69,79,90,

93,94,97,98,106,109-112,117-119,121-123,126,134,136,137,140-144,148-150,153,159,172,175,188

Epigenetic control 73,160Epigenetic regulation 108,111, 159E-protein 148-153Ezh2 68,69,77, 134, 140

F

Fluorescence in situ hybridization (FISH)63,64,109,111,112,126,141,144,188,189

Fluorescence resonance energytransfer(FRET) 5,7

G

ybT cell 126,152,166,168,182,184,189,190

Gene targeting 66,94, 116, 117, 119,120,122,128,137

Genomicinstability 16, 19,33,37,53,59,186

Germlinetranscription 62,63,65,66,68,69,76,93-95,97,105,106,109-111,123,124,127,135,139,142-144,149,160

GTP 25

V(D)]Recombination

H

H3K4me3 27, 108,158, 160-162H3FtZme2s/K4me3 27Haemophilus influenzae type b (Hib) 74,75Heavychainexclusion 169HEB 122,148,149-153,188Histone acetylation 66,68,69,76-78,96,

97-99,137-139,142,143,160Histone methylation 96,98,161Histone modification 35,62,63,65,66,69,

77,80,96,99,107-109,112,135 ,137,139,160,161

HMGB1 1-8,10,11,33HMGB2 1Homologousrecombination 38,94,97,158Human primaryimmunedeficiency 47

I

Id protein 148IgHenhancer 137,150Igkenhancer 149,153IgITCR 123,126,171,180-182,184-190IL-7Rsignaling 138,140,141Immunoglobulin (Ig) 1,16,22,32-34,39,

46,48,59,60,68,73-76,98,103-106,108,109,111,116,117,123,126,127,133-136,140-144,148-150,152,153,159-161,166-176,180-182,184-190,196

Immunoglobulin (Ig) gene 75,104,105,109,133 ,135 ,136,144,167,168 ,170,171,174,175,187

Immunoglobulin heavychain (Igh) 39,59-69,94,99 ,116,122,126,140,148,161,169

Inner nuclearmembrane(INM) 109,135,136,140,141 ,144

Intergenictranscription 62-66,68,69,94,137

IRF4 134,136,142,143,153IRF8 142IlL 'supergene' 63, 66, 68

JJasegment 97,98,107

Index

K

Ku 2,3,17,18.27,36-39,51

L

Leukemia 32,33,35.39, 103, 180. 184,185.189

Locus 10.19,26.35-37,47,51,59-64.6669.73 .75,77-79.82-90.93-95,97-99,103-109,111-113,116-120.122-128 ,137-144,149-151.153. 157,159, 161,162.169,171 -174.182,186.188,189

accessibility 35.59,60,93.106.124.142-144

compaction 140contraction 35.67-69.79, 111.113, 126.

128,139-141.169.172-174Lymphocyte 1,10,24,32.35,39,40,46-51.

53.54.59,60,67.68.76,82,83.85,90.103-106,109,111.116,117,122.126.133,144.148.149,153,157,166,168,171-173,176,184

Lymphocyte development 1,10.32, 76, 103.105,109.116.133 ,144,148,149

Lymphoid cancer 180.181,188Lymphoid clonality 180.181,183.184.190Lymphoid clonalityanalysis 180,181,183,

190Lymphoma 26.32,33,37,39,40,54,103,

180,185,188.189,190

M

Mechanism 1.3,5-7,9,11 .16,24,25,27,32-37,39,46,59,60.62-64.68,69.76,79,83.90,93,95-99,103-105.107,109113.116.119.123.126,128.133-135,137-144,150.151 ,153.157,158.162,168,171,173,174,176.186.188,190

Microhomology 37,38Mutagenesis 4,8,16,19,21,24,25,52, 117

N

NoncodingRNA transcription 61-63.69Non-Hodgkins lymphoma(NHL) 180,

185,188

197

Nonhomologousend joining (NHEJ) 2, 3.16-20,22.26.27,32-34.36-38.46,51.53,54.116.158.170

alternative 32,37,38deficiency 37

Nuclearcompartmentalization 109,133.134-136.144

Nuclearlamina 135,137,139-141Nuclearorganization 67. 138Nucleosome 4,35,62,66,76,79,95-97,99,

107,108,110,124.134,135.138,139,153,159,160

Nucleosome phasing 159,160

oOct-I 138, 139Omenn syndrome(OS) 48-50Oncogenesis 16,19,22.24,32-37,117.180,

181,1~186,188-190

p

Pathway 2.3.16,18-24.26,27.32.34,3639.46,47,122.127.133,134,138,139.142-144.167,168,170

Pax5 10,35,67-69.77,79.80,94.99,134,135,139-142,151 ,153.160

PDbl 95,99Pericentromeric heterochromatin 69, 112,

127.135,136.142.144Peripheral Tdymphocyre 82. 85PHD finger 35,96,99, 161,162Plant homeodomain 27,35.108,158.161Post-cleavage complex 9Pre-B-cell checkpoint 141Pre-BCRsignaling 141-143Pre-To 150,151Pre-Ta expression 151Promoter 22.32.35,60.62-65,79,80,90,

93-99,106-111,117-119,121-123,127,135.138-140.149,153,159,188,189

Proteasomal degradation 158Protein-DNA complex 3,38


RNA-FISH 63,64RNApolymerase II 63RORyt 151,153RS-SCm 47,49,51-54

scm with microcephaly 49Secondary IgL rearrangement 1SOSevere combined immune deficiency (SCm)

37,46-54Size and diversity T cellreceptor repertoire

90Skp2-SCF 158Stat5 94,139,141,143Surrogate lightchain 104, 134,150,151Synaptic complex 1-7,9-11,33,36

s

Ragl and Rag2leaky mutation 48Rearrangement 10,11,24,32-36,40,46-49,

69,73-80,82-90,94,95,97,98,104,105,107,109-112,116-119,121,122,125-127,133,134,137-144,148-153,1~1~1~1~16~1~18~1~

188-190Recombinase 6,16,17,33-37,60,63,68,

79,93,94,97,99,103-109,111-113,116,117,119,123,128,134,135,137139,141,143,149,157,159 -162,166,167,176,180,181,184-190

Recombination 1-6,8-11,16-20,22,24 ,27,32~,46-48,50-54,59-69,73-79,82

85,87,90,93-95,97-99,103-113,116,117,119,121-128,133-144,148-153,157-162,166-176,180,186

Recombination activating gene(RAG) 1-11, T16-27,32-40,48,59,60,76,79,80,94, Target capture 19,20,24-2795,99,104-106,108,113,123,134, Target site 19,21,24,25,27,35,158135,137,140-143,148-151,157-162, T-B-SCm 47-50166-168,170·174,176,186,188,190 T-cellreceptor{TCR) 1,10,16,32-34,39,

expression 149,151 46,48-50,59,60,74-76,82-85,87,90,transposon 167 98,103-106,108,109,111,116,118,

Recombination activating gene1 (RAG-I) 126,127,148-153,166,168,171 ,172,1,3-9,11,16-21,24·27,33,35,37-39, 180-19046-51,60,77,79,104,116,117,122, gene 16,32,75,90,103-106,149,150,123,133-135,139,149,150,157,159, 153,168,187167,168,186,188·190 loci 34,50,60,74,76,98,106,108,109,

Recombination activating gene2 (RAG-2) 148-150,152,153,181,185,186,1,3-5,7,9,10,16,17,19,24,25,27,33, 18935-39,46-51,60,77,79,99,104,108, TCRa 104,106,107,111,149·151,153116,127,128,133-135,139,149, ISO, TCRf3 10,34,36,76,83,94,95,104-157-159,161,162,167,168 106,109-113,117,119-121,126,

C-terminus 24,25,27,99 127,149-151,159,160,172,182,deficiency 39,47,48 184,185,188-190

Recombination signal sequence (RSS) 1-11, Tcra 60,66,69,97,98,117,121,12616-21,23,24,26,33-36,46 ,59,60,73- TCRAD locus 82,83-86,88,8976,78·80,93,94,95,98,99,104,105, Tcrb 60,69 ,94,95,99,116·119,121·128108,110,111,116-121,124,126,128, TdT expression ISO134,135,137-140,149,157,159,160, T early a (TEA) 90,97-99166-170,175,186,188 Thymocyte 34,48-50,94,95,104,108-112,

Recombination substrate 9,51,74,75,94 116,119,123-127,151,153,158,174Regulation 16,24,25,27,32·34,40,52,59, Transcription 10,11,33,35,61-69,73,

60,62,64,85,90,93,97,98,103,104, 76-80,84,90,93-99,105-107,109-113,107-109,111·113,116,117,123,126, 117,119-124,127,133-144,148-151,128,136,141-143,148-153,157·159, 159,160,162,174161,162,168,189

R

Index

Transcriptionalelongation 96-99Transcriptionalinterference 98, 128Transcriptionalregulation 153Transcription factor (TF) 10,11,35,62,67,

73,77-80,94,96,110,119,122,123,127,133-144,148,151,160

Transcription factory 68Transcription terminator 97, 98Transposition 8, 11, 16, 18-27Transposon 6,24,27,167

vVariation 33,73, 74, 79, 80, 87,117,159,

171V(D)] 16-19,22,24,26,27,32-40,46-54,

59-62,~67,69,75,79,83,93,94,97,

99,103-109,111-113,116,117,119,122-125,126,128,134,135,138,144,148-153,157-162,166-168,170-173,176,180,181,185,186,188,190

199

V(D)J rearrangement 2,10,11,75, 122,172,173,176,181,185,188

V(D)] recombination 1-6,8-11,16-19,22,24,27,32-40,46-48,50-54,59-62,~

67,69,79,83,93,94,97,99,103-109,111-113,116,117,123-126,128,134,144,148-153,157-162,166,168,170,171,173,176,180

VH gene 10,60,62-65,67-69,73-75,77,79,94,112,1~141,169,172,173

VKgene 94, 111V region 60,63,64,66-69,82,84,166,171

X

XRCC4 2,3,17,18,26,27,36-38,51,53,54

y

ITI 67-69,79,134,136,140, 141

Documents

V(D)J Recombination (Advances in Experimental Medicine and Biology Vol 650)