Assembly of IgH CDR3: Mechanism, Regulation, and Influence on Antibody Diversity

Intern. Rev. Immunol. Vol. 8 , 1992, pp. 123-133 Reprints available directly from the publisher Photocopying permitted by license only O 1 9 9 2 Hanvood Academic Publishers GmbH Printed in the United States of America

Assembly of IgH CDR3: Mechanism, Regulation, and Influence on Antibody Diversity LINDA VANDYK and KATHERYN MEEK The Harold C. Simmons Arthritis Research Center and the Department of Internal Medicine, Division of Rheumatology, University of Texas Southwestern Medical Center, Dallas, Texas 75235, USA

The most variable portion of immunoglobulin molecules is the third complementarity determining region (CDR3 J of the heavy chain. This is simply because CDR3 encompasses the region of the rearranged gene where the three gene segments (V,-D,-J,) are joined. Since imprecisions exist in the recombinase reaction, significant differences can be generated at the sites of recombination. This results in the generation of antigen receptor molecules which can differ in their antigen specificity even though they derive from the same germline information. In sum, the significance of the inaccuracy in recombination is that antibodies which are reactive to different antigens can be derived from identical genetic information. This explains how the immune system (using only a limited amount of genetic information) can generate antibodies to virtually any antigen.

Though the basic phenomenon of V,-D,-JH assembly has been appreciated for years, two recent findings demonstrate that the generation of CDR3 is more complex than originally believed. First, junctional modification is not a stochastic process as was initially presumed, but is in part developmentally regulated. Second, it has now been well documented that more complex recombinations (for example V,-DH-D,-J,, V,-D,-invD,-J,, etc.) are involved in generating the third hypervariable region of the heavy chain. Not only do these unusual rearrangements-which break the so-called “12123” recombination rule-occur, but interestingly, certain predicted rearrangements (even some which do follow the “12123” recombination rule) cannot be demonstrated and apparently do not occur. To date, there is no adequate explanation for the lack of these predicted recombinations. These results have important implications for both the generation of antibody diversity and the recombinase reaction itself.

KEYWORDS: immunoglobulin, recombination. complementarity determining region, exonuclease, terminal deoxynucleotidyltransferase

INTRODUCTION

The rearrangement process of immunoglobulin gene segments in B cells is an ordered process in that D,-J, rearrangement (usually on both alleles) precedes V,-D,J, rearrangements (for review see reference 1 and references therein). Isolated VH-DH arrangements are generally not found. V, to D,J, rearrangements are thought to continue until a functional rearrangement is generated on one allele.

The recombination reaction involves two double-stranded DNA cuts and subsequent re- ligations. This results in the formation of two new DNA joints -coding joints (which contain the coding information) and signal joints (which contain the two recombination

123

Int R

ev I

mm

unol

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 11

/28/

14Fo

r pe

rson

al u

se o

nly.

124 L. VANDYK AND K. MEEK

signal sequences, RSS). Coding joints are modified by several processes prior to ligation resulting in remarkable variation at the sites of VH-DH and D,-J, joining.

Significant deletions (up to twelve nucleotides or more) of the coding region sequences can occur prior to coding sequence ligations [2]. This is presumably mediated by an exonuclease though no candidate enzyme has been isolated. In contrast, the two heptamer sequences are ligated without (or with very little) modification [3, 41.

It has recently been hypothesized that the two nucleotides of the 5’ ends of the DNA strands in the coding segments are excised and then ligated to the opposite DNA strand generating a short 3‘ overhang which is palindromic (termed “P” nucleotides for pal- indrome nucleotides). This 3 ‘ overhang provides a suitable template for exonuclease activity [ 5 ] . Usually, nucleotides are cleaved from one or both coding strand ends, but occasionally, no “nibbling” occurs and the overhang is converted to double stranded DNA (presumably by a DNA polymerase) and the two coding strands ligated. In this situation the P nucleotides are retained in the variable region exon. Numerous recent reports have substantiated the P nucleotide theory [6-91.

Finally, non-germline encoded nucleotides (“N” segments) are often found at the coding joint juncture probably due to the action of the enzyme terminal deoxynucleotidyltransferase (Tdt) [lo]. This enzyme is expressed in both rearranging T and B cells. N segments are often GC rich-a reflection of Tdt’s preference for certain nucleotides as substrates. It has long been appreciated that terminal deoxynucleotidyltransferase is not expressed equivalently in all rearranging lymphocytes. Since expression of Tdt in developing B cells diminishes drastically from the time of heavy chain rearrangement to the time of light chain rearrangement, K and A chains, unlike the other immune receptor molecules, generally lack N segments [ll].

REGULATION OF JUNCTIONAL MODIFICATION AT THE MURINE IGH LOCUS

The first suggestion that junctional modification of immune receptor molecules might be developmentally regulated was by Lafaille et al. when they noted that fetal and newborn T cell receptors had only limited junctional diversity in comparison to adult T cell receptors [5]. Since then, numerous laboratories have examined the impact of developmental regulation on junctional diversity in B cells [6-91.

We have approached this question by analyzing bulk populations of rearrangements amplified by PCR. In most situations, we have amplified rearrangements which should not be affected whatsoever by antigen selection-i .e., DH-JH intermediates or rearrangements involving pseudogenes. Thus, these experiments examine the characteristics of immune receptor recombination and not actual expressed antibody repertoires.

Analysis of Exonuclease Activity

We have analyzed exonuclease nibbling of coding segments within bulk populations of PCR amplified rearrangements by utilizing the restriction enzyme sites which fortuitously occur near the recombination signal sequences in various immunoglobulin gene segments (6, and below). By amplifying specific rearrangements from spleen DNA, and subsequently digesting the amplified DNA with the appropriate restriction enzyme, the proportion of the pool of amplified rearrangements which lack a particular enzyme site can be determined.

Int R

ev I

mm

unol

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 11

/28/

14Fo

r pe

rson

al u

se o

nly.

ASSEMBLY OF IgH CDR3 125

An example of this type of experiment is demonstrated in Figure 1 where exonuclease nibbling of the V, segment in VH-DHJH rearrangements has been examined. Specifically, VH33 to JH3 rearrangements were amplified and then digested with Rsa I. VH33 is a member of the 5558 family and is an NPb-like gene. To ensure that only a single V, segment is amplified, the oligonucleotide primer was made complementary to a unique region of the gene (CDR2), and an annealing temperature just below the melting point for the oligonucleotide was used in the PCR. Furthermore, there is a second Rsa I site within CDR2 (immediately following the position of the PCR primer) which is fairly specific for

Rsa I VH33 N D N JH3

L

1

I i 24bp 117bp -1 11 bp?

I I

-228bp

FIGURE 1 Top, Map of amplified VH33-DHJH3 rearrangements, Positions of Rsa I restriction sites are indicated by the vertical arrow. Bottom, Southern hybridization experiment of VH33-DHJH3 amplifications from newborn and 10 week spleen DNA-either uncut or restricted with Rsa I. Positions of the -228 bp and 117 bp fragments have been indicated with arrows. The hybridization probe was an oligonucleotide complementary to the third framework of VH33

Int R

ev I

mm

unol

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 11

/28/

14Fo

r pe

rson

al u

se o

nly.


this particular gene. Thus, the presence of the second Rsa I site should confirm that only VH33-DHJ, rearrangements are being amplified as well as verifying that the restriction digest is complete.

This particular V, segment was selected not only because it has an appropriate restriction enzyme site (Rsa I) close to the recombination signal sequence, but also because it is a pseudogene by virtue of an internal stop codon. Therefore, this gene should recombine normally, but should not be subject to antigen selection. In Figure 1, newborn and 10 week mouse spleen DNA has been amplified for vH33 to JH3 rearrangements, digested with Rsa I, and hybridized to a JH coding probe. The last nucleotide of the Rsa I site is five nucleotides from the recombination signal sequence. In both adult and newborn VH-DHJH rearrangements, the Rsa I site (represented by the presence of the 228 bp band) is occasionally absent-indicating that the exonuclease occasionally removes at least five nucleotides from the VH coding segment- which is somewhat surprising because the last two amino acids are almost always maintained in expressed heavy chain variable regions. There is no difference in the proportion of products lacking the Rsa I site when comparing newborn versus adult rearrangements. Since the amplified rearrangements were not subject to antigenic selection, the absence of the enzyme site should directly reflect the degree of exonuclease activity which occurred when the particular gene segments recombined. We have examined various DH-JH rearrangements and V,-J, rearrangements in a similar fashion and our data demonstrate that the deletion mechanism affects all types of immunoglobulin recombination -rearrangements occurring early (heavy chain) and late (light chain) in lymphocyte differentiation, and rearrangements from both newborn and adult lymphocytes. Hence, the enzyme responsible for nucleotide deletion is either an integral part of the recombinase machinery or its expression is coincident with that of the recombinase.

Analysis of Terminal lkansferase Activity

To assay for differences in N segment addition in DH-JH rearrangements, polymerase chain reaction experiments have been done using 32P labeled amplification primers. Specifically, DSP2 to JH1 rearrangements from adult spleen DNA were amplified and then cleaved with two restriction endonucleases (Taq I and Rsa I) which cut within the JH1 coding sequence. In Figure 2, asymmetric labeling of PCR products was accomplished using 32P labeled oligonucleotides flanking the germline D, and JH gene segments. The labeled products were subsequently analyzed by polyacrylamide gel electrophoresis.

When amplifying with a labeled J, oligonucleotide and digesting with either Rsa I or Taq I, the digested fragment which is labeled should not be variable since it does not contain junctional sequences. Indeed, when cutting with either enzyme, a homogeneous 86 (Rsa I) or 81 (Taq I) base pair fragment is apparent in the amplifications from both newborn and 10 week old spleen DNA. However, when amplifying with labeled DSP2 oligonucleotide, the labeled fragment will be heterogeneous depending on the amount of N segment addition and exonuclease deletion of nucleotides. In fact, when cutting with Taq I or Rsa I, the digested fragments which are labeled are not homogeneous. The fragments from amplifications of 10 week spleen DNA are consistently larger than are those from newborn spleen DNA, demonstrating that there are more N segment additions in the rearrangements from older mice. Finally, since these experiments have been done on DHJH intermediates (and not

Int R

ev I

mm

unol

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 11

/28/

14Fo

r pe

rson

al u

se o

nly.


DSPP N7 JHi DSPP N7 JHi

w . 81bp . -89bp Taqt * R w i * -84bp 8% . c *

-170bp -170bp , c .

F'IGURE 2 Top, Map of amplified DSP2-1,I rearrangements utilizing either 32 P labeled D, or J, oligonucleotides. Positions of Rsa I restriction sites are indicated by the vertical arrow. Sizes of potential 32P labeled restriction fragments have been indicated with horizontal arrows. The possible sizes of those fragments are as indicated. When utilizing the 32P labeled D, amplification primer, the size of the restriction fragments will be variable depending upon the degree of exonuclease and terminal transferase activity. Thus, the approximate lengths of these fragments have been indicated. Bottom, Polyacrylamide gel electrophoresis of DSP2-JHI amplifications from newborn and 10 week spleen DNA-either uncut or digested with Rsa I or Taq I. Radiolabeled oligomers were utilized as molecular weight markers.

Int R

ev I

mm

unol

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 11

/28/

14Fo

r pe

rson

al u

se o

nly.


expressed immunoglobulins), it is clear that the differential regulation of N segment addition is at the level of recombination and not antigen or repertoire selection.

Implications of Regulated Junctional Modification

Several other laboratories have come to similar conclusions-using different techniques, i.e., by sequencing PCR amplified transcripts or rearrangements [7-91. Thus, the concept that junctional modification is at least partially developmentally regulated is now firmly established. The most conspicuous consequence of developmental regulation of junctional modification is that there is a relative lack of N segment additions in fetal immunoglobulin molecules which will obviously limit to some extent the early fetal repertoire.

An unexpected finding from these studies was the observation that the bias in murine immunoglobulin molecules (at least in fetal and newborn receptors) for one DH segment reading frame occurs at the level of the DH-JH intermediate. The explanation for this bias is the presence of short sequence homologies at the ends of the two coding sequences [7,12]. In instances where there has been no N segment modification of coding strands, the coding strand ligation reactions are facilitated by the presence of short sequence homologies (even a single base pair) in the protruding single strands-effectively forcing the utilization of a single DH segment reading frame. This results in an even greater limitation of the early fetal repertoire. There is also biased DH segment reading frame utilization in adult mice [14], but this is accomplished via a different mechanism [16].

Compounding the apparent restriction of the fetal antibody repertoire is that only certain VH segments are employed by fetal antibodies. It has been well documented that there is overexpression of the VH segment genes which are located relatively JH proximal 115, 161. Furthermore (though there is some controversy concerning this point), there also appears to be skewing of DH and JH utilization in fetal and newborn antibodies [7-91. In sum, the fetal repertoire is effectively restricted by three different mechanisms: (1) limited utilization of germline gene segments (VH, as well as D, and JH); (2) limited N segment addition; and (3) utilization of a single DH segment reading frame.

To date, it is unclear what the consequence(s) or purpose of this relative lack of diversity is. However, another well documented feature of the fetal antibody repertoire is the high incidence of autoreactive clones. It seems possible, if not probable that these properties are in some way related. Determining whether or not there is any connection between these two characteristics of the fetal antibody repertoire remains to be elucidated, and is currently an area of active research.

NOVEL REARRANGEMENTS AT THE MURINE IGH LOCUS

During our studies of expressed antibody sequences we observed that very often CDR3 cannot be easily explained by simple VH-DH-JH recombination with the addition of random N segments at the VH-DH or DH-JH joints. Instead, the core of CDR3 could be explained by a germline DH element, but the N segments were atypical in that they were very long and not GC rich as would be expected from terminal transferase activity [17]. In fact, the N segments had some homology with other germline D, elements. Thus, we proposed that DH-DH rearrangements sometimes occur. Theoretically, DH-DH rearrangements should be precluded by the 12/23 recombination rule. We formally demonstrated that these rearrangements actually occur by isolating both direct and inverted DH-DH rearrangements via polymerase chain reaction experiments [HI. Since that time, there have been numerous

Int R

ev I

mm

unol

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 11

/28/

14Fo

r pe

rson

al u

se o

nly.


reports of expressed antibodies recognizing a variety of antigens and derived from a number of different species corroborating the concept that VH-DH-DH-JH rearrangements contribute to the generation of antibody diversity.

We have extensively analyzed rearrangement at this locus, and our data have implications for the rearrangement process itself. Theoretically, if inverted rearrangements and rearrangements which break the 12/23 rule are considered, there are numerous types of rearrangements that can occur at the IgH DH-JH locus. These include regular 12/23 D,-J, joinings (either direct or inverted), hybrid D,-J, joinings [ 191, and various DH-DH joinings. Though many “novel” rearrangements are easily demonstrable by PCR, certain predicted rearrangements are more difficult to detect and apparently do not occur. For example, since DH gene segments are flanked by equivalent recombination signal sequences, there is no apparent reason why DH-JH rearrangement should not occur by either direct or inverted rearrangement. However, as demonstrated below, inverted D,-J, rearrangement (a rearrangement which adheres to the 12/23 recombination rule) is extremely rare if not completely absent.

One difficulty in examining rearrangements involving DH segments is that since DH segments are flanked by RSS, every pair of amplification primers can potentially amplify two different joints. For example, when amplifying for inverted D,-J, rearrangements, hybrid joints of direct DH-JH rearrangements are also potentially amplified (illustrated in Fig. 3A). Still, inverted DH-JH joints and hybrid joints of direct D,-J, recombination can be distinguished by hybridization to a DH coding probe, since the hybrid joint would not contain the DH coding region. The actual product from such an amplification hybridizes to a J, coding probe, but not to a DH coding probe (lanes 3 and 4, panels 2 and 3). Thus, the amplified rearrangements are hybrid DH-JH joints. This was a surprising result in that hybrid joints are thought to occur at a significantly lower frequency than normal recombinations-especially those which maintain the 12/23 recombination rule [19]. Thus to confirm this finding, we cloned and sequenced multiple products from this type of PCR (utilizing several different D,/J, primer pairs) and indeed all of the recombinations represented hybrid D,-J, joints; we found no inverted D,-J, joints (data not shown). Similarly, when amplifying for the signal joints of potential inverted D,-J, recombinations, hybrid joints of direct D,-J, recombinations are also potentially amplified. These two products should be distinguishable not only by hybridization to DH coding probes, but also by sensitivity to the restriction endonuclease Sno I. (A Sno I site is created by the ligation of two heptamer sequences.) In this case, the actual product hybridizes to the DH coding probe, but does not hybridize to the J, coding probe, and is not sensitive to Sno 1 (lanes 5 and 6 , panels 2 and 3). Thus, the amplified rearrangements are hybrid D,-J, joints-in fact, they represent the complementary hybrid joints of the joints described above (see Fig. 3A). We find no evidence for signal joints of inverted DH-JH recombinations. In summary, when amplifying D,-J, rearrangements, we detect signal, coding and both types of hybrid joints of rearrangements that utilize the JH proximal D, segment RSS. Although we have repeated these experiments numerous times amplifying for rearrangements utilizing all the different murine D, and J, genes, w e j n d no evidence of DH-JH rearrangements utilizing the 5‘ DH segment RSS-i.e., inverted DH-JH joints. A compilation of these data (and similarly derived data from other combinations of DH-JH rearrangements), is shown in Figure 3C.

From these data, the most obvious conclusion would be that (if given a choice), deletional rearrangement is overwhelmingly favored. There are numerous arguments against this explanation. In vivo, the relative transcriptional orientation of the two rearranging genes dictates whether or not rearrangement is by deletion or inversion. If the two rearranging

Int R

ev I

mm

unol

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 11

/28/

14Fo

r pe

rson

al u

se o

nly.


a m " .

B Probe 5'D RSS Jcod DWd 8 D RSS

Primers Potential Products Hybrldlzation Probes Sno I Conclusion

5DS/3JH dDJcod/lDJhybl + f dDJcod 3DS/3JH IDJcodld DJhybl 5DS/!jJH

Dcod Jcod

+ dDJhyb1 lDJslgldDJhyb2 + dDJhyb2

c 3DW5JH dDJsl@DJhyb2 + dDJslg

FIGURE 3 (A) Schematic representation of the murine D$J, locus and potential rearrangement products. Heavy arrows above boxes depict amplification primers. Light arrows underneath boxes depict orientation of the various gene segments. (B) Southern hybridization data of PCR products which are either undigested or digested with Sno I. Amplification and hybridization primers are as indicated. (C) Summary of D,-J, amplification data. Abbreviations for products: d = deletional; i = inverted; hyb = hybrid; cod = coding; sig = signal.

gene segments are in the same transcriptional orientation, rearrangement is via deletion and the reciprocal joint is deleted from the chromosome on a closed circular DNA byproduct. If the two gene segments involved are in opposite transcriptional orientation, rearrangement is by inversion of the intervening DNA [20]. At the murine IgH locus, most VH gene segments are in the same transcriptional orientation as the JH gene segments and rearrangement is usually via deletion [21]. In contrast, many V, gene segments are oriented opposite of J,, and rearrangement at this locus is often via inversion [22] . There is no evidence suggesting a bias for rearrangement of V, gene segments oriented in one direction over the other orientation. Furthermore, using artificial recombination substrates, several groups have addressed the relative efficiency of direct and inverted rearrangement and have shown that both are efficient processes, though direct rearrangement is two to three fold more efficient than inverted rearrangement [23]. Finally, this apparent bias for deletional rearrangement is not a feature of the murine DH-JH locus, since our analysis of DH-DH rearrangement

Int R

ev I

mm

unol

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 11

/28/

14Fo

r pe

rson

al u

se o

nly.


demonstrates that inverted rearrangement does occur at this locus and that in certain situations, inverted DH-D, rearrangement is actually favored over direct D,-D, rearrangement.

More specifically, we have completed a similar analysis of D,-D, rearrangements as demonstrated above for DH-JH rearrangements. In these experiments, we have found both the signal and coding joints of inverted DH-DH rearrangement when the rearrangement is via the two JH proximal RSS; however, we cannot find either signal or coding joints from inverted DH-DH rearrangements when the rearrangement utilizes the two 5’ or V, proximal RSS. The preference of inverted DH-DH rearrangements which utilize the J, proximal RSS, is apparent for all DH genes tested (DQ52, DFL16, DSP2, as well as human D, genes DLR and DXP).

Direct DH-DH rearrangement has been slightly more difficult to interpret. In our analysis of rearrangements at this locus, direct DH-D, rearrangement is the only situation in which only one of the two joints formed during a particular rearrangement can be efficiently amplified. In this case, the signal joints of the rearrangement are easily demonstrated; whereas it has been virtually impossible to detect direct DH-DH intermediates. Since direct D,-D, rearrangements are observed in expressed antibodies, and since the signal joints of these rearrangements can be demonstrated, these rearrangements must occur. The best explanation for our relative inability to amplify direct D,-D, coding joints is the following. These rearrangements probably occur, such that an unrearranged DH segment rearranges to a D,-J, intermediate (i.e., the 3’ RSS of one DH segment rearranges to the 5’ RSS of a second, only after the 3’ RSS has rearranged to JH). Thus, the amplified reciprocal joints result from D, to D,-JH rearrangements. Obviously, in this situation, the coding joint would be amplified along with “normal” DH-JH rearrangements.

The observations that both coding and signal joints of inverted DH-DH rearrangements via the 3‘ RSS are easily demonstrable, direct DH-DH (not DH-DHJH) rearrangements are virtually undetectable, and inverted D,-D, rearrangements via the 5’ RSS are completely undetectable argue that the recombinase machinery is not simply biased in favor of direct (over inverted) rearrangement. Instead, these experiments argue that at the time when this locus is undergoing rearrangement, there is an inherent preference for utilization of the J , proximal D, segment.

One possible explanation for this bias is that there are factors involved in the recombination process which bind DNA motifs other than the recombination signal sequences. For example, it is known that germline D, segments have conserved promoter-like sequences which function in the production of D,-J, transcripts and in some cases DHJHk proteins [24]. These D,-J,p proteins may deliver a negative signal to developing B cells [14]- although the exact function of these proteins is unclear. The conservation of the D H

promoters implies some important functional role for these sequences. One possibility is that these promoters facilitate D,-J, recombination. Indeed, numerous studies (including studies of this locus) have demonstrated that transcription of rearranging gene segments precedes recombination [25]. The asymmetric position of the DH promoters may be responsible for the extreme bias of direct (versus inverted) rearrangement. It is possible that DNNprotein interactions involving these sequence elements may in some way hinder utilization of the 5 ’ recombination signal sequences. Alternatively, there may be other sequence elements (unrelated to the D, promoters) flanking the D, segments which predicate rearrangement of the 3’ recombination signal sequence first. Finally, even though the two pairs of recombination signal sequences appear by all criteria to be equivalent, it cannot be formally ruled out that subtle differences may significantly affect the relative RSS

Int R

ev I

mm

unol

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 11

/28/

14Fo

r pe

rson

al u

se o

nly.


utilization-though extensive mutational analysis [26] would seem to exclude this possibility.

Recently, we have isolated V,-D, rearrangements and in this case the V, RSS appear to be able to utilize either DH RSS. Thus, there is not an absolute necessity to rearrange the JH proximal RSS first. Instead, these data might suggest a model in which certain factors specifically drive DH-JH rearrangement (possibly by changing local chromatin structure in some way so that the recombinase enzymes have access to recombination signal sequences); these factors would obviously be expressed during DH-JH rearrangement but may be turned off once the V, locus becomes active. Similarly, other factors (expressed only after DH-JH rearrangement) may drive V,-D,J, rearrangement. In sum, these data are compatible with the view that V-D-J recombination is a targeted phenomenon in that to generate functional immune receptor molecules-not only is it necessary for a lymphocyte to express the basic recombination machinery, but it must also tightly regulate the accessibility of each immune receptor locus to that machinery.

References

1. Max, E. Immunoglobulins: Molecular Genetics. In: Fundamental Immunology, Second Edirion, edited by William E. Paul, New York: Raven Press Ltd., 1989, pp. 235-290.

2. Alt, E W. and Baltimore, D. Joining of immunoglobulin heavy chain gene segments. Implications from a chromosome with evidence of three D-JH fusions. Proc. Nutn. Acud. Sci. USA 79, 4118-4122, 1982.

3. Lewis, S., Gifford, A,, and Baltimore, D. DNA elements are asymmetrically joined during the site specific recombination of Kappa immunoglobulin genes. Science 228, 677-685, 1985.

4. Okazaki, K., Davis, D. D., and Sakano, H. Tcell receptor B gene sequences in the circular DNA of thymocyte nuclei: direct evidence for intramolecular DNA deletion in V-D-J joining. Cell 49, 477-485, 1987.

5. Lafaille, J. J., DeCloux, A., Bonneville, M., Takagake, Y., and Tonegawa, S. Junctional sequences of T cell receptor y8 genes: implications for y8 T cell lineages and for a novel intermediate of V-(D)-J joining. Cell 59,

6. Meek, K. Analysis of junctional diversity during B lymphocyte development, Science 250, 820-823, 1990. 7. Gu, H., Forster, I., and Rajewsky, K. Sequence homologies, N sequence insertion and J, gene utilization in

V,DJ, joining: implications for the joining mechanism and the ontogenetic timing of Lyl B cell and B-CLL progenitor generation. EMBO J. 9, 2133-2140, 1990.

8. Feeney, A. J. Lack of N regions in fetal and neonatal mouse immunoglobulin V-D-J junctional sequences. J. Exp. Med. 172, 1377-1390, 1990.

9. Bangs, L. A., Sanz, I. E., and Teale, J. M. Comparison of D, J,, and junctional diversity in the fetal, adult, and aged B cell repertoires. J. Immunol. 146, 1996-2004, 1991.

10. Desiderio, S. V, Yancopoulos, G. D., Paskind, M., Thomas, E., Goss, M. A., Landau, N., Alt, E W., and Baltimore, D. Insertion of N regions into heavy-chain genes is correlated with expression of terminal deoxytransferase in B cells. Nature 311, 752-755, 1984.

11. Kabot, E. A., Wu, T. T., Reid-Miller, M., Perry, H. M., and Gottesman, K. S. Sequences of Proteins of Immunologic Inreresr, 804pp. 4th ed., U.S. Department of Health and Human Services, Public Health Service, National Institute of Health, 1987.

12. Ichihara, Y., Hayashida, H., Miyazawa, S., and Kurosawa, Y. Only Dni6, D,,,, and DQ,,gene family exist in mouse immunoglobulin heavy chain diversity gene loci, of which Dn,6 and D,, onginate from the same primordial D, gene. Eur. J. Immunol. l9, 1849-1854, 1989.

13. Jeong, H. D., and Teale, J. M. Comparison of the fetal and adult functional B cell repertoires by analysis of V, gene family expression. J. Exp. Med. 168, 589-603, 1988.

14. Gu, H., Kitamura, D., and Rajewsky, K. B cell development regulated by gene rearrangement: arrest of maturation by membrane-bound D p protein and selection of D, element reading frames. Cell. 65, 47-54, 1991.

15. Yancopoulos, G. D., Desiderio, S. V, Paskind, M., Kearney, J. E, Baltimore, D., and Alt, E W. Preferential utilization of the most J, proximal V, gene segments in pre-B-cell lines. Nature (Lond.) 311, 727-733.1984.

16. Decker, D. J., Boyle, N. E., Koziol, J. A., and Klinman, N. R. The expression of the IgJ3 chain repertoire in developing bone marrow B lineage cells. J. Immunol. 146, 350-361, 1991.

17. Meek, K., Hasemann, C., Pollok, B., Alkan, S. S., Brait, M., Slaoui, M., Urbain, J., and Capra, J. D. Structural characterization of anti-idiotypic antibodies: evidence that Ab2s are derived from the germline differently than Abls. J. Exp. Med. 169, 519-533, 1988.

859-870, 1989.

Int R

ev I

mm

unol

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 11

/28/

14Fo

r pe

rson

al u

se o

nly.


18. Meek. K. D., Hasemann, C. A,, and Capra, J. D. Novel rearrangements at the immunoglobulin D locus: inversions and fusions add to somatic diversity. J. Exp. Med. 170, 39-57, 1989.

19. Lewis, S. M., Hesse, J. E., Mizuuchi, K. , and Gellert, M. Novel strand exchanges in V(D)J recombination, Cell. 55, 1099-1107, 1988.

20. Desiderio, S. V, and Wolff, K. R. Rearrangement of exogenous immuno-globulin V, and DJ, gene segments after retroviral transduction into immature lymphoid cell liens. J. Exp. Med. 167, 372-389, 1988.

21. Cory, S., and Adams, J. Deletions are associated with somatic rearrangement of immunoglobulin heavy chain genes. Cell. 19, 37-46, 1980.

22. Van Ness, B. G., Weigert, M., Coleclough, C., Mather, E. L., Kelley, D. E., and Perry, P. R. DNA between variable and joining gene segments of immunoglobulin K light chain is frequently retained in cells that rearrange the K locus. Proc. Natl. Acad. Sci. USA 79, 262-266, 1982.

23. Hesse, J. E., Lieber, M. R., Gellert, M., and Mizuuchi, K. Extrachromosomal DNA in pre-B cells undergo inversion of deletion at immunoglobulin V-(D)-J joining signals. Cell. 49, 775-783, 1987.

24. Reth, M. G. and Alt, M. W. Novel immunoglobulin heavy chains are produced from DJH gene segment rearrangements in lymphoid cells. Nature 3l2, 418-423, 1984.

25. Alessandrini, A,, and Desiderio, S. V Coordination of immunoglobulin DJ, transcription and D-to-J, rearrangement by promoter-enhancer approximation. Mol. Cell. Biol. 11, 2096-2107, 1991.

26. Hesse, J. E., Lieber, M. R., Mizuuchi, K., and Gellert, M. V(D)J recombination: a functional definition of the joining signals. Genes Dev.3, 1053-1061, 1989.

Int R

ev I

mm

unol

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 11

/28/

14Fo

r pe

rson

al u

se o

nly.

Documents

Assembly of IgH CDR3: Mechanism, Regulation, and Influence on Antibody Diversity