Identification of selected small molecule targets in living cells using
Capture Compounds™ through a bio-orthogonal chemical ligation method
Inaugural Dissertation
to obtain the Academic Degree
Doctor rerum naturalium (Dr. rer. nat.)
Submitted to the Department of Biology, Chemistry and Pharmacy
of the Freie Universität Berlin
by
João André Banha Oliveira
from Lisbon, Portugal
November, 2015
2
The following experimental work was developed between December 2010 and August 2014 atcaprotecbioanalyticsGmbHunderthesupervisionofProf.Dr.HubertKöster
1stReviewer:Prof.Dr.HubertKöster2ndReviewer:Prof.Dr.MarkusWahlDateofdefense:11April2016
3
AcknowledgmentsIwouldliketoexpressmyappreciationtoProf.HubertKöster,FounderandManagingDirector
of caprotec bioanalytics GmbH, for recruiting me and giving me this challenging but excitingopportunity towork in a start-upenvironment. Thank you for your support andpositive feedbackduringourprojectdiscussions.
I would also like to acknowledge Prof. Markus Wahl at the Freie Universität Berlin for hisacceptancetoco-supervisemyresearchprojectandaidmeinnavigatingthebureaucracy.
AspecialthankyouisdedicatedtoDr.MathiasDreger,HeadoftheBiochemistryDepartmentatcaprotecthatguidedmethroughoutthisbumpyroad.Ideeplyappreciateyoualwayshavingyourdooropenfordiscussionandgivingyourbesttomotivatemetomoveforward.Ialsoneedtothankyouforyourcommitmentinreviewingthisdissertationandsharingyourthoughts.
Toallmycolleaguesat caprotecbioanalytics, some thatwere there from thebeginningandothers thatcameandwent, thankyou for thesupportaroundthe labandmeetingsbutespeciallythankyouforthefunnymoments.Laughteralwaysmakesthejobeasier.Youmademefeel lessofan outsider addressing me in English but still teaching me some German. I will never forgetLuftfeuchtigkeitandStreichholzschächtelchen.
ThereareafewcolleaguesIneedtosingleoutinmyacknowledgements.LisavonKleistfortheveryhelpfulbrainstormingsessionsbutmostofallforyourpersonalityandworkethics.YanLuoforgettingme started in the lab and later collaborating in thephenobarbital project. The guys at theMedicinal Chemistry Lab, SimonMichaelis, Frank Polster and Daniel Ohlendorf, as well as HenrikDieks for synthesizingall compounds Iused in thiswork.MatthiasHakelbergandChristianDalhofffromtheAnalyticsDepartmentforthecontinuoussupportinconfirmingthestabilityandqualityofmycompounds.KathrinBarthofortakinggoodcareofmymassspectrometrysamplesandanalyzingthemasquicklyaspossibleaswellasthefriendlydiscussionsaroundExceltablesonhowtoextractthe information from theMSdata. Finally,Michael Sefkow for the supportandenthusiasm in thiswholeproject.Thankyouall.
This work was supported by the Initial Training Network, Bio-Orthogonal Chemo-SpecificLigation (BioChemLig), funded by the FP7Marie Curie Actions of the European Commission (FP7-PEOPLE-2008-ITN-238434). Within this consortium I had the opportunity to establish greatconnectionstomyfellowcolleagues,whoIwouldliketosaluteforthefriendshipcreatedduringthemeetingsandputtingupwiththeonebiologistinthemidstofallchemists.ToMarcoandValentina,who chose to come toBerlin for their secondments, thank you for lettingme teach youaroundabiochemistrylab.Itrulyenjoyedit.Also,totheP.I.softhegroups,thankyouforyourcommentsandguidanceintheprogressreportsmeetings.
A warm appreciation goes out tomy family who have supportedme inmy journey acrossbordersandawayfromhome.Alwaysconcernedbutalsomotivatingmetokeepongoingandaimforhappiness.
Finally,IwouldnothavecomethisfarwithouttheloveandsupportofmypartnerChris.Thesepast twoyearshavebeenachallengebutablessingbecauseofyou.Weovercomeour limitationsdrawingstrengthfromeachotherandthatmakesusundefeatable.
ThankyouformakingHomewhereverandwheneverweare.
4
PrefaceTheoverallaimofthisstudywastheintroductionofanovelapproachfortheidentificationof
small-moleculetargetsfromlivingcellsusingcapturecompoundmassspectrometry.Toachievethisgoal, the research was performed in a step-wise comprehensive manner that is reflected in thestructureofthisdissertation.
A general introduction covers the subjects underlying all the work that was developed,followed by three chapters of specific research, each containing an explanatory introduction, theobtainedresultsandthespecificmethodsusedforthoseexperiments.
Thefirstspecificchapterdescribestheestablishmentofthenewmethodologyandnecessaryoptimization steps, which were done in a simple system containing the small molecule capturecompound and a purified recombinant protein that are known in the literature to interact, in thiscasethecyclicAMPandproteinkinaseA.Thefollowingchapteradvancestheestablishedprotocoltobe used with a clinically relevant drug, dasatinib, in order to identify targets in more complexsamples,suchascell lysates,and implementthe initialparameters forcapturing in livingcells.Thelast of the experimental chapters explains how this novel workflow can shed some light on thehepatotoxic side-effect of phenobarbital, an old anti-epileptic drug. Finally, a summary of thecomplete research is outlined, where the achievements are discussed and future prospectspresented.
Thisresearchprojectwas integrated inaEuropeanUnionMarieCurieActions InitialTrainingNetwork – BioChemLig. Within this consortium we had regular meetings where progress reportswerepresentedandseveralscientificandsoft-skillsseminarswereoffered. Ipresentedatotalof6communications in the3-yearperiodof theproject.Additionally, Ihad theprivilege tocollaboratecloser and host at caprotec two fellow members of BioChemLig: Valentina Bevilacqua, from thegroupofFrédéricTaranattheCEASaclayinParis,whosynthesizednewcopper-chelatingmoleculesthat were introduced in some of the capture compounds I used for my research; and MarcoBartoloni,fromthegroupofJean-LouisReymondattheUniversityofBern,whomItaughtandaidedinthecaptureoftargetsfromcelllysatesusinghisbicyclicpeptides.
The collaboration with the University of Bern resulted in a publication that I co-authored:Bartoloni,M.,Jin,X.,Marcaida,M.J.,Banha,J.etal.(2015)Bridgedbicyclicpeptidesaspotentialdrugscaffolds:synthesis,structure,proteinbindingandstability.ChemicalScience6(10):5473-5490.
DuringthecourseofthisresearchprojectIwasabletoparticipateinanEMBOPracticalCourseinChemicalBiologyfrom27Marchto2April2011attheEMBLinHeidelberg,Germanyandattendthe3rdEuropeanChemicalBiologySymposiumthattookplace1-3July2012,inVienna,Austria.
Youseethings;andyousay“Why?”ButIdreamthingsthatneverwere;andIsay“Whynot?”
GeorgeBernardShaw
5
ContentsListofTables.....................................................................................................................................7ListofFigures....................................................................................................................................8Abbreviations................................................................................................................................10Summary.........................................................................................................................................11Zusammenfassung........................................................................................................................121.GeneralIntroduction..............................................................................................................131.1.ChallengesinDrugDiscovery.............................................................................................................................131.2.Drugtargetdeconvolution...................................................................................................................................15
1.2.1.Insilicoapproach.....................................................................................................................161.2.2.AffinityChromatography.........................................................................................................161.2.3.Activity-basedproteinprofiling(ABPP)...................................................................................171.2.4.Capturecompoundmassspectrometry(CCMS)......................................................................18
1.3.Clickchemistryandbio-orthogonalligations..............................................................................................211.3.1.Staudingerligation..................................................................................................................221.3.2.Copper-catalyzedazido-alkynecycloaddition(CuAAC)...........................................................231.3.3.Metal-freeclickcycloadditions................................................................................................25
2.Labelingaproteinusingclickchemistry.........................................................................282.1.Introduction................................................................................................................................................................282.2.Results...........................................................................................................................................................................31
2.2.1.Establishingandoptimizingclick-captureofPKARIusingacAMPCC.....................................312.2.2.MassSpectrometryanalysisofcapturedproteinsfromHEK293andHepG2lysatesusingaclassicalandclickablecAMPcapturecompound..............................................................................42
2.3.SpecificMethodsforthecAMP/PKARIsystem...........................................................................................482.3.1.Structuresofusedcompounds................................................................................................482.3.2.Establishingandoptimizingclick-captureofPKARIusingacAMPCC.....................................502.3.3.PreparationofHEK293lysates................................................................................................502.3.4.ClickcaptureofcAMP-bindingproteinsfromcelllysates........................................................53
3.IdentifyingDasatinibtargetsincells................................................................................543.1.Introduction................................................................................................................................................................543.2.Results...........................................................................................................................................................................58
3.2.1.Introducinganewcopper-chelatingazideintheCuAACreaction...........................................583.2.2.CapturingdasatinibtargetsinK562celllysateusingCuAAC..................................................603.2.3.CellviabilityassayfordasatinibcompoundsinK562cells......................................................683.2.4.CapturinginlivingK562cells...................................................................................................70
3.3.Specificmethodsforthedasatinibsystem....................................................................................................773.3.1.Structuresofusedcompounds................................................................................................773.3.2.OptimizingCuAACreactionparameterswithnewcopper-chelatingazide.............................783.3.3.CapturingdasatinibtargetsinK562celllysateusingCuAAC..................................................793.3.4.CellviabilityassayfordasatinibcompoundsinK562cells......................................................793.3.5.CaptureofdasatinibtargetsinlivingK562cellsusingCuAAC................................................80
6
4.Phenobarbital:newbindersforanolddrug..................................................................814.1.Introduction................................................................................................................................................................814.2.Results...........................................................................................................................................................................84
4.2.1.Determiningpermeabilityofphenobarbital-alkynecapturecompounds................................844.2.2.Developingaphenotypicassaytoevaluatephenobarbitaleffect...........................................854.2.3.CapturingphenobarbitaltargetsinMH1C1celllysate.............................................................974.2.4.CaptureofphenobarbitaltargetsinlivingMH1C1cells.........................................................1014.2.5.CapturingphenobarbitaltargetsintissuewithhighEGFRexpression..................................103
4.3.Specificmethodsforthephenobarbitalsystem......................................................................................1054.3.1.Structuresofusedcompounds..............................................................................................1054.3.2.TreatmentofMH1C1cellswithphenobarbital.......................................................................1064.3.3.SubcellularfractionationofMH1C1cells................................................................................1064.3.4.TreatmentofMH1C1cellswithhighconcentrationofphenobarbital....................................1064.3.5.Treatmentofratprimaryhepatocyteswithphenobarbital..................................................1074.3.6.RT-PCR...................................................................................................................................1074.3.7.CaptureofphenobarbitaltargetsinMH1C1celllysate..........................................................1094.3.8.CaptureofphenobarbitaltargetsinlivingMH1C1cells.........................................................1104.3.9.Captureofphenobarbitaltargetsinhumanplacentalysate.................................................111
5.ConclusionsandOutlook....................................................................................................1126.GeneralMethodsandMaterials.......................................................................................1166.1.CellCulture...............................................................................................................................................................116
6.1.1.HEK293cellculture................................................................................................................1166.1.2.K562cellculture....................................................................................................................1166.1.3.MH1C1cellculture..................................................................................................................1176.1.4.Ratprimaryhepatocytesculture...........................................................................................117
6.2.BiochemistryGeneralMethods.......................................................................................................................1186.2.1.Proteinconcentrationdetermination....................................................................................1186.2.2.SDS-PAGE...............................................................................................................................1186.2.3.PolyacrylamideGelStaining..................................................................................................1196.2.4.WesternBlotting...................................................................................................................120
6.3.MassSpectrometry...............................................................................................................................................1216.3.1.Sampleacquisitionanddatabasematchidentification........................................................121
6.4.MolecularStructuresofCompounds............................................................................................................1226.4.1.Capturecompounds..............................................................................................................1226.4.2.Competitors...........................................................................................................................1256.4.3.Ligands..................................................................................................................................125
6.5.Materials...................................................................................................................................................................1266.5.1.Chemicalreagents.................................................................................................................1266.5.2.KitsandConsumables............................................................................................................1276.5.3.Buffers,mediaandsolutions.................................................................................................1286.5.4.Equipment.............................................................................................................................129
References...................................................................................................................................130
7
ListofTablesTable2–1:SummaryofidentificationsfromcAMPcaptureexperimentsinHEK293andHepG2lysates............42Table2–2:ProteinsidentifiedasspecificallycompetedwithcAMPacrossthedifferentcapturemethodsineach
typeofcellularlysate.........................................................................................................................43Table2–3:NumberofproteinsidentifiedascompetedbycAMPincommonbetweenthedifferentsamples...44Table2–4:Calculatedintensityfold-changes(assayvs.competition)andMScountsforidentifiedPKAsubunits
ineachsample...................................................................................................................................45Table2–5:SummaryofidentificationsfromcAMPcaptureexperimentsinHEK293lysates...............................46Table2–6:Calculated intensity fold-changes (assayvs. competition)andcorresponding significancevalue for
identified PKA subunits from capture experiments using full and clickable cAMP capturecompounds........................................................................................................................................47
Table2–7:SummaryofparametersutilizedinPKARIlabelingandcapturingwithacAMPcapturecompoundviaaCuAACreaction...............................................................................................................................52
Table3–1:SummaryofidentificationsfromdasatinibcaptureexperimentsinK562lysate................................60Table3–2: Listof competedand significanthits indasatinib captureexperiments inK562 lysateusingapre-
clickedworkflow................................................................................................................................61Table3–3:Listofcompetedandsignificanthits inDasatinibcaptureexperiments inK562 lysateusingaclick
workflow............................................................................................................................................64Table3–4:ValuesforexpectedtargetsofDasatinibinK562lysateusingaclickworkflow.................................65Table 3–5: Gene Ontology term enrichment from significantly competed proteins in Dasatinib capture
experimentinK562lysateusingaclickworkflow.............................................................................66Table3–6:Effectofdasatiniborclickabledasatinib-alkynecapturecompoundonK562cellviability................69Table3–7:SummaryofidentificationsfromdasatinibcaptureexperimentsinK562livingcellsinnineseparate
experiments.......................................................................................................................................70Table 3–8: Excerpt of potential hits for dasatinib in K562 living cell capture using a click workflow from 9
individualexperiments......................................................................................................................72Table3–9:Top20list(sortedbyincreasingpvalue)ofdasatinib-competedproteinswhencalculatingvaluesfor
allrunsfromallexperiments.............................................................................................................73Table 3–10: Gene Ontology term enrichment from significantly competed proteins in dasatinib capture
experimentinK562livingcellsusingaclickworkflow......................................................................75Table4–1:PAMPAresultsofphenobarbitalalkynecapturecompounds.............................................................84Table4–2:SummaryofidentificationsfromphenobarbitalcaptureexperimentsinMH1C1celllysate...............98Table4–3:Potentialhits forphenobarbital inarathepatomacell line(MH1C1) from3 individualexperiments
usingaMS-basedanalysis...............................................................................................................100Table4–4:SummaryofidentificationsfromphenobarbitalcaptureexperimentsinMH1C1livingcells............101Table4–5:ListofcompetedandsignificanthitsinphenobarbitalcaptureexperimentsinMH1C1livingcells..102Table4–6:Summaryofidentificationsfromphenobarbitalcaptureexperimentinhumanplacentalysate.....103Table 4–7: List of competed and significant hits in phenobarbital capture experiments in human placenta
lysate...............................................................................................................................................103
8
ListofFiguresFigure1-1:DrugResearchandDevelopmenttimeline.........................................................................................13Figure1-2:Schematicrepresentationofanaffinitychromatographypull-downprobe......................................16Figure1-3:Schematicrepresentationofanactivity-basedproteinprofiling(ABPP)probe.................................17Figure1-4:Schematicrepresentationofacapturecompound............................................................................18Figure1-5:Schematicworkflowsforthreedifferenttargetdeconvolutiontechniques.......................................19Figure1-6:Workflowforidentificationofsmall-moleculetargetsinlivingcells..................................................20Figure1-7:Illustrationofbio-orthogonalligation.................................................................................................21Figure1-8:Classical(A)andtraceless(B)Staudingerligation...............................................................................22Figure1-9:Mechanismofthecopper-catalyzedazido-alkynecycloaddition(CuAAC).........................................23Figure1-10:CuAAC-acceleratingligandsofchoice...............................................................................................25Figure1-11:Strain-promotedazido-alkynecycloaddition.ExamplereactionbetweenDIFOandanazide.........26Figure1-12:Tetrazineligation..............................................................................................................................27Figure2-1:cAMPproductionandactivationofPKA.............................................................................................29Figure2-2:LabelingofrecombinantPKARIusingCuAACinaserialdilutionofcAMP-alkynecapturecompound.
...........................................................................................................................................................32Figure2-3:LabelingofrecombinantPKARIusingCuAACinaserialdilutionofcAMP-alkynecapturecompound
withcorrespondingcompetitionsamples.........................................................................................32Figure 2-4: Effect of copper sulfate concentrationon the click labeling of recombinant PKARI using a cAMP-
alkynecapturecompound.................................................................................................................33Figure 2-5: Effect of copper ligand TBTA on the click labeling of recombinant PKARI using a cAMP-alkyne
capturecompound............................................................................................................................34Figure2-6:LabelingofrecombinantPKARIusingacAMP-alkynecapturecompoundatdifferentCuAACreaction
times..................................................................................................................................................35Figure2-7:LabelingofrecombinantPKARIusingCuAACwithdifferentbiotin-azideconcentrations..................37Figure2-8:Click-captureofPKARIusingacAMPalkynecapturecompound.......................................................38Figure2-9:LabelingofPKARIwithacAMP-alkynecapturecompoundclickedtoabiotin-azideusingCuAACwith
BTTEascopperligand........................................................................................................................40Figure2-10:EffectoftimeofCuAACreactiononthelabelingofPKARIwithacAMP-alkynecapturecompound
andaTAMRA-azide...........................................................................................................................41Figure3-1:SchematicrepresentationofthemodulardomainsoftheABLkinases.............................................55Figure3-2:Representationofauto-inhibited(closed)andactive(open)ABLkinases.........................................55Figure3-3:Proteinsinhibitedbykinaseinhibitorswithanaffinityoflessthan30nMforABL1andABL2.........56Figure 3-4: Labeling of recombinant SRC and K562 cell lysate with a clickable dasatinib-alkyne capture
compoundusinganewcopper-chelatingazidewithaTAMRAlabel................................................58Figure3-5:CCMSchartof significantlycompetedhits inDasatinibcaptureexperiment inK562 lysateusinga
pre-clickedworkflow.........................................................................................................................61Figure 3-6: STRING protein interaction analysis of significantly competed proteins in dasatinib capture
experimentinK562lysateusingthepre-clickedworkflow...............................................................62Figure 3-7: STRING protein interaction analysis of significantly competed kinases in Dasatinib capture
experimentinK562lysateusingtheclickworkflow..........................................................................64
9
Figure3-8:CCMSchartofsignificantlycompetedkinasesinDasatinibcaptureinK562lysateusingCuAAC......65Figure 3-9: STRING protein interaction analysis of significantly competed proteins in dasatinib capture
experimentinK562lysateusingtheclickworkflow..........................................................................67Figure 3-10: Scheme of multi-well plate for cell viability assay of dasatinib (Dasa) and dasatinib clickable
compound(CPT392)inK562cells.....................................................................................................68Figure3-11:Luminescencedetectionofcellviabilityassayofdasatinibandclickabledasatinib-alkynecapture
compoundinK562cells.....................................................................................................................69Figure3-12:CCMSchartsofCSK,ALDOC,ABLandBCRproteinsinDasatinibcaptureexperimentinK562living
cellsusingaclickworkflow................................................................................................................74Figure 3-13: STRING protein interaction analysis of significantly competed proteins in dasatinib capture
experimentinK562lysateusingtheclickworkflow..........................................................................75Figure4-1:Luminal(phenobarbital)asitwaspackagedin1940/1945................................................................81Figure4-2:PhenobarbitalinducesCYP2BexpressionviaEGFRinhibition............................................................83Figure4-3:Synthesizedphenobarbital-alkynecapturecompounds.....................................................................84Figure4-4:WesternBlotanalysisofCARinisolatednucleiandcytosolfromMH1C1cells...................................85Figure4-5:EffectofphenobarbitalonCARtranslocationfromcytosoltonucleusinMH1C1cells.......................86Figure 4-6: Effect of phenobarbital onAMPKphosphorylation (pAMPK) inMH1C1 cells treated in serum-free
medium.............................................................................................................................................88Figure 4-7: Comparative analysis of the effect of phenobarbital on AMPK phosphorylation for each time
incubationpointinMH1C1cellstreatedinserum-freeDMEM..........................................................88Figure 4-8: Effect of serum supplementation in AMPK phosphorylation (pAMPK) from non-treated (0 µM
phenobarbital)MH1C1cells...............................................................................................................89Figure4-9:EffectofphenobarbitalonAMPactivatedproteinkinase(AMPK)phosphorylation(pAMPK)inarat
hepatomacelllinetreatedincompletemedium..............................................................................90Figure 4-10: Comparative analysis of the effect of phenobarbital on AMPK phosphorylation for each time
incubationpointinMH1C1cellstreatedinFBS-supplementedDMEM.............................................90Figure4-11:EffectofphenobarbitalinCYP2B1/2detectioninarathepatomacellline......................................91Figure4-12:EffectofhighconcentrationphenobarbitalonthedetectionofCYP2B1/2inMH1C1cellsusingrat
livermitochondriamatrixfractionaspositivecontrolforantibodyreactionquality.......................92Figure 4-13: Effect of high concentration phenobarbital on the detection of CYP2B1/2 and AMPK
phosphorylationinarathepatomacellline......................................................................................92Figure4-14:DetectionofD-aminoacidoxidase(DAO)inpurifiedfractionsandMH1C1lysate...........................93Figure 4-15: Effect of phenobarbital in the detection of D-amino acid oxidase (DAO), CYP2B1/2 and AMPK
phosphorylation(pAMPK)fromratprimaryhepatocytes.................................................................94Figure4-16:AgarosegelelectrophoresisofRNA,cDNAandPCRproducts..........................................................95Figure4-17:EffectofphenobarbitalonCYP2B1mRNAexpressioninarathepatomacellline...........................96Figure4-18:EffectofphenobarbitalonCYP2B1mRNAexpressioninratprimaryhepatocytes..........................96Figure 4-19: STRINGprotein interaction analysis of significantly competed proteins in phenobarbital capture
experimentsinMH1C1lysate.............................................................................................................99Figure 4-20: STRINGprotein interaction analysis of significantly competed proteins in phenobarbital capture
experimentsinhumanplacentalysate............................................................................................104
10
Abbreviations
(p)AMPK (phosphorylated)AMP-activatedkinase IR Infrared
AKAP A-kinaseanchoringprotein LC Liquidchromatography
ATP Adenosinetri-phosphate LTQ Lineartrapquadrupole
BCA Bicinchoninicacid MS Massspectrometry
BSA Bovineserumalbumin nLC Nanoflowliquidchromatography
BTTE Bis(tert-butyltriazoly)ethanol PAGE Polyacrylamidegelelectrophoresis
cAMP Cyclicadenosinemono-phosphate PB Phenobarbital
CAR Contitutiveandrostranereceptor PBREM Phenobarbitalresponsiveenhancermodule
CC Capturecompound PBS Phosphate-bufferedsaline
CCMS Capturecompoundmassspectrometry PCR Polymerasechainreaction
CID Collision-induceddissociation PI Proteaseinhibitorcocktail
CuAAC Copper-catalyzedazido-alkynecycloaddition
PKA ProteinkinaseA
CYP CytochromeP450 RNA Ribonucleicacid
DDM n-Dodecylβ-D-maltoside RP Reversephase
DMEM Dulbecco’smodifiedEaglemedium SA-MB Streptavidin-coatedmagneticbeads
DMSO Dimethylsulfoxide SDS Sodiumdodecylsulfate
DNA Deoxyribonucleicacid TAMRA Tetramethylrodamine
DTT Dithiotreitol TBS Tris-bufferedsaline
ECL Enhancedchemiluminescence TBS-T Tris-bufferedsalinewith0.1%(v/v)Tween
EDTA Ethylene-diamine-tetra-aceticacid TBTA Tris(benzyltriazolylmethyl)amine
EGFR Epidermalgrowthfactorreceptor THPTA Tris(3-hydroxypropyltriazolylmethyl)amine
FBS Fetalbovineserum UV Ultraviolet
HEPES 4-(2-Hydroxyethyl)piperazine-1-ethanesulfonicacid
v/v Volumepervolume
HRP Horseradishperoxidase w/v Weightpervolume
11
SummaryThedevelopmentofnoveldrugtargetdeconvolutionmethodologiesisagrowingneedinthe
pharmaceuticalindustry.Theincreasedregulatoryscrutinydemandsthatbetterandsafermedicinesreachtheconsumermarketand,forthat,acomprehensivecharacterizationofthemultipletargetsandoff-targetsofthedrugearlyinthedevelopmentprocessisakeysuccessfactor.
Chemicalproteomicsadvanceshaveprovidedthetoolstostudydrugmechanismsofactioninamoredirectandunbiasedwaythantarget-basedscreeningstrategies,byanalyzing theeffects inthecontextofthefullproteomefromthetargetcellortissueofinterest.
Capturecompoundmassspectrometry (CCMS) isa robustandversatilechemicalproteomicstechniqueduetotheoriginaldesignoftheprobes.Capturecompoundsaretrifunctionalmoleculesthattraditionallycarryaselectivityfunction,consistingofthesmallmoleculeunderstudy,attachedtoascaffoldbearingatag,suchasafluorophoreorbiotin,andaUV-reactivemoietyprotrudingfromthecentralcorethatcovalentlybindstheprobetothetargetprotein.Althoughcapturecompoundsareveryefficientandspecificintargetingofevenlowaffinityandlowabundanttargets,theirusehasbeenlimitedsofartoworkingwithcellularlysatesortargetingcell-surfaceproteins.
ThemainobjectiveofthisworkwastoadvanceCCMStoimplementandoptimizeworkflowsfortheuseofnewlydesignedpermeablecapturecompoundstoprofilesmallmoleculedrugsinlivingcells,wherethetargetsareintheirunalterednaturalenvironment.Bio-orthogonal“click”chemistryprovidestheprerequisitefortheuseofpermeablecapturecompoundsbearingonlytheselectivityandreactivityfunctions,andasmall“clickable”chemicalhandletobindandcross-linkthetargetsinlivingcellsandafterwardsspecificallyattachthesorting/detectionfunctionbyacontrolledchemicalclickreactionsuchasthecopper-assistedazide-alkynecycloaddition(CuAAC).
Initial reactionparameters for theclickreactionweretestedusingaclickablecAMPcapturecompoundinasolutionofrecombinantPKARI.Theoptimizationeffortsincopperionconcentration,ligandnature,durationandpHofthereaction,azideandalkyneconcentrationsyieldedsubstantialimprovementsfromthestartingpoint.CapturingPKAsubunitsfromcell lysateswiththeoptimizedparametersprovedtobepossibletousethisworkflowbuttheresultswerenotfullysatisfactory.
Acapturecompoundwithanon-naturalselectivityfunctionwasusedtofurtherdeveloptargetidentificationincelllysatesandlivingcellsusingthenewclickreaction.Thekinaseinhibitordasatinibwasprofiledusingthisstrategyandbonafidetargetsweresuccessfullyidentifiedusingtheclickablecompound. ABL, BCR and SRC were specifically identified in lysates and CSK, a known dasatinibbinder,waspositivelyidentifiedforthefirsttimebychemicalproteomicsinlivingcells.
Thenewlyestablishedclickworkflowallowedtobettercharacterizephenobarbital,adrugthatwas developed as an anti-epileptic but elicits unexplained hepatocarcinogenic effects in somespecies. The resultsof captureexperiments suggest that this smallmolecule interactswith severalproteinsintheepidermalgrowthfactorreceptorsignalingpathwayandthattheprimarytargetcouldbethetranscriptionfactorYBX1,andnotthereceptoritselfasitwassuggestedrecently.Acell-basedassaywasdevelopedtovalidatethisfindingbutwasnotfurtherexploredduetotimeconstraints.
In summary, the strategy to identify small molecule targets in living cells using capturecompoundsandbio-orthogonalchemo-selectiveligationmethodswassuccessfullyestablished.Thismethodprovides the tools forabettercharacterizationof the targetprofilesofdrugsandstillhaspotentialtobefurtherinnovated.
12
ZusammenfassungDie Entwicklung neuer Methoden zur Dekonvolution der Profile von Zielproteinen für
pharmazeutische Wirkstoffe ist von wachsender Bedeutung für die Pharmaindustrie. Eineumfassende Charakterisierung der Zielproteine einschließlich sogenannter „Off-targets“ ist einSchlüsselfaktorumbessereundsicherereMedikamenteaufdenMarktzubringen.
Fortschritte in der Chemischen Proteomanalyse bieten Werkzeuge für die direktere undbezüglich der Zielproteine hypothese-freie Untersuchung der Wirkungsmechanismen vonMedikamenten, bei der die Bindung imKontext des vollständigen Proteomsder Zielzelle oder desZielgewebesuntersuchtwird.
Die Capture CompoundMassenspektrometrie (CCMS)-Technologie ist eine robuste und sehrflexibel einsetzbare Technik in der chemischen Proteomanalyse. Capture Compounds sindtrifunktionelleMoleküle,dieinihrerklassischenVersioneineSelektivitätsfunktiontragen,diedemzuuntersuchenden Kleinmolekül-Wirkstoff entspricht, die mit einer Gerüststruktur verbunden ist,welche außerdem eine Isolierungsfunktion sowie eine UV-aktivierbare Reaktivitätsfunktion für diekovalente Verbrückung zum gebundenen Zielprotein enthält. Obwohl Capture Compounds sehreffizientundspezifischZielproteinebindet,warihrEinsatzbislangwegenfehlenderPermeabilitätaufZell-Lysatebzw.aufOberflächenproteinelebenderZellenbeschränkt.
Ziel der vorliegenden Arbeit war die Weiterentwicklung der CCMS-Technologie durchImplementierung und Optimierung des Einsatzes permeabler Capture Compounds neuartigenDesignszurProfilierungvonKleinmolekül-Wirkstoffen in lebendenZellen, indenendieZielproteinein ihrernatürlichenUmgebungvorliegen.Bio-orthogonale „Click“-ChemiebietetdieVoraussetzungfür den Einsatz permeabler Capture Compounds, die lediglich die Selektivitätsfunktion, dieReaktivitätsfunktion sowie eine kleine, „Click“-fähige chemische Gruppe enthalten. Bindung undVerbrückungzudenZielproteinenkannindenlebendenZellenerfolgen,understdanachdieSortier-/Detektionsfunktion in einer kontrollierten chemischen „Click“-Reaktion (wie z.B. die Kupfer-katalysierteAzid-Alkin-Cycloaddition,CuAAC)angebrachtwerden.
Initiale Reaktionsbedingungen für die „Click“-Reaktionwurdenmit einer Capture Compoundmit cAMP als Selektivitätsfunktion in einer Lösung einer rekombinanten PKARI getestet. DieOptimierung der Reaktion brachten substantielle Verbesserungen in der Reaktionsausbeute imVergleichzudenStartbedingungen.
Um die Identifizierung von Zielproteinen in Zell-Lysaten und lebenden Zellen unterVerwendung der „Click“-Reaktion weiter zu entwickeln, wurde eine Capture Compoundmit eineranderen Selektivitätsfunktion verwendet. Das Kinaseinhibitor-Medikament Dasatinib wurde mitdieser Strategie profiliert. Diebona fide Zielproteine ABL, BCR, und SRCwurden spezifisch in Zell-LysatenundzumerstenMalCSKmitchemischerProteomanalyseinlebendenZellenidentifiziert.
Der neu etablierte methodische Arbeitsablauf ermöglichte es, Phenobarbital besser zucharakterisieren, weil es in einigen Spezies hepatocarcinogen wirkt. Die Ergebnisse der Capture-Experimentelegennahe,dassderprimäreBinderderTranskriptionsfaktorYBX1seinkönnte.
InsgesamtkonntedieStrategie,ZielproteinevonKleinmolekül-WirkstoffeninlebendenZellenmitHilfeneuartigerCaptureCompoundsundbio-orthogonalerchemo-selektiverLigationerfolgreichetabliert werden. Diese Methode bietet ein Werkzeug für die bessere Charakterisierung derZielprotein-ProfilevonMedikamentenundhatdasPotential,nochweiterverbessertzuwerden.
Chapter1:GeneralIntroduction
13
1. GeneralIntroduction1.1. CHALLENGESINDRUGDISCOVERY
Thepharmaceutical industry is facingunprecedented challenges to its businessmodel.Overthepastdecade,concernsabouttheindustry’sintegrityandtransparencyregardingdrugsafetyandefficacyhavebeenraised,andleadtoanincreasedregulatoryscrutiny1.
The traditional drug development timeline (Figure 1-1) consists of a discovery phase,whichincludesthetarget-to-hit,screening,hit-to-lead,andleadoptimizationstagesthatprovideaclinicalcandidate for the development phase consisting of pre-clinical, phases I, II and III clinical trials,culminatinginasubmissiontothecompetentauthoritiesforadruglaunch.
Figure1-1:DrugResearchandDevelopmenttimeline.
Chapter1:GeneralIntroduction
14
In the initial stageof thediscovery phase a target is selected either by actively pursuing anunmetmedicalneedorbyscreeningliteraturewithreportsonnovelassociationswithdiseasestates,always guided by the size of the potential market. In the target-to-hit phase it is important todetermine the differences and similarities across species for future safety studies translation.Following the target identification, a screening for potential hits is performed using novel highthroughput technologies scanning small molecules or biologics, such as antibodies or RNA/DNA-based compounds. Opposite to the large-scale profiling, amore direct approach in drug-design issometimes utilized via computational molecular modeling prediction systems. Once a potentialcandidate is identified, a series of analogues are synthesized and structure-activity relationships(SAR) are measured as well as initial pharmacokinetics (PK) studies are performed. In the leadoptimization stage, thepotential drug is refined and improved for efficacy andpotency aswell asidentification of undesired off-targets. At this stage a drug candidate can enter the developmentphase and initiate the pre-clinical studies, which include drug formulations compatible with safeadministration, further PK studies andADME (Absorption,Distribution,MetabolismandExcretion)analysis in disease models. Safety and toxicity studies are focused on unwanted effects incytochromeP450(CYP)metabolizingenzymes,orspecific ionchannelsthatcan leadtodeleteriousconsequences.With safety and toxicity parameters in check, the phase I clinical trial consists of asmall groupof individuals,usuallyhealthyvolunteers,wheredosagesand frequencyare tested. InPhase II the cohort is increased and the effectiveness of the compound is provedbydoubleblindstudies,whichcontinuesontoPhaseIIIwithmoresubjectsandalsodiversifiedgeneticbackgroundstodetermineobservedsideeffectsandestablishthegeneralbenefitsofthenewmolecularentityincomparisontothecurrentstandardofcare.
AsFigure1-1shows,navigatinganewmedicinetothemarketisalongandstrenuouspath.Alotofpotentialdrugsdonotmoveforwardinthepipelinebecauseofundesiredeffectsorproblemsofformulation.Astudyonthesuccessratesforinvestigationaldrugs2covering835drugdevelopers,includingbiotechcompaniesaswellasspecialtyandlargepharmaceuticalfirmsfrom2003to2011,reportsthatfromthe7.300independentdrugdevelopmentpathsanalyzed,asmuchas10.4%ofallindication development paths in phase Iwere approved by theUS Food andDrug Administration(FDA).Thebiggesthurdlealong thedevelopmentchain is the transition fromphase II tophase III,with only 32.4% of the drugs advancing in the pipeline, making companies carefully analyze therisk/costbenefitsoftheseundertakings.Hayetal.2suggestonewaytoimprovethesuccessratesisto improve the predictive animal models, perform earlier toxicology evaluation, biomarkeridentificationandnewtargeteddeliverytechnologiesthatmayincreasefuturesuccessintheclinic.
Chapter1:GeneralIntroduction
15
1.2. DRUGTARGETDECONVOLUTIONOne way to improve the success rates in the drug development pipeline, discussed in the
previousparagraph,istoimplementtechniquesofdrugtargetdeconvolutionearlyonintheprocess.Drug target deconvolution is a process where the action of a drug, a small molecule, is
characterized by identifying the proteins binding the drug and initiating the biological effect. Thebiological relevant targethas tobeextracted,ordeconvoluted, froma listofproteins identified insuch an approach. Besides the medically desired action of the drug, the identification of otherproteinsbindingthedrugcanhelpto identifysideeffectsandtoxicityataveryearlystageofdrugdevelopment3.
Traditionally,newdrugshavebeendeveloped followinga target-based screening inwhichalargelibraryofcompoundsisscreenedagainstasingletargetproteinthoughttoberesponsibleforthedisease,andsubsequentlytheactivehitscanbeoptimizedthroughmedicinalchemistryefforts.Onemajor limitationof thisapproach is the fact thatmanycompoundsare found to interactwithmultiple targets. The “one drug, one target” paradigm, thought to be the cornerstone of target-basedmethods,oftendoesnotholdtrueforcompoundsidentifiedusingthisstrategy4.
Oneofthemajoradvantagesofphenotype-basedapproachesisthattheyprovideanunbiasedway to find active compounds in the context of complex biological systems. Because phenotypicscreening takes place in a physiologically relevant environment of cells or whole organism, theresults fromsuchscreensprovideamoredirectviewof thedesired responsesaswellashighlightpotentialsideeffects.Moreimportantly,phenotypicscreenscanleadtotheidentificationofmultipleproteinsorpathwaysthatmaynothavebeenpreviouslylinkedtoagivenbiologicaloutput.Becausetargetidentificationfromphenotypicscreensisexpectedtogenerateaspectrumofpossibletargets,theterm“targetdeconvolution”wascoinedtomoreaccuratelydefinetheprocess4.
Recent developments in chemical proteomics, a multidisciplinary research area integratingbiochemistryandcellbiologywithorganic synthesisandmass spectrometry,haveenabledamoredirect and unbiased analysis of a drug’s mechanism of action in the context of the proteome asexpressedinthetargetcellorthetissueofinterest.Advancesinquantitativemassspectrometrynowenablethedetectionandquantificationofdrug-inducedchangesinproteinexpressionandactivationwith unprecedented proteome coverage. There aremultiple approaches that can be employed inchemical proteomic workflows. These include small molecule affinity-based and activity-basedprobes that canbeused to isolate targets andmore recently, in silico techniques topredict smallmoleculebindingproteins.Moreover,affinityenrichmentstrategiesusingimmobilizeddrugsortoolcompounds now allow to characterize the expressed “interactome” of a drug directly from cellextracts5.
Indrugdevelopment,chemicalproteomicscanshedanunbiasedlightontheinteractionofthedrugmoleculewiththeproteomeandidentifybothmoleculartargetsresponsibleforbenigneffectsand unwanted potentially toxic interactions. Furthermore, a different attachment position of thedrug to the rest of the probe results in different interaction profiles with the proteome. Thisfavorable “sideeffect”of usingmulti-functional probesmaygive insight into the structure activityrelationship(SAR)ofthedrugandmayserveasstartingpointforchemicalmodifications6.
Chapter1:GeneralIntroduction
16
1.2.1. InsilicoapproachComputeraideddrugdesign isamajorpillarof target-baseddrugdiscovery.Onthebasisof
docking studiesandvirtual screening,drugcandidateswithoptimalpotencyandselectivity canbepredicted. With the recent renaissance of phenotype-based drug discovery, these in silicotechnologieshavefoundanimportantnewroleintheprocessoftargetprediction.Overthelasttwodecades,extensive informationregardingactivity,structuresandtargetsofsmallmolecule librarieshas beendeposited into public databases such as ChEMBL,DrugBank andChemBank.Using thesetools, targetsof active compounds canbepredictedbasedon similarities in structurebetweenanactivehitandwell-characterizeddrugsinthesedatabases.Thiscomputeraidedtargetpredictionhasbeenwidelyusedtoidentifynewtargetsofknowndrugs,topredictthetargetsofactivehitsfromalibrary screening, and to investigate themechanism of action of hits discovered from phenotypicscreens4.
Thepreviouslydescribedstrategyreliesonmoleculesthathavealreadybeensynthesizedandstudied inaspecific system, fromwhichresultsareextrapolated foranewprotein/small-moleculepair. The drug-like chemical space, comprising organic small molecules (MW < 500 Da) ofintermediate polarity, is very large andmight contain on the order of 1060 possiblemolecules.Denovo drug design consists in exploring this chemical space in search for bioactive compounds bygeneratingandscoringvirtualmoleculespriortotheirsynthesis,wherebyscoringreliesonavarietyofvirtualscreeningapproachessuchasligand-basedandtarget-basedmethods.Theadvantageofdenovodrugdesignisthatmanymoremoleculescanbeevaluatedthanwhatispossibleifconsideringonly already synthesized molecules, a strategy which might help to uncover lead structures andeventuallyimproveclinicalsuccessfornewdrugs7.
1.2.2. AffinityChromatographyAffinitypurificationisthemostwidelyusedtechniquetoisolatespecifictargetproteinsfroma
complex proteome. Smallmolecules identified in phenotypic screens are immobilized onto a solidsupportthatcanbeusedtoisolateboundproteintargets(Figure1-2). The small molecules used as selectivity functions in thesestudies display binding to their target proteins with lownanomolar affinity. This approach relies on extensive washingsteps to remove non-binders, followed by specific methods toelutetheproteinsofinterest.Theelutedproteinscantheneitherbe directly identified using ‘shotgun’ type sequencing methodswith multidimensional liquid chromatography or be furtherseparated by gel electrophoresis and analyzed by massspectrometry.Theidentifiedpeptidesequencescanthenbeusedindatabasesearchestoidentifythetargetprotein4.
Representativestudiesutilizingthistechniqueproducedso-called kinobeads to study the kinase subproteome from eukaryotic cell lines. These beads carry anumberofdifferent smallmoleculekinase inhibitorsandhavenotonlybeenused tomapkinasesfromcomplexbiologicalsamples,butalsotoassesstheselectivityprofilesofkinaseinhibitordrugs8.Daubandcolleaguesusedadifferentsetofkinaseinhibitorstoisolatelargenumbersofkinasesfrom
Figure1-2: Schematic representationofanaffinitychromatographypull-downprobe.Bi-functional molecule bearing atargeting/selectivity function (redrectangle), corresponding to the smallmolecule under study, and a sortingfunction to allow attachment to a solidmatrix,usuallyviabiotin/avidininteraction.ReproducedfromLenzetal.6
Chapter1:GeneralIntroduction
17
celllysates,andalsofocusedonthesubsequentanalysisofthephosphorylationstateoftheenrichedproteins9.
Themajor problemof affinity chromatography and its “pull-down” approach consists in theevaluation of nonspecific binding especially if we take into consideration that some intracellularproteins (e.g. glyceraldehyde-3-phosphate dehydrogenase) are detected as components of varioussubproteomes.Itisalsopossiblethatroughelutionconditionsmayremoveproteintargetsexhibitingmoderateaffinitytoligandsused.Affinity-basedapproachesarealsoapplicableforidentificationofsmall-molecule protein targets, but the previously described limitations demand a subsequentanalysisofidentifiedproteinswithaparticulartargetinhibitionassay10andaremostlyunusableinlivingcells.
1.2.3. Activity-basedproteinprofiling(ABPP)Activity-based probes are small molecule tools that can be used to monitor the activity of
specificclassesofenzymes.ABPPprobesarecharacterizedbyaselectivityfunctionthatiscapableofcovalentlyreactingwiththeactivesitesoftargetenzymesdrivenbythecatalyticmechanismoftheenzyme. By directly modifying the active site amino acid residues, ABPP probes act as suicideinhibitors. Over the past decade and a half, various ABPP probes have been designed to studyproteases,hydrolases,phosphatases,histonedeacetylases,andglycosidases,andtheseprobeshaveproven to be valuable in investigating enzyme-related disease mechanisms including cancer andmetabolic disorders. Typical activity-based protein profiling follows a similar overall workflow asaffinity chromatography, including probe binding, protein separation, sequence analysis, anddatabasesearching.However,becausemostABPPprobesaredesignedtotargetaspecificenzymeclass, they aremostly useful for the development of small-moleculeswhere a specific enzyme orenzymefamilyissuspectedtobeinvolvedinacertaindiseasestateorpathway4,6.
ABPP probes have three components: a reactiveelectrophile for covalent modification of enzyme active site, alinker or a specificity group for directing probes to specificenzymes,andareporteroratagforseparatinglabeledenzymes(Figure1-3).CovalentenzymeinhibitorscanbereadilyconvertedtoABPPprobesbyattachinga tagginggroup,and thecombineduse of covalent inhibitors and ABPP probes in phenotypicscreeninggreatlyfacilitatestargetidentificationandalsooffersapowerful tool to study function and mechanism of identifiedproteins. In order to covalently attach ABPP probes to targetproteins, an active sitenucleophile suchas cysteineor serine isrequired4,6.
ABPP probes are not only useful for target identification,butalsoarepowerfultoolsforthediscoveryofdiseaserelatedproteinsandthetechnologyhasbeenusedtoidentifyserineandcysteinehydrolasesfromhumancancersamplesorproteolyticenzymesof the proteasome, for example.Mechanism-based ABPP-probes have also been reported for theprofiling of native ATP-binding proteins, in particular kinases. These probes exploit the catalyticreaction of the kinase to form a chemical group that can covalently target lysine residues in the
Figure 1-3: Schematic representation of anactivity-basedproteinprofiling(ABPP)probe.Typically a bi-functional probe where thesmallmolecule as selectivity function reactswiththeactivesiteofthetargetproteinandestablishes a covalent bond. The attachedsorting function can either be a reporter,such as a fluorophore for visualization, or atag, such as biotin for isolation andenrichment.ReproducedfromLenzetal.6
Chapter1:GeneralIntroduction
18
catalytic sites. Othermechanism-based ABPP-probeswere successfully applied for the profiling oftyrosine phosphatases, phospholipases, glycosidases, cytochrome P450 proteins, and the list isgrowingatafastpace6.
ThelimitationofABPPprobesinpharmaceuticaltargetresearchisthatmostsmall-moleculesarenon-covalentbindersandmodifying theselectivity functionwhere it interactswith theproteinwould most probably alter its binding affinity. A new design of probe was developed where thesmall-moleculeinteractswiththetargetproteininanaturalreversiblemanner,butanindependentphoto-reactivemoiety forms a covalent ligation to the target allowing for the necessary stringentwashingstepstoreduceunspecificbinders–theCaptureCompound11.
1.2.4. Capturecompoundmassspectrometry(CCMS)Smallmoleculeprobes termedCaptureCompoundsare tri-functionalmoleculescontaininga
selectivityfunction,aphotoactivatablereactivityfunction,andasortingfunctionprotrudingfromacommoncoreatom(Figure1-4).Bindingoftheselectivityfunctionoccursnon-covalentlydrivenbythe thermodynamic equilibrium. Covalent cross-linking to the targetprotein isaccomplished throughphoto-activationofaseparatereactivityfunctionthat,incontrasttoclassicphotoaffinitylabeling,isnotpartoftheselectivityfunction,butislocatedonadifferentbranch of the probe, intentionally at a certaindistance from the selectivity function 11. As a result,all three functional entities exhibit a high degree offreedomindependentofeachother.Inparticular,dueto the linker between selectivity function and coreatom of the probe, the selectivity function binds tothe small molecule binding pocket of the targetproteins, whereas the reactivity function stays at the protein surface. Therefore, the reactivityfunctioncantargetsurface-exposedstructuresofthetargetproteins,andisnotstructurallyconfinedwithinthesmallmoleculebindingpocketoftheprotein.Thisflexibilityofthephoto-reactivefunctionallowsforhighcross-linkyields12.
In a typical CCMS experiment, the selectivity function accomplishes an equilibrium-drivenaffinity binding to target proteins. In a second step, this equilibrium is “frozen” through photo-activationof the reactivity function. This results in a covalentbond formationbetweenprobeandproteintargetthroughthereactivityfunction.Becausetheequilibriumbindingofthetargetproteinsis“frozen”bythecovalentcross-linkofthereactivityfunctiontotheprotein,weak,butstillspecificbindersareretainedattachedtotheprobeduringsubsequent isolationandwashingstepsandcanbe identified. The thirdmodule is a sorting function for isolation of the targeted proteins, usuallybiotin that can be enriched with streptavidin-coated magnetic beads for subsequent proteomicanalysis.ThebasicexperimentalworkflowofCCMSandthepreviouslydescribedaffinitypull-downandABPPareillustratedinFigure1-5,onpage19.
Capture compounds with different selectivity functions have been used for the isolation ofprotein familiessuchaskinases 13,14,GTPases 15,cAMP-bindingproteins 12,methyltransferases 16,17,
Figure1-4:Schematicrepresentationofacapturecompound.Tri-functional molecule consisting of a selectivity function,formed by the small molecule under study, a reactivityfunction that establishes a covalent bond with the targetproteinuponUV irradiation,anda sorting function that canbeafluorophoreforvisualizationoratag,suchasbiotin,forenrichmentandisolation.ReproducedfromLenzetal.6
Chapter1:GeneralIntroduction
19
and 2-oxoglutarate oxygenases 18 from complex biologicalmixtures. In each case, specific binderswereisolatedandidentifiedfromcell lysatesorsubcellularfractionsoftissues.Ineachofthecitedcapture compound studies, the yield of identified specific binders was shown to be substantiallyhigher than inparallelaffinitypull-downexperimentswithoutphoto-activatedcross-linkingvia therespectivereactivity functions6.Particularly inthecaseofmethyltransferasesthatwereaddressedbyacapturecompoundwiththemethyltransferaseproductinhibitorS-adenosyl-L-homocysteineasselectivityfunction,theresultsdemonstratedthatacapturecompoundbasedapproach,duetotheadditionalphoto-activatedcross-linktotheboundtargetproteinsviathereactivityfunction,candealwithsmallmolecule-protein interactions inarangeofrelatively lowaffinity, i.e., inthemicromolarrange.Theseinteractionsareusuallydifficulttoaddressbysimpleaffinitypull-down.
Figure1-5:Schematicworkflowsforthreedifferenttargetdeconvolutiontechniques.In affinity chromatography pull-down experiments (panel A), probes reversibly interact with target proteins defined by a dynamicequilibrium.Non-bindingproteinsarewashedaway,andretainedproteinsareanalyzedafterelutionusingdenaturingbufferoranexcessoffreeaffinity ligand.Thechoiceofbufferconditionsrequiresatradeoff: lowstringencywill leadtohighunspecificbackgroundbinding,while high stringency will lead to loss of weakly, but still specifically interacting proteins. In activity-based protein profiling (ABPP)experiments (panelB), the selectivity functionof theprobebinds toand chemically reactswith theactive siteof targetenzymes.Harshwashingconditionscanbeappliedtoremoveproteinsnonspecificallyboundtothesolidphase.Asalimitation,ABPPcanonlybeusedwithsmall molecules that chemically react with their targets. In Capture Compound Mass Spectrometry (CCMS) experiments (panel C) theCaptureCompoundselectivity function reversibly interactswith targetproteinsdefinedbyadynamicequilibrium,and thisequilibrium isfixed through a photo-induced cross-linking of the reactivity function. Harsh washing conditions can be applied to remove unspecificbackground.Somenonspecificallycross-linkedproteinsareretainedaswell.ReproducedfromLenzetal.6
Chapter1:GeneralIntroduction
20
Theprerequisiteforthesuccessfulfunctional isolationofproteinsfrombiologicalsamplesbysmall molecule probes is that the proteins are in their native (functional) state. Some proteinstoleratetransfer fromthecellularcontext toacellular lysate,whileothersmaynot. It is thereforeimportant to keep the proteins in a state as native as possible in order to reveal physiologicallyrelevantsmallmolecule–proteininteractions.
Thevastmajorityofchemicalproteomicsstudieshavebeenconductedonlysatesamples16,19–21. Ideally, probing of small molecule–protein interactions on intracellular proteins should beconductedinintactcellswithmembrane-permeableprobes.Thepermeabilityoftheprobescanbeincreasedbyonlyaddingthetagafterthesmallmoleculehasfoundthetargetandisfixedtoit.Thisis achieved using chemo-selective ligatable probes. The smaller probes comprise only a combinedselectivity/reactivityfunction(inABPPprobes)orseparateselectivityandphoto-reactivityfunctions(in capture compounds), but no sorting or detection function. Instead, the probes contain a smallfunctional group, such as an alkyne or azidemoiety. This enables chemical ligation of the sortingfunction later in theworkflow. The sorting function is attached to the probeonly after the targetrecognition and (photo-) cross-linking reaction has taken place and cell integrity is no longernecessary. This reaction workflow is called chemo-selective bio-orthogonal chemical ligation.Permeable probes consisting only of selectivity and reactivity function or combinedselectivity/reactivity functions have been successfully used to address targets of small moleculeswithinlivingcells22–24.
The in-cell workflow for the application of either a chemically reactive or a photo-reactiveprobeisdepictedinFigure1-6.
Figure 1-6: Workflow for identification of small-moleculetargetsinlivingcells.Permeable small molecule probes carry aselectivity function and a reactivity function or acombinedselectivity/reactivityfunction,butlackasortingfunction.Instead,theprobecarriesasmallfunctionalgroupthatenablesadditionofasortingfunction througha specific ligation reaction laterin the workflow. Chemical or photoactivatedcross-link can be carried out, and only after thatthecellislysedforligationtothesortingfunction.ReproducedfromLenzetal.6
Anessentialconditionforthis
approach is that the functions onthe small probe and the sortingfunctiononly reactwitheachotherand not to any other intracellularcomponent.
In essence, it needs to allowforabio-orthogonalchemoselectiveligation.
Chapter1:GeneralIntroduction
21
1.3. CLICKCHEMISTRYANDBIO-ORTHOGONALLIGATIONSIn the advent of successfully using green fluorescent protein (GFP) tags to visualize protein
dynamicsandfunctioninlivingsystems,bycleverlyintroducingagenethatconcomitantlytranslateswith the target protein and produces an active fusion protein, it quickly became clear that thisstrategydidnotallowtostudypost-translationalmodificationsorotherrelevantbiologicalmoleculessuchasglycans, lipids,metabolitesandnucleicacids 25,26.Toovercomethis limitation,a reactivity-basedbio-orthogonalchemistryapproachwassuccessfullydeveloped.
Bio-orthogonal reactant pairs, which aremostsuitableforsuchapplications,aremoleculargroupswith the followingproperties: (i) theyaremutually reactive but do not cross-react orinteract in noticeable ways with biologicalfunctionalities or reactions in a cell, (ii) they andtheir products are stable and non-toxic inphysiological settings, and (iii) ideally, theirreaction is highly specific and fast 25,27. Anillustration from Sletten et al. 26 depicts thechallenge of “fishing” out a specific molecularentity froma “sea”of functional groups thatarepresentincells(Figure1-7).
Bio-orthogonal chemical reactions haveemergedashighlyspecifictoolsthatcanbeusedfor investigating the dynamics and function ofbiomolecules in living systems 27–29. Clickchemistry,inspiredbynature’suseofsimpleandpowerfulconnectingreactions,describesthemostspecificbio-orthogonalreactionsthatarewideinscope,easytoperform,andusuallyemployreadilyavailablereagentsthatareinsensitivetooxygenandwater30–32.Invivobio-orthogonalchemistryandclick chemistry therefore overlap quite a bit, reflecting the same underlying chemical principlesappliedinsomewhatdifferentwaystowardthediscoveryordevelopmentofmolecularfunctionandinformation25.
With over 20 bio-orthogonal transformations now reported in the literature and new onesbeing discovered at a rapid pace, selecting the “best fit” for a given application is non-trivial. Thechemistries vary widely in terms of their selectivities and biocompatibilities, and many of theirperceivedstrengthsandweaknessesremainanecdotal29.
Oneofthemostpopularbio-orthogonalfunctionalgroupsistheazide.Thesmallsize,coupledwithitsinertnaturetowardsendogenousbiologicalfunctionalities,setstheazideapart.Additionallytheazidehasuniquechemistrieswithotherbio-orthogonalfunctionalities,mostnotablyphosphinesand alkynes 28. Because of this versatility, old andnewlymodified reactions basedon azides havebeenthebasisfortheexpansionofclickchemistryinthebiologicalcontext,suchastheStaudingerligation,[3+2]cycloadditionswithorwithoutmetalcatalystsand,morerecently,atetrazineligation.
Figure1-7:Illustrationofbio-orthogonalligation.The bio-orthogonal reactant pairs should react exclusively witheachotherformingastableandnon-toxicproducttothecellularenvironment.ReproducedfromSlettenetal.26
Chapter1:GeneralIntroduction
22
1.3.1. StaudingerligationTheStaudingerligationessentiallylaunchedthefieldofbio-orthogonalchemistry,notbecause
itwas the first bio-orthogonal reaction per se, but because itwas the first among entirely abioticfunctionalgroupsandthereforehadthepotential fortranslationto liveorganisms27. Itsprototypereaction was the Staudinger reduction of azides with triphenylphosphine and water, reported byHermann Staudinger in 1919. It appeared well suited as a prototype for bio-orthogonal reactiondevelopment because the two participantswere abiotic,mutually and selectively reactive,mostlyunreactivewithbiologicalfunctionalities,andtolerantofwater.In2000,SaxonandBertozzi33wereabletoadaptthisreaction intoa ligationmethodwheretheysuccessfully labeledmodifiedglycansonthesurfaceofcellsviathenewlydevelopedStaudingerligation(Figure1-8,panelA).
Figure1-8:Classical(A)andtraceless(B)Staudingerligation
The Staudinger ligation was promising but there were some limitations. The phosphinereagentsslowlyunderwentairoxidationwithinbiologicalsystemsandwereprobablymetabolizedbycytochromeP450enzymes inmice.Additionally, the kinetics of the reactionwasnot fast enough,whichnecessitatedtheuseofhighconcentrationsofphosphinereagent.
Bertozzi’s groupovercame this limitation later in the same yearby developing a “traceless”Staudingerligation(Figure1-8,panelB)wherethepositionofthereportermoduleinthephosphinereagentwasmodified.Thisallowedthephosphineoxidetobeexpelledduringthereaction leavingthetwogroupslinkedviaanamidebond.
Sincethistime,themechanismofthereactionandotherdetailshavebeenprobedindepth34–36, andanumberof iterationshavebeen reported, suchas fluorogenic versions, includinga FRET-basedsystem,aswellasreagentswithincreasedwatersolubility37.
While still the reaction of choice for awide range of bio-conjugation applications, the slowkineticsoftheStaudingerligation,togetherwiththeoxidationliabilityofitsphosphinecompounds,remains an unsolved problem and an obstacle for in vivo chemistry. Consequently,most researchgroupsturnedtheirattentiontotheothermodeofbio-orthogonalreactivityexhibitedbytheazide:its1,3-dipolarcycloadditionwithalkynes.
Chapter1:GeneralIntroduction
23
1.3.2. Copper-catalyzedazido-alkynecycloaddition(CuAAC)The1,3-dipolar [3+2]cycloadditionofazidesandalkyneswas first reported in189338.But it
wasonly70yearslaterthatthekineticsandmechanismofthereactionwasdescribedbyHuisgen39,whohadhisnamedbestoweduponthisazido-alkynecycloaddition(AAC).
In2001,Sharplessandhiscollaboratorssetouttogeneratesubstancesbyjoiningsmallunitstogetherwith heteroatom links (C-X-C)with the purpose to develop an expanding set of selectiveand modular “blocks” that would reliably work in both small and large-scale applications. Theytermedthisapproachas“clickchemistry”anddefinedastringentsetofcriteria forthisprocess30.The reaction should be modular, wide-scoped, generate high yields with little or inoffensivebyproducts40,stableunderphysiologicalbufferconditionsandinsensitivetooxygen.Therefore,clickreactionsachievetheir requiredcharacteristicsbyhavingahighthermodynamicdriving force,as ifbeing“spring-loaded”forasingletargetinasingletrajectory.Amongthesewerethecycloadditionsofunsaturatedspecies,especiallythe1,3-dipolarcycloadditionreaction40.
Beingsmall,incapableofsignificanthydrogenbondingandrelativelynon-polar,bothazideandalkyneareunlikely to significantly change thepropertiesof structures towhich theyareattached.Additionally,azidesandalkynesarestableinthepresenceofsurroundingnucleophiles,electrophilesandsolventscommontostandardreactionconditionsandcomplexbiologicalmedia.However, therate of this reaction is quite slow, requiring prolonged heating for inactivated alkynes. The AACprocesswasacceleratedbycatalysiswithcopper(I)orruthenium(II)complexes.Becauseofitshighrate and use of the small and easily synthesized azide and terminal alkyne groups, the copper-catalyzed(CuAAC)versionhasbeenthemostwidelyused40.
ThereactionmechanismwasinitiallyproposedbyHimoandSharpless41andrecentlyrefinedby Worrell and Fokin 42. Copper catalysts change the mechanism and outcome of the thermalreaction of terminal or internal alkynes with organic azides, converting it to a sequence of stepswhichculminatesintheformationofadi-copperintermediate(Figure1-9).
Figure1-9:Mechanismofthecopper-catalyzedazido-alkynecycloaddition(CuAAC)
Chapter1:GeneralIntroduction
24
TherateofCuAACisincreasedbyafactorof107relativetothenon-catalyzedprocess,makingitconvenientlyfastandbelowroomtemperature43.
By 2010, copper stood out as the only metal for the reliable, facile and 1,4-regiospecficcatalysis of the azide-alkyne cycloaddition. Fokin’s group at the Scripps Institute researchedcomplexes of all first-row transition elements and complexes of Ag(I), Pd(0/II), Pt(II), Au(I/III), andHg(II),amongothers,haveallfailedtoproducetriazolesinsyntheticallyusefulyields;theireffectontherateandselectivityofthecycloadditionwasatbestmarginallynoticeable43.
In2005, rutheniumcyclopentadienylcomplexeswere foundtocatalyzethe formationof thecomplementary 1,5-disubstituted triazole from azides and terminal alkynes, and also to engageinternal alkynes in the cycloaddition 44. While the scope and functional group compatibility ofruthenium azido-alkyne cycloaddition (RuAAC) are excellent, the reaction ismore sensitive to thesolventsandthestericdemandsoftheazidesubstituentsthanCuAAC43.ApplicationsofRuAACarestillscarce44.
Meldal and Tornøe 45 published an extensive review on CuAAC showing the effects of thecoppersourceinthereactionandpotentialsidereactionsthatmaypreventthisapproachtobeusedin living cells, which ultimately lead to the introduction of copper ligands to perform thecycloadditionsafelyinintactcells.
1.3.2.1. Coppercatalystsandligands
A wide range of experimental conditions for the CuAAC have been employed since itsdiscovery, underscoring the robustness of the process and its compatibility with most functionalgroups,solvents,andadditivesregardlessofthesourceofthecatalyst.Thechoiceofthecatalystisdictated by the particular requirements of the experiment, and usually many combinations willproduce desired results. Copper(I) salts (iodide, bromide, chloride, acetate) and coordinationcomplexeshavebeencommonlyemployed43.
Copper(II) salts and coordination complexes are not competent catalysts. Cupric salts andcoordinationcomplexesarewell-knownoxidizingagents fororganic compounds.Alcohols,amines,aldehydes,thiols,phenols,andcarboxylicacidsmayreadilybeoxidizedbythecupricion,reducingitto the catalytically active copper(I) species in the process 43. Still, performing the reaction in thepresenceofmild oxidants allows capture of the triazolyl copper intermediates. Ascorbate, amildreductant,wasintroducedbyRostovtsevetal.46asaconvenientandpracticalalternativetooxygen-free conditions. Its combination with a copper(II) salt, such as the readily available and stablecopper(II)sulfatepentahydrate(CuSO4�5H2O)orcopper(II)acetate(Cu(CH3COO)2),hasbeenquicklyacceptedasthemethodofchoiceforpreparativesynthesisof1,2,3-triazoles.Waterappearstobeanideal solvent capable of supporting copper(I) acetylides (Cu2C2) in their reactive state, especiallywhentheyareformedinsitu.The‘‘aqueousascorbate’’procedureoftenfurnishestriazoleproductsinnearlyquantitativeyieldandgreaterthan90%purity,withouttheneedforligandsoradditivesorprotection of the reactionmixture from oxygen 43. Of course, copper(I) salts can also be used incombination with ascorbate, wherein it converts any oxidized copper(II) species back to thecatalyticallyactive+1oxidationstate.
Chapter1:GeneralIntroduction
25
Manyothercoppercomplexesinvolvingligandshavebeenreportedascatalystsormediatorsof the CuAAC reaction.Many reported reactions are performed underwidely differing conditions,makingquantitativecomparisonsofligandsperformancedifficult.ThelargestandmostsuccessfulclassofligandsforCuAACcatalysis istheone containing heterocyclic donors. The need for these ligandswasparticularly evident for reactions involving biological molecules thatare handled in water in low concentrations and are not stable toheating.
Despite the experimental simplicity and efficiency of the‘‘ascorbate’’procedure,therateoftheCuAACreactionintheabsenceof an accelerating ligand is simply not high enough whenconcentrationsofreactantsare low.Thefirstgeneral solutiontothebio-conjugation problem was provided by tris(benzyltriazolyl)methylamine ligand (TBTA, CC 16 on page 125), prepared using the CuAACreactionandintroducedshortlyafteritsdiscovery47.Afteritsutilityinbio-conjugationwasdemonstratedby theefficient attachmentof60alkyne-containing fluorescent dye molecules to the azide-labeledcowpea mosaic virus 48 it was widely adopted for use with suchbiological entities as nucleic acids, proteins, E. coli bacteria, andmammaliancells43.
ThepoorsolubilityofTBTAinwaterpromptedthedevelopmentofmorepolaranalogues suchas tris(3-hydroxypropyltriazolylmethyl)amine (THPTA, CC 17 on page 125) 49 and a bis(tert-butyltriazoly)ethanolligand(BTTE,CC18onpage125)50thatshowedtremendousimprovement in reaction kinetics compared to the ligand-freecycloadditioninbiologicalconditions.
Despitetheacceleratedkineticsandreducedcoppertoxicitybytheintroductionof ligandsintheCuAACreaction,theBertozzigroupdevelopedmorebio-friendly copper-free cycloadditionsbyusing theintrinsicenergyofstrainedcyclicalkynestodrivethereaction51.
1.3.3. Metal-freeclickcycloadditionsExogenousmetalscanhavemildtoseverecytotoxiceffectsandcanthusdisturbthedelicate
metabolicbalanceof the systemsbeing studied.Copper, forexample,exerts its toxicitymainlyviafreeradical-inducedoxidativedamage52.
Basedonknowledgethatcyclooctyneandphenylazide“proceededlikeanexplosiontogiveasingle product” – the triazole, Bertozzi’s lab explored this new click chemistry field to circumventmetaltoxicityoftheclassicalcycloadditions.
Figure 1-10: CuAAC-acceleratingligandsofchoice.
TBTA
THPTA
BTTE
Chapter1:GeneralIntroduction
26
1.3.3.1. Strain-promotedazido-alkynecycloaddition(SpAAC)
The first cyclooctyne that Bertozzi evaluated underwent cycloadditionwith azides to give atriazole, and although its reaction kineticswas vastly accelerated compared to the linear alkynes,therewasroomforimprovement51.Theysoughttoreducetheelectrondensityofthealkynebondbyintroducingafluorineatomtotheringsystem,andthisincreasedthereactionrateconstantby3-fold.Withtheadditionofasecondfluorineatomtothering,tocreatethecompoundnamedDIFO(Figure1-11),theyobservedamarkedincreaseinrateconstant,makingthereaction60timesfasterthantheparentcompound28.WhiletherateofreactionforDIFOwasexceptional,thecompound’ssolubility inwaterwas less than ideal. They then started to fine-tune themoleculeswithdifferentheteroatomswithin the ring structure to achieve a good solubility/reactivity balance, and found agood candidate with DIMAC 53. At the same time other groups increased the reactivity of thecyclooctynesvia introductionofneighboringbenzyl rings 54.Researchadvancedrapidlyanda largemyriad of strained alkynes surfaced where in most cases the increased reactivity meant lesshydrophilicityandvice-versa28,55,56.
Figure1-11:Strain-promotedazido-alkynecycloaddition.ExamplereactionbetweenDIFOandanazide.
Therearenumerousapplicationsofcyclooctynesinthecontextofbiologicalsystems.Bertozzi
et al. proved its applicability on modification of purified proteins 51, later on the method wasextended to in vivo experiments 55. Furthermore, SpAACwas used to image tumors in livingmicewiththehelpofnanoparticles57and18Fpositronemissiontomography(PET)58.
Despite some problems in the earlier development of SpAAC such as slow kinetics andspontaneousdefluorinationofreagentswhichwereovercomebyreagentoptimization,thismethodhasprovenitssuitabilityforinvivolabelingseveraltimes.TheissueofincreasedunspecificlabelingcomparedtoStaudingerligationremainstobesolved.
1.3.3.2. Tetrazineligation
While others were working with alkynes to form triazoles upon cycloaddition with azides,Rutjesandco-workers59developedanalkynesurrogatethatformsa1,2,3-triazoleproductinatwo-stepreactionthatincludedaretro-Diels-Alderreaction*.
*TheDiels-Alderreactionisa[4+2]cycloadditionofaconjugateddieneandadienophile(analkeneoralkyne).E.g.:
Chapter1:GeneralIntroduction
27
In 2008, the Fox 60 and Hildebrand 61 groups put the inverse electron demand Diels-Alderreaction to use in bio-orthogonal ligation chemistry using a tetrazine to react with a trans-cycloocteneandnorbornene,respectively(Figure1-12).
Figure1-12:Tetrazineligation.Example reactions between tetrazines and a trans-cyclooctene (A) as published in Fox et al. 60 and a norbornene (B) as published byHildebrandandco-workers61.
Electron-richtetrazinesarelesssusceptibletohydrolysisandattackbybiologicalnucleophiles,
andare thusmoredesirable foruse invivoand intracellularenvironments, inparticular.Stabilizedtetrazines have recently been introduced into recombinant proteins via unnatural amino acidmutagenesis 62. The heterocyclic amino acid derivatives integrated biologicalmolecules that werethendetectedwithstrainedalkenes63.Ahead-to-headcomparisonofthereactionsofthesevariousalkenes and tetrazines has not been reported, but combining the best feature of each approachcouldaddvaluabletoolstothefield28.
The tetrazine click is showingpromising results in thebio-orthogonal chemical ligation field,particularly in the detection of fast biological processes. However, the bulkiness of themoleculesinvolvedcouldhindersomeapplicationsofthetechnique.
Veryrecentlytheuseofclickreactionsforchemicallytriggeredreleaseofdrugs invivobasedontetrazinesandtrans-cyclooctenesisbeingstudied64,65.
Manyotherchemical ligationshavebeenexplored tobeused inbiological settings, suchas
cycloadditionsinvolvingnitrileiminesornitrileylides;azido-alkynecycloadditionsusingothermetalsas catalyst; thiol-ene coupling;palladium-catalyzed coupling reactions; andphoto-inducibledipolarcycloadditions. In fact, a cooperation partner of the BioChemLig network in which the researchpresented in this dissertation was included, published recently novel chemical reactions withpotentialtooriginatenewbio-orthogonalligations66.
The ligationspresented in thepreviousparagraphs represent themostusedand relevant inbio-orthogonalclickchemistryresearch.
Chapter2:Labelingaproteinusingclickchemistry|Introduction
28
2. Labelingaproteinusingclickchemistry2.1. INTRODUCTION
Cyclic adenosine monophosphate (cAMP) is an important biological second messengermolecule that regulates several cellular events and complex biological processes such asmemoryconsolidation,immunesystemregulation,insulinsecretionandcardiacfrequency67.cAMPwasfirstdescribedin1956byRalletal.whentheyobservedthattheincreasedproductionofphosphorylaseinliverhomogenatesuponhormonalstimulationwasmediatedbyaheat-stablefactor68.Soonafterthatdiscovery, the samegroupcharacterized that factor tobeanadenine ribonucleotide 69,70 thatwasidenticaltoaproductisolatedatthesametimebyCooketal.fromthealkalinedegradationofadenosinetriphosphate(ATP)71.
cAMPisproducedincellsfromATPbyadenylylcyclasesthataremainlylocatedintheplasmamembrane and these can be regulated in different ways such as via G protein-coupled receptors(GPCRs) that respond to external stimuli in the form of hormones or neurotransmitters 72,73;intracellularCa2+andbicarbonateregulatecAMPproductionviaasolubleadenylylcyclase74,75;andpost-translational modifications such as phosphorylation can also affect the activity of adenylylcyclases 76,77. The effector function of cAMP is terminated by specific phosphodiesterases thatcatalyze thehydrolysisofcAMP intoAMP69,controlling thepoolofavailablecAMPandregulatingthe signaling cascade. Additional mechanisms of cAMP production and clearance have beendescribed 72,76, suchas the targetingof specific adenylyl cyclases tomembranemicrodomains thatcontributetotheformationofcAMPcompartmentswithclinicalrelevance78,andtheeffluxofcAMPtotheextracellularspacethroughamultidrugresistancetransportsystem79.
ProteinkinaseA(PKA)wasthefirstdown-streamtargetforcAMPtobeidentifiedbyWalshetal.in196880.PKAisaheterotetramercomposedoftworegulatoryandtwocatalyticsubunitswhosemultiplesubunits(RIα,RIβ,RIIα,RIIβandCα,Cβ,Cγ)possessdistinctphysicalandbiologicalproperties,are differentially expressed, and are able to form different isoforms of PKA holoenzymes 81.Activation of PKA involves the cooperative binding of four cAMPmolecules to the two regulatorysubunits,inducingaconformationalchangeandreleasingbothcatalyticsubunitsthatarethenabletophosphorylatetheirproteinsubstrates82.TheregulatorysubunitsofPKAarealsoresponsibleforthe binding of scaffolding proteins calledA-kinase anchoring proteins (AKAPs) that localize PKA tospecific sites in close proximity to its protein substrates 82,83. A scheme of cAMP synthesis andsignalingviaPKAisillustratedinFigure2-1,onpage29.
AdditionaltargetsforcAMPhavebeendescribedthataffecttheelectricalactivityofcells,suchashyperpolarization-activatedcyclicnucleotide-gated (HCN) channels 84 anddirectlymodulatedbycyclic nucleotides (CNG) channels 85.More recently, exchangeproteins activatedby cAMP (EPACs)thatmodulatetheactivityofthesmallGTPasesRap1andRap2,havebeenestablishedasafamilyofproteinsthatdirectlybindcAMP86,87.
Chapter2:Labelingaproteinusingclickchemistry|Introduction
29
Figure2-1:cAMPproductionandactivationofPKA.LigandbindingtoGprotein-coupledreceptors(GPRCRs)activatesneighboringadenylylcyclasesandgeneratespoolsofcAMPfromATP.ThedistributionandconcentrationgradientofcAMPislimitedbyphosphodiesterases.BindingofcAMPtotheregulatory(R)subunitsofproteinkinaseA(PKA)releasesthecatalytic(C)subunitsoftheheterotetramerallowingthemtophosphorylatetargetsubstrates.CertainadenylylcyclasesandGPCRsareconfinedtomembranemicrodomainsinassociationwithintracellularorganellesorcytoskeletalconstituents.ThesubcellularstructuresharborspecificisoformsofPKAthatarelocalizedinthevicinityofthereceptorandcyclasethroughanchoringviaAkinaseanchoringproteins(AKAPs).Thesemechanismsservetolocalizeandlimittheassemblyofthepathwaytoadefinedareaofthecellclosetothesubstrate.AdaptedfromTaskénetal.81usingServierMedicalArttemplates.
Massspectrometry-basedstudiesoncAMP-bindingproteinshaveidentifiedfurthertargetsfor
thesmallmolecule,inadditiontothepreviouslydescribed.Pull-downstudiesperformedonHEK293cellsanddifferentrattissuelysates88–90usedcAMPimmobilizedonagarosebeadstoenrich,isolateand identify the directly interacting proteins and their interaction partners. The results show thatdifferent isoforms of the regulatory subunit of PKA are enriched as well as a set of AKAPs andphosphodiesterases.Oneparticularenrichedprotein,theoncoproteinsphingosinekinaseinteractingprotein (SKIP) that was not known at the time to be involved in the cAMP pathway, was laterestablished as a new AKAP 91. The authors also concluded that the cyclic nucleotide proteome iscontaminatedbyabundantproteins88andsotheysequentiallyelutedthepulled-downproteinswithdifferent nucleotides to further enrich the true cAMP binders. Pull-down strategies with bead-immobilizedsmall-moleculesarepronetopresenthighunspecificbackgroundbecausethewashingconditions have to be mild and usually consist of only several washes with lysis buffer. Thesestrategies are also little effective in the enrichment of membrane proteins with multiple trans-membranesegments.
Chapter2:Labelingaproteinusingclickchemistry|Introduction
30
In the CCMS approach, as described earlier (1.2.4 Capture compound mass spectrometry(CCMS),onpage18),washingconditionscanbemorestringent(highsaltanddetergent)becauseofthecovalentcrosslinkbetweenthemodifiedsmallmoleculeandthetarget,increasingthesignal-to-noiseratioofthetruebindersofthesmallmolecule.Havingaparallelcompetitionassay,inadditiontothestringentwashing,allowsalsoforamoreaccuratedeterminationofrealtargets.
ThecAMP-interactingproteomewas studiedby classicalCCMS inHepG2cells lysateand ratbrainsynaptosomalfractions12.Aspreviousstudieshaveshown,isoformsoftheregulatorysubunitsofPKAwere identifiedusingCCMSalongwithseveralAKAPs,PKGandEPACs.Thenoveltywasthesuccess in identifying endogenously expressing HCN channels from the synaptosomal fractions,whichwas a knownbinding partner of cAMPbut not demonstrated before bymass-spectrometryproteomics. Some of the enriched proteins in these experiments were also competed with othernucleotides, confirming the promiscuous binding observed before and theymay aswell be directbindersofcAMPbut lack thecyclicnucleotidebindingdomainpresent in theclassic targetsof thesmallmolecule.
With the knowledge of the strong interaction between cAMP and PKA via its regulatorysubunits, a purified recombinant human PKARIwas acquired and a cAMP analogwas synthesizedbearingaphoto-reactivemoietyandaterminalalkynereactivityfunction(CC2,onpage122)fortheestablishmentofanewCCMSworkflowusingatwo-stepbio-orthogonalchemicalligationmethod.
Chapter2:Labelingaproteinusingclickchemistry|Results
31
2.2. RESULTS2.2.1. Establishingandoptimizingclick-captureofPKARIusingacAMPCC
As previously described in the General Introduction (1.3 Click chemistry and bio-orthogonalligations,onpage21),thereareanumberofbio-orthogonalligationsdescribedintheliterature,buttodatethemostbroadlyutilizedhasbeenthecopperassistedazido-alkynecycloaddition(CuAAC).This “click” reaction was chosen to establish a labeling method of a protein using a capturecompoundcontainingtheselectivityandreactivityfunctionsbutwithanalkyneterminalfunctioninplaceofthesortingfunctionasinaclassicalcapturecompound.
To achieve a successful labeling, different parameters such as capture compoundconcentration, time of the click reaction, copper concentration and copper ligands wereindependentlyvaried.
Thecapturecompounds,competitorand ligandsutilized in thischapterare illustrated in theSpecificMethodssection(2.3.1Structuresofusedcompounds,onpage48)
As a starting point to establish this method in a CCMS workflow, a cAMP-alkyne capturecompound(CC2)waschosenandallreactionoptimizationstepswereperformedinarecombinantPKARIsolution.Theprobewasclickedtoanazide-biotin (CC3) forvisualizationandcaptureofthetarget.Asareference,aclassicalcAMP-capturecompound(CC1)wasusedinparallel.
2.2.1.1. EffectofcAMP-alkynecapturecompoundonclickyieldsinaPKARIsolution
Initialclick reactionconditionswerebasedonapublicationbySpeersetal. 23.The firststepwas a concentration-dependent assessment starting from the generic condition to determine theeffectoftheconcentrationofcAMP-alkynecapturecompoundonthelabelingofrecombinantPKARIusingbiotinasalabel.
16pmolofPKARIwasincubatedwithcAMP-alkynecapturecompound(CC2)at5µM,2.5µM,1µM,0.5µMornoneforunspecificlabelingcontrol.Asamplecontaining5µMofcAMPfullclassicalcompound(CC1)wasusedaslabelingreference.After30minequilibration,sampleswereirradiatedfor10minat310nminaCaproBox®.Next,tothecAMP-alkynecontainingsamplesthefollowingwasadded:10µMbiotin-azide(CC3),50µMCuSO4,250µMTBTA(CC16)and2.5mMsodiumascorbate.The click reactionwas performed for 30minwith shaking. To a duplicate sample of cAMP-alkynecapture compound at 5 µM, all click reagents except for the biotin-azidewere added. The resultswerevisualizedbySDS-PAGEfollowedbyBlottingusingastreptavidin-horseradishperoxidase(HRP)solution for chemiluminescence readout detected on x-ray film (hereafter referred as biotin blot).BandintensitywasmeasuredusingImageJsoftware.
ThebiotinblotshowninFigure2-2,onpage32,revealsthattheclickreactiontakesplaceandisdependentontheconcentrationofthecAMP-alkynecapturecompound.Figure2-2suggeststhatthe click reaction yield with these initial parameters is less than 50% when using the equivalentconcentrationof cAMP-alkynecapturecompoundandcomparing it to the labeling intensityof theclassical full capture compound (Figure 2-2, lane 1 vs. lane 4). Decreasing the clickable capturecompound’s concentration decreases the intensity of the biotin signal, suggesting the biotin-azidecounterpart is in excess and does not cross-link to the protein directly (Figure 2-2, lanes 5 to 7).Additionally, it can be observed that no unspecific reaction occurs in the absence of the alkyne
Chapter2:Labelingaproteinusingclickchemistry|Results
32
capture compound (Figure2-2, lane2), confirming thebio-orthogonal characteristicof theCuAAC,noristhereanunspecificsignalintheabsenceofthebiotinlabel(Figure2-2,lane3).
Figure 2-2: Labeling of recombinant PKARI using CuAAC in a serialdilutionofcAMP-alkynecapturecompound.16 pmol of PKARI was incubated with: lane 1 – 5 µM full cAMPcapturecompound(CC); lane2–nocapturecompound; lane3andlane 4 – 5 µM cAMP-alkyne CC; lane 5 – 2.5 µM cAMP-alkyne CC;lane 6 – 1 µM cAMP-akyne CC; lane 7 – 0.5 µM cAMP-alkyne CC.CuAACreactionwascarriedout in lanes2to7using50µMCuSO4,250µMTBTA,2.5mMsodiumascorbateand10µMbiotin-azide(inlane 3 equivalent volume of water was added instead). Bar plotdisplaysrelative intensityofbandsmeasuredbydensitometryusingfullcapturecompound(lane1)asreference.
This initial experiment seems to suggest
that either the new cAMP-alkyne CC binds withless affinity than the original full CC or that theclickreactioninducesproteindegradationorloss.To control the latter, subsequent experimentswereprobedforPKARIusingaspecificantibody.
Inanewexperiment,PKARIsamplescontainingthedifferentconcentrationsofcAMP-alkyne
capturecompoundwereduplicatedandpre-incubatedwith1mMfreecAMPtoactascompetitor(Figure2-3).Subsequenttreatmentwasperformedsimilarlyacrosssamplesasdescribedpreviously.InFigure2-3 intensitybarswereplottedasanormalizationofthebiotin intensityagainstthetotalPKARIsignal ineachlane.Thefirstobservationisthatproteinamountisvariableacrosstheclickedsamples(Figure2-3,PKARIpanel,lanes3–12).Moreover,consistentcompetitionwasnotobservedin this experiment, except perhaps in samples containing the least amount of cAMP-alkynecompound (Figure2-3, lanes11and12),but thedifferenceof theamountofPKARIbetweenbothsamples does not allow a correct evaluation of different biotin signals which allows for theassessmentofcompetition.
Figure 2-3: Labeling ofrecombinant PKARI usingCuAAC in a serial dilution ofcAMP-alkyne capturecompoundwithcorrespondingcompetitionsamples.16 pmol of PKARI wasincubatedwith (C) or without(A) 1 mM cAMP competitor.Samples in lane 1 and lane 2were labeled with 5 µM fullcAMPcapturecompound(CC);samples in lanes3to12wereincubatedwith serialdilutionsof cAMP-alkyne CC asindicated.CuAACreactionwascarried out in lanes 3 to 12using 250µMCuSO4, 250µMTBTA, 2.5 mM sodiumascorbate and 10 µM biotin-azide. Bar plot displaysintensity of biotin normalizedto intensity of PKARI incorrespondingsample.
Chapter2:Labelingaproteinusingclickchemistry|Results
33
Theexperiments illustrated inFigure2-2andFigure2-3,onpage32,demonstrated that thelabelingsignalobtainedwithafullclassicalcapturecompoundismoreintensethantheoneobtainedwith the new CuAACmethod using the cAMP/PKARI system and the original Speers et al. CuAACconditions23.Furtheroptimizationsontheindividualcomponentsoftheclickreactionarenecessaryandtheresultsfromthoseexperimentsaredescribedahead.
2.2.1.2. EffectofcoppersulfateconcentrationonclickyieldsinaPKARIsolution
TodeterminetheeffectofcoppersulfateconcentrationonthebiotinylationofPKARIusingacAMP alkyne compound a new experiment was designed with the knowledge obtained from theinitialexperiments.
16pmolofPKARIwasincubatedwith5µMcAMP-alkynecapturecompound(CC2).Asamplecontaining5µMofcAMPfullclassicalcompound(CC1)wasusedaslabelingreference.After30minequilibration, samples were irradiated for 10min at 310 nm in a CaproBox®. Next, to the cAMP-alkynecontainingsamplesthefollowingwasadded:10µMbiotin-azide(CC3),250µMTBTA(CC16),2.5mMsodiumascorbateanddifferentCuSO4concentrations(0,50µM,250µM,500µMand750µM).Theclickreactionwasperformedfor30minwithshaking.ToanadditionalsamplecontainingthecAMP-alkynecapturecompound,allclickreagentsexcept forthebiotin-azidewereadded.Theresults were visualized by SDS-PAGE followed by Blotting using a streptavidin-HRP solution forchemiluminescencereadoutonx-rayfilm.BandintensitywasmeasuredusingImageJsoftware.
Figure 2-4: Effect of copper sulfateconcentration on the click labeling ofrecombinant PKARI using a cAMP-alkynecapturecompound.16 pmol of PKARI was incubated with 5µM full cAMP capture compound (CC)(lane1)or5µMofcAMP-alkyneCC(lanes2–7).CuAACreactionwascarriedout inlanes2to7using250µMTBTA,2.5mMsodiumascorbate, 10µMbiotin-azide (inlane 2 equivalent volume of water wasadded instead) and increasingconcentrationsofCuSO4(lanes3–7).Barplot displays relative intensity of bandsmeasured by densitometry using fullcapturecompound(lane1)asreference.
As depicted in Figure 2-4, in the conditions tested, the increasing concentration of copper
sulfateintheclickreactiononlyyieldsthehighestsignalat250µMofCuSO4.Evenatthismaximum,the signal intensityonlyaccounts for≈10%of the labelingobtainedwith theclassical full capturecompound(Figure2-4,lane5).
TodetermineifthelowyieldreactionwasduetothepresenceofthecopperligandTBTAoritsratiotocoppersulfate,newsampleswerepreparedandanalyzedaspreviouslydescribed.
Chapter2:Labelingaproteinusingclickchemistry|Results
34
Figure2-5:EffectofcopperligandTBTAontheclicklabelingofrecombinantPKARIusingacAMP-alkynecapturecompound.16pmolofPKARIwas incubatedwith (C)orwithout (A)1mMcAMPcompetitor.Sampleswere incubatedwith5µMfull cAMPcapturecompound(CC)(lanes1and2),5µMofcAMP-alkyneCC(lanes3to8andlanes11and12),ornone(lanes9and10).CuAACreactionwascarriedoutinlanes3to12using2.5mMsodiumascorbate,10µMbiotin-azide(exceptlanes7-8),250µMTBTA(exceptlanes5and6)andeither250µMCuSO4(lanes3to10)or50µMCuSO4(lanes11and12).Barplotdisplaysintensityofbiotin(panelA),PKARI(panelB)andnormalizedbiotinintensitytoPKARIincorrespondingsample(panelC).
In the biotin blot illustrated in Figure 2-5, we can observe that labeling is reduced in thecompetition sample of the classical cAMP capture compound (Figure 2-5, panel A, lane 2) whencompared to the corresponding assay (Figure 2-5, panel A, lane 1) as is expected in a CCMSexperiment. The samples containing copper sulfate and TBTA at 1:1 ratio present a similarcompetitionpattern,althoughtheoverallintensitiesofthebiotinsignalarelower(Figure2-5,panelA, lanes 3 and 4). With higher labeling intensities, although no visible competition, we have thesamples reacted with CuSO4:TBTA at 1:5 (Figure 2-5, panel A, lanes 11 and 12) and even higher,comparabletothefullclassicalcapturecompound,thesampleswithoutTBTA(Figure2-5,panelA,lanes5and6).
The signal for total PKARI is variable across the samples (Figure 2-5, panel B), as seenpreviouslyinFigure2-3,onpage32.Duetothisvariability,thesignalobtainedinthebiotinblotwasnormalized to the intensity of PKARI detected in the corresponding sample. The results of thisnormalizationareplotted inFigure2-5,panelC.Theoutcomeof thenormalizationshowsthat theclickedsampleswithoutTBTApresentthehighestbiotinylationyieldsbutnocompetitionisobserved(Figure2-5,panelC,lanes5and6).Thiscouldbeaskewedresultfromthemuchlowerabundanceofthe signal from the total protein in these samples (Figure 2-5, panel B, lanes 5 and 6). Theseobservationssuggestthatfreecoppersulfate,intheabsenceofaligand,promotesthecycloadditionreactionbutatacostofdegradingprotein.ThepresenceofTBTAminimizestheproteindegradationbutalsodiminishestheyieldoftheCuAACreaction.
Overall,theseresultsseemtoindicatethatTBTAisnotanadequatecopperligandforourclickapproachworkflow. New ligands should be investigated if they produce amore efficient reactionyieldintheCuAAC.
Chapter2:Labelingaproteinusingclickchemistry|Results
35
2.2.1.3. TestingtimeofCuAACreactiononclickyieldsandproteinstabilityinaPKARIsolution
The yields of the click reaction in a semi-complex system such as the cAMP-alkyne capturecompoundandbiotin-azideinarecombinantPKARIsolutioncouldbetime-dependent,butaproteinin solutionwill tend to degrade throughout this process. It is therefore important to determine atime-point for theclickreactionwherethe labelingyieldsaresatisfactoryandthetargetprotein ismostlyintactinthereactionvial.
Previousresultsdescribedinthepreviousparagraph2.2.1.2andillustratedinFigure2-5,havesuggestedthatthecopper ligandsofarutilizedwasnotadequateforthenewclickworkflowtobeimplemented.Hongetal.producedawater-solublecopperligand,THPTA(CC17),andshowedthatithashigherkineticsintheCuAACreactionwhencomparedtothetraditionallyutilizedTBTA(CC16)49.WesetouttodeterminetheoptimaltimefortheclickreactionofacAMP-alkynecapturecompoundtoabiotin-azideinarecombinantPKARIsolutionusingthenewligandTHPTA.
RecombinantPKARIwasincubatedwithcAMP-alkynecapturecompound(CC2)at5µM,cAMPfull classical compound (CC 1) at 5 µM or none for unspecific labeling control. After 30 minequilibration, samples were irradiated for 10min at 310 nm in a CaproBox®. Next, to the cAMP-alkynecontainingsamplesthefollowingwasadded:10µMbiotin-azide(CC3),250µMCuSO4,250µMTHPTA(CC17)and2.5mMsodiumascorbate.ToanadditionalsamplecontainingcAMP-alkynecapture compound at 5 µM, all click reagents except for the biotin-azide were added. The clickreaction was performed up until 180 min shaking and aliquots were collect at determined timepoints. The results were visualized by SDS-PAGE followed by Blotting using a streptavidin-HRPsolution for chemiluminescence readout on x-ray film. Band intensitywasmeasured using ImageJsoftware.
Figure2-6:LabelingofrecombinantPKARIusingacAMP-alkynecapturecompoundatdifferentCuAACreactiontimes.PKARIwasincubatedwith5µMfullcAMPcapturecompound(CC)(lanes1and2),5µMofcAMP-alkyneCC(lanes5–14)ornone(lanes3and4).AliquotswererecoveredatdescribedtimepointsoftheCuAACreactioncontaining250µMCuSO4,250µMTHPTA,10µMbiotin-azide(inlanes5and6equivalentvolumeofwaterwasaddedinstead)and2.5mMsodiumascorbate.SampleswiththeclassicalfullcAMPCCwere treated similarly andalso collectedatdifferent timepoints. Barplot displays intensityof biotin (panelA), PKARI (panelB) andnormalizedbiotinintensitytoPKARIincorrespondingsample(panelC).
Chapter2:Labelingaproteinusingclickchemistry|Results
36
Figure2-6,onpage35,showsthatdetectionofPKARIdecreasesasthetimeofCuAACreactionincreases(Figure2-6,panelB),mostlikelyduetoproteindegradation.Wecanalsoobserveintheseresults that the degradation is independent of the presence of either the cAMP-alkyne CC or thebiotin-azidecounterpart(Figure2-6,panelB,lanes3–6).InpanelAofFigure2-6,wecanseethatthebiotinlabelingobtainedviatheCuAACreactionreachesamaximumat60min(Figure2-6,panelA,lane11),althoughthissignalisstillafractionoftheoneobtainedbyusingtheclassicalfullcapturecompound(Figure2-6,panelA,lanes1and2).WhenthebiotinsignalisnormalizedtotheamountofPKARIdetected in thesample (Figure2-6,panelC)wecanobserveanapproximationof thesignalintensitybetweentheclassicalfullcapturecompoundandtheclickablecompoundat,e.g.,60minofreaction which was previously mentioned as the maximum biotin signal in the CuAAC reactionsamplesmeasured (Figure2-6,panelC, lane11).Highernormalized intensities canbeobservedat120minof reaction (Figure2-6, panelC, lane13), but these results areoverestimateddue to thedecreasingintensityofthetotalPKARIdetectionwhencomparedtotheinitialtimepoint(Figure2-6,panelB,lanes7and13),probablyduetoproteindegradation,precipitationordenaturation.
ThechosenCuAACreactiontimeforfutureexperimentswasbetween30and60min.
2.2.1.4. Optimizationofbiotin-azideconcentrationinCuAACreactionfordetectionofPKARI
Inordertoachievemaximumreactionyield,weresearchedtheinfluenceoftheconcentrationof biotin-azide in the CuAAC reaction using a recombinant PKARI solution and the cAMP-alkynecapturecompoundastheprimaryprobe.
PKARIsamplesforcompetitionassayswerepre-incubatedwith1mMfreecAMP(CC13)andtonormalassaysamplestheequivalentvolumeofwaterwasadded.Competitionandnormalassaysamples were then incubatedwith with 5 µM cAMP-alkyne capture compound (CC 2). Assay andcompetitionsamplescontaining5µMofcAMPfullclassicalcompound(CC1)wereusedaslabelingreference.After30minequilibration,sampleswereirradiatedfor10minat310nminaCaproBox®.Next,tothecAMP-alkynecontainingsamplesthefollowingwasadded:250µMcoppersulfate,250µMTHPTA(CC17),2.5mMsodiumascorbateanddifferentbiotin-azideconcentrations(0,5µM,10µM, 20 µM and 30 µM). The click reaction was performed for 30 min shaking. To an additionalcontrol sample containing the cAMP-alkyne capture compound, all click reagents except for thebiotin-azide were added. The results were visualized by SDS-PAGE followed by Blotting using astreptavidin-HRP solution for chemiluminescence readout on x-ray film. Band intensity wasquantifiedusingtheImageJsoftware.
Chapter2:Labelingaproteinusingclickchemistry|Results
37
Figure2-7:LabelingofrecombinantPKARIusingCuAACwithdifferentbiotin-azideconcentrations.16pmolofPKARIwasincubatedwith(C)orwithout(A)1mMcAMPcompetitor.Samplesinlane1andlane2werelabeledwith5µMfullcAMPcapturecompound(CC);samplesinlanes5to14wereincubatedwith5µMcAMP-alkyneCCandsamplesinlanes3and4wereusedascontrol.CuAACreactionwascarriedout in lanes3to14using250µMCuSO4,250µMTHPTA,2.5mMsodiumascorbateandvaryingconcentrationsofbiotin-azideasindicated.Barplotdisplaysintensityofbiotin(panelA),PKARI(panelB)andnormalizedbiotinintensitytoPKARIincorrespondingsample(panelC).
Figure2-7showsthebio-orthogonalcharacteristicoftheCuAACreactiondemonstratedbythe
absenceofabiotinylationsignalinsampleslackingeitherthecAMP-alkynecapturecompoundorthebiotinazide(Figure2-7,panelA,lanes3to6).Thesameblotalsoshowsthebiotinsignalincreasingwithmorebiotin-azide in solutionand saturatingat20µMofazide,which is still a fractionof thesignalobtainedwiththeclassicalfullcapturecompoundbuthigherthaninanypreviouslypresentedresult.Figure2-7alsoshowsthatincreasedamountsofbiotin-azideinsolutionhasaneffectonthedetection of PKARI but the intensity of these samples is always higher than 50% of the signaldetected in the samplewithoutbiotin-azide (Figure2-7,panelB, lanes5and6),except in lane11where detection is considerably lower andwas considered anoutlier. Analyzing the plot obtainedwhenthebiotin labeling isnormalizedtothetotalPKARI(Figure2-7,panelC)wecanobservethatthepre-incubationwithexcessfreecAMPcompetitordidnotproducetheexpectedreductioninthebiotin signal for those sampleswhencompared to theassaypair (e.g., Figure2-7,panelsAandC,lanes1and2).Also,reactingmorethan10µMofbiotin-azideapparentlydoesnotimprovefurtherthebiotinsignalreadout.
Chapter2:Labelingaproteinusingclickchemistry|Results
38
2.2.1.5. CaptureofPKARIwithaclickablecAMPcapturecompoundusingCuAAC
AftertheconditionsforthebiotinlabelingofPKARIusingaCuAACreactionwereoptimized,asdescribed in the previous paragraphs of this chapter, we set out to capture it using magneticstreptavidinbeads.
RecombinantPKARIwasinitiallyincubatedwith(C)orwithout(A)excessfreecAMP(CC13)for10minandthenaddedtothesamplescAMP-alkynecapturecompound(CC2)at5µMorcAMPfullclassicalcompound(CC1)at5µM.After90minequilibration,sampleswereirradiatedfor10minat310nminaCaproBox®.Next,tothecAMP-alkynecontainingsamplesthefollowingwasadded:10µMbiotin-azide(CC3),50µMCuSO4,250µMTHPTA(CC17)and2.5mMsodiumascorbate.Tothefullclassicalcompoundsamples,thesamevolumeofwaterwasadded.Theclickreactionwasrunfor30min.Analiquotwasremovedandlabeledasinputforvisualizationongel.Totheremainderofthereaction,streptavidinmagneticbeadswereaddedandmixedfor1hour.Intheend,thebeadswerecollectedwith a CaproMag®,washed and boiled in Laemmli buffer to extract the bound proteins.Several fractionsof thesampleswere recovered foranalysis: (1) thesampleaftercross-linking thecapturecompoundbeforebeadswereadded,(2)thebeadswiththecapturedcompoundand(3)theremainder supernatant after beads were recovered. All fractions were visualized by SDS-PAGEfollowedbyBlottingusingastreptavidin-HRPsolutionforchemiluminescencereadoutonx-rayfilm.BandintensitywasmeasuredusingImageJsoftware.
Figure2-8:Click-captureofPKARIusingacAMPalkynecapturecompound.PKARI was pre-incubated with (C) or without (A) excess free cAMP. The samples were then incubated with 5 µM full cAMP capturecompound(CC)(lanes3–8),5µMofcAMP-alkyneCC(lanes9–14)ornone(lanes1and2).Fractionsanalyzed:beforeadditionofbeads(input),beadcapture(beads)andremainingsampleafterbead incubation(supernatant).TheeffectofUV irradiationtoPKARIdetectionwasanalyzed(PKARI@310nm).Barplotdisplaysintensityofbiotin(panelA)andPKARI(panelB).
Chapter2:Labelingaproteinusingclickchemistry|Results
39
Inthisinitialclick-captureexperimentwealsoanalyzedtheeffectoftheUVradiationusedtocross-link the capture compounds to the targetproteinon thedetectionof PKARI. In fact, theUVradiation decreases the detection of PKARI by 10% (Figure 2-8, panel B, lanes 1 and 2), not byinducingdegradationoftheproteinbutbypromotingdimerizationofPKARI,whichwasobservableasabandnextto130kDamarkeronlyintheirradiatedsample.
When looking at the results obtained using the classical full capture compound (Figure 2-8,lanes3 to8) it is clear thatpre-incubationwithexcesscAMPcompetitorprevents thebiotinylatedcapturecompoundtobindtoPKARIandhencereducethedetectionofbiotininthatsample(Figure2-8, panel A, lane 4) when compared to the normal assay (Figure 2-8, panel A, lane 3). ThiscompetitioneffectisonceagainobservedinthecapturedbeadsfractionandsuccessisconfirmedbythedifferentialdetectionoftotalPKARIbetweenassayandcompetition(Figure2-8,panelB,lanes5and 6). Some PKARI remains in solution after the bead capture with a reversed intensity profile(Figure2-8,panelB, lanes7and8),mainlyduetothefactthat itwasnotcarryingthebiotinylatedcapturecompound(Figure2-8,panelA,lanes7and8)andthereforedidnotbindtothestreptavidinmagneticbeads.
Results fromthenewclick captureworkflowarealso shown (Figure2-8, lanes9 to14).Thebiotinsignalintheassaysampleandalmostundetectableinthecompetitionsamplebeforeaddingthebeads(Figure2-8,panelA,lanes9and10)suggeststhathavingatwostepapproachprovidesfora more evident competition effect. The corresponding PKARI blot for these samples (Figure 2-8,panel B, lanes 9 and 10) shows that the amount of protein present is similar between them andtherefore a true competition effect on the labeling using the new click workflow was obtained.Analyzing thecapturedbeads fractionof theclick samplesnobiotinylation isdetected (Figure2-8,panelA,lanes11and12),whichwouldsuggestthatthetargetproteinwasnotsuccessfullycaptured,but looking at thePKARIblot for the same samples (Figure2-8, panelB, lanes11 and12)we canobservethattheproteinwasindeedsuccessfullycapturedintheassaysampleandisnotdetectedinthecompetition.Thisdiscrepantobservationwouldsuggestthatdetectionofbiotin isnotsensitiveenoughtodeterminewithcertainty thesuccessofacaptureexperimentandshouldbe taken intoaccountinfutureexperiments.Itisalsoimportanttoconsiderthatnon-cross-linkedcompoundsarestill able to pull-down the target and after the sample treatmentnobiotin signal is detected. Theoverall capture yield in thenewclick chemistryworkflow is less thanwhenusing theexhaustivelyoptimizedclassicalcapturebutitappearstobe“cleaner”.
Overall, thenewCuAACcapturemethodofPKARIusingaclickablecAMPcapturecompoundpresented in this experimentwas successful and thisworkflowcanbeusedas a startingpoint forcapturing in more complex systems such as a cellular lysate or living cells. Before increasing thecomplexityofthesystem,anewcopperligandwasintroducedintheworkflowtoincreasetheyieldsoftheCuAACreaction.
Chapter2:Labelingaproteinusingclickchemistry|Results
40
2.2.1.6. UsingBTTEascopperligandandadjustingreactionpHincreasesPKARIclicklabelingyield
Soriano del Amoet al. demonstrated BTTE (CC 18), a new copper ligand, to have increasedreaction kinetics in comparison to the previously used TBTA (CC 16) and THPTA (CC 17) 50. Weinvestigated the effect of this new ligand on the efficiency of click labeling PKARIwith the cAMP-alkynewiththepreviouslyestablishedparameters.
Upuntilthisexperiment,labelingefficiencyoftheclickmethodwasdirectlycomparedtothenormal classical compound and themaximum labeling achievedwith the cAMP-alkyne compoundwasaround50%oftheoneobtainedwiththeoriginal fullcompound. Inordertodiscardpotentialaffinity differences towards PKARI between the classical and the clickable cAMP compounds, thecAMP-alkyne compound was pre-clicked to the biotin-azide (CC 4) before incubation with therecombinantproteinandusedasanadditionalcontrol.
As seen in Figure 2-9, the pre-clicked
cAMP compound labels PKARI similarly to thefull classical capture compound (Figure 2-9,panel C, lanes 1-2 and5-6) confirming that theaffinity of these molecules to the recombinantproteinisguidedbytheselectivityfunctionandthattherestofthecompoundhasnoimpactonthelabeling.However,whentheCuAACreactionis performed after the cAMP-alkyne is cross-
linked to the PKARI protein there is a substantial decrease in biotin labeling (Figure 2-9, panel C,lanes3and4).Theseobservationssuggestthatthecompoundsremainintactaftertheclickreactionbutwithinaproteinsolutiontheyieldsaredrasticallyreduced.
2.2.1.7. IntroducingTAMRAasadetectionfunctionusingaclickablecAMPcapturecompoundforPKARIlabeling
Biotinhasbeenusedasa labelwithintheprobesusedsofarmainlyduetothefactthattheoriginalcAMPcapturecompound(CC1)carriesitanditwasusedasreferencefortheoptimizationofPKARI labeling using the CuAAC reaction. This molecule is widely used for pull-down/captureexperiments due to its high-affinity towards streptavidin.Workingwith a purified protein such asPKARI presents little challenges when using biotinylation as a readout system, but when movingtowardslabelingproteinswithincomplexmedia,suchascellularlysatesandwholelivingcells,itcanbe a hindrance because these proteomes contain endogenously biotinylated proteins which willinterferewiththereadout.
Figure 2-9: Labeling of PKARI with a cAMP-alkyne capturecompound clicked to a biotin-azide using CuAAC with BTTE ascopperligand.PKARI was pre-incubated with (C) or without (A) excess freecAMP. The samples were then incubated with 5 µM full cAMPcapturecompound(CC)(lanes1and2),5µMofcAMP-alkyneCC(lanes 3 and 4) or 5 µMof pre-clicked cAMP-alkyne-azide-biotin(lanes 5 and 6). Bar plot displays intensity of biotin (panel A),PKARI (panel B) and normalized biotin intensity to PKARI incorrespondingsample(panelC).
Chapter2:Labelingaproteinusingclickchemistry|Results
41
Introducing a non-natural stable molecule with fluorescence properties such as TAMRA forlabeling/capturingincomplexmediareducesthebackgroundoriginatedbybiotin.
Recombinant PKARI was labeled as previously established using the optimal determinedconditions: 5µMcAMP-alkyne capture compound (CC2), 100µMCuSO4, 200µMBTTE (CC18), 5mM sodium ascorbate and 10 µM TAMRA-azide (CC 5). Aliquots of samples were taken out atdifferenttimepointsduringthecourseoftheCuAACreactionandanalyzedbySDS-PAGE/W.Blot.
Figure2-10,showsthatusingthenewTAMRA-azide iscompatibletotheCuAACreactionfor
PKARI labeling. The TAMRA label is observable only after 15 min of reaction and the intensitycontinuesto increaseuntil60minofCuAACreactionalwayswiththeexpectedcompetitioneffect,wheretheintensityislowerinthesamplespre-incubatedwithexcesscAMP.Infact,theresultsalsoshowthattheproteinintegrityisnotaffectedduringtheCuAACreactionusingtheTAMRA-azideandthe described conditions, as it was previously observedwhen running the equivalent time courseexperimentwiththebiotin-azidecounterpart(seeFigure2-6,onpage35).
ThisapproachusingtheTAMRAlabelshouldbeusedforfutureexperimentsincomplexmediasuchascellularlysateswheretheclickreactionefficiencycanbemoreeasilydeterminedwithouttheconfoundingbackgroundofbiotinylatedproteinswhenusing thebiotin-azide for theCuAAC.Aftertheparametersforthereactionaredeterminedforthespecifictargetbeingevaluatedinaparticularlysate,thebiotinclickwillhavetobeusedforthecaptureandenrichmentofthetargetsbecauseoftheavailablestreptavidin-magneticbeadssystem.
Figure 2-10: Effect of timeof CuAAC reaction on the labeling ofPKARI with a cAMP-alkyne capture compound and a TAMRA-azide.PKARI was pre-incubated with (C) or without (A) excess freecAMP. The samples were then incubated with 5 µM of cAMP-alkyneCC,10µMTAMRA-azide,200µMCuSO4,400µMBTTEand5mM sodium ascorbate. Aliquotswere taken at 15, 30 and 60min of CuAAC reaction. Bar plot displays intensity of TAMRA(panel A), PKARI (panel B) and normalized TAMRA intensity toPKARIincorrespondingsample(panelC).
Chapter2:Labelingaproteinusingclickchemistry|Results
42
2.2.2. Mass Spectrometry analysis of captured proteins fromHEK293 andHepG2lysatesusingaclassicalandclickablecAMPcapturecompound
Withalloptimization stepsput inplace,wesetout to findwhetherPKARwasenrichedandcapturedfromacellularlysateusingCuAACreaction.
Inan initialexperimentdifferent lysisbufferswereusedtosolubilizecellularpellets.HEK293cellswere lysed inabuffercontainingornotn-Dodecylβ-D-maltoside (DDM).Alternatively, ready-madeHepG2lysateswereobtainedfromInVivo.ThedifferentlysateswereincubatedwiththecAMPcapture compounds in optimized conditions as previously described in the labeling experiments.After the CuAAC reaction proteins were either precipitated or not to remove excess non-reactedbiotin-azide and re-solubilized before incubation with the streptavidin-coated magnetic beads. Adetailed description of themethod can be found in 2.3.4 Click capture of cAMP-binding proteinsfromcelllysates,onpage53.
Table2–1gathersthedataobtainedfromtheproteomicsanalysisoftheexperiment.
ClassicalcAMP-CC
ClickablecAMP-CC(noprecipitation)
ClickablecAMP-CC(withprecipitation)
Assay Competition Assay Competition Assay Competition
Totalproteinsidentified(LFQ>0)
• HEK293(+DDM) 901 1354 640 1448 1435 1467
• HEK293(-DDM) 1004 1045 991 1333 1394 1389
• HepG2 892 1256 1417 345 1459 1545
Assay/Competition fold-change > 2(%oftotalproteinsinAssay) 24 15 2
• HEK293(+DDM)170
(18.9%)98
(15.3%)124
(8.6%)
• HEK293(-DDM) 209(20.8%)
189(19.1%)
163(11.7%)
• HepG2 118(13.2%)
1226(86.5%)
178(12.2%)
Assay/Competition fold-change > 2acrossthedifferentworkflows 0
• HEK293(+DDM) 5
• HEK293(-DDM) 1
• HepG2 3Table2–1:SummaryofidentificationsfromcAMPcaptureexperimentsinHEK293andHepG2lysates.ListofproteinswasobtainedusingMaxQuantv1.2.2.4.Numbersforpositiveproteinidentificationsineachsamplearegivenandintensity-basedcalculated fold-changesbetweenassaysandcorrespondingcompetitionsaswellasacrossall sampleswithineach lysatearealsopresented.
Chapter2:Labelingaproteinusingclickchemistry|Results
43
Table2–1,onpage42, shows thatusingdifferent lysisbuffers to solubilize theproteomeofthe samecell linewillproducedifferent resultsnotonly in specifically competedproteins, rangingfrom 98 up to 209 depending on the capturemethod, but also in the total of identified proteinswhich was of only 640 in the click-capture without precipitation method and as high as 1467differentproteins inthecompetitionsampleoftheclick-capturefollowedbyprecipitationmethod.In the commercial HepG2 lysate there was a low value of 345 identified proteins in one of thesampleswhichbydefaultgreatlyincreasesthecalculatednumberofspecificallycompetedproteins,but that could have been an experimental error.When the calculated competed proteins in eachlysate subset are compared between them according to the capture method, 24 proteins areidentifiedascommonlycompetedfortheclassicalcapturecompoundacrossthedifferentlysates,15for the clickable cAMP capture compound and only 2 when an additional step of proteinprecipitationisaddedtotheclickablecapturecompound.Theseproteinsarenotcommonacrossallsamplesandmethodsbuttherearetwoproteinsthatarecompetedinalllysatesidentifiedbothwiththe classical capture compound and the clickable cAMP capture compound without proteinprecipitation–RPS8,aribosomalproteinandSRRM2,aproteininvolvedinpre-mRNAsplicing.
The different methods utilized in the experiment did not successfully produce a commonspecific binder across all proteomes and capture methods although some commonly competedproteinswere identified across the differentmethodswithin one type of lysate. These commonlycompeted proteins were mostly ribosome-associated proteins, which are usually consideredunspecificbackground(Table2–2).
Table 2–2: Proteins identified asspecifically competed with cAMPacross the different capturemethods in each type of cellularlysate.
ProteinGeneName HEK293(+DDM) HEK293(-DDM) HepG2
H1F3 ✓
RPS19 ✓
RPL27A ✓ ✓
RPS18 ✓
HABP1 ✓
SERPINB6 ✓
HRNPG ✓
RPL21 ✓
Chapter2:Labelingaproteinusingclickchemistry|Results
44
Comparing the competed proteins between each lysate within the same capture methodranges between 37 and 54 for the classical cAMP capture compound, 20 and 82 for the clickablecAMPcapturecompoundwithouttheproteinprecipitationstepandwhenthisproteinprecipitationstep is introduced it ranges between 18 and 25. These values are represented in Table 2–3highlightedinbold. It is interestingtonoticethatthenumberofcommonly identifiedascompetedproteinsusingtheclickablecAMPcapturecompoundwithaprecipitationstepisgreatlyreducedincomparisontotheothertwomethodsbuttheoverallnumberof identifiedproteins issimilar.Alsohighlighted in Table 2–3 in italic are thenumberof commonly identifiedproteins as competedbycAMPfromthesamelysatebutusingdifferentcapturemethods.
ClassicalcAMP-CC
ClickablecAMP-CC(noprecipitation)
ClickablecAMP-CC(withprecipitation)
HEK293(+DDM)
HEK293(-DDM) HepG2 HEK293
(+DDM)HEK293(-DDM) HepG2 HEK293
(+DDM)HEK293(-DDM) HepG2
ClassicalcAMP-CC
HEK293(+DDM)
HEK293(-DDM) 54
HepG2 49 37
ClickablecAMP-CC(noprecipitation)
HEK293(+DDM) 42 25 31
HEK293(-DDM) 37 20 19 20
HepG2 118 140 68 55 82
ClickablecAMP-CC(precipitation)
HEK293(+DDM) 18 18 7 8 14 70
HEK293(-DDM) 14 15 7 6 10 103 25
HepG2 25 22 5 8 15 102 18 24
Table2–3:NumberofproteinsidentifiedascompetedbycAMPincommonbetweenthedifferentsamples.Highlightedvaluesinboldrepresentthecompetedproteinsbetweendifferentlysatesusingthesamecapturemethod;highlightedvaluesinitalicrepresentthenumberofproteinsfromoneparticularlysatecompetedacrossthedifferentcapturemethods.
Chapter2:Labelingaproteinusingclickchemistry|Results
45
Focusing on the results obtained for PKA, Table 2–4 compiles the calculated fold-change ofassayandcompetitionintensitiesandthenumberofMSspectraobtainedforeachidentifiedsubunitofPKA(threeregulatorysubunits–PRKAR1A,PRKAR2AandPRKAR2B–andonecatalyticsubunit–PKACA).
Protein
ClassicalfullcAMP-CC ClickablecAMP-CC(noprecipitation)
ClickablecAMP-CC(withprecipitation)
Genename Intensityfold-
change(log2)
MSCounts Intensityfold-
change(log2)
MSCounts Intensityfold-
change(log2)
MSCounts
(Assay) (Comp.) (Assay) (Comp.) (Assay) (Comp.)
PRKAR1A
• HEK293(+DDM) 14.4 10 0 0.0 0 0 0.0 1 0
• HEK293(-DDM) 7.8 9 0 0.1 0 0 -0.1 0 1
• HepG2 13.9 10 0 -8.9 0 0 2.8 2 1
PRKAR2A
• HEK293(+DDM) 6.9 12 0 4.1 0 0 0.9 2 1
• HEK293(-DDM) 6.7 10 0 -0.9 0 0 0.9 2 0
• HepG2 8.7 16 0 7.7 0 0 -0.3 4 2
PRKAR2B
• HEK293(+DDM) 8.2 1 0 0.0 0 0 0.0 0 0
• HEK293(-DDM) 8.4 2 0 0.0 0 0 0.0 0 0
• HepG2 8.2 3 0 0.0 0 0 0.0 0 0
PKACA
• HEK293(+DDM) 0.0 0 0 0.0 0 0 7.8 0 0
• HEK293(-DDM) 0.0 0 0 -7.9 0 0 0.0 0 0
• HepG2 0.0 0 0 9.1 2 0 0.0 0 0
Table2–4:Calculatedintensityfold-changes(assayvs.competition)andMScountsforidentifiedPKAsubunitsineachsample.
As seen on Table 2–4, the classical full cAMP capture compound specifically captures the
regulatory subunitsofPKAacrossall typesof cell lysate, ranging from6.7 to14.4 log2 fold-changebetweentheassayandthecorrespondingcompetition.Additionally,noMSspectrawereobtainedinthe competition samples using the classical full capture compound, confirming the successfulblockageofthecapturecompoundbyfreecAMPcompetitor.
Chapter2:Labelingaproteinusingclickchemistry|Results
46
Resultsontheclickapproacharemorevariable.Table2–4,onpage45,showsthatthereisabigreductiononthenumberofidentifiedMSspectrainassaysusingtheclickablecapturecompoundwhen compared to the classical full capture compound,meaning that there is little enrichmentofthese subunits of PKA. Althoughwith only fewMS spectra, and sometimes none, PKA regulatorysubunit 2A (PRKAR2A) is successfully identified as competed using the new developed CuAACapproach in both HEK293 and HepG2 cells with or without protein precipitation. Additionally,PRKAR1AisalsospecificallyidentifiedwithpositiveMSspectrainHepG2lysatewhenusingaproteinprecipitationstepaftertheCuAACreactionwiththeclickablecAMPcapturecompound.
Fromanoverallobservationof the resultspresented inTable2–4,eitherHEK293cells lysedwith a DDM-containing buffer or HepG2 lysate obtained externally can be used for furtherexperimentstoimprovethemethodofcapturingcAMP-bindingproteinsincellularlysatesusingtheCuAACreaction.Noexactdataonefficiencyorreproducibilityofthisnewmethodcouldbeextractedfrom this experiment due to its design lacking technical replicates; therefore, a new captureexperimentofcAMP-bindingproteinsusingCuAACreactionwasperformed.
HEK293cells lysedinabuffercontainingDDMwereincubatedwitheitherclassicalfullcAMPcapturecompoundorwithclickablecAMP-alkynecapturecompound.Tocorrespondingcompetitionsamples free cAMP was added prior to the capture compounds. Additionally, control samplescontaining no capture compound were processed in parallel. After photo-crosslinking and CuAACreactionallsamplesweresubmittedtomethanolprecipitationandreconstitutedbeforeincubationwith streptavidin-coated magnetic beads. All samples were processed in technical triplicates toestimatereproducibilityandcalculatesignificanceofidentifications.
Table2–5belowpresentsthesummaryofidentificationsandcalculatedspecificandsignificantproteinsinacaptureexperimentusingcAMPcapturecompoundsinHEK293lysate.
ClassicalcAMP-CC ClickablecAMP-CC Assay Competition Assay Competition
Totalproteinsidentified 1462 1429 1591 1637
Student’st-test<0.05(%oftotalproteinsinAssay)
27(1.8%)
50(3.1%)
Assay/Competitionfold-change>2(%oftotalproteinsinAssay)
382(26.1%)
177(11.1%)
Competedandsignificantproteins(%oftotal)
21(1.4%)
17(1.1%)
Table2–5:SummaryofidentificationsfromcAMPcaptureexperimentsinHEK293lysates.ListofproteinswasobtainedusingMaxQuantv1.2.2.4.Numbersforpositiveproteinidentificationsineachsamplearegivenandintensity-basedcalculatedfold-changesbetweenassaysandcorrespondingcompetitions.
Having replicate samples allows for the calculation of the significance of the calculated
intensity fold-change between assays and competitions. This statistical analysis method narrowsdown382potentialcompetedhitstoonly21with95%confidenceinthecaptureexperimentsusingthe classical full cAMP capture compound. In the CuAAC reactionmethod using the cAMP-alkynecompoundthepotentialhitswithsignificantvaluesare17intotal(Table2–5).
Chapter2:Labelingaproteinusingclickchemistry|Results
47
Inanon-biasedapproach the listof significantly competedproteins forbothsetsof capturecompoundswouldbepresented,butthepurposeofthisexperimentdesignistofocusonthecAMP-bindingproteinPKA.TheresultsfocusedonPKAsubunitsarepresentedinTable2–6.
Protein
Genename Intensityfold-change(log2)
Student’st-test(pvalue)
SumofSpectralCounts
(Assay) (Comp.)
PRKAR1A
• fullcAMP-CC 13.1 0.048 5 0
• clickablecAMP-CC -0.1 0.932 2 1
PRKAR2A
• fullcAMP-CC 2.4 0.027 5 0
• clickablecAMP-CC -0.1 0.927 2 4
PRKAR2B
• fullcAMP-CC - - 2 0
• clickablecAMP-CC - - 0 0
PRKACA
• fullcAMP-CC - - 0 0
• clickablecAMP-CC -11.1 0.192 2 4Table2–6:Calculatedintensityfold-changes(assayvs.competition)andcorrespondingsignificancevalueforidentifiedPKAsubunitsfromcaptureexperimentsusingfullandclickablecAMPcapturecompounds.Sumofspectralcountsoftriplicateassaysandcompetitionsisalsoshown.
The classical full cAMP capture compound was able to specifically capture two different
regulatory subunits of PKA – PRKAR1A and PRKAR2A – from the HEK293 lysate. Intensity fold-changesof13.1and2.4with5identifiedMSspectraforbothregulatorysubunitswere,respectively,obtained(Table2–6).
ThemethodusedwiththeclickablecAMPcapturecompoundwasabletoproducemoreMSspectraforPRKAR1Aintheassayscomparedtothecontrols,butthecorrespondingintensitieswhenfold-changeswerecalculatedgiveresulttoanegativevalue(-0.1)althoughnon-significant(pvalue=0.932).ForPRKAR2A,moreMSspectrawere identified in thecompetitionsamplesof theclickablecapturecompoundandintensityfold-changewasalsonegative(Table2–6).Thelackofsignificanceof these results does not allow any conclusive analysis on the performance of themethod to bemade.
Previousexperimentspresentedinthischapterhaveshownalackofreproducibilityregardingthe competition effect expected when excess free cAMP competitor is added to the samples (asexample, see Figure 2-7 and Figure 2-9, on pages 37 and 40, respectively). Even if this lack ofcompetitionwouldbefurtherinvestigatedandresolved,thenaturalhighconcentrationofcAMPinacellularenvironmentisnottheoptimalsystemtofurtherproceedwithcapturingtargetsinanintactcellenvironment.
A non-natural selectivity function such as the drug dasatinib was chosen for subsequentexperiments.
Chapter2:Labelingaproteinusingclickchemistry|SpecificMethods
48
2.3. SPECIFICMETHODSFORTHECAMP/PKARISYSTEM2.3.1. StructuresofusedcompoundsCC1:ClassicalfullcAMPcapturecompound
CC2:ClickablecAMPalkynecapturecompound(containsselectivityandphotoreactivefunctions)
CC3:Clickablebiotin-PEG-azide(containslabelingfunction)
CC4:ClickedcAMP-biotincapturecompound.ReactionproductofCC2andCC3
Chapter2:Labelingaproteinusingclickchemistry|SpecificMethods
49
CC5:ClickableTAMRA[tetramethylrhodamine]azide(containslabelingfunction)
CC13:cAMPcompetitor
CC16:Copper-chelatingligandTBTATris(benzyltriazolylmethyl)amine
CC17:Copper-chelatingligandTHPTATris(3-hydroxypropyltriazolylmethyl)amine
CC18:Copper-chelatingligandBTTEbis(tert-butyltriazoly)ethanol
Chapter2:Labelingaproteinusingclickchemistry|SpecificMethods
50
2.3.2. Establishingandoptimizingclick-captureofPKARIusingacAMPCC16pmolof recombinanthumanPKARIwaspipetted into200µLPCR-tubesanddilutedwith
CaptureBufferandwatertoafinalconcentrationof0.1µM.Tothecompetitionsamples,1,2or4mMofcAMP(CC13)wasaddedand incubatedat least10minon ice.Next, cAMP-alkynecapturecompound(CC2)wasaddedtothesamplesatdifferentconcentrationsrangingfrom0.31µMto10µM.Thesampleswereincubatedatleast30minrotatingat4°CandthenUV-irradiatedfor10minat310 nm in a CaproBox™ to promote photo crosslinking. To the samples containing cAMP-alkyne,reagentsfortheCuAACreactionwereaddedandtestedatdifferentconcentrations:biotin-azide(CC3)waspresentbetween5µMand30µM;CuSO4between50µMand750µM;andthreedifferentcopperligandsusedindividually–TBTA(CC16)at250µM,THPTA(CC17)at150µMor250µM,orBTTE(CC18)usedat150µMor200µM.Toinitiatetheclickreaction,sodiumascorbatewasalwaysadded lastly at a final concentration of 2.5 mM. The CuAAC time of reaction was performed atdifferent times. A new TAMRA-labeled azide (CC 5) was also clicked to the cAMP-alkyne capturecompound in one set of samples at a final concentration of 5 µM.All parameters tested for eachexperimentandcorrespondingobtainedresultsarespecifiedinTable2–7onthefollowing2pages.
Insomeexperiments,paralleltotheclicksamples,thefullclassicalcAMPcapturecompound(CC 1) was added at a final concentration of 5 µM to be used as a reference control. To thesecontrols,allclickreagentsexceptfortheazidewereadded.
TheresultswerevisualizedbySDS-PAGEfollowedbyBlottingwithchemiluminescencereadoutonx-rayfilmusingeitherastreptavidin-HRPsolution(forbiotin-labeledsamples)orIgGanti-TAMRAfollowedbyanti-IgG-HRPsecondaryantibody(forTAMRA-labeledsamples).
BandintensitywasmeasuredusingImageJsoftwareforfold-changecalculations.
2.3.3. PreparationofHEK293lysatesHEK293 cell pellets were lysed in 2X their volume with a hypotonic buffer (10 mM HEPES
pH7.9, 1.5 mM MgCl2, 10 mM KCl) containing (+) or not (-) 0.5%(w/v) DDM. Both buffers weresupplementedwithproteinaseinhibitorsandbenzonaseatthetimeoflysis.After30minincubationon ice, the lysatewasdounced20Xusinga tightpestleandthenrunthrougha21Gneedlewithasyringe.Remainingdebriswaspelletedbyrefrigeratedcentrifugationat15.000rpmfor30min.Thesupernatantwas recovered and the leftover pelletwas discarded. The lysateswere aliquoted andsnap-frozeninliquidnitrogen.Sampleswerestoredat-80°Cuntiluse.
ProteinconcentrationoflysateswasdeterminedbytheBCAassayasdescribedintheGeneralMethodssection(6.2.1Proteinconcentrationdetermination,onpage118).
Chapter2:Labelingaproteinusingclickchemistry|SpecificMethods
51
PARAMETEROPTIMIZATIONFORLABELINGANDCAPTURINGOFPKARIUSINGCLICKCHEMISTRY cAMP-alkyne
CCcAMP
competitorCuAACreaction
biotin-azide TAMRA-azide CuSO4Copperligand sodium
ascorbate TBTA THPTA BTTE
Experiment1
Results2.2.1.1
(Figure2-2)
5µM -
30min 10µM - 50µM 250µM - - 2.5mM2.5µM -1µM -0.5µM -0µM -
Experiment2
Results2.2.1.1
(Figure2-3)
5µM
1mM 30min 10µM - 250µM 250µM - - 2.5mM2.5µM1.25µM0.63µM0.31µM
Experiment3
Results2.2.1.2
(Figure2-4)
5µM - 30min 10µM -
0µM
250µM - - 2.5mM50µM250µM500µM750µM
Experiment4
Results2.2.1.2
(Figure2-5)
5µM 2mM 30min 10µM -250µM
0µM250µM - - 2.5mM
50µM 250µM
Experiment5
Results2.2.1.3
(Figure2-6)
5µM -
2min
10µM - 250µM - 250µM - 2.5mM
10min20min30min60min90min120min180min
Chapter2:Labelingaproteinusingclickchemistry|SpecificMethods
52
PARAMETEROPTIMIZATIONFORLABELINGANDCAPTURINGOFPKARIUSINGCLICKCHEMISTRY cAMP-alkyne
CCcAMP
competitorCuAACreaction
biotin-azide TAMRA-azide CuSO4Copperligand sodium
ascorbate TBTA THPTA BTTE
Experiment6
Results2.2.1.4
(Figure2-7)
5µM 2mM 30min
0µM
- 250µM - 250µM - 2.5mM5µM10µM20µM30µM
Experiment7
Results2.2.1.5
(Figure2-8)
5µM 2mM 20min 10µM - 50µM - 150µM - 2.5mM
Experiment8
Results2.2.1.6
(Figure2-9)
5µM 2mM 10min 10µM - 50µM - - 150µM 2.5mM
Experiment9
Results2.2.1.7
(Figure2-10)
5µM 2mM
15min
- 5µM 100µM - - 200µM 2.5mM30min
60min
Experiment10
Results2.2.2
10µM 4mM 30min 20µM - 100µM - - 200µM 2.5mM
Table2–7:SummaryofparametersutilizedinPKARIlabelingandcapturingwithacAMPcapturecompoundviaaCuAACreaction
Chapter2:Labelingaproteinusingclickchemistry|SpecificMethods
53
2.3.4. ClickcaptureofcAMP-bindingproteinsfromcelllysatesHEK293andHepG2cellularlysateswerepre-clearedofendogenouslybiotinylatedproteinsby
incubating with streptavidin-coated magnetic beads prior to the capture experiment. The lysates
wereincubatedwiththebeadsfor45minrotatingat4°Candthenthesewereremovedusingthe
caproMag™.
The cleared lysatewas distributed in 200µL PCR-tube strips and dilutedwithwater and 5X
Capture Buffer. To the competition samples, 4 mM of cAMP (CC 13) was added and the
correspondingvolumeofwaterwasaddedtotheassaysamples.Thevialswereincubatedfor10min
oniceandthen10µMofeithercAMP-alkyne(CC2)orclassicalfullcAMP(CC1)capturecompounds
wereaddedtothedesignatedtubes.Thesampleswere incubatedfor1hrotatingat4°Candthen
irradiatedfor10minat310nminacaproBox™topromotephoto-crosslinkingofthecompoundsto
thebindingtargets.TothevialscontainingthecAMP-alkynecapturecompoundtheCuAACreaction
wasperformedbyadding20µMbiotin-azide (CC3),100µMCuSO4,200µMBTTE (CC18)and2.5
mMsodiumascorbate.TothevialscontainingtheclassicalfullcAMPcapturecompound,thesame
volumeofwaterwasadded.Theclickreactionranfor30minshakingatroomtemperature.
Theclickedsamplesweredividedintwoandonesetwasmixedwithmethanolpre-cooledat-
20 °C to precipitate proteins and leave non-clicked biotin-azide in solution. These samples were
storedover-nightat-80°Ctoallowforprecipitation.Theotherfractionofclickedsamplesandthe
samplescontainingtheclassicalfullcapturecompoundweresnap-frozeninliquidnitrogenandkept
at-80°Cforthedurationoftheprecipitationprocedure.Theprecipitatedsampleswerepelletedfor
20min spinning at 15.000 rpm. The supernatantwas discarded and the tubes let to dry at room
temperature.Theprecipitatewaswashedthreetimesbyressuspendingitinfreshcoldmethanoland
vigorouslyvortexingtohomogenizethepellet.Afterthelastwash,thepelletswereresuspendedin
4%SDSandsonicateduntilaclearsolutionwasobtained.ThesamplesweredilutedwithPBS toa
finalconcentrationof0.4%SDSandincubatedwithstreptavidin-coatedmagneticbeadsforcapturing
of biotinylated proteins. The non-precipitated samples were thawed and also incubated with the
beadsforcapturing.
Themagneticbeadswereincubatedwiththelysatesforatleast1hrotatingat4°Candthen
recoveredusingacaproMag™.Thebeadswerethenwashedsixtimeswithwashbufferfollowedby
threetimeswashwith80%acetonitrileandafinalwashwithMS-gradewater.
0.5µgoftrypsininammoniumbicarbonatewasaddedtothebeadsandincubatedover-night
shakingat37°C.Thesupernatantcontainingthetrypticdigestwasrecoveredtocleantubes;next,
the beads were washed with MS-grade water and the wash solution pooled with the original
supernatant. All samples were lyophilized and analyzed by LC-MS/MS following the protocol
describedintheGeneralMethodssection(6.3MassSpectrometry,onpage121).
Chapter3:IdentifyingDasatinibtargetsincells|Introduction
54
3. IdentifyingDasatinibtargetsincells3.1. INTRODUCTION
Having implementedthebasicstepsforanovelclick-capturesysteminacell lysatewiththe
cAMPcapturecompound,thenextstepwouldbetoproceedtoalivingcellcapture.cAMPisanon-
permeablehighlyabundantmoleculeincells, inthemicromolarrange,whichmakesitnon-idealto
establishanewworkflowforalivingcellcapturingsystem.Ideally,thecapturecompoundtobeused
for target identification in cell culture should be a non-naturally occurringmolecule.With this in
mind,acollaborationwasestablishedwithProf.YaofromtheUniversityofSingaporewhopublished
apaperonthecellulartargetsofdasatinibusingaclickablecapturecompound92.
Dasatinib (CC 14) is a protein tyrosine kinase inhibitor initially developed by Bristol-Meyers
Squibb(commercializedasSprycel)totreatimatinib-resistantchronicmyeloidleukemia93.
Chronicmyeloidleukemia(CML)accountsfor15%ofallnewcasesofleukemiaintheWestern
hemispherewithaprevalenceof1/100.000peryear,whichuntreatedhasamediansurvivalof5to7
years94. InCML thehematopoietic stemcells possess an abnormal chromosome, thePhiladelphia
chromosome,whichoriginates fromthereciprocal translocationofaregionofchromosome9that
containstheAblesononcogene(ABL)andthebreakclusterregion(BCR)genefromchromosome22.
This translocation produces a new fusion gene that ends up being translated into the BCR-ABL
protein.TheABLfamilyofproteintyrosinekinases,whichiscomprisedofABL1andABL2(alsoknown
as ARG), link diverse extracellular stimuli to signaling pathways that control cell growth, survival,
invasion,adhesionandmigration.Thereisastrongrelationshipbetweenderegulationofthisfamily
of protein kinases and cancer, from leukemia to solid tumors95. Whilst in most leukemia the
deregulation originates from chromosome translocations that produce fusion proteins, in solid
tumors the activation of the ABL kinases is driven by their over-expression, chemokine receptors,
oxidativestressand/orinactivationofnegativeregulatoryproteins95.
ABL1 and ABL2 function redundantly in some cellular contexts, but also have unique roles,
owing to distinct and conserved sequences and structural domains (Figure 3-1, onpage 55).ABL1encodesanon-receptortyrosinekinasethatphosphorylatessubstrateproteinswithcrucialcellular
activities such as increased proliferation, loss of stromal adhesion, and resistance to apoptosis94.
Abl1-knockout mice exhibit diverse abnormalities leading to premature senescence and neonatal
morbidity.Bycontrast,Abl2-knockoutmiceareviablethoughtheyexhibitneuronaldefects96.
UnlikePKAwithitssingleactivatingmoleculecAMP(asdiscussedinthepreviouschapter),the
ABLkinaseisactivatedbymanydifferentextrinsicligands,suchasepidermalgrowthfactorreceptor
(EGFR),platelet-derivedgrowthfactorreceptor(PDGFR),vascularendothelialgrowthfactorreceptor
(VEGFR),andothers95.TheABLkinaseisalsoactivatedbyintrinsicsignalssuchasDNAdamageand
oxidativestress97.
Chapter3:IdentifyingDasatinibtargetsincells|Introduction
55
Figure3-1:SchematicrepresentationofthemodulardomainsoftheABLkinases.Alternative splicing of ABL1 and ABL2 produces several isoforms, including the 1a isoforms (as shown by the straight line) and the 1bisoforms (jagged line); the 1b isoforms are targeted for amino-terminal myristoylation. The ABL N termini are comprised of the SRChomology 3 (SH3), SH2 and SH1 (tyrosine kinase) domains. The ABL carboxyl termini contain a conserved filamentous (F)-actin-bindingdomain(BD).ABL1hasaglobular(G)-actin-bindingdomainandaDNA-bindingdomain,whereasABL2hasasecondinternalF-actin-bindingdomainandamicrotubule(MT)-bindingdomain.ABL1hasthreenuclearlocalizationsignal(NLS)motifsandonenuclearexportsignal(NES)initsCterminus.BothABL1andABL2haveconservedPXXPmotifstomediateprotein–proteininteractions.Phosphorylation(P)ofABL1atY412withintheactivationloop(Y439inABL2)andY245intheSH2-kinasedomainlinker(Y272inABL2)stabilizestheactiveconformation.ReproducedfromGreuberetal.95
The catalytic activities of the ABL kinases are tightly regulated by inter and intramolecular
interactions, and post-translational modifications (Figure 3-2). The intramolecular inhibition is
obtained by the formation of a clamp structure involving a SRC homology (SH) 3 domain*, a SH2
domainandaSH1domain.ABLactivitycanbenegativelyorpositivelyregulatedby intermolecular
interactionswith distinct binding partners. In general, the intermolecular interactions that disrupt
the autoinhibitory interactions and stabilize the active (open) conformation of the ABL kinases
promoteincreasedenzymaticactivity.Forexample,thecatalyticefficiencyofABLcanbeenhanced
byinteractionswithadaptorproteinssuchasRasandRabinteractor1(RIN1),whichinteractsboth
withtheSH3andSH2domains95.
Figure3-2:Representationofauto-inhibited(closed)andactive(open)ABLkinases.IntramolecularinteractionsplayaroleintheregulationofABLkinaseactivity.TheSH3domainbindstothelinkersequenceconnectingtheSH2andthekinasedomains,andtheSH2domain interactswithC-terminal lobeof thekinase(SH1)domainforminganSH3–SH2clampstructure.ThemyristoylatedresidueintheNterminusofthe1bABLisoformsbindstoahydrophobicpocketwithintheC-lobeofthekinasedomain,stabilizingtheauto-inhibitedconformation.TheconfigurationandinteractionsoftheABLC-terminalsequencesarenotincludedinthemodel.DashedlinerepresentsABLN-terminalsequencesupstreamoftheSH3domainthatfoldovertopromotebindingofthemyristoylgrouptoapocketintheC-lobeofthekinasedomain.AdaptedfromGreuberetal.95
*SRChomology(SH)domainsaredividedinthreecategories:SH1domainreferstothetyrosinekinasedomainwhichwas
firstidentifiedintheSRCkinase;SH2domainisaproteinmodulethatbindstotyrosinephosphorylatedsitesinasequence-
specificcontext;SH3domainisaproteinmodulethatbindstoproline-richsequences.
Chapter3:IdentifyingDasatinibtargetsincells|Introduction
56
Among the proteins reported to negatively regulate ABL kinase activity are peroxiredoxin 1
(PRDX1;alsoknownasPAG)98,thetumorsuppressorproteinFUS1(alsoknownasTUSC2)
99andthe
phosphatase-interacting adaptor protein PSTPIP1100. Finally, the enzymatic activity of the ABL
kinasesisalsoregulatedbytyrosinephosphorylation,whichstabilizestheactiveconformation,and
canoccurbytrans-phosphorylationbetweenABLkinasesormaybemediatedbySRCfamilykinases95.
Inhibitors of the ABL kinases are classified into three main classes on the basis of their
mechanism of action: type 1 inhibitors target the active conformation of the kinase domain
(dasatinibandbosutinib)101,102
,whereastype2 inhibitorsstabilizetheinactiveconformationofthe
kinasedomain,preventingitsactivation(imatinib,nilotinibandponatinib)103–105
.Thethirdcategory
of inhibitors includesallosteric inhibitors,whichdonot compete forATPbindingbutwhich rather
bindtoregulatorydomainstoinhibitkinaseactivity(GNF-2andGNF-5)106,107
.IncontrasttotheATP-
competitiveinhibitorsthattargetmultiplekinases,theallostericinhibitorsarehighlyselectiveforthe
ABLkinases95.
In2011,Davisetal.108publishedanextensiveanalysisontheselectivityof72kinaseinhibitorsagainst442kinases,bymeansofacompetitionbindingassay.Fromthose72kinaseinhibitors,only7
presentedacalculatedmedianaffinityoflessthan30nMforthetestedvariationsofABL1andABL2:
imatinib103,dasatinib
109,bosutinib
102,nilotinib
104,VX-680/MK-0457
110,PD-173955
111andAST-487
112.Allthesekinaseinhibitorsinteractedwithotherproteinsatthesame30nMcutoff,rangingfrom
only2morewithnilotinibto43additionaltargetswithAST-487.Basedonthesupplementalresults
fromDavisetal. 108, Figure3-3wasproduced to illustrate theproteins that are targetedby theseselectedkinaseinhibitorsandhowtheyoverlapbetweenthem.
Figure3-3:Proteinsinhibitedbykinaseinhibitorswithanaffinityoflessthan30nMforABL1andABL2.Sevenkinaseinhibitors(imatinib,dasatinib,VX-680/MK-0457,bosutinib,nilotinib,PD-173955andAST-487)fromatotalof72presentedanaffinityagainstABL1andABL2oflessthan30nM.Thisthresholdwasappliedtotheothertargetswithpositiveresultsandlisted.Amapoftheoverlappingtargetsbyeachofthekinaseinhibitorswasdrawn.ThisfigureisbasedresultsfromDavisetal.108
Chapter3:IdentifyingDasatinibtargetsincells|Introduction
57
Dasatinib is an ATP-competitive protein tyrosine kinase inhibitor, which was originally
identifiedasapotentinhibitorofSRCfamilykinases(includingSRC,LCK,HCK,YES,FGR,LYNandFYN)
andwas subsequently found to have activity against ABL, KIT, themacrophage colony stimulating
factorreceptor(FMS),theplatelet-derivedgrowthfactorreceptor(PDGFR)andtheephrinreceptor
family members EPHB1, EPHB2 and EPHB4113. The receptor tyrosine kinase DDR1 and the
oxireductaseNQO2havealsobeenrecently identifiedasspecificbindersofdasatinib19. Theexact
mechanismofdasatinibactioninCMLisstillunknownandtherearemanytheories114.Somestudies
have shown that dasatinib induces defects in spindle formation, cell cycle arrest, and centrosome
alterations in leukemic cells, tumor cell lines, as well as in normal cells115. The mechanisms of
dasatinib and its role in CML as well as how to decrease the toxicity of dasatinib require further
investigation114,soanunbiasedfullproteomeanalysisofdasatinibbindersmightelucidatesomeof
thosemechanisms.
In2012,Shietal.92characterizedtheproteomeofdasatinibtargetsbyuseofclickablecapture
compounds. The uniqueness of this approachwas to address those targets directly in living cells,
where theprotein-ligand, aswell asnativeprotein-protein interactionswouldbemorenatural, as
opposed to potential artifacts and alternative protein conformations created during cell lysate
preparations that are commonly used in binding/inhibition assays. Although they identified some
protein andnon-protein kinases in livingK562 cells*, noneof themwere theABLor SRC familyof
proteins;thesewereonlyidentifiedwhenthecapturewasperformedinlysate(supplementarydata
tables from Shi et al. 92). Also lacking in this study is a parallel competition assay to correctly
determinespecificbindersofdasatinib.
BasedontheworkfromShietal.92andhavingadirectcollaborationwiththatgroup,wesetout to capture dasatinib targets in living cells using our established and optimized click capture
workflowdescribedinthepreviouschapter.
*K562cells:celllinecontainingthePhiladelphiachromosomeandthereforecarryingtheBCR-ABLfusionprotein.Usedas
cellmodelofchronicmyeloidleukemia(CML)
Chapter3:IdentifyingDasatinibtargetsincells|Results
58
3.2. RESULTS3.2.1. Introducinganewcopper-chelatingazideintheCuAACreaction
The experiments done so far to optimize the CuAAC reaction were all performed using a
commerciallyavailablebiotin-azide (CC3,onpage122)asdescribed in thepreviouschapter (2.2.1
Establishingandoptimizingclick-captureofPKARIusingacAMPCC,onpage31).
Valentina Bevilacqua, a fellow member of the BioChemLig consortium, synthesized new
copper-chelating azides to be used in bioconjugation experiments116. The most efficient azide
regardingreactionkineticsatthetimewaschosenandabiotin(CC7)orTAMRA(CC8)anchorswere
attached to it to perform labeling and capture experiments. The structures of the compounds are
illustratedontheSpecificMethodssection(3.3.1Structuresofusedcompounds),onpage77.
An initialexperimentwasdesigned to test theefficiencyof thisnewazide in the labelingof
recombinant SRC protein solution or in K562 cells lysate using a clickable dasatinib capture
compound (CC 6). The previously optimized click reaction parameters with the cAMP capture
compound were used as reference and the conditions used to measure the reaction kinetics in
Bevilacqua’s system were also used. The parameters used in the reaction kinetics experiments
seemedfromthestarttobetooharshforabiologicalsystem;hence,anewsetofparameterswas
designedandincludedinthelabelingexperiment.
Figure 3-4: Labeling of recombinant SRC and K562 cell lysatewith a clickable dasatinib-alkyne capture compound using a new copper-chelatingazidewithaTAMRAlabel.Recombinant SRC (lanes 1 – 6) or K562 cell lysate (lanes 7 – 12)were pre-incubatedwith (C) orwithout (A) excess free dasatinib. Thesampleswerethenincubatedwithdasatinib-alkyneCCandclickedwithclassicalbiotin-azideandBTTEH(lanes1,2,7and8),Bevilaqua’sinitialconditionswiththecopper-chelatingazide(lanes3,4,9and10)andanewoptimizedCuAACreactionwiththenewcopper-chelatingazide(lanes5,6,11and12).BarplotdisplaysintensityofTAMRAfromthebandswithinthedrawnbox.
Chapter3:IdentifyingDasatinibtargetsincells|Results
59
AsseenonFigure3-4,onpage58,thepreviouslyestablishedCuAACreactionconditionsshow
the faintestsignalwhencomparedtotheothersamples (Figure3-4, in-houseCuAAC lanes inboth
recombinantSrc-GSTandK562lysategroups).Ontheotherhand,theinvitroconditionsdeveloped
bytheauthorsofthenewmoleculeshowanintenselabelingbutalsounspecific,asrevealedbythe
smeared lanes (Figure 3-4, initial CuAAC conditions on both recombinant Src-GST and K562 lysate
groups). Thenewly establishedparameters for theCuAAC reactionwith thenew copper-chelating
azidepresentedoptimalresults,withaclearSrc-GSTbandintheassaylaneandclearcompetitionin
the adjacent lane (Figure 3-4, optimized CuAAC in the Src-GST group) aswell as in the full lysate
labeling(Figure3-4,optimizedCuAACintheK562lysategroup).
The newly established parameters for the CuAAC reaction with 20 µM TAMRA copper-
chelating-azide(CC8),100µMCuSO4and2.5mMsodiumascorbateatpH8.5wereusedforfuture
labeling experiments. Further adjustments in concentrations of the corresponding biotin copper-
chelatingazide(CC7)werefine-tunedforcaptureexperiments.
Chapter3:IdentifyingDasatinibtargetsincells|Results
60
3.2.2. CapturingdasatinibtargetsinK562celllysateusingCuAACProteins from100μgofK562cell lysatewereenrichedusingaclickableandphotoreactive
dasatinibcapturecompound(CC6)producedintheYaolaboratoryandtheirinitialresultspublished
in Shiet al. 92. The probeswere clicked to the newbiotin copper-chelating azide (CC 8) using theconditions described in the SpecificMethods (3.3.3Capturing dasatinib targets in K562 cell lysateusingCuAAC,onpage79).
Capture experiments were performed following two different workflows: click and pre-
clicked.
3.2.2.1. Pre-clickedcompoundworkflow
In the pre-clicked workflow, competition samples were pre-incubated with free dasatinib
competitor and assay samples with equivalent volume of DMSO. The complete dasatinib capture
compoundcontainingthesortingfunction(CC9)wasproducedaprioriandaddedtoallsamples.The
sampleswerethenphoto-crosslinkedandenrichedwithstreptavidin-coatedmagneticbeads.
Followingtheenrichment,trypsinwasaddedtothesamplesandtheresultingpeptideswere
recoveredandanalyzedbyLC-MS/MS.
Table3–1showsthesummaryofidentificationsonbothsetsobtainedfromtriplicatesofeach
typeofsample.
Pre-clicked Click
Alignment-
based
MS-based Alignment-
based
MS-based
Totalproteinsidentified 1764 1355 1764 1582
Student’st-test<0.05(%oftotal)
158
(9.0%)
136
(10.0%)
246
(13.9%)
198
(12.5%)
Assay/Competitionfold-change>2(%oftotal)
69
(3.9%)
50
(3.7%)
186
(10.5%)
155
(9.8%)
Competedandsignificantproteins(%oftotal)
7
(0.4%)
7
(0.5%)
48
(2.7%)
47
(3.0%)
Totalkinasesidentified(%oftotalproteins)
93
(5.3%)
77
(5.7%)
93
(5.3%)
78
(4.9%)
Student’st-test<0.05(%oftotalkinases)
14
(15.1%)
14
(18.2%)
17
(18.3%)
12
(15.4%)
Assay/Competitionfold-change>2(%oftotalkinases)
15
(16.1%)
14
(18.2%)
15
(16.1%)
11
(14.1%)
Competedandsignificantkinases(%oftotalkinases)(%oftotalcompetedandsignificantproteins)
6
(6.5%)
(85.7%)
6
(7.8%)
(85.7%)
5
(5.4%)
(10.4%)
5
(6.4%)
(10.6%)
Table3–1:SummaryofidentificationsfromdasatinibcaptureexperimentsinK562lysate.ListofproteinswasobtainedusingMaxQuantv1.3.0.3.Numbers for identificationsbasedonspectralcounts (MS-based)andalignment-basedaregiven.
Chapter3:IdentifyingDasatinibtargetsincells|Results
61
As mentioned previously, dasatinib inhibits the BCR-ABL fusion protein but other proteins,
mainlykinases,havebeendescribedtobetargetedbythissmallmolecule19,117,118
.Infact,inthepre-
clickedworkflowbothBCRandABL1areclearly identifiedassignificantlycompetedproteins inthe
K562lysate,alongsidefiveotherkinases(ABL2,SRC,YES1,RIPK2andEPHB4).Table3–2andFigure
3-5compilethedataforthissetofproteins.
Protein TotalMScounts Enrichment(Assay/DMSO) Specificity(Assay/Competition)Genename
DMSO Assay Comp.Intensityfold-
change(log2)
Student’st-test
(pvalue)Intensityfold-
change(log2)
Student’st-test
(pvalue)
ABL1k 2 55 4 5.4 0.021 3.8 0.022
ABL2k 0 16 0 6.7 0.009 4.0 0.011
BCR 0 48 6 5.8 0.007 3.0 0.008
EPHB4k 0 12 0 4.0 0.036 7.6 0.035
RIPK2k 0 5 0 10.5 0.017 9.4 0.017
SRCk 0 9 0 12.1 0.036 11.1 0.047
YES1k 0 19 5 8.0 0.038 3.5 0.044
Table3–2:ListofcompetedandsignificanthitsindasatinibcaptureexperimentsinK562lysateusingapre-clickedworkflow.TotalMS counts from technical replicates (n = 3) of DMSO control, assay and competition (comp.) samples. Intensity fold-change andsignificancelevelsaregivenforenrichment(assayversusDMSOcontrol)andspecificity(assayversuscompetition)ofhits.k=kinase
Figure3-5:CCMSchartofsignificantlycompetedhitsinDasatinibcaptureexperimentinK562lysateusingapre-clickedworkflow.Boxes: minimum and maximum intensity values for DMSO control (D), assay (A) and competition (C) samples. Black bar: Average ofintensity (n=3) ± SEM. Green triangle: log2 fold-change between average intensities from assay and competition samples. Proteins aredisplayedbyincreasingspecificitypvaluefromlefttoright.
AsseenintheCCMSchart(Figure3-5,above),allhitswereidentifiedinlowabundanceinthe
DMSO control samples and significantly enriched in the assay samples. BCR and ABL1, have been
extensively described as strong dasatinib binders19,108,109,113,114
andwere here identifiedwith high
intensitiesintheassaysrelativelytomostoftheotheralsosignificantlycompetedproteins.YES1was
identifiedwithhighintensityintheassaysamplesaswell;itbelongstothetyrosineproteinkinaseas
ABL1andtotheSRCproteinkinasessubfamily.
Chapter3:IdentifyingDasatinibtargetsincells|Results
62
SRCisalsoidentifiedasasignificantlycompetedproteininthisset.ABL2istheothertyrosine
proteinkinasefromtheABLfamilywithanover-lappingroleinsomebiologicalprocessescarriedout
byABL194,95
andwasalso identifiedasa significanthit fordasatinibusing thismethod.EPHB4 is a
receptor tyrosine kinase previously reported to be inhibited by dasatinib as well113. Lastly, the
serine/threonineproteinkinaseRIPK2 isalsospecificallycaptured in thisworkflow,whichwasalso
previouslydescribedasbeinginhibitedbydasatinib117.
When the targets are surveyed for their interactions using the STRING database119, an
intricatenetworkofbindingandactivitiescanbeobservedbetweenthem(Figure3-6,below).
Figure3-6:STRINGproteininteractionanalysisofsignificantlycompetedproteinsindasatinibcaptureexperimentinK562lysateusingthepre-clickedworkflow.Blue:binding;black:reaction;pink:post-translationalmodification;purple:catalysis;gray:associationincurateddatabases.
With this experiment we have shown that bona fide dasatinib targets can be identified inwholecelllysatesfromK562cellsusingaclickablephoto-reactivedasatinibprobethatwasreactedapriori with a new biotin copper-chelating azide, by a bio-orthogonal chemical ligationmethod, to
enrichandidentifytheinteractionpartnersviaaLC-MS/MSshotgunanalysis.Interestingly,Shietal.92wereunabletoidentifythesetargetsintheirexperimentswithK562lysates,allowingforapossible
interpretationthattheprobedidnothavethesameaffinityasfreedasatinibbutwecannowassure
thatknowndasatinibtargetscanbeaddressedwiththedasatinib-alkynecapturecompound(CC6).
Withtheseresults,wesetouttoevaluatetheproteinsidentifiedwiththeclickabledasatinib-
alkyne capture compound in a K562 lysate but performing the CuAAC reaction after the capture
compoundisphotocross-linkedtothebindingpartner.
Chapter3:IdentifyingDasatinibtargetsincells|Results
63
3.2.2.2. Clickworkflow
In theclickworkflow, thedasatinib capturecompound (CC6wasadded to the lysate in the
assay sample as well as the competition sample that was previously incubated with excess free
dasatinib (CC14).Afterequilibration, the sampleswere irradiatedwithUV light topromotecross-
linking of the probe to the interacting proteins. Next, the counter-probe containing the sorting
function(CC7)andthereagentsfortheCuAACreactionwereaddedand,lastly,thelabeledproteins
wereenrichedwithstreptavidin-coatedmagneticbeads.
Followingtheenrichment,trypsinwasaddedtothesamplesandtheresultingpeptideswere
recoveredandanalyzedby LC-MS/MS.Adetailedmethoddescription canbe found in theSpecific
Methodssection(3.3.3CapturingdasatinibtargetsinK562celllysateusingCuAAC,onpage79).A total of 48 of potential hitswere identified. Fold-changes and correspondingp values for
intensitydifferencesaresummarizedinTable3–3,below.
Protein Enrichment(Assay/DMSO) Specificity(Assay/Competition)Genename Intensityfold-
change(log2)
Student’st-test
(pvalue)Intensityfold-
change(log2)
Student’st-test
(pvalue)ADH5 10.7 0.007 1.8 0.006
AMMECR1;AMMECR1L 8.1 0.011 2.2 0.044
ANXA6 7.4 0.008 6.5 0.008
ATIC 12.2 0.002 1.4 0.001
BRIX1 0.8 0.096 1.2 <0.001
CALML5 1.8 0.071 1.1 0.032
CBR1;CBR3 7.8 0.019 2.3 0.033
CLIC1 6.6 0.004 1.1 0.010
DBT -1.6 0.004 3.9 0.044
DDX42 9.5 0.013 1.1 0.040
DLST -4.0 0.002 2.5 0.007
ENO1 4.3 0.022 1.4 0.038
EPB41L2 9.6 0.011 1.0 0.031
ETFA 8.1 0.026 1.4 0.047
FTSJ3 0.9 0.013 1.2 0.006
GSTP1 8.0 0.014 1.5 0.022
HARS 4.8 <0.001 1.1 0.001
HIST1H1B 2.9 0.007 1.1 0.023
HIST1H1C 3.1 0.028 1.4 0.045
LCP1 6.0 <0.001 1.9 <0.001
MAPK14k 10.0 0.010 1.7 0.010
MDH2 4.9 <0.001 1.2 <0.001
MMTAG2 -1.4 0.027 6.7 0.012
MYH9 6.2 0.014 1.4 0.020
MYL12B;MYL12A;MYL9 6.1 0.007 1.1 0.019
NANS 10.7 0.005 1.2 0.004
OGDH -5.4 <0.001 1.3 0.001
PFAS 9.5 0.040 2.6 0.036
PFN1 6.3 0.002 1.1 0.002
PGK1k 4.1 0.001 1.1 0.001
PMPCA 10.6 0.007 1.5 0.006
Chapter3:IdentifyingDasatinibtargetsincells|Results
64
Protein Enrichment(Assay/DMSO) Specificity(Assay/Competition)Genename Intensityfold-
change(log2)
Student’st-test
(pvalue)Intensityfold-
change(log2)
Student’st-test
(pvalue)PREP 4.1 0.001 1.2 0.011
PRPF4Bk 2.1 0.006 1.3 0.022
PSAT1 6.9 0.003 3.2 0.003
TAOK3k 7.8 0.007 1.5 0.046
TOMM5 10.6 0.010 1.4 0.021
TPI1 3.4 0.001 1.5 0.002
TPM3 2.2 0.012 1.2 0.026
TSFM 10.0 0.007 1.2 0.047
TXLNG 9.0 0.001 1.8 0.046
TXNRD1 8.4 0.009 1.1 0.014
U2SURP 9.2 0.027 1.8 0.040
UBA1 11.4 0.015 1.0 0.037
UBA6 8.8 0.006 1.1 0.006
UBE2L3 5.2 0.002 1.4 0.002
UQCRFS1 1.8 0.079 1.5 0.046
UQCRQ 4.5 0.006 1.1 0.013
VRK1k 9.7 0.009 4.5 0.007
Table3–3:ListofcompetedandsignificanthitsinDasatinibcaptureexperimentsinK562lysateusingaclickworkflow.Intensity fold-change and significance levels are given for enrichment (assay versus DMSO control) and specificity (assay versuscompetition)ofhits.k=kinase
Somekinasesare identifiedusingtheclickworkflowin lysate:PGK1,VRK1,MAPK14,PRPF4B
andTAOK3.NointeractionsbetweenthemareknownaccordingtotheSTRINGdatabase(Figure3-7).
Figure3-7:STRINGproteininteractionanalysisofsignificantlycompetedkinasesinDasatinibcaptureexperimentinK562lysateusingtheclickworkflow.
Chapter3:IdentifyingDasatinibtargetsincells|Results
65
According to a recent publication by Kitagawa et al. 118, where tyrosine kinase inhibitors,includingdasatinib,wereprofiledagainstapanelof310kinases,MAPK14showsanIC50of330nM
for dasatinib whereas the small molecule does not inhibit PGK1. TAOK3 was not included in this
studybuttheauthorsprofiledTAOK2,whichwasalsonot inhibitedbydasatinib. Inthestudyfrom
Davis et al. 108, some additional kinases were profiled against dasatinib and results showed that
TAOK3 presented a Kd of 2300µM for dasatinib, VRK2 (not VRK1) a Kd of 3200µM and PRPF4B
showednodata. TheCCMSdata for these significantly identifiedkinases in the clickworkfloware
presentedinFigure3-8.
Figure3-8:CCMSchartofsignificantlycompetedkinasesinDasatinibcaptureinK562lysateusingCuAAC.Boxes: minimum and maximum intensity values for DMSO control (D), assay (A) and competition (C) samples. Black bar: Average ofintensity (n=3) ± SEM. Green triangle: log2 fold-change between average intensities from assay and competition samples. Proteins aredisplayedbyincreasingspecificitypvaluefromlefttoright.
BCR,ABLandSRCwerenotidentifiedassignificantlycompetedintheclickworkflow.Table3–
4presentsthevaluesfortheexpectedtargetsofdasatinibaspreviouslyobservedinthepre-clicked
workflow.
Protein TotalMScounts Enrichment(Assay/DMSO) Specificity(Assay/Competition)Genename
DMSO Assay Comp.Intensityfold-
change(log2)
Student’st-test
(pvalue)Intensityfold-
change(log2)
Student’st-test
(pvalue)ABL1k 2 10 3 2.1 0.003 0.4 0.018
ABL2k 0 0 0 2.2 0.009 0.2 0.531
BCR 0 6 7 3.3 0.016 0.3 0.262
SRCk 0 0 0 7.3 0.068 6.4 0.069
Table3–4:ValuesforexpectedtargetsofDasatinibinK562lysateusingaclickworkflow.TotalMS counts from technical replicates (n = 3) of DMSO control, assay and competition (comp.) samples. Intensity fold-change andsignificancelevelsaregivenforenrichment(assayversusDMSOcontrol)andspecificity(assayversuscompetition)ofhits.k=kinase
Chapter3:IdentifyingDasatinibtargetsincells|Results
66
Allproteinspresented inTable3–4,onpage65,areenrichedwhencompared to theDMSO
controls. Even in theabsenceof spectral counts,MaxQuant software is able to retrievedata from
otherrunsandsearchfortheintensityofthateventinthesamplebeinganalyzed120.Thisexplains
thepositiveintensityvaluewhennoMScountforaspecificproteinisobservedinthatsample.
Thespecificityfold-changeobtainedforBCRandABL2arepurelyrandom,assuggestedbythe
highpvaluesof0.262and0.531,respectively.However,thelowintensityfold-changebetweenassayandcompetitionforABL1isvalid(p=0.018),suggestingthattheclickabledasatinibprobeaddressesthis target but the overall capture yields are smaller compared to the pre-clicked compound
workflow.Infact,whenlookingatthespectralcountsforthisprotein,theassayclearlyshowsmore
identifiedpeptides(10)thantheDMSO(0)andcompetition(3)controls. Inanoppositesituationis
SRC,wherenoMScountswereidentifiedbutalignmentdatarevealshighenrichmentandspecificity
(log2fold-changeof7.3and6.4,respectively),andeventhoughitdoesnotmeettheinclusioncriteria
ofp<0.05,thedifferencetothethresholdislessthan0.02.When the 48 specific targets obtained in the click workflow are submitted to the STRING
protein interactiondatabase, the loosenetworkdepicted inFigure3-9 (onpage67) isobtained. In
fact, by analyzing further this network searching for enrichment of Gene Ontology terms, “small
molecule metabolic process” and “cellular nitrogen compound metabolic process” are obtained
(Table3–5),suggestingthehitsshownonTable3–3,onpage64,arenotspecifictargetsofdasatinib
butgeneraldrugmetabolizingenzymes.
GeneOntologyTerm–BiologicalProcess NumberOfGenes p-value p-value_FDR p-value_bonferroni
cofactormetabolicprocess 6 3.86E-06 2.61E-02 4.87E-02
cellularaminoacidmetabolicprocess 7 4.14E-06 2.61E-02 5.22E-02
smallmoleculemetabolicprocess 13 3.45E-04 1.00E+00 1.00E+00
cellular nitrogen compound metabolicprocess 18 1.84E-03 1.00E+00 1.00E+00
Table3–5:GeneOntology termenrichment fromsignificantlycompetedproteins inDasatinibcaptureexperiment inK562 lysateusingaclickworkflow.
Chapter3:IdentifyingDasatinibtargetsincells|Results
67
Figure3-9:STRINGproteininteractionanalysisofsignificantlycompetedproteinsindasatinibcaptureexperimentinK562lysateusingtheclickworkflow.Blue:binding;gray:associationincurateddatabases.
Theresultsobtainedusingthedasatinib-alkynecapturecompoundinapost-irradiationclick
usingK562 lysatedonot reflect thepreviouslyobtainedknowntargetsofdasatinib.However, this
probewasdevelopedtobecellpermeableandusedinalivingcellsystem.Therefore,wemovedon
establishing the click capture conditions in living K562 cells knowing the probe is able to identify
bonafidedasatinibtargets,asseenonthepre-clickworkflow.
Chapter3:IdentifyingDasatinibtargetsincells|Results
68
3.2.3. CellviabilityassayfordasatinibcompoundsinK562cellsBeforeperformingcaptureexperimentsinlivingcells,itisimportanttodetermineifdasatinib
and the clickable dasatinib probe affect cell viability at the used concentrations in capture
experimentssinceitcouldintroduceabiasinthesubsequentdataanalysisduetodifferencesincell
numbersandconsequentproteinamounts.
To determine the cell viability of K562 cells after incubation with dasatinib and dasatinib
clickable probe, the CellTiter-Glo® Luminescent Cell Viability Assay from Promega™ was used.
Generally,thismethoddeterminesthenumberofviablecellsinculturebasedonquantitationofthe
ATPpresent,whichsignalsthepresenceofmetabolicallyactivecells.Thekitincludesathermo-stable
luciferasethatprovidestheluminescentsignalthatcanbemeasured.
The experimentwas performed on a 96-well plate,where a dilution series of K562 cells (in
triplicate)wasused toestablisha calibration curve for the correspondingATPpresent ineach cell
dilution. A regression curvewas determined based on the average of themeasured signal. Assay
wellswerepreparedcontaining100.000cellsandeitherdasatinib (CC14–Dasa inFigure3-10)or
clickable dasatinib-alkyne capture compound (CC 6 – CPT392 in Figure 3-10) at different
concentrations and incubation times. The luminescent signals measured in the assay wells were
convertedtoATPconcentrationvaluesandcorrespondingviablecellnumbersbyinterpolationfrom
thecalibrationcurve.Thepercentageofviablecellswasthencalculatedbasedontheinitialinputof
cells.Additionalwellscontainingonlycompoundswithoutcellswerealsopreparedtoruleoutany
effect of the compound on the luminescent signal. Figure 3-10 illustrates the multi-well plate
scheme.
Figure3-10:Schemeofmulti-wellplateforcellviabilityassayofdasatinib(Dasa)anddasatinibclickablecompound(CPT392)inK562cells.WellsA1toC12:DilutionseriesofK562cells.WellsD1toG8:100.000cellswereincubatedfor30or90minwitheitherdasatinib(Dasa)ordasatinibclickablecompound(CPT392)at100nM,500nM,2.5μMand12.5μM.WellsD10toG11:onlycompoundwasadded.
Luminescence signal was obtained using an Imaging System (Syngene G:BOX XT4) and the
imageobtainedwasanalyzedwithImageJsoftwaretoextracttheintensitiesofeachwell.Datawas
furtherprocessedinMicrosoftExceltocalculatetheeffectofthecompoundsontheviabilityofK562
cells(Figure3-11,onpage69).
1 2 3 4 5 6 7 8 9 10 11 12
A 100.000 50.000 25.000 12.500 6.250 3.125 1.563 781 391 195 98 0
B 100.000 50.000 25.000 12.500 6.250 3.125 1.563 781 391 195 98 0
C 100.000 50.000 25.000 12.500 6.250 3.125 1.563 781 391 195 98 0
D100nMDasa
500nMDasa
2.5μMDasa
12.5μMDasa
100nMDasa
500nMDasa
2.5μMDasa
12.5μMDasa
100nMDasa
100nMCPT392
E100nMDasa
500nMDasa
2.5μMDasa
12.5μMDasa
100nMDasa
500nMDasa
2.5μMDasa
12.5μMDasa
500nMDasa
500nMCPT392
F100nMCPT392
500nMCPT392
2.5μMCPT392
12.5μMCPT392
100nMCPT392
500nMCPT392
2.5μMCPT392
12.5μMCPT392
2.5μMDasa
2.5μMCPT392
G100nMCPT392
500nMCPT392
2.5μMCPT392
12.5μMCPT392
100nMCPT392
500nMCPT392
2.5μMCPT392
12.5μMCPT392
12.5μMDasa
12.5μMCPT392
H
#cells(calibrationcurve)
30min1h30 nocells
Chapter3:IdentifyingDasatinibtargetsincells|Results
69
A B
Figure3-11:Luminescencedetectionofcellviabilityassayofdasatinibandclickabledasatinib-alkynecapturecompoundinK562cells.(PanelA)WellsA1toC12:DilutionseriesofK562cells.WellsD1toG8:100.000cellswereincubatedfor30or90minwitheitherdasatinibor clickable dasatinib capture compound compoundat 100nM, 500nM, 2.5 μMand12.5μM.Wells D10 toG11: only compoundwasadded.(PanelB)Plotofnumberofcellsversusluminescenceintensity(rawvolume)fromnon-incubatedcells.Linearregressionofpointsispresented.
Calculated values for the viability of K562 cells after incubationwith dasatinib or dasatinib-
clickablecompoundarepresentedinTable3–6.
%ofviableK562cells
dasatinib dasatinib-alkynecapturecompound
Compoundconcentration 30min 90min 30min 90min
12.5μM 75% 100% 74% 97%
2.5μM 90% 99% 68% 95%
500nM 95% 98% 90% 95%
100nM 99% 93% 95% 93%
Table3–6:Effectofdasatiniborclickabledasatinib-alkynecapturecompoundonK562cellviability.
AsobservedinTable3–6,increasingconcentrationsofbothcompoundsreducecellviabilityto
74-75%whenincubatedforashortperiodof30min.At90minincubations,thiseffectisnolonger
visibleandcellviabilityremainsover90%.Thiscouldindicateacellularresponsemechanismtoboth
small molecules that leads to the increase of ATP production after the initial toxic effect. Also
interesting to notice is that both compounds have the same effect pattern indicating a similar
mechanismofaction.Atlowerconcentrations(100nMor500nM),bothdasatinibandthedasatinib-
alkynecapturecompounddonotappeartobetoxictothecellsdespitetheincubationtime.
Thelackofcytotoxicityisrathersurprisingbutthisabsenteffectcouldbeduetoanacquired
resistance to dasatinib from the particular strain of K562 cells used for the experiments or,more
likely,duetotheshortincubationtimesthatareinsufficienttotriggercelldeath.
Nevertheless, the identificationof dasatinibbinders in living cells by clickCCMSwas carried
outsincethetargetswereconfirmedinthelysateofthesamestrainofcells.
Living cell capture of dasatinib targets was then performed on K562 cells using low
concentrationsofcompoundsandshortincubationtimes.
Chapter3:IdentifyingDasatinibtargetsincells|Results
70
3.2.4. CapturinginlivingK562cellsCurrentsmall-moleculetarget interactionprofiling ismainlydoneinsolution.Certainprotein
conformations are modified during the preparation of lysates and, consequently, putatively
importantprotein-protein interactions are lost.Devising anew strategy to address small-molecule
targets in their natural cellular environments allows for a more accurate profiling of targets and
partner interactions as well as directly demonstrating target engagement. Using a small clickable
compoundallowstoovercomethelimitationsofacompleteprobeintermsofcellularpermeability
andtoreducenon-specificscaffoldinteractions.ACCMSworkflowstrategyforcaptureofdasatinib
targets in living K562 cells was developed using the click conditions previously established and
optimizedinlysates.
Briefly,thecellsaregrowninmulti-wellplatestoafixedconcentrationandsupplementedwith
mediumcontainingdasatinib for the competition samplesorDMSO for the assays.After that, the
clickabledasatinibprobeisaddedtoallwellsandfurtherincubatedinacontrolledatmosphere.The
mediumisremovedandthecellsarewashedwithPBS.TheplatesarethenirradiatedwithUVlight
topromotethecrosslinkingofthereactivityfunctioninthecapturecompoundtotheboundprotein.
Thecellsarerecovered,lysedandthenclickedtothecounter-probecontainingthesortingfunction
using the reagents for the CuAAC reaction and, lastly, the labeled proteins are enriched with
Streptavidin-coatedmagneticbeads.Followingtheenrichment,trypsinisaddedtothesamplesand
theresultingpeptidesarerecoveredandanalyzedbyLC-MS/MS.
A totalofnine individualexperimentseachwith triplicatesampleswereanalyzed.Table3–7
summarizestheresultsobtained.
Experiment A B C D E F G H I
Totalproteinsidentified 1322 1322 884 1096 1430 1653 921 921 1428
Student’st-test<0.05 99 33 3 33 73 30 21 70 197
(%oftotal) (7.5%) (2.5%) (0.3%) (3%) (5.1%) (1.8%) (2.3%) (7.6%) (13.8%)
Assay/Competitionfold-change>2 577 118 16 89 73 65 83 40 160
(%oftotal) (43.6%) (8.9%) (1.8%) (8.1%) (5.1%) (3.9%) (9%) (4.3%) (11.2%)
Competedandsignificantproteins 93 5 0 14 5 4 9 2 14
(%oftotal) (7%) (0.4%) (0%) (1.3%) (0.3%) (0.2%) (1%) (0.2%) (1%)
Totalkinasesidentified 55 55 31 50 69 89 30 30 64
(%oftotalproteins) (4.2%) (4.2%) (3.5%) (4.6%) (4.8%) (5.4%) (3.3%) (3.3%) (4.5%)
Student’st-test<0.05 4 1 0 2 2 3 0 4 16
(%oftotalkinases) (7.3%) (1.8%) (0%) (4%) (2.9%) (3.4%) (0%) (13.3%) (25%)
Assay/Competitionfold-change>2 31 8 1 3 9 7 4 2 6
(%oftotalkinases) (56.4%) (14.5%) (3.2%) (6%) (13%) (7.9%) (13.3%) (6.7%) (9.4%)
Competedandsignificantkinases 3 1 0 1 1 1 0 0 1
(%oftotalkinases) (5.5%) (1.8%) (0%) (2%) (1.4%) (1.1%) (0%) (0%) (1.6%)
(%oftotalcompetedandsignificantproteins) (3.2%) (20%) (0%) (7.1%) (20%) (25%) (0%) (0%) (7.1%)
Table3–7:SummaryofidentificationsfromdasatinibcaptureexperimentsinK562livingcellsinnineseparateexperiments.(A–I)eachconsistingofatleastthreetechnicalreplicates.ListofproteinswasobtainedusingMaxQuantv1.3.0.3.
Chapter3:IdentifyingDasatinibtargetsincells|Results
71
As seen in Table 3–7, on page 70, the number of significantly competed proteins varies
between0and93,withacalculatedmedianof5.Theidentifiedproteinswithinthesenumbersare
alsovariablebetweeneachexperiment,whichhindersthecorrectidentificationofpotentialtargets
fordasatinibusingthisworkflow.
One possible way to extract significant information from this combined set of data is to
develop a scoringmethod to rank the proteins according to the frequency theywere significantly
competed in each experiment. A value of 1 is attributed to each protein that was significantly
competedintheexperimentandthefinalscoreisthesumacrossallexperiments.Usingthisanalysis
method,fourproteinshaveascorehigherthan1,meaningtheyweresignificantlycompetedinmore
thanoneexperiment.Table3–8,onpage72,summarizesthevaluesforspecificityfold-changeand
correspondingpvalueforthesetargetsineachexperiment.Additionally,resultsforABLandBCRare
alsopresentedinTable3–8.SRCwasnotidentifiedinanyexperimentusingthisworkflow.
TheproteinwiththehighestscorewasCSK,anon-receptortyrosine-proteinkinase.CSKwas
identified as specifically competed in 3 out of 9 experiments and is then the highest potential
candidate to be inhibited by dasatinib in a natural cell environment. CSK phosphorylates tyrosine
residues located in theC-terminal tails of Src-family kinases regulating their activity and therefore
plays an important role in the regulation of cell growth, differentiation, migration and immune
response121. It. According to Kitagawa et al. 118, dasatinib has an IC50 of 3.2 nM against CSK.
Additionally, dasatinib binding to CSK in living cells was previously observed in Kim et al. usingmagneticbiotinparticles
122.
AnotherpotentialtargetfordasatinibwasWARS,Tryptophan--tRNAligasethatregulatesERK,
Akt,andeNOSactivationpathwaysthatareassociatedwithangiogenesis,cytoskeletalreorganization
andshearstress-responsivegeneexpression123,124
.Interestingly,anassociationbetweenWARSand
dasatinib was previously reported in a PhD thesis by BrendanMartin Corkery, entitled “Tyrosine
KinaseInhibitorsinTripleNegativeBreastCancer”125.Inthiswork,itwasobservedahighlyconfident
increase ofWARS (annotated as IPF53 in themanuscript) in dasatinib-treated versus non-treated
cellsusingaDIGE-MALDI-TOF/TOFanalysisapproach.
TheothertwopotentialcandidatesfordasatinibtargetswerePSMC4(26Sproteaseregulatory
subunit 6B) andRPS17 (40S ribosomal protein S17).No literature about any interactions between
theseproteinsanddasatinibwasfound.
Insummary,theprofilediffersfromwhatisobservedinthelysate.Thiscouldbeattributedto
a lower efficiency of the method or that living cells are a more complex system, though more
realistic,makingtheanalysismoredifficult.
Chapter3:IdentifyingDasatinibtargetsincells|Results
72
Specificity(Assay/Competition) Intensityfold-change(log2) Student’st-test(pvalue)
ExperimentA B C D E F G H I A B C D E F G H I
Genename Score
CSK 3.87 9.19 #N/I 0.09 1.02 0.63 #N/I #N/I 3.86 0.210 0.005 #N/I 0.950 0.040 0.685 #N/I #N/I 0.001 3
WARS 1.86 6.21 -1.08 1.18 0.23 0.16 0.65 -0.18 0.64 0.031 0.425 0.480 0.048 0.633 0.645 0.196 0.599 0.012 2
PSMC4 2.59 0.69 0.13 0.68 -0.23 -1.02 -0.01 0.31 1.24 0.014 0.287 0.786 0.241 0.261 0.572 0.973 0.073 0.042 2
RPS17;RPS17L 4.60 0.54 #N/I 0.75 -0.61 -1.26 0.32 0.20 1.06 <0.001 0.461 #N/I 0.471 0.209 0.608 0.545 0.279 0.020 2
ABL1;ABL2 6.42 0.24 #N/I -0.76 0.05 #N/I -0.21 -0.09 1.34 0.022 0.797 #N/I 0.405 0.685 #N/I 0.099 0.926 0.156 1
BCR 3.18 -0.13 -0.89 0.50 0.43 #N/I #N/I #N/I 0.68 0.003 0.883 0.482 0.591 0.356 #N/I #N/I #N/I 0.351 1Table3–8:ExcerptofpotentialhitsfordasatinibinK562livingcellcaptureusingaclickworkflowfrom9individualexperiments(A–I)eachconsistingoftriplicatesforassaysandcompetitions.Valuesinboldmarkspecific(log2fold-change>1)andsignificant(pvalue<0.05)resultsforagivenexperiment.Scoreiscalculatedbythesumofnumberofresultsthatfulfilledboththecriteriaofspecificityandsignificance.#N/I:notidentified
Chapter3:IdentifyingDasatinibtargetsincells|Results
73
Analternativeanalysismethodtobypassthevariabilityobservedbetweeneachexperimentisto analyze all samples together as technical replicates, calculating a unique fold-change andassociated p value between the average normalized intensity obtained from all assays andcompetitions.
With this method of analysis only two proteins are identified as significantly competed bydasatinib:ALDOC(Fructose-bisphosphatealdolaseC)and,onceagain,CSK.
Table3–9liststhefold-changeandpvaluesforthetop20competedproteins(approximatelytop1%ofallidentifications)whenthesearesortedbyincreasingstatisticalsignificance.
Protein Specificity(Assay/Competition) Protein Specificity(Assay/Competition)
Genename Intensityfold-change(log2)
Student’st-test(pvalue)
Genename Intensityfold-change(log2)
Student’st-test(pvalue)
ALDOC* 0.97 0.008 NHP2 4.48 0.112CSK*,k 1.33 0.044 AZGP1 1.72 0.121HBA1 2.19 0.050 SARS 0.55 0.122GSTO1 0.85 0.073 NAP1L1 1.19 0.123ESF1 1.92 0.084 RAB21 1.20 0.125FAU 7.55 0.087 GAR1 1.78 0.134ABL2;ABL1k 3.33 0.088 BCR 0.75 0.136KRI1 5.27 0.099 AHCY 0.45 0.139TSR1 2.86 0.104 GARS 0.54 0.140EEF2 0.63 0.112 SLC1A5 0.36 0.145
Table3–9:Top20list(sortedbyincreasingpvalue)ofdasatinib-competedproteinswhencalculatingvaluesforallrunsfromallexperiments
26assaysand26competitions.*=significantandcompeted;k=kinase
The results obtained for ALDOC, CSK, ABL and BCR in each individual experiment can be
observed in Figure 3-12, on page 74. ALDOCwas the protein from this listwhichwasmore oftenidentified across experiments but reduced intensity values in the competitionwere only observedsignificantly in two experiments (A and F) as opposed to CSK which, as said previously, wassignificantlycompeted in threeseparateexperiments (A,Eand I).BCRandABLareonlycompetedwithasignificantpvalueinoneexperiment(A).
Analyzingtheoverallresults, itbecomesmoreapparentthatdasatinibprimarilyengagesCSKinK562cells.Infact,CSKwasoneofthetoptyrosinekinasesidentifiedbyShietal.whenperformingthe invitroclickcapture inK562 lysate 92.Also,CSKhasbeenpreviouslycapturedwithadasatinibcapture compound in HepG2 lysate using the classical CCMS workflow 13 and in the analysisperformedbyDavisetal.108dasatinibinhibitsCSKwithaKdof1nM(thisproteinisrepresentedinFigure 3-3, on page 56, as being inhibited by dasatinib but not by the other represented tyrosinekinaseinhibitors).
Chapter3:IdentifyingDasatinibtargetsincells|Results
74
Figure3-12:CCMSchartsofCSK,ALDOC,ABLandBCRproteinsinDasatinibcaptureexperimentinK562livingcellsusingaclickworkflow.
Boxes:minimumandmaximumintensityvaluesforassay(A/blue)andcompetition(C/red)samples.Blackbar:Averageofintensity(n=3)
±SEM.Greentriangle:log2fold-changebetweenaverageintensitiesfromassayandcompetitionsamples.*significantfold-change(pvalue
<0.05)
Chapter3:IdentifyingDasatinibtargetsincells|Results
75
When the protein list from Table 3–9, on page 73, is analyzed with STRING, a network ofenriched interactions is obtained (Figure 3-13).Most of the interacting proteins are associated toribosomesortothecellulartranslationalmachinery.Theothersmallernetworkobserved inFigure3-13 isbetweenABL1,ABL2andBCRwhileallother identifiedproteinshave little interactionwitheachotheraccordingtothisdatabase.
Figure3-13:STRINGproteininteractionanalysisofsignificantlycompetedproteinsindasatinibcaptureexperimentinK562lysateusingthe
clickworkflow.
Blue:binding;black:reaction;pink:post-translationalmodification;purple:catalysis;gray:associationincurateddatabases.
Furtheranalysisof thenetwork forGeneOntology terms revealsa significantenrichmentof
protein tyrosine kinases as well as other functions related to general small molecule bindingproperties (Table 3–10). Lastly, correlation analysis of this network for related diseases identifieschronicmyeloidleukemiaasatopcandidate,whichisinaccordanceofdasatinibtherapeuticuseandK562cellbiology.
GeneOntologyTerm–MolecularFunctionNumberOfGenes
pvalue pvalue_FDR pvalue_bonferroni
poly(A)RNAbinding 8 3.94E-06 7.62E-03 1.51E-02
heterocycliccompoundbinding 14 6.71E-06 7.62E-03 2.57E-02non-membrane spanning protein tyrosinekinaseactivity
3 6.75E-06 7.62E-03 2.59E-02
organiccycliccompoundbinding 14 7.94E-06 7.62E-03 3.05E-02Table3–10:GeneOntologytermenrichmentfromsignificantlycompetedproteinsindasatinibcaptureexperimentinK562livingcellsusing
aclickworkflow.
Chapter3:IdentifyingDasatinibtargetsincells|Results
76
Althoughthefirstsystemestablishedtoidentifysmallmoleculetargetsinlivingcellsviaabio-orthogonalchemicalligationmethoddidnotdelivertheexpectedresults,importantinformationcanbe retrieved. Living cell systems have inherent high variability in several structural, biological andbiochemicalcharacteristicsduetodifferentcellcyclestages,whichcaninfluencetheuptakeofsmallmoleculesandintroduceunforeseenbias.Onewaytotacklethisissuecouldbetosynchronizecellspriortoexperimentationbutthiswouldbelessrepresentativeofhowdynamicabiologicalsystemisandthemainobjectiveofintroducingthisnewapproachtoidentifysmallmoleculetargetsistobeasclosetothenaturalsystemaspossible.Analternative,whichwasconfirmedinthefinalanalysis,istoincreasethenumberofbiologicalandtechnicalreplicatestohavestrongstatisticsandconfidenceintheobtainedresults.
With this inmind, the next stepwas to identify targets from a smallmolecule with poorlyunderstoodmodeofaction.
Chapter3:IdentifyingDasatinibtargetsincells|SpecificMethods
77
3.3. SPECIFICMETHODSFORTHEDASATINIBSYSTEM3.3.1. StructuresofusedcompoundsCC5:ClickableTAMRA[tetramethylrhodamine]azide(containslabelingfunction)
CC6:ClickableDasatinibalkynecapturecompound(containsselectivityandphotoreactivefunctions)
CC7:Clickablebiotin-PEGcopper-chelatingazide(containslabelingandcopperligandfunctions)
CC8:ClickableTAMRA-PEGcopper-chelatingazide(containslabelingandcopperligandfunctions)
Chapter3:IdentifyingDasatinibtargetsincells|SpecificMethods
78
CC9:Clickeddasatinib-biotincapturecompound.ReactionproductofCC6andCC7
CC14:Dasatinibcompetitor
CC18:Copper-chelatingligandBTTEbis(tert-butyltriazoly)ethanol
3.3.2. Optimizing CuAAC reaction parameters with new copper-chelatingazide
160pmolof recombinanthumanSRCor50µgofK562 cell lysatewasdistributed in200µLPCR-tubestripsanddilutedwithwaterand5XCaptureBuffer.Tothecompetitionsamples,500µMofdasatinibcompetitor(CC14)wasaddedandthecorrespondingvolumeofwaterwasaddedtotheassay samples. The vials were incubated for 10 min on ice and then 20 µM of dasatinib-alkynecapturecompound(CC6)wasadded.Thesampleswere incubated for90minrotatingat4 °Candthen irradiated for 10 min at 350 nm in a caproBox™ to promote photo-crosslinking of thecompounds to the binding targets. The CuAAC reaction was performed in three different sets tocompareefficiencies:(i)previouslyestablishedconditionswith20µMTAMRA-azide(CC5),100µMCuSO4, 200 µMBTTE (CC 18) and 2.5mM sodium ascorbate at pH 8.5; (ii) Bevilacqua’s chemistryparameters with 10 µM TAMRA copper-chelating-azide (CC 8), 200 µM CuSO4 and 5mM sodiumascorbate; (iii) newlydesignedmixturewith20µMTAMRAcopper-chelating-azide (CC8), 100µMCuSO4and2.5mMsodiumascorbateatpH8.5.Theclick reaction ran for30minshakingat roomtemperature.ConcentratedLaemmlibufferwasaddedtothesamplesandthenboiledfor10minat100°C.
TheresultswerevisualizedbySDS-PAGEfollowedbyBlottingwithchemiluminescencereadoutonx-rayfilmusingIgGanti-TAMRAfollowedbyanti-IgG-HRPsecondaryantibody.
BandintensitywasmeasuredusingImageJsoftwareforfold-changecalculations.
Chapter3:IdentifyingDasatinibtargetsincells|SpecificMethods
79
3.3.3. CapturingdasatinibtargetsinK562celllysateusingCuAACK562 cell lysate was prepared as described in the General Methods (6.1.2.1 K562 lysate
preparation,onpage116).A full dasatinib-biotin capture compound (CC 9) was prepared a priori to be used as a
reference:500µMofbiotincopper-chelating-azide(CC7)wasequilibratedwith5mMofCuSO4for10minatR.T.andthenreactedwith500µMofdasatinib-alkynecapturecompound(CC6)inPBSatpH8.5byadding2.5mMofsodiumascorbate.Thevialwasshakenvigorouslyfor10minatR.T.
100 µg of K562 cell lysate was distributed in 200 µL PCR-tube strips and either 125 µM ofdasatinibcompetitor(CC14)wasaddedtothecompetitionsamplesorthecorrespondingvolumeofDMSOwasaddedtotheassaysamples.Thevialswereincubatedfor30minoniceandthenadded5µM of dasatinib-alkyne capture compound (CC 6) for post-click samples or 5 µM of pre-clickeddasatinib-biotin capture compound (CC 9) to the pre-click samples. To additional control sampleswithorwithoutdasatinibcompetitor,theequivalentvolumeofwaterwasadded.Allsampleswereincubated for90min rotatingat4 °Cand then irradiated for10minat350nm inacaproBox™topromote photo-crosslinking of the compounds to the binding targets. The CuAAC reaction wasperformedinthesamplescontainingthedasatinib-alkynecapturecompound(CC6)byadding5µMofbiotincopper-chelating-azide(CC7) (pre-equilibratedwithtwo-timesexcessCuSO4)and2.5mMofsodiumascorbate.Tocontrolandpre-clickedsamples,theequivalentvolumeofPBSwasadded.Theclickreactionranfor30minshakingatroomtemperature.
Streptavidin-coatedmagnetic beadswerewashedwith PBS and incubatedwith the samplesfor1h rotatingat4 °C.Thebeadswere recoveredusingacaproMag™and thenwashedsix timeswithwashbufferfollowedbythreetimeswashwith80%acetonitrileandafinalwashwithMS-gradewater.
0.5µgoftrypsininammoniumbicarbonatewasaddedtothebeadsandincubatedover-nightshakingat37°C.Thesupernatantcontainingthetrypticdigestwasrecoveredtocleantubes;next,the beads were washed with MS-grade water and the wash solution pooled with the originalsupernatant. All samples were lyophilized and analyzed by LC-MS/MS following the protocoldescribedintheGeneralMethodssection(6.3MassSpectrometry,onpage121).
3.3.4. CellviabilityassayfordasatinibcompoundsinK562cellsTo determine the cell viability of K562 cells after incubation with dasatinib (CC 14) and
clickable dasatinib-alkyne capture compound (CC 6), the CellTiter-Glo® Luminescent Cell ViabilityAssayfromPromega™wasused.
CellTiter-Glo® Reagent was prepared as described by the manufacturer and the assay wasperformed in triplicate inanopaque96-wellmultiplate.Toestablishacalibrationcurve,100µLofK562cellsataconcentrationof1x106cells/mLinRPMImediumwasdispensedinthestartingwellanddilutedsequentiallyinthesamemediumwithadilutionfactoroftwo.Controlwellscontainingmedium without cells were also prepared to determine the value for background luminescence.Assay wells were prepared with 100.000 cells and either dasatinib (CC 14) or dasatinib-alkynecapturecompound(CC6)wereaddedandincubatedfor30or90minat37°C(5%CO2).Theplate
Chapter3:IdentifyingDasatinibtargetsincells|SpecificMethods
80
anditscontentswasequilibratedatroomtemperaturefor30min.100µLofCellTiterGlo®Reagentwasaddedtothewellsandtheplatewasmixedfor2minonanorbitalshaker.Theplatewasthenallowedtoincubatefor10minatroomtemperature.
The luminescence signal was recorded in an Imaging System (Syngene G:BOX XT4) and theimageobtainedwasanalyzedwiththeImageJsoftwaretoextracttheintensitiesofeachwell.DatawasfurtherprocessedinMicrosoftExceltocalculatetheeffectofthecompoundsontheviabilityofK562byinterpolationfromthecalibrationcurve.
3.3.5. CaptureofdasatinibtargetsinlivingK562cellsusingCuAACK562cellswereressuspendedinRPMI1640mediumwithoutphenolred,supplementedwith
10%FBSand1%penicillin/streptomycin,whenusedforexperiments.1x106K562cellswereplatedonto 6-wells multiplates and left over-night in a cell incubator at 37 °C (5% CO2). The cells werepelleted for 5 min at 1.000 xg, ressuspended in serum-free RPMI medium and returned to theiroriginalwells.
10 µM of dasatinib competitor (CC 14) or equivalent volume of DMSO was added to thecompetition wells or to the assay, respectively. The plates were returned for 15 min to the cellincubator.Then,1µMofdasatinibalkynecapturecompound(CC6)wasaddedtoallwellsandthecellswere further incubated for90minat37 °C (5%CO2).Theplateswereplaced inaCaproBox™and irradiated for10minat 350nm.The cellswere recovered to tubes andpelleted for5minat1.000xg,andwashedwithD-PBS.
The treated cellswere lysed for 30min on icewith 100 µL of RIPA buffer (50mM Tris-HClpH8.0,150mMNaCl,1%IGEPALCA-630,0.5%Sodiumdeoxycholate,0.1%SDS)supplementedwith0.2mMDTT,0.5µL/mLbenzonaseandproteinaseinhibitors.Thecrudelysatewastransferredto500µLeppiesandcentrifugedfor10minat15.000r.p.m.toremovecelldebris.
The supernatant containing the solubilized proteinswas transferred to 200µL PCR strips towhichwasadded1µMbiotincopper-chelatingazide(CC7)(previouslyincubatedwith10-foldexcessCuSO4) and 2.5 mM sodium ascorbate. The click reaction ran for 30 min shaking at roomtemperature.
Streptavidin-coatedmagnetic beadswerewashedwith PBS and incubatedwith the samplesfor1h rotatingat4 °C.Thebeadswere recoveredusingacaproMag™and thenwashedsix timeswithwashbufferfollowedbythreetimeswashwith80%acetonitrileandafinalwashwithMS-gradewater.
0.5µgoftrypsininammoniumbicarbonatewasaddedtothebeadsandincubatedover-nightshakingat37°C.Thesupernatantcontainingthetrypticdigestwasrecoveredtocleantubes;next,the beads were washed with MS-grade water and the wash solution pooled with the originalsupernatant. All samples were lyophilized and analyzed by LC-MS/MS following the protocoldescribedintheGeneralMethodssection(6.3MassSpectrometry,onpage121).
Chapter4:Phenobarbital–newbindersforanolddrug|Introduction
81
4. Phenobarbital:newbindersforanolddrug4.1. INTRODUCTION
Phenobarbital (PB)was first produced in 1912 by Bayer and commercialized as Luminal®. Itwas widely used as a hypnotic/sedative for the treatment ofepilepsyuntilthelaterintroductionofbenzodiazepines,almost50years later.Tothisday,PBcontinuestobeadministeredtopatients with established status epilepticus, the most severeformofepilepsyusuallyresistanttobenzodiazepinetreatment,despite phenobarbital’s suboptimal safety profile 126. PB isknown to promote hepatocellular carcinomas in both rat andmice livers but, interestingly, the same effect has not beenobserved in humans 127, and the pathway associated to thismechanismofactionremainstobeelucidated.
It is largely known that PB induces the expression ofspecific cytochrome P450 (CYP) genes, amongst othermetabolizing enzymes present in liver microsomes 127–130. TheCYP superfamily of proteins, comprised of at least 27 distinctgenefamilies,isacollectionofstructurallyrelatedhaemoproteinmono-oxygenaseenzymesthathydroxylatea largenumberofsteroidhormones,fattyacids,drugs,carcinogens, and environmental chemicals 129. In rat liver, PB induces mainly the expression ofCYP2B1andCYP2B2butonlywithhighconcentrationsofthedrug.ThePBreceptor,ifittrulyexists,has lowaffinity forPB127.EventhoughCYP2B1andCYP2B2aretheprimarymembersof theP450superfamily inducedbyPB,theydonotappeartometabolizethedrug127.ThisknowninductionofCYPs by PB is actually used as reference to determine the hepatotoxicity of other drugs andcategorizethemas“PB-like”induceriftheresponseispositive130.
There are several reasons to think that “PB-type” induction of CYPs is involved in hepatictumorpromotionbyPB.AdetailedanalysisonPBmechanisticdataand riskassessmenthasbeencovered byWhysner et al. 127 Induction by PB of P450 enzymes is part of a complex, pleiotropicresponse.PBhasmanyeffectsonliverthatmayormaynotbemediatedbyP450induction,suchaseffectsongapjunctions,effectsonapoptosis,effectsonEGFreceptors,andeffectsonproliferation.Infact,thereareparadoxicaleffectsofPBonmousehepatocarcinogenesis131,makingitdifficulttoextrapolateexperimentaldatafromlaboratoryanimalstohumanriskassessment.Inratsandmicepre-initiatedwithgenotoxiccarcinogens,PBadministration increases thenumberofhepatocellulartumors by approximately 5-fold despite its non-genotoxicity. However, in mice PB occasionallyexhibits strong inhibitory effects on hepatocarcinogenesis initiated with the potent carcinogendiethylnitrosamine.Bothpositiveandnegativeeffectsofphenobarbitalonhepatocyticproliferationandapoptosis,whicharemechanistically involved in thepromotionstageofhepatocarcinogenesis,havebeendescribed.
Understanding the activation mechanism through which PB induces CYP expression mightprovecrucialtoabetterclarificationoftheobservedinter-specieseffectvariance.
Figure4-1:Luminal(phenobarbital)asitwas
packagedin1940/1945.
Picture from product on display at the
StadtgeschichtlichesMuseumLeipzig.
Chapter4:Phenobarbital–newbindersforanolddrug|Introduction
82
Several studies have demonstrated that intracellular phosphorylation events control the PBinduction of CYP2B genes 132–134, with some suggesting that both energy status and nutritionalenvironmentof the cell can influencePB regulationofCYP2Bgeneexpression, although theexactmechanismwasnotunderstoodat the time. In2005,Rencureletal.providedevidence thatAMP-activatedkinase(AMPK)mediatesthePBinductionofCYP2Bgenesexpression135,whichtheylaterconfirmedwithAMPKknockoutexperiments136.AMPKisaubiquitouslyexpressedmetabolic-stress-sensing protein kinase that regulates metabolism in response to energy demand and supply bydirectlyphosphorylatingrate-limitingenzymesinmetabolicpathwaysaswellascontrollinggeneandproteinexpression.Iftheenergystockdecreases,AMP/ATPratioincreasesfollowedbyactivationofAMPKwhich subsequently turns off ATP-consuming pathways such as fatty acid, triglycerides andcholesterol synthesis as well as protein synthesis and transcription, and switches on catabolicpathways suchas glycolysis and fatty acidoxidation.AMPKactivity is activatedbyawidearrayofmetabolicstresses,includinghypoxia,ischemia,andoxidativeandhyperosmoticstresses137.
In the proposed mechanism by Rencurel et al. 135,136, PB increases the AMP/ATP ratio thatactivatesAMPKthroughphosphorylationofitsThr172,areactionlaterproposedtobecarriedoutbythe LKB1enzyme 138.A connectionbetweenAMPand itsmainbinderPKA (alreadydescribed in apreviousChapter)and thePBeffectonCYPexpressionhadbeenestablishedby Joannardetal. 139who observed a competition between the two systems, suggesting that the activation of theAMP/PKApathwaydown-regulatedtheinductionofCYP2B1/2byPB.
In 1999, Kawamoto et al. 132 discovered that the constitutive androstane receptor (CAR),sometimesalsoreferredtoasconstitutivelyactivereceptor,translocatedfromthecytoplasmtothenucleus upon PB stimulus. CAR was originally described as a regulator of gene expression fromretinoic acid-responsive elements and is now associated to a series of regulation mechanismsinvolvedinthemetabolismandtransportofvariousxenobioticsandendogenouschemicals140.UponPBtreatment, thesignalingcascade leadstoarecruitmentofCAR inthenucleuswhere itbindstothe phenobarbital responsive enhancer module (PBREM) present in the CYP2B gene family andinduces the expression of P450 enzymes. In 2007, Shindo et al. 141, established the missing linkbetween AMPK activation/phosphorylation and CAR nuclear translocation to enhance CYP2Bexpression, confirming thatPButilizes theAMPKpathwayas signalingmolecules leading toCYP2Bgene expression, although not exclusively with data suggesting the presence of other activationpathways.
A more cohesive theory for the phenobarbital activation and signaling pathway has arisenduringthedevelopmentoftheworkherepresented.Mutohandco-workers142discoveredthatPBindirectlyactivatesCARbyinhibitingtheepidermalgrowthfactorreceptor(EGFR).TheactivationofCAR through dephosphorylation is catalyzed by phosphatase PP2A when associated todephosphorylatedRACK1.DephosphorylatedCARtranslocatestothenucleus,interactswithretinoidXreceptor(RXR),andactivatesthetranscriptionofCYP2BthroughthePBREMenhancer.InhibitionofEGFRwithPBprevents the receptor tobeactivatedand impedesRACK1phosphorylation,possiblythroughtheSrckinase,allowingitsassociationwithPP2AandsubsequentCARactivation.
TheinteractionofCARwiththeEGFRsignalingpathwaywaspreviouslydescribedbyMeyeretal. in 1989 143,where the results showed the affinity of EGF to its receptorwasdiminished in PB-treatedratprimaryhepatocytesalthoughtheabundanceofEGFRwasunchanged.Theexperimentsperformed at the time indicated the inhibitionwas not a direct competition for the receptor but
Chapter4:Phenobarbital–newbindersforanolddrug|Introduction
83
acted through some other underlying mechanism. Even though earlier results contradicted thecurrent findings, the author of the initial research has published a letter in support of Mutoh’sachievements144.
ArecentpublicationprovidesevidenceforthemissinglinkbetweenMutoh’sproposalandtheobservedAMPKphosphorylationinassociationwithCARtranslocation.Josephetal.145haveshownthatAMPKisdephosphorylatedbyPP2A,thesamephosphataseshowntocoordinatewithRACK1todephosphorylateCARand lead to the translocationof the latter towards thenucleus. PBbinds toEGFR preventing its activation and RACK1 phosphorylation, which binds to PP2A, preventing thephosphatase to continueactingonAMPKand thus leads toan increaseonAMPKphosphorylationlevels.
AschemedepictingtheproposedmechanismisillustratedinFigure4-2.
Figure4-2:PhenobarbitalinducesCYP2BexpressionviaEGFRinhibition
Epidermalgrowthfactor(EGF)bindingtoitshigh-affinityreceptor(EGFR)inducesdimerizationandautophosphorylationofitsintracellular
tyrosines.Theactivatedreceptorgeneratesadaptingproteindockingsitestotriggerseveraldownstreameffectorpathways,includingthose
stemmingfromthetyrosine-specificphosphorylationofSrc.ActivatedSrcinturnphosphorylatesRACK1,thusinactivatingRACK1-assisted
dephosphorylationofCAR. IncreasingAMP:ATP ratiosactivates cytosolicAMPK, viaphosphorylationby LKB1, increasingATP-generating
intracellular processes such as glycolysis and fatty-acid oxidation. TheAMPK activity is regulated by PP2A,which dephosphorylates the
kinase. Phenobarbital (PB) binding to EGFR prevents the receptor’s activation, suppressing RACK1 phosphorylation and promoting the
binding toPP2A. TheRACK1:PP2A complex recruits CARand catalyzes its dephosphorylation.DephosphorylatedCAR translocates to the
nucleus, interacts with the retinoid X receptor (RXR), and activates the transcription of CYP2B through the phenobarbital responsive
enhancermodule (PBREM).Adapted fromMeyerand Jirtle144, basedonMutoh et al.
142 and Josephet al.
145 using ServierMedicalArt
templates.
Chapter4:Phenobarbital–newbindersforanolddrug|Results
84
4.2. RESULTS4.2.1. Determiningpermeabilityofphenobarbital-alkynecapturecompounds
Three phenobarbital alkyne capture compounds bearing different photo-reactivity moietieswere synthesized by the Medicinal Chemistry group at caprotec bioanalytics (Figure 4-3). Thepermeability of the compounds was evaluated externally by Pharmacelsus (in Saarbrücken,Germany)usingaPAMPA*assay.Theconcentrationofeachcompoundismeasuredintheacceptoranddonorwellsatthebeginningandendoftheassaytocalculatethepercentageofflux.TheassaywasperformedintriplicateandtheresultsaredetailedinTable4–1.
Figure4-3:Synthesizedphenobarbital-alkynecapturecompounds
%Flux Mean[%]
SDStartingSolution[nM]
Meanrecovery
[%]Classification
Compound Assay1 Assay2 Assay3
A(CC10) 74.1 30.0 53.8 52.7 22.1 7467.1 5.0 Highpermeability
B(CC11) 39.6 52.2 41.5 44.4 6.8 8897.1 57.7 Mediumpermeability
C(CC12) 5.7 8.0 6.2 6.6 1.2 6894.8 6.3 Lowpermeability
Table4–1:PAMPAresultsofphenobarbitalalkynecapturecompounds.
Thepermeability of phenobarbital alkyne capture compoundswas calculatedbymeasuring its concentration in thedonor andacceptor
wells(n=3).Flux(%)=(acceptorwell)/sum(donorwell+acceptorwell)x100x2
CompoundAhasameanfluxof52.7%,whichclassifiesitasbeinghighlypermeable.However,
themeanrecoverywasonly5%,suggestingthatitistrappedintheartificialmembrane.CompoundBhasasmallerpermeability(44.4%)comparedtothefirstcompoundbutthehighermeanrecoveryof57.7%makesitagoodcandidatetobeusedinbiologicalexperiments.Finally,compoundCshowedlittle permeability with a mean flux of 6.6% and a recovery of 6.3%, which suggests that thiscompoundwasdegradedduringtheassay.
AlthoughcompoundBappearedtobethebestcandidatetobeusedinCCMSexperiments,thelowyieldsinitssynthesisdidnotallowforittobeproducedinenoughquantityforallexperiments.Therefore, compound A was selected to be used in the identification of phenobarbital targets inlivingcellsusingaclickablephenobarbitalalkynecapturecompound(CC10).
*PAMPA:ParallelArtificialMembranePermeationAssayisamethodthatdeterminesthepermeabilityofsubstancesfromadonorcompartmentdiffusingthroughalipid-infusedartificialmembraneintoanacceptorcompartment.
Chapter4:Phenobarbital–newbindersforanolddrug|Results
85
4.2.2. DevelopingaphenotypicassaytoevaluatephenobarbitaleffectSelection of adequate experimental systems is key in understanding the underlying
mechanismsofactionofaspecificdrug.In1984,Ferroetal.146showedtheinductionofcytochromeP450 (CYP) by phenobarbital in a rat hepatoma cell line (MH1C1),making it a good candidate as abiological system to identify phenobarbital targets and characterize its phenotypic effect. A laterstudy on the effect of phenobarbital on some CYPs and other liver metabolizing enzymes in sixdifferentrathepatomacelllinesconfirmedthatMH1C1cellscomparedbettertoprimaryhepatocytesintheirexpressionofbiotransformationactivities147.Thegoalofimplementingaread-outsystemfortheeffectofphenobarbitalistouseitlateroncepotentialtargetcandidatesareidentifiedbyCCMSandmeasureifablatingorreducingtheirexpressionproducesthesametypeofresponseobservedwiththeuseofthesmall-molecule.
To develop a phenotypic assay, three different approaches with known phenobarbitalresponses,asdescribedintheliterature,tophenobarbitalstimulationwerepursuedsimultaneously:constitutiveandrostanereceptor translocation;AMPactivatedproteinkinasephosphorylation;andcytochromeP450induction.
4.2.2.1. ConstitutiveAndrostaneReceptor(CAR)translocation
The constitutive androstane receptor (CAR) is a known responder to phenobarbital stimulus148–150viaanindirectactivationmechanism.PhenobarbitalactivatesapathwaythattranslocatesCARfromthecytoplasmtothecellularnucleuswhereitwillinducegeneexpression132.
MH1C1 cells were submitted to cellular fractionation using a commercially available kit thatisolates nuclei from cytoplasm as described in the specific methods section (4.3.3 Subcellularfractionation ofMH1C1 cells, on page 106). The isolated fractionswere resolved by SDS-PAGE andtransferredontonitrocellulosemembraneforprobingusingaspecificanti-CARantibody(Figure4-4).
Figure 4-4: Western Blot analysis of CAR in
isolatednucleiandcytosolfromMH1C1cells.
Expected molecular weight for CAR: 46 kDa
(arrowindicatespossibleCAR)
The anti-CAR blot shown inFigure 4-4 reveals that the
antibody produces considerable background, having the strongest reaction at unexpected 55 kDa.The expected reaction for CAR at 46 kDa is very faint in the pre-fractionated samples (Figure 4-4,lanes1and2),mostlyabsentinthecytosolfraction(Figure4-4,lane3)butbecomingclearerinthenuclear fraction (Figure 4-4, lane 4). The cells used in this experiment were not treated withphenobarbital therefore thepresenceofCAR in thenucleusand its absence in the cytoplasmwasunexpected,suggestingapre-activationofthepathway.
Chapter4:Phenobarbital–newbindersforanolddrug|Results
86
AnewexperimentwasdesignedtoevaluateiftheresultpresentedinFigure4-4confirmsthepresenceofCARinthenucleusinuntreatedcellsorifupontreatmentwithphenobarbitalmoreCARisrecruitedtothissubcellularfraction.
MH1C1cellsweretreatedwithphenobarbitalatdifferentconcentrations(0,1µM,10µM,100µM,250µMor500µM)during1h,2h,4hor6h.ThedetectionofCARineachsubcellularfractionbywesternblotisshowninFigure4-5.
Figure4-5:EffectofphenobarbitalonCARtranslocationfromcytosoltonucleusinMH1C1cells.
Cellswereincubatedwith0–500µMphenobarbitalfor1h,2h,4hor6hfollowedbysubcellularfractionation.Theindividualfractions
wereanalyzedbywesternblotforCARdetection.ArrowindicatespossibleCAR.ExpectedmolecularweightforCAR:46kDa
The results presented in Figure 4-5 show once again the high background created by theantibody used for the detection of CAR. The expectedmolecularweight for CAR (46 kDa) can beattributed to several bands detected between the 55 kDa and 35 kDa markers in the cytosolicfractions independently of the different phenobarbital stimulus (Figure 4-5, lanes 1 – 6). Thestrongestreactionobserved inthepicture isabandslightlyoverthe35kDamarker inthenucleusfraction(Figure4-5, lanes7–12)which, if in factcorrespondstoCAR,confirmsthattheprotein isalreadypresentwithoutphenobarbital stimulus (Figure4-5, lane7) and the intensityof thatbanddoesnotshowlinearitywithincreasingphenobarbitalconcentrationsortimeofstimulation.
The overall results are not clear and suggest that CAR translocation is not a good readoutsystemforphenobarbitaleffectonMH1C1cells.
Chapter4:Phenobarbital–newbindersforanolddrug|Results
87
4.2.2.2. PhosphorylationofAMPactivatedproteinkinase(AMPK)
AMP-activated kinase (AMPK) is describedas a cellular energy sensor, sensitive toAMP:ATPratios,whichisactivateduponphosphorylationofitsthreonine172(T172)residue151.Rencureletal.135,136 have demonstrated a relationship between phenobarbital stimulus and AMPK activation viaT172phosphorylationinhepaticcellsfromdifferentspecies.
MH1C1 cells cultured in complete DMEMmedium (10% FBS, 1% pen/strep) were starved infreshserum-freemediumpriortotheincubationwithphenobarbital.Thecellswereincubatedfor1h,2h,4hor6hwithphenobarbitalat0,1µM,10µM,100µM,250µMor500µM.Aftertreatmentthemediumwas removed, thecellswerewashedwithD-PBSandrecovered for fractionation.ThecytosolicfractionsofeachsamplewereanalyzedbywesternblotusingantibodiesagainsttotalAMPKandspecificphosphorylationonT172(pAMPK).TheresultsarepresentedinFigure4-6,onpage88.
The results of phenobarbital effect on AMPK phosphorylation in MH1C1 cells after serumstarvationshowthatconcentrationsofphenobarbitalhigherthan10µMreducetheoverallamountofAMPKinthecytosol(Figure4-6,lanes13–24,middlepanel),onlyafter1hincubation,whichisalsoobtainedwith10µMphenobarbitalbutwith6h incubation time (Figure4-6, lane12,middlepanel). The phosphorylation of AMPK (pAMPK) fluctuates over time with a tendency to bemaintained at higher concentrations of phenobarbital but still in fewer amounts compared to thesmallertimeincubation(Figure4-6,toppanel).
Looking at the control samples, which were not stimulated with phenobarbital (Figure 4-6,lanes1–4) it is clear thatAMPKactivation/phosphorylation is cyclic.After1hof the startof theexperiment, AMPK is phosphorylated; it gets inactivated after 2 h incubation; at 4 h is againphosphorylatedandintheendoftheexperimentitisagainde-phosphorylated.
A clearer picture of the effect of phenobarbital on AMPK activation is obtained when theintensity of the phosphorylation is normalized to the intensity of the total AMPK present in thesample (Figure 4-6, bottom panel). Different patterns of the time incubation effect emergewhenusing different concentrations of phenobarbital, suggesting that an analysis of the effect ofconcentrationatspecifictimepointscouldprovemoreusefultowithdrawconclusions.
Figure4-7,onpage88,showsthephosphorylationofAMPKnormalizedtothetotalamountofAMPK in the sample and in relation to the phenobarbital-free control at each time of incubation.Incubatingthecellsfor1hshowsavariationonAMPKactivationbetweena3%decreasewith10µManda3%increasewith500µMphenobarbitalincomparisonwiththecontrol.At2hincubationthevariation is the highest with an increase of 33% of AMPK phosphorylation using 250 µM ofphenobarbital, which is also observed at 6 h with an increase of 26% in comparison to thecorresponding control.With longer incubationwithphenobarbital at6h, theoverall effecton thephosphorylationofAMPKisobservedevenatlowerconcentrationsofthedrug.
ThismethodofanalysissuggeststhatphenobarbitalhasinfactaneffectonAMPKactivation,althoughveryvariableanddependentontimeofincubation.
Chapter4:Phenobarbital–newbindersforanolddrug|Results
88
Figure4-6:EffectofphenobarbitalonAMPKphosphorylation(pAMPK)inMH1C1cellstreatedinserum-freemedium.MH1C1cellswereincubatedfor1h,2h,4hor6hinserum-freeDMEMcontainingphenobarbitalat0µM(lanes1–4),1µM(lanes5–8),10µM(lanes9–12),100µM(lanes13–16),250µM(lanes17–20)or500µM(lanes21–24).SampleswereanalyzedbywesternblotforAMPKandpAMPKdetection.Barplotdisplaysintensityofphosphorylated(p)AMPK(top),totalAMPK(middle)andnormalizedpAMPKintensitytototalAMPKintensityforeachsample(bottom).
Figure4-7:ComparativeanalysisoftheeffectofphenobarbitalonAMPKphosphorylationforeachtimeincubationpointinMH1C1cellstreatedinserum-freeDMEM.BarplotsdisplaythenormalizedpAMPKintensitytototalAMPKintensity(aspresentedinFigure4-6,bottompanel)inrelationtotheintensitycalculatedforthemediumcontrol(0µM)ateachtimepoint.
Chapter4:Phenobarbital–newbindersforanolddrug|Results
89
Considering the cyclic nature of AMPK’s activation over the time course of the experiment,
observed in the control samples (Figure 4-6, lanes 1 – 4, on page 88) and the variable results
obtained in thesamplesstimulatedwithphenobarbital,anewexperimentaldesign isnecessary to
eliminate thatvariability. In fact,Chingetal. 152 showed thatAMPKphosphorylation iselevated in
serum-starvedcellsincomparisontocontrolcells,whichcouldbethecausefortheunstablestateof
activation of AMPK observed in the control samples and, possibly, affecting the effect of
phenobarbitalstimulation.
AnewexperimentwasperformedinMH1C1cellsgrownandincubatedinserum-supplemented
medium.Anadditionalserum-freecontrolwasaddedinthisexperiment.
Asobservedpreviously,serum-freesamplespresentphosphorylationofAMPKinnon-treated
cells (Figure4-9, lanes1, 7, 13and19, toppanel,onpage90). Likewise, thenewcontrol samples
treated in complete medium but absent of phenobarbital stimulation also show an active AMPK
(Figure 4-9, lanes 2, 8, 14 and 20, top panel, on page 90). In fact, as shown below in Figure 4-8,
comparing the phosphorylation of AMPK normalized to the total amount of AMPK in the sample
betweencellsgrownincompleteorserum-freemediashownodistinctdifferencesthroughoutthe
incubationtimes(withtheexceptionofareductionin2hincubationincompletemedium).
Figure4-8:EffectofserumsupplementationinAMPKphosphorylation(pAMPK)fromnon-treated(0µMphenobarbital)MH1C1cells.
Bar plot displays the normalized pAMPK intensity to total AMPK intensity (as presented in Figure 4-9, bottom panel) in the different
incubationtime-points(1h,2h,4hor6h)forserum-deprivedorcompleteDMEMgrowthmedia.
TheresultsforthetreatmentwithphenobarbitalareshowninFigure4-9,onpage90.
Treatment with phenobarbital for 1 or 6 h show no marked differences on the relative
phosphorylation of AMPKwith increasing concentrations of the drug (Figure 4-9, lanes 3 – 6 and
lanes21–24,bottompanel,onpage90).IncubatingMH1C1cellsfor4hresultsinthesamerelative
phosphorylationusingthe lowestandthehighestconcentrationofphenobarbital,withareduction
onAMPKactivationobserved in themidconcentrationsutilized (Figure4-9, lanes15–18,bottom
panel, on page 90). The 2 h incubation timewith phenobarbital (Figure 4-9, lanes 9 – 12, bottom
panel,onpage90)presentssomedoseresponsebutlackslinearity;AMPKactivationislowestinthe
control,increasingwith1µMphenobarbitalandhavingamaximumvaluewith10µM,butdropping
tothesamephosphorylationvalueasthecontrolwhenusing100µMphenobarbitalandrisingagain
with500µM.
LookingattherelativeactivationofAMPKincomparisontothecontrolatthesametimepoint
(Figure4-10,onpage90)itispossibletosaythatattwohoursincubationwithphenobarbitalat10
µMproducesthehighestactivationofAMPKwhencomparedtothecontrol,buttheoverallpicture
suggeststhatthereisnodose-responseeffectofphenobarbitalonAMPKphosphorylation.
Chapter4:Phenobarbital–newbindersforanolddrug|Results
90
Figure4-9:EffectofphenobarbitalonAMPactivatedproteinkinase(AMPK)phosphorylation(pAMPK)inarathepatomacelllinetreatedincompletemedium.
MH1C1cellswereincubatedfor1h,2h,4hor6hinFBS-supplementedDMEMcontainingphenobarbitalat0µM(lanes1–4),1µM(lanes5–8),10µM(lanes9–12),100µM(lanes13–16),250µM
(lanes17–20)or500µM(lanes21–24).SampleswereanalyzedbywesternblotforAMPKandpAMPKdetection.Barplotdisplaysintensityofphosphorylated(p)AMPK(top),totalAMPK(middle)and
normalizedpAMPKintensitytototalAMPKintensityforeachsample(bottom).
Figure4-10:ComparativeanalysisoftheeffectofphenobarbitalonAMPKphosphorylationforeachtimeincubationpointinMH1C1cellstreatedinFBS-supplementedDMEM.
BarplotsdisplaythenormalizedpAMPKintensitytototalAMPKintensity(aspresentedinFigure4-9,bottompanel)inrelationtotheintensitycalculatedforthemediumcontrol(0µM)ateachtime
point.
Chapter4:Phenobarbital–newbindersforanolddrug|Results
91
Onceagain,thevariabilityoftheeffectsofphenobarbitalonthephosphorylation/activationof
AMPKanditsendogenouscyclicactivatedstateinMH1C1cellsprovesdifficulttointerprettheresults
and establish this method as a reliable phenotypic assay to measure the effect of phenobarbital
stimulation.
4.2.2.3. InductionofcytochromeP450(CYP)2B1and2B2
The ability of PB to induce expression of P450 genes has been investigated and reviewed
extensively129,153,154.ThemajorP450typesinducedbyPBinratliverareP450b(CYP2B1)andP450e
(CYP2B2)127.
Thesametreatmentofphenobarbitalinarathepatomacellline,asthepreviousexperiments,
was used to determine the effect of the drug in CYP2B1 and CYP2B2 detection by western blot.
MH1C1 cellswere treatedwith 0 – 500µMphenobarbital in completemedium for 1 hor 6 h and
unfractionatedcelllysateswereprobedwithanantibodythatdetectsbothCYP2B1andCYP2B2.The
resultsareshowninFigure4-11.
Figure4-11:EffectofphenobarbitalinCYP2B1/2detectioninarathepatomacellline.MH1C1cellsweretreatedfor1hor6hinFBS-supplementedDMEMcontainingphenobarbitalat0µM,1µM,10µM,100µMor500µM.Expectedmolecularweight:50kDa
CYP2B1 and CYP2B2 were not detected in MH1C1 cells after treatment with phenobarbital
(Figure4-11).Theblotshowstheabsenceofabandattheexpected50kDamolecularweight.Other
bands are detected with this antibody at higher molecular weights but none changes with
phenobarbital,suggestingacaseofunspecificantibodybinding.
ToevaluatethequalityoftheantibodyagainstCYP2B1andCYP2B2,differentfractionsofrat
liver isolated for anotherprojectwere loadedon the gel, and anewexperimentwithMH1C1 cells
treatedwithhigherphenobarbitalconcentrations(1mMand2mM)wasprobedsimultaneously(see
detailsinspecificmethods4.3.4TreatmentofMH1C1cellswithhighconcentrationofphenobarbital,
onpage106).ResultsarepresentedinFigure4-12,onpage92.
Chapter4:Phenobarbital–newbindersforanolddrug|Results
92
Figure4-12: Effectofhigh concentrationphenobarbital on thedetectionofCYP2B1/2 inMH1C1 cells using rat livermitochondriamatrixfractionaspositivecontrolforantibodyreactionquality.
Rat liver mitochondria were probed with an anti-CYP2B antibody and the presence of a
reactionbelowthe55kDamarkerbandconfirmsthepresenceofthisproteinandthecapacityofthe
antibodytodetect it (Figure4-12,arrow, lane4).Thisnewprotocolallowedfor thedetectionofa
faintbandinMH1C1cells(Figure4-12,lanes1–3),althoughstillabitlowerinmolecularweightas
comparedtotheratliversamples.Theresultsalsoshowthattreatmentwithphenobarbitalashigh
as2mMdoesnotincreasethedetectionofCYP2B1orCYP2B2comparedtothenegativecontrol.
Since the new method for treatment of MH1C1 cells with higher concentration of
phenobarbital revealed improvements in the detection of P450 proteins, the same protocol was
appliedforthedetectionofAMPKanditsphosphorylationstate.
Figure 4-13: Effect of high concentration phenobarbital on the detection ofCYP2B1/2andAMPKphosphorylationinarathepatomacellline.MH1C1 cells were treated with 0 mM (NaCl control), 1 mM or 2 mMphenobarbitalandanalyzedbySDS-PAGEfollowedbyWesternBlot.
The new protocol confirmed that even with high
concentrationofphenobarbital(1mMor2mM),MH1C1
cellsdonotshowdifferencesinAMPKphosphorylationin
comparisontothenegativecontrol(Figure4-13)
Chapter4:Phenobarbital–newbindersforanolddrug|Results
93
4.2.2.4. DetectionofD-aminoacidoxidase(DAO)inMH1C1cells
UnpublishedCCMSdatasuggestedD-aminoacidoxidaseasapossibletargetofphenobarbital
inratlivermitochondria.D-AminoacidoxidaseisaFAD-dependentenzymethatcatalyzesoxidative
deaminationofD-aminoacids.In1965,GiudittaandCasola155determinedthatphenobarbitalhadan
inhibitoryconstant(Ki)of4x10-3MagainstD-aminooxidase.
To determine if DAO could be used as a phenotypic assay to monitor response to
phenobarbitaltreatment,MH1C1celllysateinaserialdilutionwasprobedfortheproteinbywestern
blot.ResultsshowthatDAOisundetectableinMH1C1lysate(Figure4-14,lanes6–10),evenatthe
highestinputof50µgtotalprotein.Toconfirmantibodyspecificity,purifiedDAOwasloadedalsoin
serialdilution(Figure4-14,lanes1–5).
Figure4-14:DetectionofD-aminoacidoxidase(DAO)inpurifiedfractionsandMH1C1lysate.
Parallel experiments to determine the inhibitory action of phenobarbital on DAO were
performedatcaprotecbyDr.YanLuo.SheproducedrecombinantDAOandtestedforphenobarbital
inhibitionviaanAmplexRedbasedassay.Resultsconfirmedthatphenobarbitalaffinityislowforrat
DAO(bothrecombinantandfromlivermitochondria)withanIC50>400µMandinactiveagainstDAO
fromahumanhepatocytecellline(HepG2)withanIC50>10mM.
4.2.2.5. Effectofphenobarbitaltreatmentinratprimaryhepatocytes
Despite all efforts to develop a phenotypic assay for the effect of phenobarbital in a rat
hepatoma cell line, the resultswere so far negative or inconclusive. The genetic andphysiological
differences induced during the establishment of the cell line could be responsible for the
ineffectivenessoftheestablishedassays.Forthisreason,anewbiologicalsystem,lessmanipulated,
wasselectedtodevelopfurtherphenotypicassays:ratprimaryhepatocytes.
Theratprimaryhepatocyteswereobtained frozenandusedalways in the firstpassage.The
previously established phenotypic assays to determine AMPK phosphorylation, CAR translocation,
CYP2B1/2andDAOdetectioninresponsetophenobarbitaltreatmentwereperformedinratprimary
hepatocytes.
Theresultsoftheeffectof1mMphenobarbitalonratprimaryhepatocytesarepresentedin
Figure4-15,onpage94.
Chapter4:Phenobarbital–newbindersforanolddrug|Results
94
Figure4-15:Effectofphenobarbital inthedetectionofD-aminoacidoxidase(DAO),CYP2B1/2andAMPKphosphorylation(pAMPK)fromratprimaryhepatocytes.
As seen in Figure 4-15, rat primary hepatocytes treatedwith 1mMphenobarbital show no
differences to control regarding AMPK detection and the drug also does not promote the
phosphorylation of AMPK at its threonine 172. Similarly, CYP2B1 or CYP2B2 detection is also not
affectedbythephenobarbitalstimulus.Finally,DAOisnotdetectedineithercontrolortreatedrat
primaryhepatocytes.
The results here presented are sufficient to question the published data regardingwestern
blot methods as a readout system for the effect of phenobarbital in rat hepatocytes from both
primaryandhepatomaorigins.
Analternativemethodof detection for all thepreviouslydescribedmarkerswasdeveloped,
usingmRNAexpressionasareadoutsystem.
4.2.2.6. EffectofphenobarbitalonCYP2B1andCYP2B2mRNAexpressioninrathepatocytes
Reverse-transcription polymerase chain reaction (RT-PCR) is a well-established method to
evaluate the effect of compounds in living cells. A semi-quantitative RT-PCRmethodwas used to
compare treated samples to controls. Extracted RNA and synthesized cDNA was quantified and
loaded in the sameamountsateach step forall samples, and thehousekeepinggeneG3PDHwasusedasinternalcontroltoconfirmequalloadingbetweensamples.Detailedmethodsaredescribed
in4.3.6RT-PCR,onpage107.
Non-treatedMH1C1cellswereusedinitiallytoestablishtheRNAextractionandcDNAsynthesis
protocols, followedbyspecificamplificationof thehousekeepinggeneG3PDH.ThequalityofeachstepwasanalyzedbyagarosegelelectrophoresisandispresentedinFigure4-16,onpage95.
Chapter4:Phenobarbital–newbindersforanolddrug|Results
95
Figure4-16:AgarosegelelectrophoresisofRNA,cDNAandPCRproducts.QualitycontrolforRNAextractionfromMH1C1cells,cDNAsynthesisandPCRforthehousekeepinggeneG3PDH
AsseenonFigure4-16,lane1,totalRNAwassuccessfullyextractedfromtherathepatomacell
lineand the28Sand18S rRNAbandsareclearlyvisible.The reverse transcriptionofpoly-AmRNA
was also successful although some remaining RNA is still detected after purification of the cDNA
(Figure4-16,lane2),whichcanbespecificallydegradedbyaddingRNaseHafterthecDNAsynthesis
(Figure4-16,lane3).Thepurificationprocessdoneinasamplewithoutreversetranscriptase(Figure
4-16,lane4)showsthesameleftoverRNAbutnoadditionalimpurities.ToconfirmcDNAquality,the
specificamplificationofthehousekeepinggeneG3PDHwasdoneinbothcDNAsamples(Figure4-16,
lanes5and6)containingornotRNaseHandasingleintensebandattheexpected983bpconfirms
thatsamplesshouldbetreatedwiththeRNaseHenzymeaftercDNAsynthesis.Additionalnegative
controls(Figure4-16,lanes7and8)inpurewaterorlackingthereversetranscriptaseenzymewere
amplifiedwiththespecificprimersforG3PDHbutpresentednovisibleband,confirmingthequality
ofthesamplesandthespecificityofthePCR.
Having confirmed each step of the RNA extraction, cDNA synthesis and gene-specific PCR,
MH1C1cellstreatedwitheither1mMor2mMphenobarbitalweresubjectedtothesameprotocol.
Untreated MH1C1 cells and rat liver tissue were used as controls to monitor the expression of
CYP2B1, as was previously tested by western blot but yielded no discernible results (see Results4.2.2.3InductionofcytochromeP450(CYP)2B1and2B2,onpage91).
Figure 4-17, on page 96, clearly shows an increase of CYP2B1 mRNA expression upon
treatmentwithphenobarbital.
Chapter4:Phenobarbital–newbindersforanolddrug|Results
96
Figure4-17:EffectofphenobarbitalonCYP2B1mRNAexpressioninarathepatomacellline.MH1C1 cells were treated with 1 or 2 mMphenobarbital (PB) and one vehicle control(0.9%NaCl)over-nightat37°C(5%CO2).TotalRNAwasextractedfromthecellsandthemRNAwas specifically transcribed into cDNA using areverse transcriptase and oligo (dT)20 primers.RNA extraction and cDNA synthesis was alsodone from a rat liver sample. CYP2B1 wasamplifiedbyPCRusingspecificprimersandtheproduct was run on a 2% agarose gel in TBE.The expected amplification product is 204 bp(toppanel).ThehousekeepinggeneG3PDHwasusedasloadingcontrol(bottompanel).
RatlivertissueandMH1C1cellshavethesamelevelofCYP2B1mRNAexpression(Figure4-17,
lanes1and2,toppanel).Therathepatomacelllinewhentreatedwithphenobarbitalincreasesthe
productionofCYP2B1mRNA(Figure4-17,lanes3and4).Theresultsoftheeffectofphenobarbital
on the expression of the cytochrome P450 are confirmed by the unchanged intensity of the
housekeepinggeneG3PDHusedasloadingcontrol(Figure4-17,lane1–4,bottompanel).
This phenotypic assay was replicated and more genes were probed for changes with
phenobarbital treatment in primary rat hepatocytes. Figure 4-18, shows the effects of 1 mM
phenobarbital treatment on the expression of D-amino acid oxidase (DAO), cytochrome P450 2B1
(CYP2B1) and cytochrome P450 2B2 using two different pairs of primers that amplify different
regions of the gene (CYP2B2 and CYP2B2.1). Again, the housekeeping gene G3PDH was used asloadingcontrol.
Figure4-18:EffectofphenobarbitalonCYP2B1mRNAexpressioninratprimaryhepatocytes.Ratprimaryhepatocytes(1x106cells)wereseededina6-wellplatewithsupplementedWilliam’sEmediumandallowedtoadhereanddifferentiatefor6hours.Thecellswerethentreatedwith1mMphenobarbital(PB)and0.9%NaClvehiclecontrol(Ctrl)over-nightat37°C(5%CO2).TotalRNAwasextractedfromthecellsandthemRNAwasspecificallytranscribedintocDNAusingareversetranscriptaseandoligo(dT)20primers.TargetgeneswereamplifiedbyPCRusingspecificprimersandtheproductsrunona2%agarosegelinTBE.
Chapter4:Phenobarbital–newbindersforanolddrug|Results
97
Ratprimaryhepatocytesweretreatedwith1mMphenobarbitalandmonitoredfortheeffect
ontheexpressionofmRNAfromdifferentgenes.ThesimilarlevelofexpressionofG3PDHincontrol(Ctrl)andtreated(PB)samplesallowstoestablishthatanydifferenceobservedinotheramplification
productsisduetoamountofsample(Figure4-18,lanes1and2).
Previouswestern blot results (Figure 4-15, on page 94) had failed to elucidatewhether rat
primaryhepatocytescontainedDAO.InFigure4-18,itisclearthatthesecellsexpressDAOmRNAbut
itsexpression isnotaffectedby treatmentwithphenobarbital (Figure4-18, lanes3and4).Oneof
theanalyzed cytochromeP450mRNA (CYP2B2) didnot showa visible increase inexpressionafterphenobarbitalstimulusevenwhenanalyzingthegeneexpressionwithtwodifferentsetsofspecific
primers(Figure4-18,lanes7–10).
TheexpressionofCYP2B1mRNAfromratprimaryhepatocytes,however, isclearly increased
afterphenobarbitaltreatment(Figure4-18,lanes5and6),ashighlightedintheredbox,confirming
thesameeffectobservedpreviouslyinMH1C1cells(Figure4-17,onpage96).
These results confirm the establishment of a working phenotypic assay for the effect of
phenobarbitalinratlivercells.Thiseffectcanbemonitoredbysemi-quantitativeRT-PCRofCYP2B1mRNAexpression.ThisalsoconfirmsthatbothratprimaryhepatocytesandMH1C1cellscanbeused
toidentifyspecificbindingpartnersofphenobarbitalinlivingcellsthroughthepreviouslyestablished
CCMSusingclickchemistry.
4.2.3. CapturingphenobarbitaltargetsinMH1C1celllysate
Theinitialexperimentto identifyphenobarbitaltargetswasperformedin lysatefromtherat
hepatomaMH1C1cellline.OnlyphenobarbitalcapturecompoundsA(CC10)andC(CC12)wereused
because, as described previously, compound B (CC 11) did not yield sufficient amounts for a
comprehensivestudy.Also,eventhoughthepermeabilityofcompoundC(CC12)wasdeterminedas
beinglow(see4.2.1Determiningpermeabilityofphenobarbital-alkynecapturecompounds,onpage
84),theinitialapproachwasdesignedinlysate,wherecompoundpermeabilitydoesnotplayarole,
inordertocastawidernetfortheidentificationofputativephenobarbitaltargets.
Three separate experiments were performed (I, II and III) with varying compound and
competitor (CC 15) concentrations, as well as individual and combined data analyses. Detailed
informationontheexperimentprotocolisdescribedintheSpecificMethodssection4.3.7Captureof
phenobarbitaltargetsinMH1C1celllysate,onpage109.
Table4–2on the followingpage shows the summaryof the identificationson the individual
experiments or in combined analysis. More specifically, results were combined for both capture
compounds ineachexperimentor forall experiments toallow foran investigativeanalysison the
reproducibilityoftheCCMSoutcome.
Chapter4:Phenobarbital–newbindersforanolddrug|Results
98
Experiment I II III I+II+III
Compound CC10 CC12CC10+
CC12CC10 CC12
CC10+
CC12
CC10CC10 CC12
1µM 5µM
Totalproteinsidentified 1190 1181 1265 920 958 1013 976 2020 2020 1460
Student’st-test<0.05 123 26 18 25 175 126 55 17 2 2(%oftotal) (10%) (2.2%) (1.4%) (2.7%) (18%) (12%) (5.6%) (0.8%) (0.1%) (0.1%)
Assay/Competitionfold-change>2 139 74 133 48 41 51 33 70 146 98(%oftotal) (12%) (6.3%) (11%) (5.2%) (4.3%) (5.0%) (3.4%) (3.5%) (7.2%) (4.9%)
Competedandsignificantproteins 37 5 2 0 8 2 2 7 2 0(%oftotal) (3.1%) (0.4%) (0.2%) (0%) (0.8%) (0.2%) (0.2%) (0.3%) (0.1%) (0%)
Table4–2:SummaryofidentificationsfromphenobarbitalcaptureexperimentsinMH1C1celllysate.Threedifferentexperiments(I,IIandIII),eachconsistingofmultipletechnicalreplicates,wereanalyzedforputativetargetsoftwodifferentphenobarbitalcapturecompounds(CC10andCC12).TheseresultswereanalyzedwithMaxQuantv.1.3.0.3inanindependentmannerorincombination(symbolizedby+)
Table4–2showsthatthetotalnumberofidentificationsbetweenexperimentsvariedbetween
920 and 2020 but more strikingly that putative targets for phenobarbital, here referred to as
competedandsignificanthits,spansfromnoneto37usingthesamecompound.
An extensive analysis on each individual and combined result could be presented but the
variability of results would confound the analysis. To find the minimal common denominator all
experiments performed with the permeable compound CC10 were analyzed together, treating
samples as technical replicates, and 2 proteins are identified as significantly competed by
phenobarbital(Table4–2,I+II+III,CC10).
The first,Wdr1, also known asWD repeat*-containing protein 1, has a fold-change of 11.8
between assays and competitions with a calculated p-value of 0.024. According to Uniprot, thisprotein induces disassembly of actin filaments and is involved in platelet degranulation. Its
associationtoanycommonlyreportedphenobarbitaleffectisatfirstglancedistant,butithasbeen
publishedthatphenobarbitalincreasestheexpressionofWdr1mRNAinmouseliverphenobarbital-
induced tumorsand isdependentonCARsince the sameeffectwasnotobserved in tumors from
CAR knock-out mice (supplementary tables S3E and S3I in Phillips and Goodman 149). The direct
implicationofthisputativedirectbinderofphenobarbitalshouldbefurtherinvestigatedsinceWD-
repeatcontainingproteinsareknowntobeinvolvedinseveralimportantbiologicalfunctionssuchas
signaltransduction,transcriptionregulationandareassociatedwithdifferenthumandiseases156.It
is interesting to note that RACK1, a protein thought to have a role in phenobarbital effect 142, as
describedintheIntroductionsection,isaWD-repeatprotein.
Ybx1wasalsofoundtobesignificantlycompetedbyphenobarbitalwithafold-changeof2.6
andp-valueof 0.046. Ybx1, knownas nuclease-sensitive element-bindingprotein 1,would aswell
appear not to be related to known phenobarbital effects in a superficial analysis. However, an
alternativenameforthisprotein,CCAAT-bindingtranscriptionfactorIsubunitAsuggestsotherwise.
Asdescribedintheintroduction,CYP2Bgenespossessaphenobarbitalresponsiveenhancermodule
*WDrepeatproteinsaredefinedbythepresenceoffourormorerepeatingunitscontainingaconservedcoreofapproximately40aminoacidsthatusuallyendwithtryptophan-asparticacid(WD)
Chapter4:Phenobarbital–newbindersforanolddrug|Results
99
(PBREM)composedbyabindingsitefornuclearfactor-1(NF-1)flankedbytwoDR-4nuclearreceptor
(NR) binding sites for a heterodimer of constitutive androstane receptor (CAR) and retinoid X
receptor(RXR)132,157.Interestingly,NF-1bindstoaso-calledY-boxorCCAAT-box,muchlikeYbx1158–
160. Among several functions attributed to Ybx1, there is another association of this protein to a
knownphenobarbitaleffect,namelyregardingthesmallmolecule’sinteractionwithEGFR161.Ybx1,
alsoreferredtoasYB-1,wasshowntobindtotheEGFRpromoterandreducingYB-1levelsresulted
inthedown-regulationofEGFR162.ThecaptureofYbx1withthephenobarbitalcapturecompound
suggestsadirectbindingand,given the functionof thisprotein in the regulationof severalgenes,
thiscouldbeacontributingfactortothecellularpleiotropiceffectofphenobarbital.
Inordernottolooseinformationoneachindividualexperiment,acomparisonbetweenthe
different resultsusinga scoringmethod (aspreviouslydone in3.2.4Capturing in livingK562cells,
Table 3–8, on page 72) was used to retrieve the most relevant putative hits for phenobarbital
capture inMH1C1 lysate. Analyzing the datawith values based onMS-positive identifications (LFQ
values inMaxQuantsoftware),a totalof19hitswere identified inmore thanoneexperimentand
the fold-change and corresponding p-values are presented in Table 4–3, on page 100. This listincludes the previously describedWdr1 and Ybx1 but also reveals other interesting hits regarding
previously described associations with phenobarbital interacting mechanisms. Running a STRING
interaction analysis, two clusters of interacting proteins are revealed (Figure 4-19) and a pathway
analysisindicatesthatAMPKpathwayproteinsareenrichedinthislist,namelythedifferentsubunits
ofPP2AphosphatasethathasbeendescribedintheEGFRsignalingpathwayintheintroduction(see
Figure4-2:PhenobarbitalinducesCYP2BexpressionviaEGFRinhibition,onpage83).
Figure4-19:STRINGproteininteractionanalysisofsignificantlycompeted proteins in phenobarbital capture experiments inMH1C1lysateBlue: binding; black: reaction; pink: post-translationalmodification; purple: catalysis; green: association in curateddatabases.
Three proteins were identified in 3
different experiments as phenobarbital
binder candidates: Psmb1, Aqp1 and
Gsta3/Gsta5.Psmb1isaproteasomesubunit
that has been mentioned to increase its
mRNA expression upon phenobarbital
stimulus,butthiseffectwasonlyobservedin
asubsetoftheexperimentanditsrelevance
remains dubious 163. A surprising hit was
Aqp1, aquaporin-1, which regulates the influx of water in cells. It has been described that
phenobarbital increases and decreases Aqp1 mRNA expression and that CAR is involved in the
process164,165.Interestingly,Aqp1andthelargeraquaporinfamilyofproteinsplayarelevantrolein
epilepsy 166–168, the initial therapeutic use of phenobarbital. Lastly, glutathione S-transferase alpha
subunits (Gsta3/Gsta5)were targeted by phenobarbital in the capture experiments. This family of
proteins is involved in the protection against stress sensors in the cellular environment and its
stimulationbyphenobarbitaliswellestablishedintheliterature169.
Chapter4:Phenobarbital–newbindersforanolddrug|Results
100
Specificity(Assay/Competition) Intensityfold-change(log2) Student’st-test(pvalue)
Experiment I II* III I+II+III I II* III I+II+III compound CC10 CC12 CC10 CC12 CC10
CC10 CC12
CC10
+
CC12
CC10 CC12 CC10 CC12 CC10CC10 CC12
CC10
+
CC12
Genename 10µM 10µM 1µM 5µM 10µM 10µM 1µM 5µM Score
Aqp1 0.3 12.6 1.0 1.1 0.1 0.3 0.3 3.4 0.6 0.233 0.028 0.189 0.028 0.880 0.669 0.480 0.085 0.142 3Gsta3;Gsta5 2.8 0.1 #N/I #N/I 0.7 1.7 2.3 0.1 1.8 0.033 0.552 #N/I #N/I 0.219 0.030 0.175 0.917 0.170 3
Psmb1 12.9 0.2 #N/I -7.2 1.2 9.5 5.7 0.1 1.6 0.014 0.900 #N/I 0.423 0.627 0.036 0.100 0.933 0.189 3Aldoa 2.0 -0.8 -1.0 -0.9 0.7 1.6 1.6 -0.8 0.8 0.001 0.489 0.238 0.022 0.400 0.061 0.225 0.533 0.398 2Arcn1 0.2 -0.1 -0.3 -0.6 0.4 8.7 0.2 -0.1 0.1 0.312 0.772 0.088 0.029 0.016 0.005 0.858 0.922 0.936 2
Etfb 11.8 #N/I #N/I #N/I 9.8 9.6 10.5 #N/I 10.2 0.500 #N/I #N/I #N/I 0.003 0.188 0.082 #N/I 0.082 2Got2 3.0 0.1 -0.5 -0.6 0.9 1.3 1.9 -0.2 1.5 0.036 0.879 0.012 0.047 0.142 0.244 0.118 0.763 0.125 2
Mdh1 1.0 0.1 -0.5 -0.6 1.2 1.5 1.0 #N/I 0.6 0.247 0.961 0.041 0.035 0.048 0.092 0.415 0.997 0.464 2Me1 3.5 -1.0 -0.6 -0.9 1.4 2.5 2.4 -1.0 1.5 0.063 0.716 0.199 0.032 0.091 0.010 0.117 0.575 0.175 2Pklr #N/I #N/I -1.0 -0.7 0.8 1.3 0.6 -0.7 0.4 #N/I #N/I 0.272 0.165 0.152 0.016 0.211 0.473 0.398 2Ppp2cb;Ppp2ca 1.7 12.4 #N/I #N/I -5.6 6.5 1.6 11.0 2.5 0.255 0.046 #N/I #N/I 0.391 0.389 0.389 0.179 0.107 2Ppp2r1b -11.5 -11.1 0.6 8.7 -4.5 0.4 -2.9 -1.9 -2.4 0.500 0.500 0.702 0.001 0.391 0.391 0.340 0.509 0.224 2Rab8a;Rab8b 11.2 1.0 #N/I #N/I #N/I -12.0 11.2 1.0 2.1 0.024 0.482 #N/I #N/I #N/I 0.391 0.164 0.532 0.145 2
Rtn3 #N/I #N/I 0.6 2.4 0.8 10.2 1.0 2.4 1.2 #N/I #N/I 0.162 0.035 0.437 0.391 0.258 0.146 0.139 2Sdad1 -0.3 -0.1 1.2 7.7 #N/I 0.9 -0.3 -0.1 -0.2 0.071 0.891 0.132 0.013 0.912 0.211 0.839 0.960 0.854 2
Syngr1 -0.3 12.3 1.6 2.7 -0.9 0.4 -0.2 6.0 0.6 0.395 0.109 0.162 0.013 0.434 0.391 0.854 0.133 0.490 2Tpp2 1.5 1.4 -0.3 -0.2 1.3 1.0 1.2 1.0 1.1 0.317 0.342 0.050 0.353 0.015 0.179 0.268 0.375 0.150 2Wdr1 #N/I #N/I #N/I #N/I 12.1 11.2 11.8 #N/I 11.8 #N/I #N/I #N/I #N/I 0.001 0.391 0.024 #N/I 0.026 2
Ybx1 #N/I #N/I -6.7 5.3 7.9 3.3 2.6 4.6 2.7 #N/I #N/I 0.423 0.423 0.067 0.023 0.041 0.374 0.038 2Table4–3:Potentialhitsforphenobarbitalinarathepatomacellline(MH1C1)from3individualexperimentsusingaMS-basedanalysis
(I-III)eachconsistingof2–4replicatesforassaysandcompetitionsusingtwopre-clickedphenobarbitalcapturecompounds(CC10andCC12).Valuesinboldmarkspecific(log2fold-change>1)andsignificant(p
value<0.05)resultsforagivenexperiment.Scoreiscalculatedbythesumofnumberofresultsthatfulfilledboththecriteriaofspecificityandsignificance.#N/I:notidentified.*CC10subsetofexperimentIIwas
analyzedindividuallywithMaxQuantsoftwareandwasaddedtothescorecalculations.
Chapter4:Phenobarbital–newbindersforanolddrug|Results
101
4.2.4. CaptureofphenobarbitaltargetsinlivingMH1C1cells
The capture of phenobarbital targets in living cells was done to improve upon the alreadysuccessful results obtained in the capture experiments from lysate samples and establish a newmethodfortheidentificationofsmall-moleculebindingpartnersthatbetterresemblesthedeliverysystemofdrugsinlivingorganisms.
TheexperimentwasdonewithlivingMH1C1cellsusingtriplicatesofassaysandcompetitionsandthedetailedprotocolisdescribedintheSpecificMethodssection4.3.8CaptureofphenobarbitaltargetsinlivingMH1C1cells,onpage110.
Table4–4gathersthesummaryoftheidentificationsineachsubsetoftheexperiment.
Pre-clicked Click
Alignment-based
MS-based Alignment-based
MS-based
Totalproteinsidentified 740 735 1363 1362
Student’st-test<0.05(%oftotal)
41(5.5%)
43(5.9%)
52(3.8%)
44(3.2%)
Assay/Competitionfold-change>2(%oftotal)
16(2.2%)
33(4.5%)
21(1.5%)
49(3.6%)
Competedandsignificantproteins(%oftotal)
2(0.3%)
1(0.1%)
1(0.1%)
1(0.1%)
Table4–4:SummaryofidentificationsfromphenobarbitalcaptureexperimentsinMH1C1livingcells.
ListofproteinswasobtainedusingMaxQuantv1.3.0.3.Numbers for identificationsbasedonspectralcounts (MS-based)andalignment-
basedaregiven.
One immediate observation of Table 4–4 is that a larger number of proteins is identified in
samples where the click reaction was done after the photo-crosslink to the target, 740 vs. 1363where the capture compoundwaspre-clicked to thebiotin anchorduring the cell incubation. Thedifference innumberscouldbeexplainedby theprobable impermeabilityof the largerpre-clickedcompound,preventing it todiffuse into thecelland thereforemainly targetingsurfaceproteinsoronlyafterendocytosis.Adifferentperspectivesuggeststhatthe largernumberof identifications inclick samples could be due to unspecific capture inherent to the CuAAC reaction, although initialexperiments(describedinChapter2.Labelingaproteinusingclickchemistry)showthattheCuAACisrathercleanbutwithlowyields.
Even thougha largenumberofproteinswere identified,only fourpass thecompetitionandreproducibilitycriteria.ThevaluesfortheseputativephenobarbitalbindersarepresentedinTable4–5,onthefollowingpage.
Chapter4:Phenobarbital–newbindersforanolddrug|Results
102
Specificity(Assay/Competition)
Protein TotalMScounts Alignment-based MS-based
Genename
Assay Comp. Intensityfold-change(log2)
Student’st-test(pvalue)
Intensityfold-change(log2)
Student’st-test(pvalue)
Pre-click
Them6 11 5 1.6 0.002 1.4 0.014
Anxa4 1 1 1.2 0.022 9.6 0.423
Click
Arf5 3 0 4.5 0.032 -7.9 0.423
Dnm2 3 0 0.1 0.761 12.7 0.008Table4–5:ListofcompetedandsignificanthitsinphenobarbitalcaptureexperimentsinMH1C1livingcells.
TotalMScountsfromtechnicalreplicates(n=3)ofassayandcompetition(comp.)samples.Intensityfold-changeandsignificancelevelsare
givenforspecificity(assayversuscompetition)ofhits.ValuesinboldindicateeitherFC>2orpval<0.05.
The list ofproteins identified in the captureexperimentusingphenobarbital in livingMH1C1
cells, presented in Table 4–5, do not coincide with the previously described putative bindersobtainedwiththecaptureinlysateofthesamecellline.Thereis,however,acommondenominatortoalltheseidentifiedproteins:EGFR.
Them6isaratheruncharacterizedthioesterasewithonlysevenreferencesonPubmedand9known protein interactions. The thioesterases, or thioester hydrolases, comprise a large enzymegroupwhosemembershydrolyzethethioesterbondbetweenacarbonylgroupandasulfuratom.Interestingly, Them6 has been co-immunoprecipitated with EGFR 170 and was also previouslyidentified as a potential phenobarbital target in Experiment B of the capture in MH1C1 lysatedescribed in the previous paragraph (the alignment-based values for Them6 revealed a log2 fold-changebetweenassaysandcompetitionsof1.3withanassociatedp-valueof0.002).
AnnexinA4,Anxa4,isacalcium/phospholipid-bindingproteinthatpromotesmembranefusionand is involved in exocytosis. Anxa4 has clinical significance in cancer 171, and even if this specificmemberof theannexin familyhasnotbeendescribed inEGFRtransductioncontext, severalotherannexinshavebeenassociatedwithendocytosisandEGFRsignaling172,173.
Arf5,alsoknowasADP-ribosylationfactor5isaGTP-bindingproteinthatisInvolvedinproteintrafficking174andrecentlyfoundtoco-localizewithEGFRinA431cells175.
Dnm2,Dynamin-2, isamicrotubule-associatedforce-producingprotein involved inproducingmicrotubulebundlesandabletobindandhydrolyzeGTP.Dnm2playsanimportantroleinvesiculartrafficking processes, in particular endocytosis 176. Association of Dnm2 to EGFR has also beendescribedmultiplepublicationsasanEGF-inducedEGFR-associatedprotein177,178.
Onceagain,resultsseemtopointtowardsaconnectionwithEGFR,inlinewithMutho’srecentfindings142,butthequestionthatremainsiswhyisn’tEGFRcapturedintheseexperiments.Arethesethe true phenobarbital binders that establish the connection between EGFR signaling andphenobarbitalphenotypiceffectorarethese interactingproteinscapturedduetounspecificcross-linking to neighboring proteins inherent to the design of capture compounds? It can also bepostulated that EGFR peptides are not identified in the CCMS experiments because the attachedcapturecompoundaltersthemassofthepeptidesandaffectstheanalysisinthedatabasesearch.
Chapter4:Phenobarbital–newbindersforanolddrug|Results
103
4.2.5. CapturingphenobarbitaltargetsintissuewithhighEGFRexpressionWithpilingevidencethatphenobarbitaltargetsproteinsrelatedtoEGFR,andgiventherecent
publicationofMutohandco-workers142thatconfirmsEGFRasatargetforphenobarbital, itwouldbeinterestingtoverifyifcaptureexperimentsdeterminethereceptorasaprimarybinder,oriftheassociatedproteinsidentifiedinthepreviousparagraphsarestillpreferentialtothedruginatissueexpressinghighlevelsofEGFR.
According to the Human Protein Atlas, a tissue-based map of the human proteome 179,trophoblasticcellsfromplacentapresentahighexpressionofEGFRprotein.
Humanplacentalysatewasusedassampletoidentifyphenobarbitaltargetsusingacompletepre-clicked phenobarbital capture compound. Detailed protocol is described in Specific Methodssection4.3.9Captureofphenobarbitaltargetsinhumanplacentalysate,onpage111.
Table 4–6: Summary of identifications from
phenobarbital capture experiment in human
placentalysate.
List of proteins was obtained using MaxQuant
v1.3.0.3. Numbers for identifications based on
spectralcounts(MS-based)andalignment-based
aregiven.
From over 2000 protein identifications, only 6 were competed by phenobarbital in areproduciblemanner. These identifications, summarized in Table 4–6, yield different resultswhenanalyzing the values basedon alignment features or exclusively on sequencedpeptides,with onlyonesignificantlycompetedtargetidentifiedinbothmethods.
The list of significantly competed proteins with phenobarbital in human placenta lysate ispresentedinTable4–7.
Specificity(Assay/Competition)
Protein TotalMScounts Alignment-based MS-based
Genename
Assay Comp. Intensityfold-change(log2)
Student’st-test(pvalue)
Intensityfold-change(log2)
Student’st-test(pvalue)
SRP72 52 37 2.4 <0.001 0.2 0.087
YBX1/3/5 4 3 1.5 0.001 13.3 <0.001
MICALL1 4 1 1.7 0.011 1.4 0.378
TMEM97 3 4 1.1 0.045 1.3 0.073
MESDC2 3 1 0.1 0.673 12.4 <0.001
GAK 3 2 0.8 0.047 12.2 <0.001
EGFR 33 30 0.1 0.761 0.14 0.129Table4–7:Listofcompetedandsignificanthitsinphenobarbitalcaptureexperimentsinhumanplacentalysate.
TotalMScountsfromtechnicalreplicates(n=3)ofassayandcompetition(comp.)samples.Intensityfold-changeandsignificancelevelsare
givenforspecificity(assayversuscompetition)ofhits.EGFRisaddedseparatelyforfailingcriteria.Values inbold indicateeitherFC>2or
pval<0.05.
Pre-click
Alignment-based
MS-based
Totalproteinsidentified 2054 2047
Student’st-test<0.05(%oftotal)
80(3.9%)
64(3.1%)
Assay/Competitionfold-change>2(%oftotal)
51(2.5%)
78(3.8%)
Competedandsignificantproteins(%oftotal)
4(0.2%)
3(0.1%)
Chapter4:Phenobarbital–newbindersforanolddrug|Results
104
The first observation of the phenobarbital capture experiment in a tissue expressing highlevelsofEGFR is that thereceptor is indeedcapturedbut thecompetitionwith freephenobarbitalcompetitor does not exceed the two-fold ratio between assays and competitions, nor is theconfidenceleveloftheidentificationsacceptable.
Interestingly,theonlypositiveidentificationusingeitheranalignmentorMS-basedanalysisisYBX,thepreviouslydescribedY-box-bindingproteinidentifiedasatargetinrathepatomacelllysate.The results in placenta do not allow for a definite identification of the YBX isoform since thequantifiedpeptide is common toYBX1,YBX3andYBX5.Data suggests it tobeYBX1, justas in theMH1C1celllysateexperiment,becauseanadditionalYBX1peptidewasidentifiedalthoughwithoutasignificantcompetitioneffect.
The other identified putative targets for phenobarbital are also described to have anassociation toEGFR,with theexceptionof the signal recognitionparticle subunit SRP72.MICALL1,MoleculeInteractingwithCasL(MICAL)-like1protein,hasbeendescribedtobeinvolvedinreceptor-mediated endocytosis and, in complex with Rab protein, regulates EGFR trafficking 180. Cyclin G-associatedkinase(GAK)isaserine/threoninekinasewitharelevantroleincellulartrafficofclathrin-coated vesicles and it was shown that its down-regulation resulted in an enhancement of EGFRsignaling181.AlthoughnotdirectlyassociatedwithEGFR,MESDC2isachaperoneproteinspecificallyassisting thebeta-propeller/EGFmoduleswithin the familyof low-density lipoprotein receptors 182andtransmembraneproteinTMEM97wascorrelatedwithEGFRinastudyfromTongetal.178.
Toovercomethe incongruousresultsobtainedwhenusingadifferentvalueoutputfromtheMaxQuant analysis software and the way each contributing peptide generates the intensity for awholeprotein, theexperimentaldataobtained fromthephenobarbital capture inhumanplacentalysate was analyzed at the peptide level. Out of 12188 identified peptides only 710 (6%) werecompetedandoftheseonly26(0.2%)wereidentifiedwithapvalueoflessthan0.05.
These 26 unique peptides,peptides corresponding exclusivelyto a specific protein, weresubmitted to the proteininteraction database STRING andrevealed an intricate network thatshows EGFR and YBX1 as centralnodes(Figure4-20).
Gathered evidence suggeststhat phenobarbital binds primarilytoYBX1and there is ahighaffinitytowards proteins associated toEGFRsignaling.
Figure 4-20: STRING protein interaction analysis
of significantly competed proteins in
phenobarbital capture experiments in human
placentalysate.
Blue: binding; pink: post-translational
modification;purple:catalysis;green:association
incurateddatabases.
Chapter4:Phenobarbital–newbindersforanolddrug|SpecificMethods
105
4.3. SPECIFICMETHODSFORTHEPHENOBARBITALSYSTEM4.3.1. Structuresofusedcompounds
CC10:ClickablephenobarbitalalkynecapturecompoundA(containsselectivityandatrifluoromethyl-phenyl-diazirinephotoreactivefunctions)
CC11:ClickablephenobarbitalalkynecapturecompoundB(containsselectivityandamethyl-diazirinephotoreactivefunctions)
CC12:ClickablephenobarbitalalkynecapturecompoundC(containsselectivityandaazido-tetrafluorobenzenephotoreactivefunctions)
CC15:Phenobarbitalcompetitor
CC7:Clickablebiotin-PEGcopper-chelatingazide(containslabelingandcopperligandfunctions)
Chapter4:Phenobarbital–newbindersforanolddrug|SpecificMethods
106
4.3.2. TreatmentofMH1C1cellswithphenobarbital
TherathepatomacelllineMH1C1wastreatedwithincreasingconcentrationofphenobarbitalfor theanalysisofCAR translocation,AMPKphosphorylationandCYP2B1/2B2analysisbyWesternBlotting.
MH1C1 cells in DMEM (10% FBS, 1% pen/strep) were seeded in 6-well plates at 2 x 106cells/well and incubated over-night at 37 °C (5% CO2). Confluent cells were starved over-night byremovingthemediumandaddingserum-freeDMEM.
Phenobarbitaldiluted inDMEMwasaddedtothewellsatdifferentconcentrations:0,1µM,10µM,100µM,250µMand500µM.Plateswereincubatedat37°C(5%CO2)for1h,2h,4hor6h.
Aftertreatment,themediumwasremovedandthecellswerewashedwith1mLD-PBS.Cellswerescrapedin1mLoffreshD-PBSandtransferredto1.5mLmicrotubesforcellularfractionation(asdescribedin4.3.3SubcellularfractionationofMH1C1cells,below).
The cellular fractions and the unfractionated samples were resolved on a SDS-PAGE andtransferred toanitrocellulosemembrane for specificantibodyprobingasdescribed in thegeneralmethodssection(6.2BiochemistryGeneralMethods,onpage118).
4.3.3. SubcellularfractionationofMH1C1cells
MH1C1cellpelletsfromculturein6-wellplatesforphenobarbitaltreatmentwerefractionatedusingtheNE-PER®NuclearandCytoplasmicExtractionkit(PierceBiotechnology™).
Cells were scraped in 1 mL D-PBS and transferred to 1.5 mL microtubes to pellet bycentrifugationat500xgfor3min.Thesupernatantwascarefullyremovedwithamicropipetteanddiscarded,leavingadrypelletwith≅15µLofpackedvolume.150µLofice-coldCERIreagentwasadded to thecellpelletand the tubewasstirred inavortexatmaximumspeed for15seconds toresuspendthecells.Thetubewasincubatedonicefor10min.8.25µLofice-coldCERIIreagentwasadded to the tube, following 5 seconds vortex and 1 min incubation on ice. The tube was againvortexed for 5 seconds and centrifuged for 5 min at 17.000 xg. The supernatant containing thecytoplasmicfractionwastransferredtoacleanpre-chilledtubeandstoredat-80°C.Theremainingpellet, containing thenuclei,was resuspended in75µLofNER reagent. The tubewas vortexedatmaximumspeedfor15secondsevery10minforatotalof40min,keepingitoniceinbetween.Thetubewascentrifugedfor10minat17.000xgandthesupernatantcontainingthenuclearextractwastransferredtoapre-chilledcleantubeandstoredat-80°C.
4.3.4. TreatmentofMH1C1cellswithhighconcentrationofphenobarbital
MH1C1 cells in DMEM (10% FBS, 1% pen/strep) were seeded in 6-well plates at 2 x 106cells/well and incubated over-night at 37 °C (5% CO2). Confluent cells were starved for 2.5 h byremovingthemediumandaddingserum-freeDMEM.Phenobarbitaldiluted in0.9%NaClata finalconcentrationof0,1mMor2mMwasadded to thewellsand incubatedover-nightat37 °C (5%CO2).Themediumwasremovedandthecellswerewashedwith1mLD-PBS.
To the wells for protein analysis, 100 µL of 2X Laemmli Buffer (containing benzonase) wasaddedtodisrupt thecells.Thesampleswere transferred tomicrotubes, sonicatedandboiled,andstored at -20 °C until analysis by SDS-PAGE and Western Blotting (as described in the generalmethodssection(6.2BiochemistryGeneralMethods,onpage118).
ForRT-PCRanalysis,thecellswereprocessedasdescribedin4.3.6RT-PCR,onpage107.
Chapter4:Phenobarbital–newbindersforanolddrug|SpecificMethods
107
4.3.5. TreatmentofratprimaryhepatocyteswithphenobarbitalRatprimaryhepatocyteswereplatedasdescribed intheGeneralMethodssection(6.1.4Rat
primaryhepatocytesculture,onpage117).The cells were incubated over-night with 1 mM phenobarbital diluted in 0.9% NaCl. The
mediumwasremovedandthehepatocyteswerewashedwith1mLD-PBS.To the wells for protein analysis, 100 µL of 2X Laemmli Buffer (containing benzonase) was
addedtodisrupt thecells.Thesampleswere transferred tomicrotubes, sonicatedandboiled,andstored at -20 °C until analysis by SDS-PAGE and Western Blotting (as described in the generalmethodssection(6.2BiochemistryGeneralMethods,onpage118).
ForRT-PCRanalysis,thecellswereprocessedasdescribedbelow.
4.3.6. RT-PCRReverse-transcription polymerase chain reaction (RT-PCR) is a commonly used technique to
semi-quantitatively measure the expression of specific genes. RNA is converted into acomplementary DNA (cDNA) library that can then be probed for target genes using appropriateprimers.Oneoftheadvantagesofthistechniqueistheneedforverylittlesamplequantitytodetectthetargetsofinterest.
4.3.6.1. RNAextractionusingRNeasyProtectMiniKit
RNAwasextracted fromeitherMH1C1cellsor ratprimaryhepatocytes inculture, treatedornotwithphenobarbital,usingacommerciallyavailablekitfromQiagen™(RNeasyProtectMiniKit).
1 x 107 cells were collected to a 15 mL Falcon tube and lysed in 600 µL RLT buffer(supplementedwithDTT).Thetubewasvortexedandthelysatewashomogenizedbypassingitfivetimesthrougha20-gaugeneedleattachedtoaRNA-freesterilesyringe.600µLof70%ethanolwasaddedtothetubeandmixedbypipetting.ThesolutionwasdividedandtransferredtotwoRNeasyspin columnsplaced in2mLmicrotubes. The sampleswere spundown for30 s at10.000xg. Theflow-throughwasdiscardedand700µLofbufferRW1wasaddedtothespincolumn.Thetubeswerecentrifuged 30 s at 10.000 xg. 500 µL of RPE buffer was added to the spin column and againcentrifuged30sat10.000xg.Fresh500µLofRPEbufferwasaddedto thecolumnandthe tubesspundownfor2minat10.000xg.Thespincolumnwastransferredtoaclean2mLmicrotubeandspundownfor1minatmaximumspeed(16.000xg).Thespincolumnwasplacedinaclean1.5mLmicrotubeandRNAwaselutedwith40µLofRNAse-freewater.Thecolumnwascentrifuged1minat10.000xg toelutetheRNA.Samplesweresnapfrozen in liquidnitrogenandstoredat -80°Cuntilfurtheruse.
Chapter4:Phenobarbital–newbindersforanolddrug|SpecificMethods
108
4.3.6.2. RNAquantificationandpuritycalculation
Nucleic acids have absorbancemaxima at 260 nm. Historically, the ratio of this absorbancemaximumtotheabsorbanceat280nmhasbeenusedasameasureofpurityinbothDNAandRNAextractions.A260/280ratioofapproximately1.8isgenerallyacceptedas“pure”forDNA;aratioofapproximately2.0isgenerallyacceptedas“pure”forRNA.
ExtractedRNAfromMH1C1cellsandratprimaryhepatocyteswasdilutedinPBS(pH7.4)andthe absorption at 260 nm and 280 nm was measured in a Varian Cary 50 Bio UV-VisSpectrophotometer.
ThepurityofRNAwascalculatedaccordingtotheformula:"#$%('()*+,)."#$%(/+(01)"#2%('()*+,).
TheconcentrationofRNAwasdeterminedby:
3 = "#$%('()*+,)."#$%(/+(01)56 ×89:;<9=>?@A<=B.
TheobtainedRNAconcentrationsfromdifferentextractionsvariedbetween0.5µg/µLand1µg/µLwithapurityalwayshigherthan2.
4.3.6.3. cDNAsynthesisfromextractedmRNA
1µgoftotalRNAextractedfromMH1C1cells,ratprimaryhepatocytesorrat livertissuewasreverse-transcribedintocDNAusingtheSuperscriptIIIreversetranscriptasekitfromInvitrogen™.
1µLofoligo(dT)20primersat50µMwasmixedwith1µgoftotalRNA,1µLofdNTPmixat10mM and water to normalize volume to 13 µL. The mixture was heated for 5 min at 65 °C in athermomixerandplacedoniceforover1min.Thetubeswerespundownandadded4µLof5xFirstStrandBuffer,1µLofDTTat0.1Mand1µLSuperScriptIIIreverse-transcriptase(at200U/µL).Thesamplesweremixedby gentlypipettingand incubated for60minat 50 °C in a thermomixer. Thereactionwasinactivatedbyheating15minat70°C.1µLofRNAseHwasaddedfor20minat37°CinsomecasestoremoveRNAcomplementarytothefirststrandcDNA.
4.3.6.4. PolymeraseChainReaction(PCR)ofspecificgenes
ThecDNAlibraryfromMH1C1cells,ratprimaryhepatocytesorrat livertissue,treatedornotwithphenobarbital,wasusedastemplatetosearchforspecificgenes:CYP2B1 (usingtwodifferentpairsofprimers),CYP2B2,DAOandG3PDHasahousekeepinggenecontrol.Thesequencesfortheprimersaredescribedbelow:
Forwardprimer Reverseprimer Expectedproduct
G3PDH 5’-TGAAGGTCGGTGTCAACGGATTTGGC-3’ 5’-CATGTAGGCCATGAGGTCCACCAC-3’ 983bp
CYP2B1 5’-CTGTGGGTCATGGAGAGCTG-3’ 5’-TCACACCGGCTACCAACCCT-3’ 201bp
CYP2B2 5’-CTGTGGGTCATGGAGAGCTG-3’ 5’-TCTCACAGGCACCATCCCT-3’ 201bp
CYP2B2.1 5’-GCTCTCCTTGTGGGCTTCTT-3’ 5’-AGGACTCACTTTCTCCATGCG-3’ 804bp
DAO 5’-ATCCTGCTGGAGACAGGTTC-3’ 5’-TCTCCTCCCATCACTCCATC-3’ 1510bp
Chapter4:Phenobarbital–newbindersforanolddrug|SpecificMethods
109
The optimization of the different parameters of PCR, such as cDNA input amount, MgCl2concentration, primer concentration, and number of cycles, was done according to the Taguchimethod183.
Thefinalparametersusedfortheexperimentspresentedintheresultssection(4.2.2.6EffectofphenobarbitalonCYP2B1andCYP2B2mRNAexpressioninrathepatocytes,onpage94)were:
cDNA F.primer R.primer
10xPCRbuffer
dNTPmix
T.aqpolymerase MgCl2 H2O
Volume(µL) 1.5 2 2 5 1 0.2 1.5 38.3
4.3.6.5. Agarosegelelectrophoresis
RNAqualityandPCRproductswereanalyzedbyhorizontalagarosegelelectrophoresis.0.8%–1.2%(w/v)agarosegelswerepreparedbymixingtheweighedamountofagarosein50
mL of Tris/Borate/EDTA (TBE) buffer. The solution was boiled in a microwave and cooled downbeforeadding3µLofSYBR®Safedyeandcastingontotheelectrophoresistray.
ThetraycontainingthegelwasimmersedintheelectrophoresisunitfilledwithTBEbufferandsamplesmixedwith6xloadingdyewereloadedinthewells.AppropriateDNAladderswereaddedasreference.Thegelelectrophoresiswascarriedoutfor45minat100VandbandsvisualizedusingaUVtrans-illuminatorinanImagingSystem(SyngeneG:BOXXT4).
4.3.7. CaptureofphenobarbitaltargetsinMH1C1celllysate
MH1C1 cell lysate was prepared as described in the GeneralMethods (6.1.3.1MH1C1 lysatepreparation,onpage117).
Fullpre-clickedcapturecompoundswerepreparedapriori:500µMofbiotincopper-chelating-azide(CC7)wasequilibratedwith5mMofCuSO4for10minatR.T.andthenreactedwith500µMofphenobarbital-alkyne capture compound (CC 10 or CC 12) in PBS at pH 8.5 by adding 2.5mM ofsodiumascorbate.Thereactionwasshakenvigorouslyfor10minatR.T.
400µgofMH1C1celllysatewasdistributedin200µLPCR-tubestripsandeither330µMor1mM of phenobarbital competitor (CC 15) was added to the competition samples or thecorresponding volumeofDMSOwas added to the assay samples. The vialswere incubated for 30minoniceandthenadded5µMor10µMofpre-clickedphenobarbital-biotincapturecompound.Allsampleswere incubated for60min rotatingat4 °Cand then irradiated for15minat310nm inacaproBox™ to promote photo-crosslinking of the compounds to the binding targets. Streptavidin-coatedmagneticbeadswerewashedwithPBSandincubatedwiththesamplesfor1hrotatingat4°C. The beads were recovered using a caproMag™ and then washed six times with wash bufferfollowedbythreetimeswashwith80%acetonitrileandafinalwashwithMS-gradewater.
0.5µgoftrypsininammoniumbicarbonatewasaddedtothebeadsandincubatedover-nightshakingat37°C.Thesupernatantcontainingthetrypticdigestwasrecoveredtocleantubes;next,the beads were washed with MS-grade water and the wash solution pooled with the originalsupernatant. All samples were lyophilized and analyzed by LC-MS/MS following the protocoldescribedintheGeneralMethodssection(6.3MassSpectrometry,onpage121).
Chapter4:Phenobarbital–newbindersforanolddrug|SpecificMethods
110
4.3.8. CaptureofphenobarbitaltargetsinlivingMH1C1cells
MH1C1 cells were seeded on 6-well plates in DMEM, supplemented with 10% FBS and 1%penicillin/streptomycin, and grown to 85% confluence in a cell incubator at 37 °C (5% CO2). ThemediumwasremovedandthecellswerewashedwithD-PBStostarvewithserum-freeDMEMfor2hat37°C(5%CO2).
A pre-clicked phenobarbital capture compound (CC 10)with a biotin-copper chelating azide(CC7)wasproducedapriori:500µMofbiotincopper-chelating-azide(CC7)wasequilibratedwith5mM of CuSO4 for 10min at R.T. and then reacted with 500 µM of phenobarbital-alkyne capturecompound (CC 10) in PBS at pH 8.5 by adding 2.5mMof sodium ascorbate. The vialwas shakenvigorouslyfor10minatR.T.
2mM of phenobarbital competitor (CC 15) or equivalent volume of DMSO diluted in freshmediumwasaddedtothecompetitionwellsor totheassay, respectively.20µMofphenobarbitalcapturecompound(CC10)asisorpre-clickedtobiotin(CC10-CC7)wasthenaddedtothewells.Theplates were returned for 35 min to the cell incubator. The medium was removed and the cellswashedwithfreshD-PBS.TheplateswereplacedinaCaproBox™andirradiatedfor4minat310nm.Thecellswererecoveredtotubeswithscrapingandpelletedfor10minat400xg.
Thecellstreatedwiththepre-clickedcompoundwerelysedfor30minshakingvigorouslywith1 mL of RIPA buffer (50 mM Tris-HCl pH 8.0, 150 mM NaCl, 1% IGEPAL CA-630, 0.5% Sodiumdeoxycholate, 0.1% SDS) supplemented with 0.2 mM DTT, 0.5 µL/mL benzonase and proteinaseinhibitors.Thesamplesincubatedwiththeunclickedphenobarbitalcapturecompoundwerelysedin500µLof Lysis Buffer E as described in 6.1.3.1MH1C1 lysatepreparation, onpage117. The crudelysates from both preparations were transferred to 500 µL eppies and centrifuged for 30 min at17.000r.p.m.toremovecelldebris.
The CuAAC reaction was performed in the samples containing the unclicked capturecompoundbyadding5µMofbiotincopper-chelating-azide(CC7)(pre-equilibratedwithtwo-timesexcessCuSO4), 10µMBTTE (CC18) pre-equilibratedwith two-timesexcessCuSO4, and2.5mMofsodiumascorbate.Theclickreactionranfor30minshakingatroomtemperature.
Streptavidin-coatedmagnetic beadswerewashedwith PBS and incubatedwith the samplesfor30minrotatingat4°C.ThebeadswererecoveredusingacaproMag™andthenwashedsixtimeswithwashbufferfollowedbythreetimeswashwith80%acetonitrileandafinalwashwithMS-gradewater.
0.5µgoftrypsininammoniumbicarbonatewasaddedtothebeadsandincubatedover-nightshakingat37°C.Thesupernatantcontainingthetrypticdigestwasrecoveredtocleantubes;next,the beads were washed with MS-grade water and the wash solution pooled with the originalsupernatant. All samples were lyophilized and analyzed by LC-MS/MS following the protocoldescribedintheGeneralMethodssection(6.3MassSpectrometry,onpage121).
Chapter4:Phenobarbital–newbindersforanolddrug|SpecificMethods
111
4.3.9. CaptureofphenobarbitaltargetsinhumanplacentalysateHuman placenta tissue was purchased from in.Vent, Berlin-Hennigsdorf, and was obtained
accordingtorelevantethicalguidelineswithinformedconsentofdonor. 400 µL of human placenta lysate, containing around 2000 µg protein, from the BioBank at
caprotec bioanalytics, were incubated with phenobarbital competitor (CC 15) at 2 mM finalconcentrationorwithequivalentvolumeofDMSOincaseofcontrolsamplesand5XCaptureBuffer.Thesampleswereequilibratedfor20minrotatingat4°C.
Fullpre-clickedcapturecompoundswerepreparedapriori:500µMofbiotincopper-chelating-azide(CC7)wasequilibratedwith5mMofCuSO4for10minatR.T.andthenreactedwith500µMofphenobarbital-alkyne capture compound (CC 10) in PBS at pH 8.5 by adding 2.5 mM of sodiumascorbate.Thereactionwasshakenvigorouslyfor10minatR.T.
Streptavidin-coated magnetic beads were pre-loaded with the pre-clicked phenobarbitalcapturecompoundusingexcesscapturecompoundtosaturatethebeads,shakingfor5minatroomtemperature.TheexcessunboundcapturecompoundwasremovedandthebeadswerewashedtwotimeswithCaptureBuffer.
The pre-loaded magnetic beads with phenobarbital capture compound were added to thesamplesand incubatedfor2hrotatingat4°C.ThebeadswerethencollectedusingacaproMag™andresuspendedin100µLofCaptureBuffer.Thesuspendedbeadswereirradiatedfor4min(twotimes2min,shakingthetubesinbetween)at310nminacaproBox™topromotephoto-crosslinking.ThemagneticbeadswererecoveredusingacaproMag™andthenwashedsixtimeswithwashbufferfollowedbythreetimeswashwith80%acetonitrileandafinalwashwithMS-gradewater.
0.5µgoftrypsininammoniumbicarbonatewasaddedtothebeadsandincubatedover-nightshakingat37°C.Thesupernatantcontainingthetrypticdigestwasrecoveredtocleantubes;next,the beads were washed with MS-grade water and the wash solution pooled with the originalsupernatant. All samples were lyophilized and analyzed by LC-MS/MS following the protocoldescribedintheGeneralMethodssection(6.3MassSpectrometry,onpage121).
ConclusionsandOutlook
112
5. ConclusionsandOutlookThe current challenges in drug discovery have strengthened the development of new
technologies for targetdeconvolution inanunbiasedapproach to identifyunwantedoff-targetsorreveal a new path for development early in the pipeline. Recent developments in chemicalproteomics associated to advances in quantitativemass spectrometry have enabled the detectionandquantificationoflowabundantbindersordrug-inducedchangesinproteinexpressionwithlargeproteomecoverage.
Capturecompoundmass spectrometry (CCMS) introducedanoveldesign for smallmoleculeprobesthatallowtheidentificationoflowabundantandloweraffinitytargetsinacomplexbiologicalsample. The capture compounds traditionally consist of a selectivity function carrying the smallmoleculeunderstudythatisattachedtoascaffoldbearingatagforvisualizationorenrichment,suchas a fluorophoreorbiotin, andalso aprotrudingphoto-reactivemoietywithenoughdistanceandflexibilityfromthecentralcoretocovalentlybindtheprobetothesurfaceofthetargetprotein.
One of themajor limitations of capture compounds in their originally published design andotheraffinity-basedprobes is thatmost cannotpermeate through cellsdue to their size, reducingthe application to cell-surface targets or the need to prepare lysates if the aimed protein isintracellular.Thedrawbackofworkingwithproteinlysatesisthatnotallcellularproteinsareeasilysolubilized,thusexcludingpartoftheproteome,andthelysisprocessmightinfactalterthenativeconformation of proteins, which can affect the interaction with the drug molecule. Subcellularfractionation instead of full lysate has been developed as a strategy to overcome some of theseobstacles.
Bio-orthogonalchemistrywasinitiallydevelopedtostudypost-translationalmodificationsbutquicklyexpandedtoabroaderusetostudydynamicsandfunctionofbiomolecules.Severalreactantpairshavebeendevelopedbuttheazideandalkynereactionstillremainthebettersuitedforlivingcellapplicationsduetotheirreducedsize.Thebiggestdisadvantageofthisbio-orthogonalligationistheneedtoaccelerateitwithmetals,copperbeingthemostwidelyused,whichcaninducestressincells.Advancesinthetechnologyweremadewiththeintroductionofchelatingagentsthatimprovethereactionkineticsandreducethemetal-inducedstress.
Establishinganewdrugdeconvolutionmethodthatallowstheidentificationoftargetsintheirnatural cell environment was set as a primary goal in this research project. The strategy was tocombinethestrengthofcapturecompounds intheenrichmentof lowabundantand loweraffinitytargetswiththeversatilityofbio-orthogonalligationchemistry.Thenewprobedesignwouldallowitto be small enough and increase cell permeability to target intracellular proteins and specificallycapture them with a controlled chemical click reaction. One main advantage of such a workflowwould be to identify drug-binding proteins directly under conditions that are used in phenotypicscreening-baseddrugdiscovery.
A modified capture compound bearing cAMP as selectivity function and a photo-reactivityfunction,buthavingareactivealkyneinsteadofthetraditionallabelingfunction,wassynthesizedtoestablishcopper-assistedazido-alkynecycloaddition(CuAAC)clickchemistryinthecontextofCCMS.The optimization steps to establish the click reaction conditions were initially established in asolutionofrecombinantPKARI,thesubunitofproteinkinaseAthatbindscAMP.Abiotin-azidewas
ConclusionsandOutlook
113
used as label to monitor this reaction and results were analyzed by western blot. The initialexperiments confirmed the specific nature of the azido-alkyne cycloaddition where no signal wasobservedinsampleslackingthecapturecompound(Figure2-2,onpage32)butalsoshowedthatthesignal was a fraction of the one obtained with a classical cAMP capture compound. DifferentparametersfromtheCuAACreactionweremodifiedtoimprovethelabelingofPKARIsuchascopperand ligands concentrations; time and pH of reaction; biotin-azide and alkyne capture compoundconcentrations; and a new fluorescent label that promised to be better suited for the labeling ofcomplexsamples,suchascellularlysates.TheoptimizationeffortsyieldedsomeimprovementfromtheinitialexperimentsbutstilldidnotachievethesameresultsaswiththeclassicalcAMPcapturecompound.
Capturing PKA subunits from cell lysates with the optimized parameters proved to be sub-optimalwhenusing thenewclickworkflowandcomparing it to resultsobtainedwith theclassicalcAMP capture compound. Although the results suggested an enrichment of PKA-related peptides,theintensityratiobetweenassaysandcompetitionsandtheirstatisticalsignificancedidnotallowfora conclusive analysis on the performance of the method. The lack of competition in severalexperiments proved to be a challenge if further efforts would be made to resolve this issue,especiallywhenthesubsequentstepwouldbetoestablishtheclickcaptureinlivingcells,wherethenatural concentration of cAMP is very high. Nevertheless, the system was useful to establishtechnical aspects of the click reaction in the context of the use of capture compounds. However,becauseoftheimpermeabilityofcAMPassuch,adifferentselectivityfunctionwasrequiredforthenextsteps.Withthisinmind,anewnon-naturalselectivityfunctionwasusedfortargetidentificationincelllysatesandlivingcellsusingthenewclickreaction.
ThekinaseinhibitordrugDasatinibwasusedforitsclinicalrelevanceandwell-knowntargets.Concomitantly to this newalkynedasatinib capture compound, a newazidewas alsoproduced incollaborationwiththegroupofTaran,fromtheCEASaclay,memberoftheBioChemLigconsortium.Thenewazide,bearingthebiotinorfluorescentTAMRAastags,alsohadacopper-chelatingmoietythatbetteracceleratedthecycloadditionreaction.Thewellcharacterizedtargetprofileofdasatinibwaswellsuitedasreferencetojudgetheeffectivenessofthenewapproach.Apreviouslypublishedstudyonlivingcellcapturewithadasatinibprobefailedtoproduceameaningfultargetprofile92,sotherewasobviousroomandneedforabetterworkflow.Usingthedasatinibcapturecompoundpre-clickedtothecopper-chelatingbiotinallowedforthesuccessfulidentificationofbonafidedasatinibtargets inaBCR-ABL-positivecell lysate,suchasABL1,ABL2,BCRandSRC,aswellasotherkinases(Table 3–2 and Figure 3-5, on page 61). Even though the new clickworkflowdid not produce thesame resultsas the full compound in lysate, theseproteinswereenrichedwhencomparing to thenegative controlbutwerenot competedenoughby freedasatinib topass the inclusion criteriaoftwo-times fold-change and p value less than 0.05. Other known dasatinib inhibited kinases weresuccessfullyidentifiedusingthisprotocol.
Searching for dasatinib targets in living cells showed that a large number of biological andtechnical replicates is necessary to achieve statistical significance when probing living cells.Individually, each experiment consisting of technical replicates yielded very different resultsregarding potential targets (Table 3–7, on page 70). Calculating the intensity differences betweenassays and competitions of all experiments as one or attributing a re-incidence score allowed abetter understanding of the results. In fact, using this strategy, known dasatinib targets were
ConclusionsandOutlook
114
competitively identified for the first time in living cells,with special relevance to the C-Src kinase(CSK),aswellasthebonafidedasatinib-binderABL1/ABL2.
Havingsuccessfullyimplementedthecaptureofsmallmoleculetargetsinlivingcells,thenewmethodologywasusedtoidentifynewbindersforphenobarbital,adrugthathasbeeninusefor100years and still has unknown mechanisms of action regarding its toxicity. The hepatocarcinogeniceffectofphenobarbitalinsomespecieswasthefocusofthisendeavor,andatthetimeexperimentswere being performed it was not known that EGFR could bind phenobarbital and elicit a cellularresponse,aspublishedbyMutohetal.in2013142.
Inordertolaterbeabletoassociatecaptureresultstophenotypiceffects,thefirstobjectivewas to implement a phenotypic assay that would serve as a validation method once thephenobarbital targetcandidatewas identifiedbymassspectrometry.Several systemswere tested,such as AMPK phosphorylation or CAR translocation that according to literature happens uponphenobarbitalstimulation,buttheresultsobtaineddidnotconfirmthatandcouldthereforenotbeused as reliable readouts. The constitutive androstane receptor (CAR) was already found in thenucleusofunstimulatedcellsandAMPKphosphorylationwascyclicmakingthempoorcandidatesasavalidationmethod.ThesuccessfulphenotypicassayturnedouttobetheexpressionofcytochromeP4502B1mRNAthatclearlyincreasesuponhighdosageofphenobarbital(Figure4-18,onpage96)in both rat hepatoma cell line and primary hepatocytes. Unfortunately, by the time massspectrometryresults indicatedpotential targets tobevalidated, therewasno furtherpossibility tocontinuetheresearchduetotimeconstraints.
The mass spectrometry analysis of the initial capture experiments using a pre-clickedphenobarbitalcapturecompoundinlysatefromarathepatomacelllineprovidedinterestingresults.Mostoftheputativehitshadbeenpreviouslyassociatedtosomephenobarbitaleffectandsomehadalso known interactions with EGFR. In those experiments, subunits from the phosphatase PP2Aresponsible for the de-phosphorylation of AMPK and RACK1, the EGFR-associated kinase that aidsthetranslocationofCARtothenucleus,wereidentifiedascandidatebindersofphenobarbital.Evenmore interestingly,Wdr1 and Ybx1were also presented as putative targets of phenobarbital. Thefirst being a WD repeat-containing protein, characteristic shared with RACK1, and the second aproteinthatbindsY-boxes,specificsequencesingenepromoters,thatarepartofthephenobarbitalresponseenhancermodule(PBREM)presentinCYP2Bgenes.
Results fromcaptureexperimentswithphenobarbitalprobes in livingMH1C1 cells continuedpointing towards a connection to the EGFR signaling pathway. Different proteins were identifieddepending on whether the capture compound was pre-clicked or not, but all have documentedassociationwithEGFR:athioesterase(Them6)andannexinA4(Anxa4)were identifiedaspotentialtargets insamplesprobedwithacompletephenobarbitalcapturecompound;andADP-ribosylationfactor 5 (Arf5) and dynamin-2 (Dnm-2) were found as binders in samples with the clickablephenobarbitalcapturecompound.
All the identifications in the phenobarbital experiments directed the attention for EGFR butthisproteinwasneverspecificallycaptured.Toanswertheposingquestionofwhetherthereceptorwas not targeted because it is not a direct binder of phenobarbital or if it was only due to lowabundanceinthesample,alastexperimentwasperformedinatissuethatexpresseshighlevelsofEGFR.Theresults fromthecapture inhumanplacenta lysateconfirmedsomepreviously identifiedproteinsasputativetargetsofphenobarbitalanditsconnectiontotheEGFRsignalingpathway,but
ConclusionsandOutlook
115
thereceptoritselfwasnotcompetedbyphenobarbital.ArecurrentcandidateastheprimarybinderofphenobarbitalwasYBX1,andgivenitsroleastranscriptionfactoranddirectbinderofEGFR,thepleiotropiceffectsand themodulationof theEGFRsignalingcascade in response tophenobarbitaltreatment could be explained. Further validation is necessary to confirm that YBX1 is inhibited byphenobarbital.
Apointof concern, inhindsight, is that thephenotypic effect elicitedbyphenobarbital onlyoccurs at high compound concentrations, close to the millimolar range, suggesting low affinitytowards the pathway under study. To further pursue CCMS in living cellswith permeable capturecompounds,thefocusshouldbeonmorepotentsmallmolecules,elicitingeffectsinthenanomolarrange.
Amongstthedrugtargetdeconvolutiontechniques,CCMSremainsthemostrobustandwithhigherpotentialtoinnovate.Atthetime,probinglivingcellswithbio-orthogonalcapturecompoundsis a reality but the results are still far from ideal. Effortsmust be put in reducing background ofcaptureandmassspectrometrymethodsaswellasexperimentingwithnewclickablepairsthatallowforabetterreactionyieldand,consequently,betteridentifications.
Inmyopinion, thebest strategypresently fordrug targetdeconvolution is to firstly identifywhich subcellular structures contain the target, using bio-orthogonal reactions and microscopyanalysis,fractionateandenrichthoseorganellesandthenuseCCMSinthosesubproteomes.
GeneralMethodsandMaterials|CellCulture
116
6. GeneralMethodsandMaterials6.1. CELLCULTURE6.1.1. HEK293cellculture
Humanembryonickidney(HEK)cells,clone293,wereusedincAMP/PKAcaptureexperiments(asdescribedin2.Labelingaproteinusingclickchemistry,onpage28).
HEK293cellswereculturedinDulbecco’sModifiedEagleMedium(DMEM)supplementedwith10% heat-inactivated fetal bovine serum (FBS) and 1% of penicillin/streptomycin in a 5% CO2incubatorat37°C.Thecellsweresubculturedata1:5–1:6dilutionevery2-3daysbymechanicaldetachment.
6.1.1.1. HEK293lysatepreparation
HEK293 cell pellets were lysed in 2X their volume with a hypotonic buffer (10 mM HEPESpH7.9, 1.5 mM MgCl2, 10 mM KCl) containing or not 0.5% (w/v) DDM. Both buffers weresupplementedwithproteinaseinhibitorsandbenzonaseatthetimeoflysis.After30minincubationon ice, the lysatewasdounced20Xusinga tightpestleandthenrunthrougha21Gneedlewithasyringe.Remainingdebriswaspelletedbyrefrigeratedcentrifugationat15.000rpmfor30min.Thesupernatant was recovered and the leftover pellet was discarded. Aliquots weremade and snap-frozeninliquidnitrogen.Sampleswerestoredat-80°Cuntiluse.
6.1.2. K562cellcultureHumanK562cellsoriginatefromachronicmyeloidleukemiainblastcrisis,andwereusedin
theidentificationofdasatinibtargetsexperiments(asdescribedin3.IdentifyingDasatinibtargetsincells,onpage54).
K562cellswereculturedinRPMI-1640mediumsupplementedwith10%heat-inactivatedfetalbovine serum (FBS) and1%penicillin/streptomycin in a5%CO2 incubator at 37 °C. The cellsweresplitevery3daysata1:3–1:5dilutionbypelletingthesuspensioncellsfor10minat1000xg.
6.1.2.1. K562lysatepreparation
K562cellpelletswerelysedineither(1)RIPAbuffer(50mMTris-HClpH8.0,150mMNaCl,1%IGEPAL®CA-630,0.5%(w/v)sodiumdeoxycholate,0.1%(w/v)SDS)supplementedwith0.2mMDTT;or(2)LysisBufferE(50mMTris-HCl,100mMNaCl,1.5mMMgCl2,0.2%(v/v)IGEPAL®CA-630,0.5%(w/v) DDM, 5% (v/v) glycerol) supplemented with 1 mM DTT. Both buffers were additionallysupplementedwithproteinaseinhibitorsand0.5μL/mlbenzonase.
After30minincubationonice,thelysatewasdounced20Xusingatightpestleandthenrunthrougha21Gneedlewithasyringe.Remainingdebriswaspelletedbyrefrigeratedcentrifugationat15.000 rpm for 30 min. The supernatant was recovered and the leftover pellet was discarded.Aliquotsweremadeandsnap-frozeninliquidnitrogen.Sampleswerestoredat-80°Cuntiluse.
GeneralMethodsandMaterials|CellCulture
117
6.1.3. MH1C1cellculture
AratclonalstrainofepithelialcellsderivedfromMorrishepatoma#7795,MH1C1cellline,wasusedinphenobarbitalexperiments(asdescribedin4.Phenobarbital:newbindersforanolddrug,onpage81).
MH1C1cellswereculturedinDulbecco’sModifiedEagleMedium(DMEM)supplementedwith10%heat-inactivatedfetalbovineserum(FBS)and1%penicillin/streptomycinina5%CO2incubatorat37°C.Thecellsweresplitusingtrypsin/EDTAonceortwiceperweekata1:2–1:4dilution.
6.1.3.1. MH1C1lysatepreparation
MH1C1cellpelletswerelysedinLysisBufferE(50mMTris-HCl,100mMNaCl,1.5mMMgCl2,0.2% (v/v) IGEPAL® CA-630, 0.5% (w/v) DDM, 5% (v/v) glycerol) supplemented with 1 mM DTT,proteinaseinhibitorsand0.5μL/mlbenzonase.
After30minincubationonice,thelysatewasdounced20Xusingatightpestleandthenrunthrougha21Gneedlewithasyringe.Remainingdebriswaspelletedbyrefrigeratedcentrifugationat15.000 rpm for 30 min. The supernatant was recovered and the leftover pellet was discarded.Aliquotsweremadeandsnap-frozeninliquidnitrogen.Sampleswerestoredat-80°Cuntiluse.
6.1.4. RatprimaryhepatocytescultureRat (male Sprague-Dawley) cryopreserved hepatocytes were used in phenobarbital
experiments(asdescribedin4.Phenobarbital:newbindersforanolddrug,onpage81).The cryopreserved hepatocyteswere thawed in a 37 °Cwater bath for less than 2min and
transferredtoapre-warmedtubecontainingWilliamsMediumEandHepatocytePlatingSupplementPack(ThermoFisherScientific).Thecellswerecentrifugedat55xgfor3minatroomtemperature.
Thesupernatantwasdiscardedand1mLper1x106cellsofplatingmediumwasadded.ThehepatocytesweredilutedandseededincollagenI-coated24-wellplatesataconcentrationof8x105cellperwell(500µL).Theplatewasincubatedfor4hat37°Ctoallowadherenceandformationofmonolayer. After incubation the platewas agitated to release unattacheddebris and themediumwasaspiratedandreplacewithfreshcompletemedium.
The ratprimaryhepatocyteswereused forRT-PCRexperiments (asdescribed in4.3Specificmethodsforthephenobarbitalsystem,4.3.6RT-PCR,onpage107).
GeneralMethodsandMaterials|BiochemistryMethods
118
6.2. BIOCHEMISTRYGENERALMETHODS6.2.1. Proteinconcentrationdetermination
TotalproteinconcentrationfromcelllysateswasdeterminedusingthecommerciallyavailableBCAProteinAssayReagent(Pierce™,Prod.#23225).TheBCAProteinAssaywasintroducedbySmith,etal.in1985184.Themainbenefitofthisassayisthecompatibilitywithawiderangeofdetergentsnormallyusedincellularlysates.Theassayreliesonacolorimetricdetectionofcopperreductionbybicinchoninicacid(BCA).
A standard curve of bovine serumalbumin (BSA) (Pierce™, Prod #23209)was prepared in adilutionseriesusingthesamelysisbufferofthecellular lysatetobequantified.Thesamplestobequantifiedwerealsodiluted(1:4–1:20)inbuffertoassurethemeasurementiswithintherangeofthestandards.
10μLofBSAstandardorsamplewerepipetted intriplicates intoa96-well flat-bottomplateand200μLofworkingreagent(50:1,reagentA:reagentB)wereaddedtothewells.Theplateswereincubated for 30min at 37°C and the light absorbance read in amultiplate reader (Anthos 2010,AnthosMikrosystemeGmbH)witha562nmfilter.
Thevalueswereinsertedinaspreadsheetandtheaverageofthestandardswasusedtoplotandcalculatealinearregressioncurve.Thetotalproteinconcentrationinthesampleswascalculatedbyinferenceandcorrectedforthedilutionfactor.
6.2.2. SDS-PAGESodiumDodecylSulfatePolyacrylamideGelElectrophoresis(SDS-PAGE)isaproteinanalytical
method established byMaizel 185 in 1963 as an advancement of the original PAGE developed byWeintraubin1959186.ThecurrentmethodmostlyusedaroundtheworldwasoptimizedbyLaemmliin1970187.ForaninsightonthedevelopmentandestablishmentofSDS-PAGEasareferenceproteinanalyticalmethodseeMaizel’sreviewpublishedin2000188.
Proteinsolutions,cellularlysatesorwholecellsweremixedwithLaemmliBuffertoafinal1Xconcentration (50 mM Tris-base, 2.5% (w/v) SDS, 10% Glycerol, 0.32 mM β-mercaptoethanol,bromophenol blue). Samples were loaded on pre-casted 4-20% Tris-Glycine gels (anamedElektrophorese GmbH, Germany) and proteins were separated for 35 min at 300V on a XCellSureLock®Mini-Cell electrophoresis chamber (Life Technologies™GmbH,Germany) filledwith SDSRunningBuffer(25mMTrisbase,192mMGlycine,0.1%(w/v)SDS).Apre-stainedproteinmolecularweightmarkerwasalwaysloadedinonewellasareference.
After separation, gels were directly stained using Coomassie Blue or Silver staining, ortransferredontonitrocellulosemembranesforWesternBlotting.
GeneralMethodsandMaterials|BiochemistryMethods
119
6.2.3. PolyacrylamideGelStainingResolvedproteinsonpolyacrylamidegelscanbevisualizedwithspecificdyes.Thetypeofdye
tousewill dependon the abundanceof theproteinsof interest. The first dyeof choice is usuallyCoomassie Brilliant Blue because of its simple protocol and the signal-to-background is, in mostcases, sufficient tomakeanassessmentof theexperiment.More sensitivedyes, suchas theonesusing Silver, usually take longer procedures but are necessary when dealing with low abundantproteins.
6.2.3.1. CoomasieBlue
The Coomassie-based gel stain used to visualize proteins on gels was the SimplyBlue™SafeStain from Invitrogen®. It contains a proprietary Coomassie G-250 stain for a fast sensitivedetectionwithoutrequiringtheuseofmethanoloraceticacid.
The protocolwas followed as described by themanufacturer. After electrophoresis, the gelwasrinsed3timesfor5minwithultrapurewatertoremoveSDSandbuffersalts.Thegelwasthenincubatedforatleast1hwiththecoomassiesolution,gentlyshakingatroomtemperature.Thegelwas thenwashedwithultrapurewater for thenecessary time to reducebackgroundand increasecontrast with stained bands. The gel was kept in water if further mass spectrometry analysis ofindividualbandswasperformed.
6.2.3.2. SilverStain
The ProteoSilver™ Silver Stain Kit (Sigma®) was used to detect low abundant proteins.ProteoSilveruses silvernitrate,whichbinds to selectiveaminoacidson theproteinsunderweaklyacidic or neutral pH conditions. The protein bound silver ions are reduced by formaldehyde atalkalinepHtoformmetallicsilverinthegel.
Theprotocolwasfollowedaccordingtomanufacturer’sguidelines.Thepolyacrylamidegelwasfixed forat least20min in fixingsolution (50%ethanol,10%aceticacid), followedbyawashwith30%ethanolandawashwithultrapurewater,10mineach.Thegelwasthenincubatedfor10mininsensitizer solution (contentsnotdescribed) and thenwashed two times for 10minwithultrapurewater.Afterthat,thegelwasequilibratedinsilversolutionfor10minfollowedbyaquickwashwithultrapurewaterandimmediatelysoakedindevelopersolution(contentsnotdescribed).Incubationtimes in this solutionvarieduntildesiredbandswerevisualized.To stop furtherdevelopmentandincreaseofbackgroundthereactionwasterminatedbyaddingstopsolution(contentsnotdescribed)directly to the immersed gel. The solution was discarded and the gel was washed for 15 min inultrapurewater. The stained gel could be kept in fresh ultrapurewater until further analysiswasnecessary,suchasmassspectrometryanalysisofindividualproteinbands.
GeneralMethodsandMaterials|BiochemistryMethods
120
6.2.4. WesternBlottingIn1979,HarryTowbindevelopedamethodtoallowfortheprobingofanyelectrophoretically
separatedproteinwithantibodiesorserum189.Theprinciplebehind itwastocreateacopyoftheelectrophoreticpatternobtainedonthegelbytransferringittoanitrocellulosemembrane189.Twoyears later, Burnette 190 made a few modifications to the transfer buffer conditions to allow acomplete andquantitative elutionof proteins from the SDS gel onto thenitrocellulosemembraneand coined it “Western Blotting”. Thismethod is still today themost commonly used for proteintransfer.
After gel electrophoresis, the resolved proteins were transferred onto a nitrocellulosemembrane (0.2 µm, Whatmann™) using a wet-transfer blotting tank (Biozym™) in either a Tris-glycine homemade transfer buffer or a proprietaryWestern Blot Transfer Buffer from Pierce™. Astack of (bottom-to-top) three layers of Whatmann paper, the nitrocellulose membrane, thepolyacrylamide gel, and another three layers of Whatman paper was assembled in a blottingchamber.All layerswere soakedwithblottingbuffer andelectrotransferwasperformedat 1mA/cm2(45mApergel).Tovisualizeproteins,thenitrocellulosemembranewasstainedwithPonceauSsolution.ThestainwasremovedwithTBS-Twashingandblockedin2%(w/v)non-fatmilk inTBS-Tforatleast1hatroomtemperatureorover-nightat4°C.
For detection of biotinylated proteins, the membrane was incubated with infrared (IR) orhorseradishperoxidase (HRP)-conjugatedstreptavidindiluted1:1000 inblockingsolutionfor2hatroomtemperatureorover-nightat4°C.BlotswerethenwashedthreetimeswithTBS-Tfor10minandkeptinTBSuntilreadout.
Forspecificproteinsdetection,themembranewaswashedthreetimeswithTBS-Tfor10minand then incubated with primary antibody solution (diluted in blocking solution) for 2 h at roomtemperatureorover-nightat4°C.Afterthree10minwashesinTBS-T,themembranewasexposedtoIR or HRP-conjugated secondary antibody solution (diluted in blocking solution) for 1 h at roomtemperature. Finally, the blot was washed three times in TBS-T for 10min and kept in TBS untilimageacquisition.
IR-labeledblotsweredirectlyacquiredintheimagingsystem(SyngeneG:BOXXT4)usingtheappropriatecombinationofexcitationlightandcutofffilters.
For HRP-labeled blots, the membranes were incubated with enhancedchemiluminescence(ECL)detectionreagent(Pierce®)andtheemittinglightwasrecordedintheimagingsysteminthedarkwithoutanyfilter.
GeneralMethodsandMaterials|MassSpectrometry
121
6.3. MASSSPECTROMETRY6.3.1. Sampleacquisitionanddatabasematchidentification
TrypticdigestswereanalyzedbyonlinenanoflowliquidchromatographytandemMS(LC-MSn)onanUltiMate3000RSLCnanoSystem(Dionex,partofThermoFisherScientific,Germany)coupledto a LTQ-Orbitrap Velos instrument (Thermo Fisher Scientific, Germany) through a Proxeonnanoelectrospray ion source (Proxeon, part of Thermo Fisher Scientific, Germany). Forchromatographic separation, samples were first loaded on a reversed phase (RP) pre-column(AcclaimPepMap100,5mm,100Å,100µmi.d.x20mm)andseparatedonaRPanalyticalcolumn(Acclaim PepMap RSLC C18, 2 µm, 100 Å, 75 µm i.d. x 150 mm, Dionex, part of Thermo FisherScientific,Germany)performinga96mingradient(5–45%acetonitrile,0.1%formicacid).
MS detectionwas performed in the data-dependentmode allowing to automatically switchbetweenOrbitrap-MSandLTQ-MS/MSacquisitioninatop20configurationat60Kresolutionforafull scanwith subsequent collision induced dissociation (CID) fragmentation. Full scanMS spectra(fromm/z300–2000)wereacquiredintheOrbitrapanalyzerafteraccumulationtoatargetvalueof1x106inthelineariontrap.Themostintenseions(uptotwenty,dependingonsignalintensity)withchargestate≥2weresequentiallyisolatedatatargetvalueof5000andfragmentedinthelineariontrapusinglowenergyCIDwithnormalizedcollisionenergyof35%.TargetionsalreadymassselectedforCIDweredynamicallyexcludedforthedurationof60s.TheminimalsignalrequiredforMS2was1000 counts. An activation q of 0.25 and an activation time of 10 ms were applied for MS2acquisitions.
AllMS/MSdatawereanalyzedusingAndromeda implemented inMaxQuant 120. Automateddatabase searching against the humanUniProtKB/Swiss-Prot databasewas performedwith 6 ppmprecursortolerance,0.5Dafragmentiontolerance,fulltrypsinspecificityallowingforupto2missedcleavages and methionine oxidation as variable modification. The maximum false discovery ratesweresetto0.01bothonproteinandpeptidelevel,themaximumPEPto1,and7aminoacidswererequired as minimum peptide length. The label free quantification option was selected with amaximalretentiontimewindowof2minforthealignmentbetweenLC-MS/MSruns.
Calculations for fold-changebetweendifferentsubsetsof samplesandStudent’s t-TestwereperformedmanuallyinExcelspreadsheetsusingbothMS-basedvalues(LFQvaluesfromMaxQuant)andalignment-basedintensityvalues.
CCMSchartsweredevelopedtocomprehensivelydisplaythemulti-parameterresultsobtainedinCCMSexperiments for individualproteins,plottingminimumandmaximumintensityvaluesasaboxwithablackhorizontalbar for themedianvalue± thestandarderrorof themean (SEM). Inaseparate axis, the fold-change between the subsets of samples is plotted as a green triangle andsignificant(p<0.05)differencesarehighlightedwithanasterisk(*).
GeneralMethodsandMaterials|MolecularStructuresofCompounds
122
6.4. MOLECULARSTRUCTURESOFCOMPOUNDS6.4.1. CapturecompoundsCC1:ClassicalfullcAMPcapturecompound
CC2:ClickablecAMPalkynecapturecompound(containsselectivityandphotoreactivefunctions)
CC3:Clickablebiotin-PEG-azide(containslabelingfunction)
CC4:ClickedcAMP-biotincapturecompound.ReactionproductofCC2andCC3
GeneralMethodsandMaterials|MolecularStructuresofCompounds
123
CC5:ClickableTAMRA[tetramethylrhodamine]azide(containslabelingfunction)
CC6:ClickableDasatinibalkynecapturecompound(containsselectivityandphotoreactivefunctions)
CC7:Clickablebiotin-PEGcopper-chelatingazide(containslabelingandcopperligandfunctions)
CC8:ClickableTAMRA-PEGcopper-chelatingazide(containslabelingandcopperligandfunctions)
GeneralMethodsandMaterials|MolecularStructuresofCompounds
124
CC9:Clickeddasatinib-biotincapturecompound.ReactionproductofCC6andCC7
CC10:Clickablephenobarbitalalkynecapture
compound A (contains selectivity and a
trifluoromethyl-phenyl-diazirine photoreactive
functions)
CC11:Clickablephenobarbitalalkynecapture
compound B (contains selectivity and a
methyl-diazirinephotoreactivefunctions)
CC 12: Clickable phenobarbital alkyne capture compound C (contains selectivity and a azido-
tetrafluorobenzenephotoreactivefunctions)
GeneralMethodsandMaterials|MolecularStructuresofCompounds
125
6.4.2. CompetitorsCC13:cAMPcompetitor
CC14:Dasatinibcompetitor
CC15:Phenobarbitalcompetitor
6.4.3. LigandsCC16:
Copper-chelatingligandTBTA
Tris(benzyltriazolylmethyl)
amine
CC17:
Copper-chelatingligandTHPTA
Tris(3-hydroxypropyltriazolylmethyl)
amine
CC18:
Copper-chelatingligandBTTE
bis(tert-butyltriazoly)ethanol
GeneralMethodsandMaterials|Materials
126
6.5. MATERIALS6.5.1. Chemicalreagents
ChemicalnameMolecularFormula
Molarmass(g/mol)
CAS SupplierRiskandSafetyStatements
(+)-SodiumL-ascorbate C6H7NaO6 198.11 134-03-2 Sigma -
Acetonitrile(HPLCgrade) C2H3N 42.05 75-05-8 SigmaR:11-20/21/22-36S:16-36/37
Ammoniumbicarbonate NH4HCO3 79.06 1066-33-7 Sigma R:22
Biotin C10H16N2O3S 244.31 58-85-5 BioMag -
BromophenolBluesodiumsalt
C19H9Br4NaO5S 692.00 34725-61-6 Serva S:22-24/25
Copper(II)sulfatepentahydrate
CuSO4•5H2O 249.69 7758-99-8AlphaCaesar
R:22-36/38-50/53S:22-60-61
Dimethylsulfoxide C2H6SO 78.13 67-68-5 Acros -
DL-Dithiothreitol C4H10O2S2 154.25 3483-12-3 FlukaR:22-36/37/38S:26-36
EDTA-Na2 C10H14N2Na2O8•2H2O 372.24 6381-92-6 Gerbu -
Formicacid CH2O2 46.03 64-18-6 FlukaR:35S:23-26-45
Glycerol C3H8O3 92.09 56-81-5 Gerbu -
Glycine C2H5NO2 75.07 56-40-6 Gerbu S:22-24/25
Hepes C8H18N2O4S 238.30 7365-45-9 Gerbu -
Hydrochloricacid HCl 36.46 7647-01-0 SigmaR:34-37S:26-45
IGEPAL®CA-630 (C2H4O)nC14H22O - 9036-19-5 FlukaR:41-50S:26-39-61
Iodoacetamide C2H14INO 184.96 144-48-9 ApplichemR:25-42/43-53S:22-36/37-45
Lutensol®GD70 - - 31799-71-0 BASFR:41S:39-26
Magnesiumacetatetetrahydrate
C4H14MgO8•4H2O 214.45 16674-78-5 Sigma -
Potassiumacetate C2H3KO2 98.14 127-08-2 Sigma -
Potassiumhydroxide KOH 56.11 1310-58-3 SigmaR:22-35S:26-36/37/39-45
Sodiumchloride NaCl 58.44 7647-14-5 Roth -
Sodiumdeoxycholate C24H39NaO4 414.55 302-95-4 Sigma R:22-37
Sodiumdodecylsulfate C12H25NaO4S 288.38 151-21-3 GerbuR:11-21/22-36/37/38S:26-36/37
Sodiumhydroxide NaOH 40.00 1310-73-2 CarlRothR:35S:26-37/39-45
Trifluoroaceticacid(99%) C2HF3O2 114.02 76-05-1 AcrosR:20-35-52/53S:9-26-27-28-45-61
Tris(hydroxymethyl)aminomethane
C4H11NO3 121.14 77-86-1 VWR R:36/38
Triton™X-100C14H22O(C2H4O)n,n=9-10
x̄=625 9002-93-1 FlukaR:22-36-51/53S:26-39-61
Trypanblue C34H24N6Na4O14S4 960.82 72-57-1 Fluka R:45S:53,45
GeneralMethodsandMaterials|Materials
127
ChemicalnameMolecularFormula
Molarmass(g/mol)
CAS SupplierRiskandSafetyStatements
Trypsin(sequencinggrade,modified,porcine)
- - 9002-07-7 PromegaR:36/37/38-42-42/43S:22-24-26-36/37-45-23
Tween®20 - 1228 9005-64-5 Sigma -
Β-Mercaptoethanol C2H6OS 78.13 60-24-2 ApplichemR:23/24/25-38-41-43-48/22-50/53S:26-36/37/39-45-60-61
N-Dodecylβ-D-maltoside C24H46O11 510.62 69227-93-6 Sigma -
6.5.2. KitsandConsumablesCategory Product ManufacturerAntibody PKARIα BDBiosciencesAntibody AMPKα1/2(H-300) SantaCruzBiotechnologyAntibody CAR1/2(M-127) SantaCruzBiotechnologyAntibody p-AMPKα1/2(Thr172) SantaCruzBiotechnologyAntibody CYP2B1/2B2(9.14) SantaCruzBiotechnologyBiotindetection HRP-conjugatedStreptavidin SigmaBlottingBuffer 10XWesternBlotTransfer
Buffer,Methanol-freePierce,Thermo
CellFractionation NE-PER®NuclearandCytoplasmicExtractionReagents
PierceBiotechnology
GelStaining SimplyBlue™SafeStain InvitrogenGelStaining ProteoSilver™SilverStainKit SigmaHorseradishperoxidasedetection
SuperSignalWestDuraChemiluminescentSubstrate
Pierce,Thermo
MagneticBeads DynaBeads MembraneStain Ponceau Molecularweightmarker PageRulerPlusPrestained
ProteinLadderPierce,Thermo
PolyacrylamideGels 4-20%Tris-GlycinGel AnamedElektrophoreseGmbHProteaseinhibitor cOmplete,Mini,EDTA-free
ProteaseInhibitorCocktailTablets
Roche
ProteinQuantificationAssay BCAProteinAssayReagent Pierce,Thermo
GeneralMethodsandMaterials|Materials
128
6.5.3. Buffers,mediaandsolutions StockSolution WorkSolutionSDSElectrophoresisRunningBuffer 10x 250mMTris
1.92MGlycine1%(w/v)SDS
1x For1L:100mLstock+900mLH2O
BlottingBuffer 10x 478mMTris386mMGlycine0.37%(w/v)SDS
1x For1L:100mLstock+700mLH2O+200mLMethanol
TrisBufferedSaline(TBS) 10x 0,5MTris1,5MNaClpH7.5(withHCl)
1x For1L:100mLstock+900mLH2O
LaemmliBuffer 4x 200mMTris-BasepH6.810%(w/v)SDS40%(v/v)Glycerol1.28Mβ-MeBromophenolblue
2x For1mL:500μLstock+500μLH2O
Radioimmunoprecipitationassay(RIPA)buffer
1x 150mMNaCl1%(v/v)IGEPAL®CA-6300.5%(w/v)sodiumdeoxycholate0.1%(w/v)SDS50mMTrispH8.0
Usestock
WashingBuffer 5x 250mMTris5mMEDTA-Na25MNaCl0.25%(v/v)LutensolGD70pH7.5(withHCl)
1x For1mL:200μLstock+800μLH2O
CaptureBuffer 5x 100mMHEPES250mMPotassiumacetate50mMMagnesiumacetatetetrahydratepH7.9(with1MKOH)50%(v/v)Glycerol
1x For1mL:200μLstock+800μLH2O
LysisBufferE 1x 50mMTris-HCl100mMNaCl1.5mMMgCl20.2%(v/v)IGEPAL®CA-6300.5%(w/v)DDM5%(v/v)Glycerol1mMDTT
Usestock
CellCultureMediaandSupplements Cellline RPMI1640 K562;HEK293 WilliamsMediumE Primaryhepatocytes DMEM MH1C1 Heatinactivatedfetalbovineserum(FBS) - Penicillin/Streptomycinsolution -
GeneralMethodsandMaterials|Materials
129
6.5.4. Equipment
Category Model/type ManufacturerCaprobox 310and350nm caprotecbioanalyticsGmbH,Berlin,DECentrifuge Microstar17R VWRInternationalGmbH,Darmstadt,DECentrifuge 5702R EppendorfAG,Hamburg,DECentrifuge MicrocentrifugeAL CarlRothGmbH,Karlsruhe,DECentrifuge MinicentrifugeMCF-2360 LMSConsultGmbH,Brigachtal,DEElectrophoresis BlottingsystemEBX-700 C.B.S.Scientific,SanDiego,CA,USA
Electrophoresis NovexMini-CellLife Technologies, Thermo Fischer ScientificGmbH,Dreieich,DE
Electrophoresis PowerSupply300V VWRInternationalGmbH,Darmstadt,DEIceMachine ZNE125 ZiegraEismachinenGmbH,Isernhagen,DE
ImagingSystem SyngeneG:BOXXT4Integrated Scientific Solutions Inc., San Diego,CA,USA
Incubator HeraCell150 ThermoFischerScientificGmbH,Dreieich,DEMagneticStirer RCTbasic IKA-WerkeGmbH,Staufen,DEMassSpectrometer LTQVelos ThermoFischerScientificGmbH,Dreieich,DEMassSpectrometer LTQXL ThermoFischerScientificGmbH,Dreieich,DEMassSpectrometer amaZonspeed BrukerCorporation,Billerica,MA,USA
Microscope InvertedMicroscopeDMILLeica Mikrosysteme Vertrieb GmbH, Wetzlar,DE
nHPLC Ultimate3000Dionex, Thermo Fischer Scientific GmbH,Dreieich,DE
pHmeter SevenEasyS20 MettlerToledoGmbH,Gießen,DEPipetting FinnpipetteNovusiMCP8 ThermoFischerScientificGmbH,Dreieich,DEPipetting ManualResearch EppendorfAG,Hamburg,DERotator Multimix115V VWRInternationalGmbH,Darmstadt,DEShaker RockingShakerRocky LabortechnikFröbelGmbH,Lindau,DEShaker OrbitalShakerOS-10 BioSan,Riga,LVSonicator Sonorex Bandelin,Berlin,DE
SpectrophotometerMicroplatereaderAnthos2010
anthosMikrosystemeGmbH,Krefeld,DE
Spectrophotometer VarianCary50Bio AgilentTechnologies,Inc.,SantaClara,CA,USASpeedVac EZ-2plus GeneVacInc,StoneRidge,NY,USAThermocycler TGradient BiometraGmbH,Göttingen,GermanyThermo-Shaker TS-100 BioSan,Riga,LVVortex Genie2 ScientificIndustries,Inc.,Bohemia,NY,USAWeighingscales SI-234A DenverInstrument,Bohemia,NY,USAWeighingscales XP105DeltaRange MettlerToledoGmbH,Gießen,DE
References
130
References
1. Paul,S.M.etal.HowtoimproveR&Dproductivity:thepharmaceuticalindustry’sgrandchallenge.Nat.Rev.DrugDiscov.9,203–214(2010).
2. Hay,M.,Thomas,D.W.,Craighead,J.L.,Economides,C.&Rosenthal,J.Clinicaldevelopmentsuccessratesforinvestigationaldrugs.Nat.Biotechnol.32,40–51(2014).
3. Raida,M.Drugtargetdeconvolutionbychemicalproteomics.Curr.Opin.Chem.Biol.15,570–5(2011).
4. Lee,J.&Bogyo,M.Targetdeconvolutiontechniquesinmodernphenotypicprofiling.Curr.Opin.Chem.
Biol.17,118–126(2013).
5. Bantscheff,M.,Scholten,A.&Heck,A.J.R.Revealingpromiscuousdrug-targetinteractionsbychemicalproteomics.DrugDiscov.Today14,1021–1029(2009).
6. Lenz,T.,Fischer,J.J.&Dreger,M.Probingsmallmolecule-proteininteractions:Anewperspectiveforfunctionalproteomics.J.Proteomics75,100–15(2011).
7. Ruddigkeit,L.,Blum,L.C.&Reymond,J.L.VisualizationandvirtualscreeningofthechemicaluniversedatabaseGDB-17.J.Chem.Inf.Model.53,56–65(2013).
8. Médard,G.etal.OptimizedChemicalProteomicsAssayforKinaseInhibitorProfiling.J.ProteomeRes.14,1574–1586(2015).
9. Daub, H. et al. Kinase-Selective Enrichment Enables Quantitative Phosphoproteomics of the KinomeacrosstheCellCycle.Mol.Cell31,438–448(2008).
10. Medvedev,A.,Kopylov,A.,Buneeva,O., Zgoda,V.&Archakov,A.Affinity-basedproteomicprofiling:Problemsandachievements.Proteomics12,621–637(2012).
11. Koster, H. et al. Capture compoundmass spectrometry: a technology for the investigation of smallmoleculeproteininteractions.AssayDrugDevTechnol5,381–390(2007).
12. Luo, Y. et al. The cAMP capture compoundmass spectrometry as a novel tool for targeting cAMP-binding proteins: from protein kinase A to potassium/sodium hyperpolarization-activated cyclicnucleotide-gatedchannels.Mol.Cell.Proteomics8,2843–56(2009).
13. Fischer,J.J.etal.Dasatinib,imatinibandstaurosporinecapturecompounds-Complementarytoolsforthe profiling of kinases by Capture CompoundMass Spectrometry (CCMS). J. Proteomics 75, 160–8(2011).
14. Fischer,J.J.etal.Comprehensiveidentificationofstaurosporine-bindingkinasesinthehepatocytecelllineHepG2usingCaptureCompoundMassSpectrometry(CCMS).J.ProteomeRes.9,806–17(2010).
15. Luo,Y.etal.GDP-capturecompound--anoveltoolfortheprofilingofGTPasesinpro-andeukaryotes
References
131
bycapturecompoundmassspectrometry(CCMS).J.Proteomics73,815–9(2010).
16. Lenz,T.etal.ProfilingofmethyltransferasesandotherS-adenosyl-L-homocysteine-bindingProteinsbyCaptureCompoundMassSpectrometry(CCMS).J.Vis.Exp.(2010).doi:10.3791/2264
17. Dalhoff, C. et al. Synthesis of S-adenosyl-L-homocysteine capture compounds for selectivephotoinducedisolationofmethyltransferases.Chembiochem11,256–65(2010).
18. Rotili, D.et al. A photoreactive small-molecule probe for 2-oxoglutarateoxygenases.Chem.Biol.18,642–654(2011).
19. Rix, U.et al. Chemical proteomic profiles of the BCR-ABL inhibitors imatinib, nilotinib, and dasatinibrevealnovelkinaseandnonkinasetargets.Blood110,4055–4063(2007).
20. Liu,Y.,Patricelli,M.P.&Cravatt,B.F.Activity-basedproteinprofiling:theserinehydrolases.Proc.Natl.Acad.Sci.U.S.A.96,14694–14699(1999).
21. Bantscheff,M. et al. Quantitative chemical proteomics revealsmechanisms of action of clinical ABLkinaseinhibitors.Nat.Biotechnol.25,1035–1044(2007).
22. Ovaa,H.etal.Chemistryinlivingcells:Detectionofactiveproteasomesbyatwo-steplabelingstrategy.Angew.Chemie-Int.Ed.42,3626–3629(2003).
23. Speers, A. E., Adam, G. C. & Cravatt, B. F. Activity-based protein profiling in vivo using a copper(i)-catalyzedazide-alkyne[3+2]cycloaddition.J.Am.Chem.Soc.125,4686–4687(2003).
24. Salisbury,C.M.&Cravatt,B.F.Optimizationofactivity-basedprobesforproteomicprofilingofhistonedeacetylasecomplexes.J.Am.Chem.Soc.130,2184–2194(2008).
25. McKay,C.S.&Finn,M.G.ClickChemistryinComplexMixtures:BioorthogonalBioconjugation.Chem.
Biol.21,1075–1101(2014).
26. Sletten,E.M.&Bertozzi,C.R.Bioorthogonalchemistry:Fishingforselectivityinaseaoffunctionality.Angew.Chemie-Int.Ed.48,6974–6998(2009).
27. Sletten,E.M.&Bertozzi,C.R.Frommechanismtomouse:ataleoftwobioorthogonalreactions.Acc.Chem.Res.44,666–76(2011).
28. Jewett, J.&Bertozzi,C.Cu-freeclickcycloaddition reactions inchemicalbiology.Chem.Soc.Rev.39,1272–1279(2010).
29. Patterson,D.M.,Nazarova,L.a&Prescher,J.a.Findingtheright(bioorthogonal)chemistry.ACSChem.
Biol.9,592–605(2014).
30. Kolb,H.C.,Finn,M.G.&Sharpless,K.B.ClickChemistry:DiverseChemicalFunctionfromaFewGoodReactions.Angew.Chem.Int.Ed.Engl.40,2004–2021(2001).
References
132
31. Kolb,H. C.& Sharpless, K. B. The growing impact of click chemistry on drug discovery.DrugDiscov.Today8,1128–37(2003).
32. Wu,P.etal.Efficiencyand fidelity inaclick-chemistry route to triazoledendrimersby thecopper(I)-catalyzedligationofazidesandalkynes.Angew.Chemie-Int.Ed.43,3928–3932(2004).
33. Saxon, E. & Bertozzi, C. R. Cell surface engineering by amodified Staudinger reaction. Science 287,2007–2010(2000).
34. Lin,F.L.,Hoyt,H.M.,vanHalbeek,H.,Bergman,R.G.&Bertozzi,C.R.Mechanisticinvestigationofthestaudingerligation.J.Am.Chem.Soc.127,2686–2695(2005).
35. Tam, A., Soellner, M. B. & Raines, R. T. Electronic and steric effects on the rate of the tracelessStaudingerligation.Org.Biomol.Chem.6,1173–1175(2008).
36. Tam,A.&Raines,R.T.Coulombiceffectson the tracelessStaudinger ligation inwater.Bioorg.Med.
Chem.17,1055–1063(2009).
37. Best,M.D.Click chemistry andbioorthogonal reactions: unprecedented selectivity in the labelingofbiologicalmolecules.Biochemistry48,6571–84(2009).
38. Michael, A. Ueber die Einwirkung von Diazobenzolimid auf Acetylendicarbonsauremethylester. J. furPrakt.Chemie48,94–95(1893).
39. Huisgen, R. Kinetics andMechanismof 1,3-Dipolar Cycloadditions.Angew. Chemie Int. Ed. English2,633–645(1963).
40. Finn,M.G.&Fokin,V.V.Clickchemistry:functionfollowsform.Chem.Soc.Rev.39,1231–2(2010).
41. Himo,F.etal.Copper(I)-catalyzedsynthesisofazoles.DFTstudypredictsunprecedentedreactivityandintermediates.J.Am.Chem.Soc.127,210–216(2005).
42. Worrell, B. T.,Malik, J. a.&Fokin,V.V.Direct evidenceof adinuclear copper intermediate inCu(I)-catalyzedazide-alkynecycloadditions.Science(80-.).340,457–460(2013).
43. Hein, J. E. & Fokin, V. V. Copper-catalyzed azide-alkyne cycloaddition (CuAAC) and beyond: newreactivityofcopper(I)acetylides.Chem.Soc.Rev.39,1302–15(2010).
44. Zhang,L.etal.Ruthenium-catalyzedcycloadditionofalkynesandorganicazides.J.Am.Chem.Soc.127,15998–9(2005).
45. Meldal, M. & Tornøe, C. W. Cu-catalyzed azide-alkyne cycloaddition. Chem. Rev. 108, 2952–3015(2008).
46. Rostovtsev,V.V.,Green,L.G.,Fokin,V.V.&Sharpless,K.B.Astepwisehuisgencycloadditionprocess:Copper(I)-catalyzedregioselective‘ligation’ofazidesandterminalalkynes.Angew.Chemie-Int.Ed.41,2596–2599(2002).
References
133
47. Chan, T. R., Hilgraf, R., Sharpless, K. B. & Fokin, V. V. Polytriazoles as copper(I)-stabilizing ligands in
catalysis.Org.Lett.6,2853–5(2004).
48. Wang,Q.etal.Bioconjugationbycopper(I)-catalyzedazide-alkyne[3+2]cycloaddition.J.Am.Chem.
Soc.125,3192–3(2003).
49. Hong,V.,Steinmetz,N.F.,Manchester,M.&Finn,M.G.Labelinglivecellsbycopper-catalyzedalkyne--azideclickchemistry.Bioconjug.Chem.21,1912–6(2010).
50. SorianoDelAmo,D.etal.Biocompatiblecopper(I)catalystsforinvivoimagingofglycans.J.Am.Chem.
Soc.132,16893–9(2010).
51. Agard,N. J., Prescher, J. a&Bertozzi, C. R. A strain-promoted [3 + 2] azide-alkyne cycloaddition forcovalentmodificationofbiomoleculesinlivingsystems.J.Am.Chem.Soc.126,15046–15047(2004).
52. Gaetke, L.M., Chow-Johnson, H. S. & Chow, C. K. Copper: toxicological relevance andmechanisms.Arch.Toxicol.88,1929–38(2014).
53. Sletten,E.M.&Bertozzi,C.R.Ahydrophilicazacyclooctyneforcu-freeclickchemistry.Org.Lett.10,3097–3099(2008).
54. Ning,X.,Guo,J.,Wolfert,M.a.&Boons,G.J.Visualizingmetabolicallylabeledglycoconjugatesoflivingcellsbycopper-freeandfasthuisgencycloadditions.Angew.Chemie-Int.Ed.47,2253–2255(2008).
55. Chang,P.Vetal.Copper-freeclickchemistryinlivinganimals.Proc.Natl.Acad.Sci.U.S.A.107,1821–6(2010).
56. Jewett, J. C., Sletten, E. M. & Bertozzi, C. R. Rapid Cu-free click chemistry with readily synthesizedbiarylazacyclooctynones.J.Am.Chem.Soc.132,3688–90(2010).
57. Koo, H. et al. Bioorthogonal copper-free click chemistry inVivo for tumor-targeted delivery ofnanoparticles.Angew.Chemie-Int.Ed.51,11836–11840(2012).
58. Evans,H.L.etal.Copper-freeclick—apromisingtoolforpre-targetedPETimaging.Chem.Commun.48,991(2012).
59. Van Berkel, S. S. et al. Metal-free triazole formation as a tool for bioconjugation. ChemBioChem 8,1504–1508(2007).
60. Blackman, M. L., Royzen, M. & Fox, J. M. Tetrazine ligation: fast bioconjugation based on inverse-electron-demandDiels-Alderreactivity.J.Am.Chem.Soc.130,13518–9(2008).
61. Devaraj, N. K., Weissleder, R. & Hilderbrand, S. a. Tetrazine-based cycloadditions: application topretargetedlivecellimaging.Bioconjug.Chem.19,2297–9(2008).
62. Kim,C.H.,Axup, J.Y.&Schultz,P.G.Proteinconjugationwithgeneticallyencodedunnaturalamino
References
134
acids.Curr.Opin.Chem.Biol.17,412–419(2013).
63. Shih,H.W.,Kamber,D.N.&Prescher,J.a.Buildingbetterbioorthogonalreactions.Curr.Opin.Chem.
Biol.21,103–111(2014).
64. Versteegen, R. M., Rossin, R., Ten Hoeve, W., Janssen, H. M. & Robillard, M. S. Click to release:Instantaneous doxorubicin elimination upon tetrazine ligation.Angew. Chemie - Int. Ed. 52, 14112–14116(2013).
65. Matikonda, S. S. et al. Bioorthogonal prodrug activation driven by a strain-promoted 1,3-dipolarcycloaddition.Chem.Sci.6,1212–1218(2015).
66. Kolodych, S.etal.Discoveryof chemoselectiveandbiocompatible reactionsusingahigh-throughputimmunoassayscreening.Angew.Chemie-Int.Ed.52,12056–12060(2013).
67. Nikolaev,V.O.&Lohse,M.J.MonitoringofcAMPsynthesisanddegradationinlivingcells.Physiology(Bethesda).21,86–92(2006).
68. Rall, T. W., Sutherland, E. W. & Berthet, J. The relationship of epinephrine and glucagon to liverphosphorylase.I.Liverphosphorylase;preparationandproperties.J.Biol.Chem.218,459–468(1956).
69. Sutherland, E.W.& Rall, T.W. Fractionation and characterization of a cyclic adenine ribonucleotideformedbytissueparticles.J.Biol.Chem.232,1077–1091(1958).
70. Rall,T.&Sutherland,E.Formationofacyclicadenineribonucleotidebytissueparticles.J.Biol.Chem.232,1065–1076(1958).
71. Cook,W. H., Lipkin, D.&Markham, R. The formation of a cyclic dianhydrodiadenylic acid (I) by thealkalinedegradationofadenosine-5’-triphosphoricacid(II).J.Am.Chem.Soc.79,3607–3608(1957).
72. Godinho, R. O., Duarte, T. & Pacini, E. S. a. New perspectives in signaling mediated by receptorscoupled to stimulatory G protein: the emerging significance of cAMP efflux and extracellular cAMP-adenosinepathway.Front.Pharmacol.06,1–9(2015).
73. Kamenetsky, M. et al. Molecular Details of cAMP Generation in Mammalian Cells: A Tale of TwoSystems.J.Mol.Biol.362,623–639(2006).
74. Jaiswal,B.S.&Conti,M.Calciumregulationof thesolubleadenylylcyclaseexpressed inmammalianspermatozoa.Proc.Natl.Acad.Sci.U.S.A.100,10676–10681(2003).
75. Litvin,T.N.,Kamenetsky,M.,Zarifyan,A.,Buck,J.&Levin,L.R.Kineticpropertiesof‘soluble’adenylylcyclase:Synergismbetweencalciumandbicarbonate.J.Biol.Chem.278,15922–15926(2003).
76. Antoni,F.a.NewparadigmsincAMPsignalling.Mol.Cell.Endocrinol.353,3–9(2012).
77. Steegborn, C. Structure, mechanism, and regulation of soluble adenylyl cyclases — similarities anddifferencestotransmembraneadenylylcyclases.Biochim.Biophys.Acta-Mol.BasisDis.1842,2535–
References
135
2547(2014).
78. Desman,G.,Waintraub,C.&Zippin,J.H.InvestigationofcAMPmicrodomainsasapathtonovelcancerdiagnostics.Biochim.Biophys.Acta-Mol.BasisDis.1842,2636–2645(2014).
79. Cheepala, S.etal. CyclicNucleotideCompartmentalization:ContributionsofPhosphodiesterases andATP-BindingCassetteTransporters.Annu.Rev.Pharmacol.Toxicol.53,231–253(2012).
80. Walsh,D.A.,Perkins,J.P.&Krebs,E.G.Anadenosine3’,5'-monophosphate-dependantproteinkinasefromrabbitskeletalmuscle.J.Biol.Chem.243,3763–3765(1968).
81. Taskén,K.&Aandahl,E.M.LocalizedeffectsofcAMPmediatedbydistinctroutesofproteinkinaseA.Physiol.Rev.84,137–167(2004).
82. Taylor,S.S.etal.SignalingthroughcAMPandcAMP-dependentproteinkinase:Diversestrategiesfordrugdesign.Biochim.Biophys.Acta-ProteinsProteomics1784,16–26(2008).
83. Edwards,A.S.&Scott,J.D.A-kinaseanchoringproteins:ProteinkinaseAandbeyond.Curr.Opin.CellBiol.12,217–221(2000).
84. Burke,D.,Howells,J.&Tomlinson,S.E.Hcnchannels:Functionandclinicalimplications.Neurology81,513–514(2013).
85. Podda,M. V.&Grassi, C. New perspectives in cyclic nucleotide-mediated functions in the CNS: Theemergingroleofcyclicnucleotide-gated(CNG)channels.PflugersArch.Eur.J.Physiol.466,1241–1257(2014).
86. Almahariq,M.,Mei,F.C.&Cheng,X.CyclicAMPsensorEPACproteinsandenergyhomeostasis.TrendsEndocrinol.Metab.25,60–71(2014).
87. Bos,J.L.Epacproteins:multi-purposecAMPtargets.TrendsBiochem.Sci.31,680–686(2006).
88. Scholten,A.etal.Analysisof thecGMP/cAMP interactomeusinga chemicalproteomicsapproach inmammalianhearttissuevalidatessphingosinekinasetype1-interactingproteinasagenuineandhighlyabundantAKAP.J.ProteomeRes.5,1435–1447(2006).
89. Aye,T.T.etal.SelectivityinenrichmentofcAMP-dependentproteinkinaseregulatorysubunitstypeIand type II and their interactors usingmodified cAMP affinity resins.Mol. Cell. Proteomics8, 1016–1028(2009).
90. Scholten,A.,VanVeen,T.aB.,Vos,M.a.&Heck,A.J.R.DiversityofcAMP-dependentproteinkinaseisoforms and their anchoring proteins in mouse ventricular tissue. J. Proteome Res. 6, 1705–1717(2007).
91. Kovanich,D.etal.Sphingosinekinase interactingprotein isanA-kinaseanchoringproteinspecificfortypeIcAMP-dependentproteinkinase.ChemBioChem11,963–971(2010).
References
136
92. Shi, H., Zhang, C.-J., Chen, G. Y. J. & Yao, S. Q. Cell-based proteome profiling of potential dasatinibtargetsbyuseofaffinity-basedprobes.J.Am.Chem.Soc.134,3001–14(2012).
93. Shah,N.P.etal.OverridingimatinibresistancewithanovelABLkinaseinhibitor.Science305,399–401(2004).
94. Apperley,J.F.Chronicmyeloidleukaemia.Lancet385,1447–1459(2014).
95. Greuber,E.K.,Smith-Pearson,P.,Wang, J.&Pendergast,A.M.RoleofABL familykinases incancer:fromleukaemiatosolidtumours.Nat.Rev.Cancer13,559–71(2013).
96. Wang,J.Y.J.ThecapableABL:whatisitsbiologicalfunction?Mol.Cell.Biol.34,1188–97(2014).
97. Shaul, Y.&Ben-Yehoyada,M.Roleof c-Abl in theDNAdamage stress response.Cell Res.15, 33–35(2005).
98. Wen, S. T. & Van Etten, R. a. The PAG gene product, a stress-induced protein with antioxidantproperties,isanAblSH3-bindingproteinandaphysiologicalinhibitorofc-Abltyrosinekinaseactivity.GenesDev.11,2456–2467(1997).
99. Lin, J.etal.Oncogenicactivationofc-Abl innon-small cell lungcancercells lackingFUS1expression:inhibitionofc-AblbythetumorsuppressorgeneproductFus1.Oncogene26,6989–6996(2007).
100. Cong,F.etal.CytoskeletalproteinPSTPIP1directsthePEST-typeproteintyrosinephosphatasetothec-AblkinasetomediateAbldephosphorylation.Mol.Cell6,1413–1423(2000).
101. Das,J.etal.2-Aminothiazoleasanovelkinaseinhibitortemplate.Structure-activityrelationshipstudiestoward the discovery of N-(2-chloro-6-methylphenyl)-2-[[6- [4-(2-hydroxyethyl)-1-piperazinyl]-2-methyl-4-pyrimidinyl]amino]-1, 3-thiazole-5-carboxamide (Dasatini. J. Med. Chem. 49, 6819–6832(2006).
102. Golas,J.M.etal.SKI-606,a4-anilino-3-quinolinecarbonitriledualinhibitorofSrcandAblkinases,isapotent antiproliferative agent against chronic myelogenous leukemia cells in culture and causesregressionofK562xenograftsinnudemice.CancerRes.63,375–381(2003).
103. Buchdunger, E. et al. Inhibition of the Abl protein-tyrosine kinase in vitro and in vivo by a 2-phenylaminopyrimidinederivative.CancerRes.56,100–104(1996).
104. Weisberg, E. et al. Characterization of AMN107, a selective inhibitor of native and mutant Bcr-Abl.CancerCell7,129–141(2005).
105. O’Hare,T.etal.AP24534,aPan-BCR-ABLInhibitorforChronicMyeloidLeukemia,PotentlyInhibitstheT315IMutantandOvercomesMutation-BasedResistance.CancerCell16,401–412(2009).
106. Adrián,F. J.etal.Allosteric inhibitorsofBcr-abl-dependentcellproliferation.Nat.Chem.Biol.2,95–102(2006).
References
137
107. Zhang, J.etal.TargetingBcr-AblbycombiningallostericwithATP-binding-site inhibitors.Nature463,501–506(2010).
108. Davis,M. I.etal.Comprehensiveanalysisofkinase inhibitorselectivity.Nat.Biotechnol.29,1046–51(2011).
109. Lombardo, L. J. et al. Discovery of N-(2-chloro-6-methylphenyl)-2-(6-(4-(2-hydroxyethyl)- piperazin-1-yl)-2-methylpyrimidin-4-ylamino)thiazole-5-carboxamide(BMS-354825),adualSrc/Ablkinaseinhibitorwithpotentantitumoractivityinpreclinicalassays.J.Med.Chem.47,6658–6661(2004).
110. Harrington, E. aetal. VX-680, apotent and selective small-molecule inhibitorof theAurora kinases,suppressestumorgrowthinvivo.Nat.Med.10,262–267(2004).
111. Moasser,M.M.,Srethapakdi,M.,Sachar,K.S.,Kraker,A.J.&Rosen,N.InhibitionofSrckinasesbyaselectivetyrosinekinaseinhibitorcausesmitoticarrest.CancerRes.59,6145–6152(1999).
112. Akeno-Stuart, N. et al. The RET kinase inhibitor NVP-AST487 blocks growth and calcitonin geneexpressionthroughdistinctmechanismsinmedullarythyroidcancercells.CancerRes.67,6956–6964(2007).
113. Vandyke,K.,Fitter,S.&Zannettino,aC.W.Thetyrosinekinaseinhibitordasatinib(SPRYCEL)inhibitschondrocyteactivityandproliferation.BloodCancerJ.1,e2(2011).
114. Chen,B.-A.&Chen,R.Theroleofdasatinibinthemanagementofchronicmyeloidleukemia.DrugDes.Devel.Ther.9,773–779(2015).
115. Fabarius,A.etal.CentrosomeaberrationsandG1phasearrestafterinvitroandinvivotreatmentwiththeSRC/ABLinhibitordasatinib.Haematologica93,1145–1154(2008).
116. Bevilacqua,V.etal.Copper-chelatingazidesforefficientclickconjugationreactionsincomplexmedia.Angew.Chemie-Int.Ed.53,5872–5876(2014).
117. Li,J.etal.Achemicalandphosphoproteomiccharacterizationofdasatinibactioninlungcancer.Nat.Chem.Biol.6,291–299(2010).
118. Kitagawa,D.etal.Activity-basedkinaseprofilingofapprovedtyrosinekinaseinhibitors.GenestoCells18,110–122(2013).
119. Szklarczyk,D.etal.STRINGv10:protein-proteininteractionnetworks,integratedoverthetreeoflife.NucleicAcidsRes.43,D447–52(2015).
120. Cox,J.etal.MaxLFQallowsaccurateproteome-widelabel-freequantificationbydelayednormalizationandmaximalpeptideratioextraction.(2014).
121. Chong, Y.-P.,Mulhern, T. D. & Cheng, H.-C. C-terminal Src kinase (CSK) and CSK-homologous kinase(CHK)--endogenous negative regulators of Src-family protein kinases. Growth Factors 23, 233–244(2005).
References
138
122. Kim,D.-J.,Yi,Y.-W.&Kim,J.H.Insitumonitoringofbindingsbetweendasatinibanditstargetprotein
kinasesusingmagneticnanoparticlesinlivecells.J.Am.Chem.Soc.130,16466–16467(2008).
123. Kisselev,L.L.Mammaliantryptophanyl-tRNAsynthetases.Biochimie75,1027–1039(1993).
124. Tzima,E.&Schimmel,P.InhibitionoftumorangiogenesisbyanaturalfragmentofatRNAsynthetase.TrendsBiochem.Sci.31,7–10(2006).
125. Corkery,B.M.TyrosineKinaseInhibitorsinTripleNegativeBreastCancer.(2010).
126. Trinka, E., Höfler, J., Leitinger,M. & Brigo, F. Pharmacotherapy for Status Epilepticus.Drugs (2015).doi:10.1007/s40265-015-0454-2
127. Whysner,J.,Ross,P.M.&Williams,G.M.Phenobarbitalmechanisticdataandriskassessment:enzymeinduction,enhancedcellproliferation,andtumorpromotion.Pharmacol.Ther.71,153–91(1996).
128. Larsen,M.&Jefcoate,C.PhenobarbitalInductionofCYP2B1,CYP2B2,andCYP3A1inRatLiver:GeneticDifferencesinaCommonRegulatoryMechanism.Arch.Biochem.Biophys.321,467–476(1995).
129. Waxman,D.J.&Azaroff,L.PhenobarbitalinductionofcytochromeP-450geneexpression.Biochem.J.281,577–592(1992).
130. Gerets,H.H.J.etal.Characterizationofprimaryhumanhepatocytes,HepG2cells,andHepaRGcellsatthemRNA level and CYP activity in response to inducers and their predictivity for the detection ofhumanhepatotoxins.CellBiol.Toxicol.28,69–87(2012).
131. Lee, G. H. Paradoxical effects of phenobarbital onmouse hepatocarcinogenesis. Toxicol. Pathol. 28,215–225(2000).
132. Kawamoto,T.etal.Phenobarbital-responsivenucleartranslocationofthereceptorCARininductionoftheCYP2Bgene.Mol.Cell.Biol.19,6318–22(1999).
133. Marc,N.etal.RegulationofphenobarbitalinductionofthecytochromeP450269/10genesinprimarymousehepatocyteculture:Involvementofcalcium-andcAMP-dependentpathways.Eur.J.Biochem.267,963–970(2000).
134. Galisteo,M.etal.Involvementofcyclicnucleotide-andcalcium-regulatedpathwaysinphenobarbital-inducedcytochromeP-4503Aexpressioninmouseprimaryhepatocytes.J.Pharmacol.Exp.Ther.290,1270–1277(1999).
135. Rencurel, F. et al. AMP-activated protein kinase mediates phenobarbital induction of CYP2B geneexpression inhepatocytes andanewlyderivedhumanhepatoma cell line. J. Biol. Chem.280, 4367–4373(2005).
136. Rencurel, F.et al. Stimulation of AMP-activated protein kinase is essential for the induction of drugmetabolizing enzymes by phenobarbital in human andmouse liver.Mol. Pharmacol.70, 1925–1934
References
139
(2006).
137. Harach,T.Crossroadsbetweendrugandenergymetabolism :RoleofconstitutiveandrostanereceptorandAMP-activatedkinase.(2010).
138. Blättler, S.M., Rencurel, F., Kaufmann,M. R. &Meyer, U. a. In the regulation of cytochrome P450genes,phenobarbitaltargetsLKB1fornecessaryactivationofAMP-activatedproteinkinase.Proc.Natl.Acad.Sci.U.S.A.104,1045–1050(2007).
139. Joannard, F., Galisteo, M., Corcos, L., Guillouzo, a & Lagadic-Gossmann, D. Regulation ofphenobarbital-inductionofCYP2BandCYP3Agenesinratculturedhepatocytes:involvementofseveralserine/threonineproteinkinasesandphosphatases.CellBiol.Toxicol.16,325–37(2000).
140. Kobayashi,K.,Hashimoto,M.,Honkakoski,P.&Negishi,M.RegulationofgeneexpressionbyCAR:anupdate.Arch.Toxicol.89,1045–1055(2015).
141. Shindo, S., Numazawa, S. & Yoshida, T. A physiological role of AMP-activated protein kinase inphenobarbital-mediatedconstitutiveandrostanereceptoractivationandCYP2Binduction.Biochem.J.401,735–741(2007).
142. Mutoh,S.etal.Phenobarbitalindirectlyactivatestheconstitutiveactiveandrostanereceptor(CAR)byinhibitionofepidermalgrowthfactorreceptorsignaling.Sci.Signal.6,ra31(2013).
143. Meyer,S.a,Gibbs,T.A.&Jirtle,R.L.Independentmechanismsfortumorpromotersphenobarbitaland12-O-tetradecanoylphorbol-13-acetate in reduction of epidermal growth factor binding by rathepatocytes.CancerRes.49,5907–5912(1989).
144. Meyer,S.a& Jirtle,R. L.Olddancewithanewpartner:EGF receptoras thephenobarbital receptormediatingCyp2Bexpression.Sci.Signal.6,pe16(2013).
145. Joseph, B. K. et al. Inhibition of AMP Kinase by the Protein Phosphatase 2A Heterotrimer,PP2APpp2r2d.J.Biol.Chem.290,10588–98(2015).
146. Ferro,M.&Bassi,A.Inductionofcytochrome(s)P450-dependentdrugmetabolisminculturedMH1C1hepatomacells.CellBiochem.Funct.2,263–268(1984).
147. Donato,M.T.etal.Evaluationofthexenobioticbiotransformationcapabilityofsixrodenthepatomacelllinesincomparisonwithrathepatocytes.InVitroCell.Dev.Biol.Anim.30A,574–80(1994).
148. Lempiäinen, H. et al. Phenobarbital mediates an epigenetic switch at the constitutive androstanereceptor(CAR)targetgeneCyp2b10intheliverofB6C3F1mice.PLoSOne6,e18216(2011).
149. Phillips, J. M. & Goodman, J. I. Multiple Genes Exhibit Phenobarbital-Induced ConstitutiveActive/Androstane Receptor–Mediated DNAMethylation Changes during Liver Tumorigenesis and inLiverTumors.Toxicol.Sci.anOff.J.Soc.Toxicol.108,273–289(2009).
150. Swales,K.&Negishi,M.CAR,drivingintothefuture.Mol.Endocrinol.18,1589–1598(2004).
References
140
151. Han,Y.,Wang,Q.,Song,P.,Zhu,Y.&Zou,M.-H.RedoxregulationoftheAMP-activatedproteinkinase.
PLoSOne5,e15420(2010).
152. Ching,J.K.,Rajguru,P.,Marupudi,N.,Banerjee,S.&Fisher,J.S.AroleforAMPKin increasedinsulinactionafterserumstarvation.Am.J.Physiol.CellPhysiol.299,C1171–9(2010).
153. Conney,aH.Pharmacologicalimplicationsofmicrosomalenzymeinduction.Pharmacol.Rev.19,317–366(1967).
154. Okey,A.B.EnzymeinductioninthecytochromeP-450system.Pharmacol.Ther.45,241–298(1990).
155. Giuditta, A. & Casola, L. Mechanism of barbiturate inhibition on purified flavoenzymes. Biochim.
Biophys.Acta(BBA)-Enzymology…110,17–31(1965).
156. Li, D. & Roberts, R. Human Genome and Diseases: WD-repeat proteins: structure characteristics,biologicalfunction,andtheirinvolvementinhumandiseases.Cell.Mol.lifeSci.58,2085–2097(2001).
157. Kim, J., Min, G. & Kemper, B. Chromatin Assembly Enhances Binding to the CYP2B1 Phenobarbital-responsiveUnit(PBRU)ofNuclearFactor-1,whichBindsSimultaneouslywithConstitutiveAndrostaneReceptor (CAR)/Retinoid X Receptor (RXR) and Enhances CAR/RXR-mediated Activation of the PB. J.Biol.Chem.276,7559–7567(2001).
158. Luc,P.V,Adesnik,M.,Ganguly,S.&Shaw,P.M.TranscriptionalregulationoftheCYP2B1andCYP2B2genesbyC/EBP-relatedproteins.BiochemPharmacol51,345–356(1996).
159. Honkakoski, P. & Negishi, M. Characterization of a phenobarbital-responsive enhancer module inmouseP450Cyp2b10gene.J.Biol.Chem.272,14943–9(1997).
160. Dolfini,D.&Mantovani,R.TargetingtheY/CCAATboxincancer:YB-1(YBX1)orNF-Y?CellDeathDiffer.20,676–85(2013).
161. Eliseeva,I.a,Kim,E.R.,Guryanov,S.G.,Ovchinnikov,L.P.&Lyabin,D.N.Y-Box-BindingProtein1andItsFunctions.Biochemistry76,1402–1433(2011).
162. Wu,J.DisruptionoftheY-BoxBindingProtein-1ResultsinSuppressionoftheEpidermalGrowthFactorReceptorandHER-2.CancerRes.66,4872–4879(2006).
163. Ren,H.etal.Evidencefortheinvolvementofxenobiotic-responsivenuclearreceptorsintranscriptionaleffectsuponperfluoroalkylacidexposureindiversespecies.Reprod.Toxicol.27,266–277(2009).
164. Phillips, J.M.,Burgoon,L.D.&Goodman, J. I.Theconstitutiveactive/androstane receptor facilitatesuniquephenobarbital-inducedexpressionchangesofgenesinvolvedinkeypathwaysinprecancerousliverandlivertumors.Toxicol.Sci.110,319–333(2009).
165. Phillips, J. M., Burgoon, L. D. & Goodman, J. I. Phenobarbital elicits unique, early changes in theexpressionofhepaticgenesthataffectcriticalpathwaysintumor-proneB6C3F1mice.Toxicol.Sci.109,
References
141
193–205(2009).
166. Zhou,S.,Sun,X.,Liu,L.,Wang,X.&Liu,K.Increasedexpressionofaquaporin-1intheanteriortemporalneocortexofpatientswithintractableepilepsy.Neurol.Res.30,400–405(2008).
167. Binder,D.K.,Nagelhus,E.A.&Ottersen,O.P.Aquaporin-4andepilepsy.Glia60,1203–1214(2012).
168. Dudek,F.E.&Rogawski,M.A.RegulationofBrainWater:IsThereaRoleforAquaporinsinEpilepsy?EpilepsyCurr.5,104–106(2005).
169. Higgins, L. G. & Hayes, J. D. Mechanisms of induction of cytosolic and microsomal glutathionetransferase (GST) genes by xenobiotics and pro-inflammatory agents.DrugMetab. Rev. 43, 92–137(2011).
170. Foerster, S. et al. Characterization of the EGFR interactome reveals associated protein complexnetworksandintracellularreceptordynamics.Proteomics13,3131–3144(2013).
171. Wei,B.,Guo,C.,Liu,S.&Sun,M.-Z.AnnexinA4andcancer.Clin.Chim.Acta.447,72–78(2015).
172. Wos,M.&Bandorowicz-Pikula, J. [Participationofannexins inendocytosisandEGFR-mediatedsignaltransduction].PostepyBiochem.60,55–61(2014).
173. Grewal,T.&Enrich,C.Annexins--modulatorsofEGFreceptorsignallingandtrafficking.Cell.Signal.21,847–858(2009).
174. Kahn,R.A.etal.Arf familyGTPases: roles inmembrane traffic andmicrotubuledynamics.Biochem.
Soc.Trans.33,1269–1272(2005).
175. Sukhanova, A. et al. Targeting C4-demethylating genes in the cholesterol pathway sensitizes cancercells to EGF receptor inhibitors via increased EGF receptor degradation. Cancer Discov. 3, 96–111(2013).
176. Cao, H., Chen, J., Krueger, E. W. &McNiven, M. A. SRC-mediated phosphorylation of dynamin andcortactinregulatesthe‘constitutive’endocytosisoftransferrin.Mol.Cell.Biol.30,781–792(2010).
177. Sousa,L.P.etal. SuppressionofEGFRendocytosisbydynamindepletionreveals thatEGFRsignalingoccursprimarilyattheplasmamembrane.Proc.Natl.Acad.Sci.109,4419–4424(2012).
178. Tong,J.,Taylor,P.&Moran,M.F.ProteomicAnalysisoftheEpidermalGrowthFactorReceptor(EGFR)InteractomeandPost-translationalModificationsAssociatedwithReceptorEndocytosisinResponsetoEGFandStress.Mol.Cell.Proteomics13,1644–1658(2014).
179. Uhlen,M.etal.Proteomics.Tissue-basedmapofthehumanproteome.Science347,1260419(2015).
180. Abou-Zeid,N.etal.MICAL-like1mediatesepidermalgrowthfactorreceptorendocytosis.Mol.Biol.Cell22,3431–3441(2011).
References
142
181. Zhang,L.,Gjoerup,O.&Roberts,T.M.Theserine/threoninekinasecyclinG-associatedkinaseregulatesepidermalgrowthfactorreceptorsignaling.Proc.Natl.Acad.Sci.U.S.A.101,10296–301(2004).
182. Culi,J.,Springer,T.A.&Mann,R.S.Boca-dependentmaturationofbeta-propeller/EGFmodulesinlow-densitylipoproteinreceptorproteins.EMBOJ.23,1372–1380(2004).
183. Cobb,B.D.&Clarkson, J.M.Asimpleprocedure foroptimising thepolymerasechain reaction (PCR)usingmodifiedTaguchimethods.NucleicAcidsRes.22,3801–3805(1994).
184. Smith,P.K.etal.Measurementofproteinusingbicinchoninicacid.Anal.Biochem.150,76–85(1985).
185. Maizel,J.V.Evidenceformultiplecomponentsinthestructuralproteinoftype1poliovirus.Biochem.
Biophys.Res.Commun.13,483–489(1963).
186. Raymond,S.&Weintraub,L.Acrylamidegelasasupportingmediumforzoneelectrophoresis.Science130,711(1959).
187. Laemmli,U.K.Cleavageof structuralproteinsduring theassemblyof theheadofbacteriophageT4.Nature227,680–5(1970).
188. Maizel,Jr,J.V.SDSpolyacrylamidegelelectrophoresis.TrendsBiochem.Sci.25,590–592(2000).
189. Towbin,H.,Staehelin,T.&Gordon,J.Electrophoretictransferofproteinsfrompolyacrylamidegelstonitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. U. S. A. 76, 4350–4(1979).
190. Burnette,W.N.‘Westernblotting’:electrophoretictransferofproteinsfromsodiumdodecylsulfate--polyacrylamide gels to unmodified nitrocellulose and radiographic detection with antibody andradioiodinatedproteinA.Anal.Biochem.112,195–203(1981).
143
Curriculumvitae
1 / 2
Forreasonsofdataprotection,theCurriculumvitaeisnotpublishedintheonlineversion.
2 / 2
Forreasonsofdataprotection,theCurriculumvitaeisnotpublishedintheonlineversion.