novel inhibitors for the protease taspase1 - core.ac.uk · m molar, mol/l maldi matrix‐assisted...

NovelinhibitorsfortheproteaseTaspase1

Inaugural‐DissertationzurErlangungdesDoktorgradesDr.rer.nat.

FakultätfürBiologiederUniversitätDuisburg‐Essen

CampusEssen

vorgelegtvonJohannesvandenBoom

ausEssen

Juni2014

2

DiedervorliegendenArbeitzuGrundeliegendenExperimentewurdeninderAbteilung

StrukturelleundMedizinischeBiochemieamZentrumfürMedizinischeBiotechnologie

derUniversitätDuisburg‐Essendurchgeführt.

1.Gutachter:Prof.Dr.PeterBayer

2.Gutachter:Prof.Dr.ShirleyKnauer

3.Gutachter:Prof.Dr.MichaelEhrmann

VorsitzenderdesPrüfungsausschusses:Prof.Dr.MarkusKaiser

TagdermündlichenPrüfung:09.09.2014

3

WIRKLICHES NEULAND IN EINER WISSENSCHAFT

KANN WOHL NUR GEWONNEN WERDEN,

WENN MAN AN EINER ENTSCHEIDENDEN STELLE

BEREIT IST, DEN GRUND ZU VERLASSEN,

AUF DEM DIE BISHERIGE WISSENSCHAFT RUHT,

UND GEWISSERMASSEN INS LEERE ZU SPRINGEN.

WERNER HEISENBERG (1901-1976)

4

FürmeineEltern,

diemiralldiesermöglichthaben

5

ListofabbreviationsÅ ÅngströmALL acutelymphaticleukemiaAML acutemyeloicleukemiaAmsil amorphoussilicananoparticleAPS ammoniumperoxydisulfateAU arbitraryunitsBioTasp artificialTaspaseactivitysensorforcellularassaysCD circulardichroismCLL chroniclymphaticleukemiaCML chronicmyeloicleukemiaCS1,CS2 Taspasecleavagesites1and2intheMLLprotein,respectivelyC‐terminus carboxy‐terminus/COOH‐terminusDa DaltonDMSO dimethylsulfoxideDOL degreeoflabelingDTT dithiothreitol,reducingagentE.coli Escherichiacoli,bacteriumEC50 effectiveconcentration,necessaryfor50%effectESI electronsprayionizationFmoc fluorenylmethyloxycarbonyl,aprotectiongroupFPLC fastproteinliquidchromatographyFRET FörsterresonanceenergytransferG,A,T,C DNAbases:guanine,adenine,thymine,cytosineGFP greenfluorescentproteinGST glutathioneS‐transferaseHela cervicalcancercelllineHox homeoboxHSQC heteronuclearsinglequantumcoherenceIC50 inhibitoryconcentration,necessaryfor50%inhibitionIPTG isopropyl‐thiogalactopyranosideKd dissociationconstantKi inhibitoryconstantKm MichaelisconstantKPi potassiumphosphateLC liquidchromatographyLDH lactatedehydrogenaseM molar,mol/lMALDI matrix‐assistedlaserdesorption/ionizationMD moleculardynamicsMLL mixedlineageleukemiaMS massspectrometryMST microscalethermophoresisNADH nicotineamideadeninedinucleotideNES nuclearexportsignalNiNTA nickelnitrilotriaceticacid,bindshexahistidine‐taggedproteinsNLS nuclearlocalizationsignalNMR nuclearmagneticresonanceN‐terminus amino‐terminus/NH2‐terminusNtn N‐terminalnucleophileOD600 opticaldensityat600nmwavelength

6

PCR polymerasechainreactionPDB ProteinDataBankPMT photomultipliertube,alightdetectorRBS ribosomalbindingsiteRMSD rootmeansquaredistancerpm roundsperminuteSDS‐PAGE sodiumdodecylsulfatepolyacrylamidegelelectrophoresisTaspase threonineaspartase1/Taspase1/TASP1Tm meltingtemperatureUV ultravioletwt wildtypeAminoacidsA Ala alanineC Cys cysteineD Asp aspartateE Glu glutamateF Phe phenylalanineG Gly glycineH His histidineI Ile isoleucineK Lys lysineL Leu leucineM Met methionineN Asn asparagineP Pro prolineQ Gln glutamineR Arg arginineS Ser serineT Thr threonineV Val valineW Trp tryptophanY Tyr tyrosine

7

ListoffiguresFigure1.1Hallmarksofcancer. 14 Figure1.2DomainsoftheMLLprotein. 17 Figure1.3MLLfusionsproteins. 18 Figure1.4ProcessingofTaspaseandMLL. 19 Figure1.5Substratebindingandproteasefamilies. 20 Figure1.6Taspasedisplaysaconservedasparaginasefold. 21 Figure1.7AutocatalyticprocessingofTaspase. 22 Figure1.8AutocatalyticactivationofTaspaseandothertypeIIasparaginases. 22 Figure1.9ProposedTaspasecleavagemechanism. 23 Figure1.10Taspaseaddictionintumorcells. 25 Figure1.11StrategiesfornewTaspaseinhibitors. 26 Figure2.1VectormapofthemodifiedpET‐41bvector. 37 Figure2.2ProteinandDNAmarker. 38 Figure2.3ConstructforactiveTaspase. 39 Figure2.4PrincipleofthefluorogenicTaspaseactivityassay. 51 Figure2.5SchematicspectraandinformationforNMRbackboneassignment. 60 Figure3.1AgarosegelafterPCRamplificationoftheTaspaseloop. 66 Figure3.2AgarosegelaftercolonyPCRforverificationofpositiveclones. 67 Figure3.3SDS‐PAGEgelsofexpressiontestforwtTaspase. 67 Figure3.4NiNTAaffinitypurificationofwildtypeTaspase. 68 Figure3.5PreparativegelfiltrationofwildtypeTaspase. 69 Figure3.6PurificationofactiveTaspasebyaffinitychromatographyandgelfiltration. 70 Figure3.7Far‐UVCDspectraofTaspase. 71 Figure3.8MeltingcurvesofTaspase. 72 Figure3.9MALDI‐TOFmassspectraofTaspase. 73 Figure3.10AnalyticalgelfiltrationofTaspase. 74 Figure3.11AutocatalyticprocessingofTaspase. 74 Figure3.121H‐15NHSQCspectrumassignment. 75 Figure3.13SecondaryandtertiarystructurepredictionoftheTaspaseloop. 77 Figure3.14PurificationoftheTaspaseloop. 78 Figure3.15SpectroscopicanalysisandMDsimulationoftheTaspaseloop. 79 Figure3.16Taspaseactivityassay. 80 Figure3.17LossofTaspaseactivityovertime. 80 Figure3.18ActivityoftheTaspasevariants. 81 Figure3.19SubstratespecificityofTaspase. 82 Figure3.20SpecificactivityofeukaryoticTaspase. 83 Figure3.21InhibitiontestwiththereportedTaspaseinhibitorNSC48300. 83 Figure3.22PreparationofTaspasemodelsforvirtualdocking. 84 Figure3.23Energydistributionofthedockedcompounds. 85 Figure3.24Dockedconformationandinhibitiontestofputativeproenzymeinhibitors. 87 Figure3.25Sharedfeaturesamongtopcandidates. 87 Figure3.26DockedconformationandinhibitionofputativeactiveTaspaseinhibitors. 88 Figure3.27NanoparticlesbindTaspase. 89 Figure3.28NanoparticlesinhibitTaspaseactivitynoncompetitively. 90 Figure3.29Nanoparticlesdonotaltersecondarystructureorstability. 91 Figure3.30LDH,chymotrypsinandproteinaseKareweaklyinhibitedbynanoparticles. 92 Figure3.31NanoparticlesinhibitTaspaseactivityincelllysates. 93 Figure3.32CellularTaspaseactivityassay. 94 Figure3.33NanoparticlesinhibitTaspaseactivityincells. 95 Figure3.34Designofthepeptidyl‐succinimidylpeptide. 96 Figure3.35Isoaspartateintermediateduringproteolysis. 98 Figure3.36Bindingofthepeptidyl‐succinimidylpeptideinhibitors. 98

8

Figure3.37Stabilityofcompetitivepeptidyl‐succinimidylpeptides. 99 Figure3.38Photolabelingandlysateinhibitionwithpeptidyl‐succinimidepeptides. 100 Figure3.39ThepeptidicinhibitorsenterthecellbutexhibitnoTaspaseinhibition. 101 Figure4.1InterfaceofthetwoTaspasesubunits. 106 Figure4.2Mappingofthe1H‐15NHSQCsignals. 107 Figure4.3ProposedstructureoftheTaspaseloop. 108 Figure4.4AsparaginasetypeIIloopmodels. 109 Figure4.5Interactionsofproteinswithnanoparticles. 112 Figure4.6Structuralchangesofproteinsboundonnanoparticles. 113 Figure4.7True‐to‐scalemodelofTaspaseboundtonanoparticles. 114 Figure4.8Proposedsubstrateorientation. 118 Figure5.1Substitutionsforamidebonds. 122 Figure7.1Mapsofbacterialexpressionvectors. 129 Figure7.2Mapsofeukaryoticvectorsusedfortransfection. 130 Figure7.3SchematicforcloningoftheTaspaseloop. 131 Figure7.4PurificationoftheinactiveTaspasemutant. 132 Figure7.5PurificationofthefirstelutionpeakafterNiNTAaffinitychromatography. 132 Figure7.6SDS‐PAGEgelsofexpressiontestforactiveTaspase. 133 Figure7.7CalibrationplotforSuperdex20010/300. 134 Figure7.8AnalyticalgelfiltrationofTaspasewith100mMNaCl. 134 Figure7.9TimedependenceoftheTaspaseNMRspectra. 136 Figure7.10ComparisonofwildtypeandmutantTaspase. 136 Figure7.11ExpressiontestfortheTaspaseloop. 137 Figure7.12InhibitionofTaspaseactivitybyNaCl. 137 Figure7.13WesternblotofTaspaseincelllysates. 138 Figure7.14ESI‐LC‐MSoftheTaspaseinhibitorNSC48300. 139 Figure7.15Modelqualityofthefull‐lengthTaspasehomologymodel. 139 Figure7.16Nanoparticlesbindparvalbuminwithlowaffinity. 140 Figure7.17Nanoparticlesdonotaltercellmorphology. 140 Figure7.18Inhibitioncurvesofthepeptidyl‐succinimidepeptides. 141 Figure7.19Taspasecanpassa50kDamembrane. 142 Figure7.20LocationoftheC‐terminiinthehetero‐tetramerofTaspase. 142 Figure7.21MolecularplugsfailtoinhibitTaspase. 143 Figure7.22Structureofthemolecularplugcompounds. 145

9

ListoftablesTable1.1SelectionofMLLfusionpartners. 16 Table2.1Chemicalsused. 30 Table2.2Peptidesused. 32 Table2.3Buffersandmediaused. 32 Table2.4Kitsused. 33 Table2.5Instrumentsanddevicesused. 34 Table2.6Consumablesused. 35 Table2.7Softwareused. 35 Table2.8Enzymesandantibodiesused. 36 Table2.9Bacterialstrainsused. 36 Table2.10PrimersusedtogeneratetheinactiveTaspasemutant. 39 Table2.11PCRcomponentstogeneratetheinactiveTaspasemutant. 39 Table2.12PCRprogramtogeneratetheinactiveTaspasemutant. 39 Table2.13PCRprimersusedtogeneratetheTaspaseloop. 40 Table2.14ApaI/XhoIrestrictionassay. 40 Table2.15LigationoftheTaspaseloopandpET‐41‐PreSc. 41 Table2.16PCRcomponentstoamplifytheTaspaseloop. 41 Table2.17PCRprogramusedtoamplifytheTaspaseloop. 41 Table2.18ColonyPCRcomponentstoamplifytheTaspaseloop. 42 Table2.19Channelsettingsusedforfluorescencemicroscopy. 45 Table2.20ProgramusedforNiNTAaffinitypurificationwithaHisTrapHP5mlNiNTA

column. 46 Table2.21ProgramusedforgelfiltrationwithaSuperdex20016/600gelfiltration

column. 46 Table2.22ProgramusedforGSTaffinitypurificationwitha20mlGSTrapcolumn. 47 Table2.23CalculatedparametersofTaspaseproteins. 49 Table2.24CompositionofpolyacrylamidegelsusedforSDS‐PAGE. 49 Table2.25Parametersusedforkineticmeasurementswiththefluorogenicassay. 52 Table2.26Parametersusedforanisotropymeasurements. 55 Table2.27Parametersusedtomeasureinhibitorstabilitywithanisotropy. 56 Table2.28Parametersusedtorecordfar‐UVCDspectrawithaJascoJ‐710. 58 Table2.29Parametersusedforrecordingoffluorescencemeltingcurves. 58 Table2.30ParametersforhomologymodelinginYASARA. 63 Table2.31Parameterstocalculatetheenergygridfordocking. 64 Table2.32ParametersusedforflexibleAnchor‐and‐Growdocking. 65 Table2.33ParametersusedforAMBERscoredocking. 65 Table3.1CatalyticparametersofTaspasetargetsequencesat37°C. 82 Table3.2Inhibitorypotentialofthesuccinimidylpeptidesandtheirderivatives. 97 Table4.1PropertiesofTaspase‐homologproteins. 104 Table4.2Parametersfordruglikelinessofthepeptidyl‐succinimideinhibitors. 120 Table7.1PulsesusedtorecordNMRspectra. 127 Table7.2RecordingparametersforNMRspectra. 127 Table7.3ProcessingparametersforNMRspectra. 128 Table7.4NMRshiftsforwildtypeTaspase. 135

10

Tableofcontents

1 Introduction 14 1.1 Mixedlineageleukemia 14 1.1.1 Cancerandleukemia 14 1.1.2 Translocationsareassociatedwithleukemia 15 1.1.3 11q23translocationscausemixedlineageleukemia 16 1.2 TheMLLprotein 17 1.2.1 MLLisamultidomainprotein 17 1.2.2 MLLfusionproteins 18 1.3 TheroleofTaspaseincancer 19 1.3.1 Proteasesandcancer 19 1.3.2 TaspaseisatypeIIasparaginase 21 1.3.3 Taspaseisanon‐oncogeneaddictionprotease 24 1.3.4 EndeavorstowardsTaspaseinhibition 26 1.4 Nanoparticles 27 1.5 Aimsofthisthesis 29

2 MaterialandMethods 30 2.1 Material 30 2.1.1 Chemicals 30 2.1.2 Purchasedpeptides 32 2.1.3 Buffersandmedia 32 2.1.4 Kits 33 2.1.5 Instruments 34 2.1.6 Consumables 35 2.1.7 Software 35 2.1.8 Enzymesandantibodies 36 2.1.9 Bacterialstrains 36 2.1.10 Plasmids 37 2.1.11 SDS‐PAGEandproteinmarker 38 2.2 Molecularbiologicalmethods 38 2.2.1 CloningofactiveTaspaseandtheinactivemutant 38 2.2.2 CloningoftheTaspaseloop 40 2.3 Microbiologicalmethods 42 2.3.1 Preparationofcompetentcells 42 2.3.2 Transformationofcompetentcells 42 2.3.3 Expressiontest 43 2.3.4 ExpressionofTaspaseconstructs 43 2.3.5 Cultivationofeukaryoticcelllines 44 2.3.6 Transfectionofeukaryoticcellsandpreparationformicroscopy 44 2.3.7 Fluorescencemicroscopyandimageevaluation 45 2.3.8 Cellpenetrationassayswithpeptidicinhibitor 45 2.4 Biochemicalmethods 45 2.4.1 Bacterialcelllysis 45 2.4.2 ProteinpurificationofHis‐taggedTaspase 46 2.4.3 ProteinpurificationoftheGST‐taggedTaspaseloop 47 2.4.4 Quantificationofprotein,peptideandDNAconcentrations 47 2.4.5 SDS‐PAGE 49 2.4.6 Analyticalgelfiltration 49

11

2.4.7 LabelingofTaspasewithAtto594‐NHS 49 2.4.8 LabelingofTaspasewithAtto488‐maleimide 50 2.4.9 AssayforautocatalyticactivationofTaspase 50 2.4.10 LimitedproteolysisassaywithproteinaseK 51 2.5 Spectroscopicmethods 51 2.5.1 Taspaseenzymeactivityassay 51 2.5.2 Bindingstudieswithdetectionoffluorescenceanisotropy 54 2.5.3 Bindingdetectedwithmicroscalethermophoresis 56 2.5.4 Determinationofinhibitorstabilitywithfluorescenceanisotropy 56 2.5.5 CDspectroscopy 57 2.5.6 Fluorescencemeltingcurves 58 2.5.7 NMRspectroscopy 59 2.5.8 RecordingandprocessingofNMRspectra 59 2.5.9 Assignmenttheory 60 2.5.10 Lactatedehydrogenaseactivityassay 62 2.5.11 Chymotrypsinactivityassay 62 2.5.12 MALDI‐TOFmassspectrometry 62 2.5.13 LC‐MSESI‐TOFmassspectrometry 63 2.6 Bioinformaticalmethods 63 2.6.1 GenerationofTaspasemodelsandpreparationfordocking 63 2.6.2 Compounddatabaseforvirtualdocking 64 2.6.3 VirtualDocking 64 2.6.4 GenerationofTaspaseloopmodels 65

3 Results 66 3.1 CloningandpurificationofTaspaseprotein 66 3.1.1 CloningofTaspaseanditsmutants 66 3.1.2 ExpressiontestofwildtypeTaspase 67 3.1.3 PurificationofwildtypeTaspaseandinactiveTaspase 68 3.1.4 PurificationofactiveTaspasemutant 70 3.2 CharacterizationofTaspaseanditsmutants 71 3.2.1 CDspectroscopicstudies 71 3.2.2 StabilityofTaspase 71 3.2.3 MALDI‐MSanalysis 72 3.2.4 Analyticalgelfiltration 73 3.2.5 AutocatalyticprocessingoftheTaspaseproenzyme 74 3.2.6 NMRspectroscopicanalysisofTaspase 75 3.2.7 CharacterizationoftheTaspaseloop 76 3.3 CatalyticactivityofTaspase 79 3.3.1 QuantitativeTaspaseactivityassay 80 3.3.2 CleavagesequencespecificityofTaspase 81 3.3.3 Taspaseactivityinhumancelllysates 82 3.3.4 TaspaseinhibitiontestusingNSC48300 83 3.4 ExplorationofTaspaseinhibitorsbyvirtualdocking 84 3.4.1 Modelpreparation 84 3.4.2 VirtualdockingofZINCdatabasecompounds 85 3.4.3 Inhibitiontestswithcompoundsobtainedbyvirtualdocking 86 3.5 InhibitionofTaspasebynanoparticles 89 3.5.1 NanoparticlesbindandinhibitTaspaseinvitro 89 3.5.2 Enzymeinhibitionisnogeneralfeatureofnanoparticles 92 3.5.3 NanoparticlesinhibitTaspaseincelllysatesandincells 93 3.6 InhibitionofTaspasebypeptidyl‐succinimidylpeptides 96

12

3.6.1 Peptidyl‐succinimidylpeptidesinhibitTaspaseinvitro. 96 3.6.2 Peptidyl‐succinimidepeptidesareresistantagainstTaspasecleavage 99 3.6.3 Peptidyl‐succinimidepeptidesbindandinhibitTaspaseincelllysates 100 3.6.4 Peptidyl‐succinimidylpeptidesenterthecell 101

4 Discussion 103 4.1 Taspaseisauniqueproteaseandrequiresuniqueinhibitors 103 4.1.1 CatalyticconstantsofTaspase 103 4.1.2 AutocatalyticprocessingofTaspasetakesseveraldays 104 4.1.3 Taspaseisahetero‐tetramerwithhighaffinity 105 4.1.4 StructureoftheTaspaseloop 107 4.1.5 InhibitionstrategiesforTaspase 109 4.2 Nanoparticlesasnovelinhibitors 111 4.2.1 Taspasebindingdependsonnanoparticlesize 111 4.2.2 Possibleinhibitorymechanisms 112 4.2.3 InvivoeffectsofnanoparticlesonTaspase 115 4.3 SuccinimidepeptidesinhibitTaspasecompetitively 117 4.3.1 Implicationsforsubstratecleavageandbindingmode 117 4.3.2 Scopeandlimitsofthesuccinimideinhibitors 118

5 Outlook 122

6 Abstract 124

7 Supplement 127 7.1 RecordingandprocessingofNMRspectra 127 7.2 Bacterialvectormaps 129 7.3 Eukaryoticvectormaps 130 7.4 Taspaseloopcloningscheme 131 7.5 PurificationofTaspase 132 7.6 ExpressiontestofactiveTaspase 133 7.7 Calibrationforanalyticalgelfiltration 134 7.8 Analyticalgelfiltrationwithlowersaltconditions 134 7.9 NMRshiftsofTaspase 135 7.10 1H‐15NHSQCspectraofTaspase 136 7.11 ExpressiontestoftheGST‐taggedTaspaseloop 137 7.12 NaClinhibitsTaspaseactivity 137 7.13 WesternblotanalysisofTaspaselevelsinhumancelllysates 137 7.14 ESI‐LC‐MSofNSC48300 139 7.15 Modelqualitystatistics 139 7.16 Bindingcontrolswithnanoparticles 140 7.17 Cellmorphologyinthepresenceofnanoparticles 140 7.18 Titrationcurvesofsuccinimidepeptides 141 7.19 Passagethrough50kDamembrane 142 7.20 StructureoftheTaspasehetero‐tetramer 142 7.21 MolecularplugstestedasTaspaseinhibitors 143

8 References 146

13

9 Acknowledgement 155

10 CurriculumVitae 156

11 Declarations 158

INTRODUCTION MIXEDLINEAGELEUKEMIA

14

1 Introduction

1.1 Mixedlineageleukemia

1.1.1 Cancerandleukemia

Cancerconstitutesagroupofdiseasescharacterizedbyuncontrolledcelldivision.Thisresultsin

unregulatedcellgrowthandtheformationofmalignanttumorsthatcanspreadtootherorgans.

Cancertypicallyoriginatesfromasinglecellwhichhasundergoneprogressivegeneticaberrations

toescapecelldeathandpromoteproliferation.

In2011, roughly450 in100000 individualswerenewlydiagnosedwith cancer [1].Themost

common formsof cancer inEuropeare lung cancer (13.3%), colorectal cancer (13.2%), and

breastcancer(13%)[2].Inprinciple,tumorscanoccurineverytissue,andcurrentlymorethan

200cancertypesaredescribed[3].

However,alltumorcellssharesomefeaturesbywhichtheycanbedistinguishedfromhealthy

cells.Essentially, they foundwaystosuccessfullyunbalancethedelicateandstrictlyregulated

balanceofproliferationandapoptosis.HanahanandWeinbergproposedsixessentialalterations

commonforthemajorityoftumors:(i)apoptosisevasion,(ii)self‐sufficiencyingrowthsignals,

(iii)insensitivitytoanti‐growthsignals,(iv)metastasis,(v)limitlessreplicativepotential,and(vi)

sustainedangiogenesis(figure1.1;[4]).

Figure1.1Hallmarksofcancer.

Mostcancershaveacquiredthesamesetoffunctionalalterationsinaseriesofsteps.Thesenewcapabilitiesaretermedhallmarksofcancer.Accordingto[4].


15

This model was later on extended by evasion of immune surveillance and mechanisms to

compensateforstressgeneratedbythetumortransformation,includingmetabolic,proteotoxic,

mitotic,oxidative,aswellasDNAdamagestress.Eachaberrationdepictedinfigure1.1represents

thebreachingofananticancermechanismandisacquiredbypointmutation,geneamplification,

deletion,orchromosomaltranslocation.Tumor‐associatedgenescanbeclassifiedasoncogenes

or tumor suppressor genes. While oncogenes, such as growth factors, can promote tumor

proliferation,tumorsuppressorsareresponsibleforcellcyclecontrolaswellasapoptosis,and

arethereforeoftenrepressedintumors.

Malignanttransformationsoccurnotonlyinsolidtissues,butalsoinbloodcells.Suchtypesof

cancer are referred to as leukemia and describe cancers of the hematopoietic system

characterizedbyarapidproliferationofwhitebloodcells.Thesecellsreplacethenormalbone

marrowcells,whichresultsinanincreasedreleaseofimmatureprogenitorcells,calledblasts,to

theperipheralbloodsystem.Eventually,thesemalignantcellssuppresstheproductionofnormal

bloodcells,whichcausesadeficiencyoffunctionalredandwhitebloodcells.Asaconsequence,

mostpatientssufferfromimmunodeficiency,impairedbloodclotting,oranemia.InGermany,26

in100000individualsarenewlydiagnosedwithleukemiaeveryyear(numberfrom2012;[5]).

Moreover,concerningchildrenunder theageof15, leukemia is themost frequentmalignancy

accountingfor31%ofallcancercases[6].

Clinically,leukemiascanbesubdividedintoacuteandchronicforms.Patientswithacuteleukemia

usuallyshowasuddenappearanceofsymptomsincombinationwitharapidprogressionofthe

disease. Acute leukemias end lethal withinweeks, if not treated immediately [7]. In contrast,

chronic leukemias typically develop over a longer period of time and progress slowly [8].

Additionally, both acute and chronic leukemia can be subdivided into lymphocytic and

myelogenousforms,accordingtothetypeofaffectedbloodcells.Whilelymphocyticformsarise

in progenitor cells of lymphocytes, myelogenous forms affect myeloid cells which normally

differentiate into erythrocytes, aswell as thrombocytes and other types ofwhite blood cells.

Altogether,thefourmajortypesofleukemiaareacutelymphaticleukemia(ALL),acutemyeloic

leukemia(AML),chroniclymphaticleukemia(CLL)andchronicmyeloicleukemia(CML).

Oneexceptionfromthisclassificationisthemyeloic‐lymphoidormixedlineageleukemia(MLL).

This aggressive leukemia affects both the lymphoid andmyeloid lineage [9] and itsmolecular

pathogenesisisexplainedindetailbelow.

1.1.2 Translocationsareassociatedwithleukemia

Althoughthecausativefactorsforleukemiaarestillunclearinmostcases,someriskfactorswere

determined.Especially ionizingradiationcontributes significantly toacute leukemias [10], but

also chemicals, such as pesticides [11] or benzenes [12] increase the risk for leukemia.

Nevertheless,even ifother factors, suchasage, smoking,obesityand familialbackground,are

considered,only15‐20%ofallleukemiacasescanbeexplainedwiththeseriskfactors[13].

Moreover,alsoatthemolecularlevelthecausesformostcasesofleukemiaareyettobeidentified.

However,severalstudiespinpointedthattranslocationsplayanimportantroleinleukemogenesis


16

andarecommonlyobservedinall fourmajortypesof leukemia,withmorethan400different

translocationsdescribedforacuteleukemias[14].Translocationsrequireadoublestrandbreak,

occurringeitherspontaneouslyor inducedby irradiationorcytotoxicagents,withsubsequent

reciprocalfusiontoanonhomologouschromosome[15].Importantly,twoseparategenescanbe

joinedtoafusiongeneinthisprocess.Ifaproto‐oncogeneisfusedtoanabundantlyexpressed

gene with a strong promoter or the fusion protein has different functions, the cell can be

transformedtoatumorcell.

Interestingly, particular translocations are associated with certain leukemias [16]. The most

prominent example is the CML‐related Philadelphia chromosome, which is the result of a

reciprocaltranslocationofchromosomes9and22(t(9;22)(q34;q11)).Thistranslocationfuses

theABLgene(Abelsonmurine leukemiaviraloncogenehomolog1)totheBCRgene(breakpoint

cluster region; [17]) and generates a fusion protein with permanently active tyrosine kinase

activity.

1.1.3 11q23translocationscausemixedlineageleukemia

11q23translocations involvingtheMLLgene,alsoreferredtoasALL‐1orHRX,areamongthe

mostcommonnon‐randomtranslocations[16].SuchrearrangementsjointheN‐terminalpartof

theMLLgenetoafusionpartnergenebynon‐homologousendjoining[18–20]andarerelatedto

mixedlineageleukemia(MLL).Incontrasttoothertranslocations,MLLisapromiscuousoncogene

with more than 60 different fusion partner genes [21]. However, the fusion partners can be

categorizedintotwogroups.ThefirstgroupcomprisingthesixmostfrequentfusionpartnersAF4,

AF9,ENL,AF10,ELL,andAF6accountsformorethan85%ofallclinicalcases[19].

FurtherstudiesshowedthatallfrequentpartnersexceptforAF6arenuclearproteins,whilethe

remaining fusion partners of the second group are predominantly localized in the cytoplasm

(table1.1).DuetothenuclearlocalizationsignalintheMLLprotein,allresultingfusionproteins

arefoundinthenucleus[9].

Table1.1SelectionofMLLfusionpartners.

Accordingto[9].

Name Features LocalizationENL BindshistoneH3,assemblesENLassociatedproteins

elongation/chromatinremodelingcomplexNucleus

AF4 ENLhomolog NucleusAF6 Dimerizationdomain CytoplasmaAF9 FounderofAF4family,memberofENLassociatedproteins NucleusAF10 InteractswithDOT1Lhistonemethyltransferase NucleusELL Elongationfactor,interactswithaAF4relatedprotein NucleusAF1p Dimerizationdomain CytoplasmaABI1 InteractswithENLwhenimportedtonucleus CytoplasmaGAS7 Dimerizationdomain Cytoplasma

InspiteofthepromiscuityofMLLwithrespecttoitsfusionpartners,noteveryproteincauses

leukemiainmicewhenfusedtoMLL[22,23].Moreover,ameretruncationoftheMLLproteinis

INTRODUCTION THEMLLPROTEIN

17

notsufficienttocauseleukemia[19].FormostMLLtranslocations,thedoublestrandbreakoccurs

in an 8 kb spanning region of theMLL gene called breakpoint cluster region [24]. SuchMLL‐

involvingtranslocationsareofhighclinicalrelevance,astheyarefoundinroughly70%ofinfant

ALLandoccurin10%ofallotherALLcases[9,25].

Thecellularmechanismsleadingtotheexhibitionoftumorfeaturesaremultifaceted[19].Current

dataindicatethatMLLtranslocationscauseincreasedexpressionofHoxgenesbyMLLitselfand

recruitmentoftranscriptioncontrolcomplexesbythefusionpartnergenes,whichalsoenhances

transcriptionofMLLtargetgenes.Altogether,thist(4;11)‐triggeredmixedlineageleukemiaisa

veryaggressivetypeofacuteleukemiawithdismalprognosis:Only40%ofallinfantssurvivefive

years afterMLL diagnosis [26],whereas the five‐year survival rate of infantswith non‐mixed

lineageleukemiaisnear90%[27].

In1996,thefirstmousemodelforMLLstudieswassuccessfullygenerated.Thesemiceexpressed

theMLL‐AF9fusionproteinunderthecontroloftheendogenousMLLpromoterandwerefound

to develop AML and rarely ALL [28]. Moreover, MLL translocations differ from other

translocationsinthatbothofthetwofusiongenesgeneratedbythereciprocaltranslocationwere

recently reported to function as oncogenes. In the respective study, the two oncoproteins

resultingfromthefusionofAF4andMLL,namelyAF4‐MLLandMLL‐AF4,contributedconcertedly

tothepathologicmechanism[29,30].

1.2 TheMLLprotein

1.2.1 MLLisamultidomainprotein

Figure1.2DomainsoftheMLLprotein.

Schematicrepresentationof themixed lineage leukemia(MLL)gene.Proteinbindingsitesare indicatedabovetherespectivedomains.ATH:AT‐hookmotifmediatingbindingtotheminorgrooveofAT‐richDNAsequences;SNL:specklednuclearlocalizationsite;TRD:transcriptionalrepressiondomaincomprisingthetwofunctionalsubunitsRD1andRD2;PHD:planthomologydomain;TA:transcriptionalactivationdomain;SET:Su(var)3‐9,enhancer‐of‐zeste,trithoraxdomain;BCR:breakpointclusterregion.Accordingto[19].

Owingtothehighclinical importanceoftheubiquitouslyexpressedMLLproteinfor leukemia,

severalstudieswereconductedtounravelitsfunctions.Consistingof3969aminoacids,theMLL

proteincomprisesmultipleconserveddomains(figure1.2):BindingtoATrichDNAregions is

mediatedbythreeN‐terminalAT‐hookmotifs[31],whicharefollowedbyarepressiondomain

with DNA methyltransferase homology [32], and zinc fingers that mediate binding to the

INTRODUCTION THEMLLPROTEIN

18

cyclophilin CYP33 [33]. In theC‐terminal region of theMLLprotein, a transactivationdomain

interactingwithCREB‐bindingprotein(CBP)isfound,aswellasaSET(Su(var)3‐9,enhancer‐of‐

zeste,trithorax)domainwithhistoneH3K4methyltransferaseactivity.

With the help of these domains, MLL functions as transcriptional regulator. In particular, it

regulates Hox gene expression and is therefore found widely expressed during embryonic

development,whilemostadulttissuesdisplayonlylowexpressionlevels[16].Ofnote,MLLseems

indispensableforvertebratedevelopmentandthusMLLknockoutmiceshowembryoniclethality

inadditiontohomeotictransformations[16,34,35].

Closer mechanistic analysis of the underlying processes revealed that especially the H3K4

methyltransferaseactivityoftheSETdomainisreportedtobeimportantforHoxgeneregulation

[36].Consequently,micewithMLLproteinlackingtheSETdomainshowedskeletaldefectsand

alteredHoxgeneexpressionsimilartoMLLknockoutmice[37].

1.2.2 MLLfusionproteins

MLLfusionproteinsresultfromgenetictranslocationsinvolvingtheMLLgene(section1.1.2).

BoththewildtypeandMLLfusionproteinsexerttheircentralroleingeneexpressionregulation

byinteractionswithaplethoraofotherproteins.Forthis,MLLandalsotheMLLfusionproteins

serve as a platform for the formation of a supercomplex composed ofmore than 30 proteins

(figure1.3;[38,39]).

Figure1.3MLLfusionsproteins.

TheAF4‐MLLfusionproteinisboundbySIAHsandsubjectedtoproteasomaldegradation.Activationbyproteolytic cleavage at defined cleavage sites (CS) allows heterodimerization and formationof a stablemultiproteinsupercomplexresistanttoSIAH‐mediateddegradation.Accordingto[38].

In2003,HsiehandcolleaguesreportedanothercriticalstepnecessaryforMLLactivity:TheMLL

proteinrequiresproteolyticcleavageandthetwosubunitsformaheterodimermediatedbythe

interactionofaFYRNdomaininthe320kDaN‐terminalfragmentandtheFYRKandSETdomains

inthe180kDaC‐terminalfragment(figure1.4;[40]).

INTRODUCTION THEROLEOFTASPASEINCANCER

19

TheuncleavedAF4‐MLLfusionproteininteractswithseveninabsentiahomologs(SIAHs)viathe

SIAH‐dependent degron located in theN‐terminal half of the AF4 protein,which subjects the

precursorfusionproteintoproteasomaldegradation.Incontrast,theactivatedMLLheterodimer

isprotectedagainstdegradationandtheMLLcomplexincludingSIAHsisstabilized(figure1.3;

[38]).ThismechanismisthoughttotrapSIAHandpreventSIAH‐mediateddegradationofother

tumor‐relevantproteins.

Figure1.4ProcessingofTaspaseandMLL.

aTheMLLproteincomprisestwosimilarTaspasecleavagesites,termedCS1(leftsequence)andCS2(rightsequence). b Upon Taspase cleavage, the 52 amino acids fragment between the two cleavage sites isdegradedandthetwolargeMLLfragmentsassembletoadimerviaaFYRNinteractiondomainintheN‐terminalfragmentandFYRCandSETdomainsintheC‐terminalfragment.

FurtheranalysesrevealedthatMLLiscleavedattwosimilarproteolyticcleavagesites53amino

acidsapartfromeachother,referredtoasCS1andCS2(figure1.4).WhileCS2ishighlyconserved

among all MLL homologs, including the human MLL2 protein and the trithorax protein in

Drosophilamelanogaster,CS1ispresentonlyinvertebrates.Subsequently,theproteaseforMLL

cleavagewasdiscovered.Duetothethreonineasitsactivesiteresidueandthespecificcleavage

after aspartate, this proteasewas named threonine aspartase 1 (Taspase). Taspase is able to

cleaveboth cleavage sites (CS1andCS2) in theMLLproteinandhasgained in importanceas

protease‐mediatedcleavageisindispensableforMLLactivation.

1.3 TheroleofTaspaseincancer

1.3.1 Proteasesandcancer

Proteasesareenzymesthatcatalyze thehydrolysisofpeptidebondsandarethereforeableto

cleaveproteinsandpeptides.Theyplaykeyrolesinfooddigestion,degradationofmisfoldedand

unneededproteins,butalsointheregulationofothercellularprocesses,suchasapoptosis,blood

coagulation,immuneresponse,andcellcycleprogression[41].

Due to the irreversibilitynatureofproteolyticcleavage,proteasesarestrictlyregulated in the

body.Besidesposttranslationalmodification[42,43],activationofaninactiveprecursorprotein

(zymogen)bylimitedproteolysisisacommonmechanism,whichrepresentsanessentialstepin

manyimportantcellularpathways:Theactivationofcaspasesduringapoptosis[44],theactivation

ofbloodclottingfactors[45],theactivationoftrypsinandotherproteasesaswellassignalingvia


20

protease‐activatedreceptors(PARs) [46]areonlya fewexamples forproteolyticactivation. In

contrasttoproteindegradation,proteolyticactivationeventsoccurinasite‐specificmannerand

therespectiveproteasesarehighlyspecific[41].

CleavageoftargetproteinsbyproteasesoccurseitherattheN‐orC‐termini(exopeptidases)or

within the polypeptide chain (endopeptidases) at protease specific cleavage sites, termed

consensussequences.TheaminoacidsN‐terminalofthecleavagesitearereferredtoasP1,P2,

etc.,whiletheaminoacidsC‐terminalofthescissilebondarenumberedP1’,P2’,etc.(figure1.5a).

Thesitesatthesurfaceoftheproteasethatcanbindasinglesubstrateresiduearecalledsubsites,

andnumberedSnorSn’accordingtoitsrespectivesubstrateposition.Consequently,thescissile

bondislocatedbetweentheP1andP1’site.

Figure1.5Substratebindingandproteasefamilies.

aSubstratebindingtoaprotease.Substrateresiduesparticipating inbindingto theprotease(gray)arenumberedP1‐Pn(non‐primedsites;blue)N‐terminalofthecleavagesiteandP1’‐Pn’(primedsites;red)C‐terminalofthecleavagesite.TherespectivebindingsitesatthesurfaceoftheproteasethataccommodatetheaminoacidsidechainsarenumberedS1‐SnandS1’‐Sn’,respectively.CleavageoccursbetweentheP1andP1’position.bHumanproteasescanbeclassifiedaccordingtotheiractivesiteintofiveclassesthatcanbeeitherintra‐(innerring)orextracellular(outerring).Numbersinsidetheringsindicatethenumberofmembersofthecolorcodedclasses.Accordingto[47].Themorethan500humanproteasescanbeclassifiedintofiveclassesaccordingtotheiractive

site,thatis,serine,cysteine,cysteine,aspartate,andmetalloproteases(figure1.5b;[47,48]).All


21

fiveclassesofproteasesareinvolvedintumorinitiation,growth,metastasisaswellasinvasion

intoothersites[49].Forthis,thetumorassociatedproteasestargetandregulatediversecellular

processes.Upregulationofcellproliferation,downregulationofcelldeath,angiogenesis,aswell

asinductionofmulti‐drugresistanceareonlyafewexamplesofcellularprocessesgovernedby

proteases [50]. In this regard, trypsin is involved in colorectal carcinogenesis, as it positively

regulatesproliferationandmetastasis[51],cathepsinsBandDarecloselyrelatedwithinvasion

andprogressionoforalsquamouscellcarcinoma[52],andcaspasederegulationisknowntocause

neuroblastomas[53].Assomeproteasesarespecificforcertaincancertypes,proteasescanalso

serveasusefulprognosticanddiagnosticmarkerstoimprovecancertreatment[54].

Some proteases possess tumor suppressive effects and loss‐of‐function mutations in the

respectivegenespromotemalignanttransformation[55].Investigatingtheroleoftumor‐related

proteasesandtheirmechanismcanthereforeyieldnewtargetsforcancertherapies.

1.3.2 TaspaseisatypeIIasparaginase

Taspasewasdiscoveredin2003astheproteaseresponsiblefortheproteolyticactivationofthe

MLL protein. As a member of the type II asparaginase family, Taspase displays the typical

asparaginase foldwithacentralβ‐sheetsandwichsurroundedonboth facesbyα‐helices [56].

However,TaspasesharesonlylowsequencesimilaritywithothertypeIIasparaginasesandthe

proteasomeasrelatedthreonineprotease(table4.1).Nevertheless,thestructureofthisfamilyis

highlyconservedwithRMSDvaluesbelow2Å(figure1.6a).

Figure1.6Taspasedisplaysaconservedasparaginasefold.

a Overlay of the crystal structure of Taspase (red; PDB 2a8j) with human asparaginase (4gdw), plantasparaginase (2gez), E. coli asparaginase (2zal) and F. meningosepticum glycosylasparaginase (1ayy)revealsaconservedasparaginasefold.ThestructureswerealignedusingPyMOL.bClose‐upviewoftheactivesiteresiduesofTaspase(red)anditshomologs(gray)showsaconservedsidechainorientation.Sidechainsaredisplayedassticksandthecatalyticthreonineishighlightedinyellow.


22

Although all other asparaginases do not act as proteases but hydrolyze the free amino acid

asparaginetoaspartate[57],theactivesiteissimilartoTaspase,includingtheorientationofthe

catalyticthreonine(figure1.6b).Despitethesmalldifferencesfoundbetweentheactivesitesof

TaspaseandothertypeIIasparaginases,thereasonforthediversereactionswhicharecatalyzed

is yet to be elucidated. To become catalytically competent, the 45 kDa Taspase proenzyme

undergoesproteolyticcleavageintotwosubunits(figure1.7).

Figure1.7AutocatalyticprocessingofTaspase.aTheTaspaseproenzymecomprises420aminoacidsandundergoesspontaneousautoproteolyticcleavagebetweenAsp233andThr234.bThisprocessyieldsa25kDaα‐subunit (aminoacids1‐233; red)anda20kDaβ‐subunit(aminoacids234‐420;orange).Inthisprocess,Thr234becomestheN‐terminalaminoacid of the β‐subunit and its free hydroxyl group renders Taspase proteolytically active. Two αβ‐heterodimersassembletoahetero‐tetramerwithαββαstructure.N:Amino‐terminus;C:carboxy‐terminus

SucharemovalofapropeptideistypicalfortypeIIasparaginasesandotherproteases,withrenin

[58]andchymotrypsin[59]aswell‐studiedexamples,andservestoinhibitenzymeactivityandas

folding helper [60]. The autocatalytic proteolytic processing of the type II asparaginase

proenzymes has been elucidated in detail [61,62]. In short, the catalytic threonine side chain

attacksthepeptidebondtotheprecedingaspartate,beforetheproteinbackboneandsidechain

bonds between threonine and aspartate are broken (figure 1.8). For Taspase, the whole

maturationprocesswasfoundtobeaspontaneousreactionandnotcatalyzedbyotherTaspase

proteinsintrans[63].

Figure1.8AutocatalyticactivationofTaspaseandothertypeIIasparaginases.The side chain OH‐group of the catalytic threonine forms a covalent bond to the CO of the precedingaspartate,yieldingatetrahedralintermediateandageneralbase(B)capturestheprotonofthethreonineOH‐group(firstreaction).Subsequently,thebackboneC‐Nbondisbrokenandageneralacid(HA)donatesaprotontothenewlycreatedN‐terminus(secondreaction).Eventually,awaterattacktotheCObreaksthebondbetweentheaspartatebackboneandthethreoninesidechain(thirdreaction).Modifiedaccordingto[61].

Eventually,thisprocessrendersafreehydroxylgroupattheN‐terminalthreonine234oftheβ‐

subunit. Hence, Taspase and other asparaginases belong to the N‐terminal nucleophile (Ntn)


23

familyofproteases.DetailedanalysisofthecrystalstructureofTaspaserevealedarotationofthe

Thr234sidechaininadditiontoabackboneshiftuponactivation[56].Mutationoftheactivesite

threonine to valine does not only prevent the autocatalytic activation event, but also renders

Taspaseunabletocleaveitssubstratesintrans[56,57].

After autoprocessing, theα‐ andβ‐subunits forma heterodimer, and twoof these dimers are

supposed to assemble to a hetero‐tetramer (figure 1.7b; [64]). The resulting αββα fold after

autoprocessingisacommonfeatureofallNtnhydrolasesdespiteverylowsequencesimilarity

and was described for glycosylasparaginase [65], plant asparaginase [66], and the only other

threonineprotease in theNtn family, the proteasome [67]. However, the formation of a αββα

shaped hetero‐tetramer in solution is controversially discussed for Taspase, as analytical gel

filtrations indicate the presence of Taspase as heterodimer and enforced dimerization could

effectivelyinhibitTaspaseincells[68].

In2005,Leeand colleaguespresentedahypothetical cleavagemechanismbasedon the close

homologytoasparaginases(figure1.9). Inthismodel,Taspasedoesnotdirectlyhydrolyzethe

peptidebondoftheproteinbackbone,butthesubstratefirstundergoesarearrangement.After

anattackoftheP1’glycinebytheP1aspartatesidechain,asuccinimide‐hydrateintermediateis

formed.Inasecondstep,theformerpeptidebondishydrolyzed,yieldinganisoaspartatewiththe

protein backbone transferred to the aspartate side chain. Finally, Taspase cleaves the newly

formedisoaspartateandreleasesthetwofragments.

Figure1.9ProposedTaspasecleavagemechanism.

Theaspartatesidechainofthesubstrateattacksthefollowingglycine(redarrow),yieldingasuccinimide‐hydrate intermediate. Subsequently, hydrolysis of the former peptide bond results in a transfer of theproteinbackbonechaintotheaspartatesidechain.ThisisoaspartateisthencleavedbyTaspase(reddashedline)andthetwofragmentsarereleased.Accordingto[69].

Furthermore,thechlorideionfoundinthecrystalstructurerevealedanothersurprisingfeature

of Taspase. It is bound to the active site and inhibits Taspase activity already at intracellular

chlorideconcentrations[56].Hence,Taspaseisassumedtobepresentasinhibitedenzymeinthe

cellanditssubstrateproteinsareexpectedtotemporarilydisplacethechlorideionforproteolytic

processing.ThefirstsubstratedescribedforTaspasewastheMLLprotein[57],whichisactivated

afterTaspasecleavageattwocleavagesites(figure1.4).Subsequentalaninescansdeterminedthe

TaspaseconsensussequenceaseitherQ‐[F,I,L,V]‐D|G‐X‐D‐XorQ‐[F,I,L,V]‐D|G‐X‐X‐D[63,70].

Withthissequenceathand,27putativeTaspasetargetproteinswereidentified[71].Bynow,the

experimentallyverifiedtargetscompriseMLL,MLL2[72],thetranscriptionfactorsTFIIA[73]and

ALF[73],upstreamstimulatoryfactor2(USF2),myosin1F(MYO1F),andMLL4[70].Strikingly,

manyTaspasesubstratesarenuclearfactorsinvolvedingeneexpressioncontrol,indicatingarole


24

ofTaspase‐mediatedcleavageintheregulationofthesefactors.However,forothertargetsthan

MLL, the role of Taspase cleavage is still elusive. For example, Taspase‐mediated cleavage of

transcriptionfactorIIa(TFIIA)seemstobeafinetuningmechanismofTFIIAactivityandwas

speculatedtobeimportantfortheexpressionofacertainsubsetofgenes[73].

LocalizationstudieswithfluorescenttagspinpointedanuclearlocalizationofTaspaseduetoits

bipartite nuclear localization signal [74]. However, closer investigation revealed a dynamic

shuttlingprocessbetweennucleusandcytoplasm.Thenuclearimportviaanuclearlocalization

signal (NLS) is mediated by importin‐α and could also be confirmed for the proenzyme. In

contrast, interactionwith the nucleolar shuttle protein nucleophosminwas only observed for

autocatalyticallyprocessedTaspase.AlthoughTaspasedoesnotpossessanuclearexportsignal,

ithasaccesstothecytoplasmandisabletoprocessitscytoplasmictargetproteins.Thisexport

processisnucleophosmin‐dependentandcanbeinhibitedbyleptomycinB[74].

TaspaseknockdowninHelacellsresultedinimpairedMLLcleavageandverifiedthatnoother

proteaseisabletoactivateMLL.TheclosecorrelationofTaspaseandMLLisalsoreflectedinthe

factthatbothgenescoevolvedinvertebrates[57].Consequently,TaspaseplaysakeyroleforMLL,

asthehistonemethyltransferaseactivityofMLLiscruciallydependentonactivationbyTaspase

cleavage[72].Accordingly,alsotheMLLfusionproteinsareinactive,unlesstheyareprocessedby

Taspase[29].

1.3.3 Taspaseisanon‐oncogeneaddictionprotease

Despite the importance of Taspase for leukemia, Taspase alone is not sufficient for the

transformationoffibroblasts.Hence,Taspaseisnotclassifiedasoncogene,butasnon‐oncogene

addictionprotease[75,76].ThismeansthatalthoughTaspaseisnooncogene,certainleukemias

heavilyrelyonTaspaseandarethereforesusceptibleforTaspaseinhibition(figure1.10).

This permissive role of Taspase in leukogenesiswas corroborated by the finding thatmouse

embryonicfibroblasts(MEFs)lackingTaspaseareresistanttooncogenictransformationinduced

byMyc, E1A, RAS, and p53 [76].Moreover, Taspase‐deficientmice exhibited decreased cyclin

levels as well as increased expression of cyclin dependent kinase (CDK) inhibitors and thus

showed a disrupted cell cycle [72]. Additionally, Taspase expression is upregulated in many

human cancers as well as cancer cell lines and required for tumor maintenance. Generally,

increasedTaspase levels correlatewithadependencyonTaspaseexpression [72,76].Ofnote,

recent studies imply a crucial role of Taspase not only in leukemia, but also in solid tumors

[57,70,76]. Consequently, Taspase knockdowns in these tumors cause a disruption of

proliferationandpromoteapoptosis [76]. Interestingly, althoughTaspase is critical formouse

embryonicdevelopment,thefewsurvivingTaspase‐deficientadultmiceexhibitedanormallife

span[63].Hence,inhibitionofTaspaseisregardedasnewapproachtotreatleukemia.


25

Figure1.10Taspaseaddictionintumorcells.

Taspase inhibition is only lethal forMLL tumor cells,while healthy cells arenot affected.Healthy cellsexpressnormalTaspaseandMLLlevels,andTaspaseinhibition(blackline)doesnotimpairviability(leftcolumn).AberrantTaspaseexpressionisnotsufficientforamalignanttransformation,aslongasMLLlevelsarenormal.Undertheseconditions,Taspaseinhibitionisnotfatal(middlecolumn).AberrantlevelsoftheoncogeneMLLcausemalignanttransformationandareoftenaccompaniedbyaberrantTaspaselevels.AsthesecellsheavilyrelyonTaspasefortheactivationofMLL,Taspaseinhibitionresultsincelldeath.

Targetingnon‐oncogeneaddictionproteinsopensnewwaysincombatingcanceranddiffersfrom

the classical approaches that either aim at suppressing oncogene activity or restoring tumor

suppressorfunction[75].Inrecentyears,severalnon‐oncogeneaddictionproteinsweredetected

anddetailedanalysesoftheseproteinsachievednovelconceptsforcancertherapy.Infact,some

inhibitorsfornon‐oncogeneaddictionproteinsarealreadyapprovedasanticancerdrugs.

Oneexampleinthisrespectistheheatshockprotein90(HSP90)inhibitorgeldanamycin™.Since

HSP90clientproteinsarecrucial for theregulationofcell survival, cellgrowth,cell cycle,and

apoptosis,tumorcellsrelyingontheseclientproteinsareeffectivelyhitbytheHSP90inhibitor

geldanamycin™ [77]. Furthermore, Luo and colleagues reported that inhibition of the non‐

oncogenepolo‐likekinase1(PLK1)istoxicforcellswithmutatedRasoncogene[78].Moreover,

alsotheproteasomeinhibitorbortezomibcanberegardedasanon‐oncogeneaddictiondrug.Itis

active against a variety of tumors, includingmyeloma, CLL, prostate and colon cancer, and is

approvedfortreatingmultiplemyelomaandmantlecelllymphoma[79].Tumorcellsacquirea

variety of genetic aberrations and thus suffer from the stress generated by mutated and

aggregatedproteins.Consequently,inhibitionoftheproteasomeexacerbatesthestressintumor

cellsmorethaninhealthycells.


26

1.3.4 EndeavorstowardsTaspaseinhibition

As non‐oncogene addiction protease, Taspase is of high interest for cancer therapy, since

inhibition of Taspase is considered a promising way to prevent leukemogenesis. However,

inhibitionofTaspaseactivityisnottrivial,asTaspaseisinsensitiveagainstinhibitionbygeneral

serine, cysteine, or metalloprotease inhibitors [57]. Hence, different approaches were

hypothesized to obtain Taspase inhibitors, including compound screening, virtual docking,

rationaldesign,andenforceddimerization(figure1.11).

Figure1.11StrategiesfornewTaspaseinhibitors.

a+bInhibitors(red)forTaspase(blue)canbefoundbydatabankscreening.Thecompoundscanbindtotheactivesite(a)orallosteric(b).cRationaldesigncanyieldsubstrateanalogcompoundsdGeneticallyenforceddimerizationwasfoundtoinhibitTaspaseactivity.

1.3.4.1 Compoundscreening

High‐throughputscreeningofsmallmolecules(figure1.11a+b)isacommonmethodtofindnew

inhibitorsandhasbeenappliedsuccessfullyformanyotherenzymes[80,81].Consequently,small

moleculescreeningsaimingatTaspaseinhibitorshavebeenpublishedpreviously[63,71,82].The

mostpromisingcandidatewastheanti‐angiogenicarsenicinhibitorNSC48300.Thiscompound

wasfoundinahigh‐throughputscreeningassayofasmallmoleculelibraryandwasreportedto

inhibit Taspase allosterically with a Ki of 4.2 µM [63]. However, the arsenic moiety of the

compoundposesahightoxicrisk,isknownforunspecificeffects[83]anditsapplicationinvivois

limited.

1.3.4.2 Virtualdocking

The available crystal structure reveals the conformation of the active site and thus allows

targetingTaspasebyvirtualdocking(figure1.11a).Knauerandcolleaguesperformedaninsilico

screeningwithadatabasecomprisingroughly180000compounds[71].However,theresulting

topcandidates(CHC‐A4andDHC‐C1)displayedonlyweakinhibitionofTaspaseactivity.

1.3.4.3 Substrateanalogs

Besides compound library screening, inhibitors canbederived from the substrate by rational

design(figure1.11c).Regardingthis, theBogyogroupdevelopedaseriesofsubstrate‐derived

vinyl sulfone and vinyl ketone derivatives [69]. However, these compounds displayed only

INTRODUCTION NANOPARTICLES

27

moderateinhibitionofTaspaseactivitywithanIC50of30µMdeterminedforthebestcompound

yzm18.

1.3.4.4 Enforceddimerization

Another way for Taspase inhibition exploits the quaternary structure by enforced protein

interaction(figure1.11d),assuggestedbyThieletal.[84].Notably,co‐expressionofactiveand

inactiveTaspasedoesnotreduceTaspaseactivity,indicatingthatcatalysisinthetwoprotomers

ofthehetero‐tetramerisindependent[85].However,BierandcolleaguesenforcedtheTaspase

dimerbythefusionofJun‐andFos‐tagstotheC‐terminusofTaspaseandfoundthattheenforced

dimerwascatalyticallyinactive[68].Sofar,Taspasedimerizationwasonlyartificiallyachievedby

geneticmanipulationandnochemicaldimerizersforTaspasehavebeendeveloped.

In summary, all these studies failed to find a potent inhibitor for in vivo applications.

Consequently,novelTaspaseinhibitorsarestillurgentlyneeded.

1.4 NanoparticlesNanoparticles are solid particles with a size of roughly 10‐100 nm [86] and have gained in

importanceduringthelastdecades.Theseparticlescaneitherbereleasedincidentally(carand

industrialexhausts,volcaniceruptions)orengineeredpurposefully[87].Thelatterarefrequently

preparedeitherbydispersionofpreformedpolymersorbuiltupbypolymerizationofmonomers

[88].Importantparametersofnanoparticlesbesidessizeincludethematerial,particleshape,zeta

potential(i.e.charge),curvature,hydrophilicity,biodegradability,andsurfacemodifications[89].

Thesurface‐to‐volumeratioofsphericalparticlesincreaseswithdecreasingparticlediameter.For

example,particlesof50nmdiameterpossessasurfaceareaof44m2pergram.Asaconsequence

ofthis,materialspossessdifferentpropertiesatnanoscalecomparedtothebulkmaterialandare

thereforeindustriallyusedforexampleforcosmetics(e.g.sunscreen),textiles(e.g.self‐cleaning

textiles)aswellascomputers(e.g.nanotransistors).

Nevertheless, the progressive usage of nanoparticles is controversially discussed, as

nanoparticles takenupvia theskin, lung,andgutaredistributedbythebloodstream [90]and

possiblehazardsandlong‐terminfluencesareinsufficientlyresearched[90].Especiallythelong

retentionhalf‐timesof inhalednanoparticles in thealveolarregionof70days inratsand two

yearsinhumans[91]poseathreattocells,sincenanoparticlesareinternalizedbyendocytosis

[92].Oncetheparticlesareinsidethecell,theycangeneratereactiveoxygenspecies(ROS)[93],

disruptmembranes [94], causeproteinaggregation [95], interferewith cellularpathways [96],

unfoldproteins[97],orintercalateintoDNA[98].

Ontheotherhand,theinteractionwithbiomoleculesandthenano‐bio‐interfaceisofhighinterest

for medical applications. In particular, the high free surface energy and therefore strong

adsorptionofbiomoleculesisexploited.Hence,nanoparticlesdonotexistasbareparticlesina

biologicalenvironment,butarecoveredbyasurfacelayerofbiomolecules,whichisnamedthe

INTRODUCTION NANOPARTICLES

28

corona[99].Theeffectofthiscoronaisessentialforbiomedicalresearchandmedicinetoimprove

thepropertiesofpharmaceuticaldrugs.Inmostcases,medicinalapplicationsofnanoparticlesaim

at cancer treatment and improved drug delivery. Examples are (i) alginate coated chitosan

nanoparticles,whichenhancetheoralabsorptionofenoxaparin,alowmolecularweightheparin

[100],(ii)specificallydesignedsilicaparticlestodeliverpeptidemimeticstotransplantedcells

[101],and(iii)recentlydescribedgoldnanorodsof5‐15nmdiameterwithpromisingresultsin

thethermaltherapyofcancerinmice[102].

Asthevasculatureoftumorsisusuallyleaky,nanoparticlescanaccumulateintumorsduetoa

phenomenon termed enhanced permeability and retention (EPR) effect [103]. Especially the

accumulationofnanoparticlesinsolidtumorswithareducedlymphaticdrainageanddeficient

vasculaturecanhelptoreversemultidrugresistanceduetoanimproveddrugdeliverytocancer

cells [104].Recently,alsoL‐asparaginasewasentrappedinhydrogel‐magneticnanoparticlesof

30nmdiameterforimprovedbioavailabilityandcellpenetration[105].Moreover,anothergroup

showedthat200nmsilicashellscanbeusedtoovercomeallergicreactionsagainstasparaginase

whileretainingtheenzymeactivity[106].

Anothermedicalapplicationfornanoparticlesisthediagnosisofdiseasesinbiochemicalassays.

Recently, a gold nanoparticle‐based immunoassay was described to detect the Alzheimer’s

disease relatedβ‐amyloidpeptide in cerebrospinal fluids [107]. Even lung cancerdiagnosis in

exhaled breath was reported. For this, chemiresistors (microsensors) based on 5 nm gold

nanoparticlescoveredwithdifferentorganicsubstanceswereusedtoquantifyvolatileorganic

compoundsintheexhaledbreath[108].

The versatile applicability of nanoparticles furthermore allows the inhibition of enzymes, as

shown for soybean peroxidase and chymotrypsin with carbon nanotubes [109]. However,

interactionsofnanoparticleswithTaspasearenotdescribedsofar.

INTRODUCTION AIMSOFTHISTHESIS

29

1.5 AimsofthisthesisAlthoughthesurvivalrates for leukemiahavesignificantly improvedinthe lastdecades [110],

leukemiasstillaccount formore thanone thirdofall cancerdeaths inchildrenand9%ofall

cancerdeaths[1].Therefore,newanticancertargetsareofhighinterest,astheyallowthedesign

ofnovelanticancerdrugs.Recently,thenon‐oncogeneaddictionproteaseTaspasehasmovedinto

focusasnewtargetforcancerdrugs,sinceTaspaseisrequiredfortheinitiationandprogression

of certain types ofmixed lineage leukemia (MLL) and is upregulated in various solid tumors

[71,76].Inthisrespect,thesetumorsarehighlydependentonTaspaseandthushypersensitiveto

alossofTaspaseactivity[76].DespitefirstachievementsinthedirectionofTaspaseinhibitors,

the compoundsoriginating from these studies exhibit eitherweak inhibition [69,71] or broad

bioactivities[63].Hence,potentTaspaseinhibitorsarestillofparamountimportance.

Todevelop novel Taspase inhibitors, a comprehensive analysis of the biochemical properties,

suchasactivity,substratespecificity,andstability, ishelpfulandcangivedeeper insights into

Taspase‐mediatedcatalysis.Tothisend,Taspaseaswellasactiveandinactivemutantswillbe

characterized.Moreover, the structure of the long loop regions lacking in the Taspase crystal

structurewill be assessedbynuclearmagnetic resonance (NMR) and circulardichroism (CD)

spectroscopy,asthestructureoftheloopcouldrevealnewstructuraltargetsforinhibitors.The

resultingdatawillbeusedtoevaluateTaspaseinhibitorsandunderstandtheirmechanisms.

Toobtainnovelinhibitors,threeapproacheswillbeemployed:

Compoundswill be derived from virtual docking of a databank comprising 12million

smallmolecules.

Silicabasednanoparticles,whichwererecentlyfoundtobindproteins,willbetestedfor

sizedependenceandinhibitionmechanism.

Substrateanalogsuccinimidepeptideswillbetestedandoptimizedincooperationwith

theKaisergroupforchemicalbiology.

First,theinhibitorswillbetestedinafluorogenicinvitroassayandsubsequentlyinvivousingcell

lysatesandacell‐basedactivityassay.TheresultingdatacanyieldbothnovelclassesofTaspase

inhibitorsandassistthedesignoffutureTaspaseinhibitors.Additionally,thesestudiescouldshed

lightonthecatalyticmechanismandmightcontributetounderstandtheinteractionofenzymes

withnanoparticlesingeneral.

MATERIALANDMETHODS MATERIAL

30

2 MaterialandMethods

2.1 Material

2.1.1 Chemicals

Allchemicalsusedforthisprojectarelistedintable2.1.Namescandifferfromtheconventional

names.Somenamesaretrademarksandunderliecopyrightrestrictions.

Table2.1Chemicalsused.

Name(abbreviation) SupplierAceticacid Caldic,DüsseldorfAcetonitrile Caldic,DüsseldorfAdenosine‐5‐triphosphatedisodiumsalt(ATP) Roth,KarlsruheAgar Roth,KarlsruheAgarose Roth,Karlsruhe15NAmmoniumchloride(15NH4Cl) CambridgeIsotopes,AndoverAmmoniumhydrogencitrate(DAC) Fluka,BuchsAmmoniumperoxydisulfate(APS) Roth,KarlsruheAmpicillinsodiumsalt(Amp) Roth,KarlsruheAmsil8,Amsil20,Amsil125 Nyacol,AshlandAntifadegold™ Lifetechnologies,DarmstadtAntifoam™ Sigma‐Aldrich,SteinheimAtto488maleimide Attotec,SiegenAtto594NHSester Attotec,SiegenBiotin AppliChem,DarmstadtBoricacid(H3BO3) Roth,KarlsruheBovineserumalbumin(BSA) NEB,IpswichBromophenoleblue Roth,KarlsruheBuffergreenDNApolymerase(10x) ThermoScientific,DarmstadtBuffergreenrestrictionenzymes(10x) Fermentas,St.Leon‐RotCalciumchloridedihydrate(CaCl2) Roth,KarlsruheCasyTonsolution™ OMNILifeScience,BremenChloramphenicol(Cam) Roth,KarlsruheCobalt(II)chloridehexahydrate(CoCl2) Fluka,BuchscOmpleteproteaseinhibitor™ Roche,BaselCoomassieBrilliantBlueG250 Fluka,BuchsCopper(II)chloride(CuCl2) Sigma,SteinheimDesoxyribonucleotidetriphosphates(dNTPs) Metabion,MartinsriedDeuteriumoxide(D2O)99,9% Sigma‐Aldrich,SteinheimDihydroxyacetonephosphatelithiumsalt(DHAP) Sigma‐Aldrich,Steinheim2,2‐Dimethyl‐2‐silaptane‐5‐sulfonate(DSS) Sigma‐Aldrich,SteinheimDimethylsulfoxide(DMSO) Roth,Karlsruhe1,4‐Dithiothreitol(DTT) Biomol,HamburgDulbecco’sModifiedEagle’sMedium(DMEM) PAN‐Biotech,AidenbachEthanol Caldic,DüsseldorfEthylendiaminetetraaceticacid(EDTA)disodiumsalt Roth,KarlsruheFetalcalfserum(FCS) PAN‐Biotech,Aidenbach


31

Name(abbreviation) SupplierFmoc‐valine Novabiochem,DarmstadtGelFiltrationHMWCalibrationKit GEHealthcare,SolingenGelFiltrationLMWCalibrationKit GEHealthcare,SolingenD‐Glucose Roth,Karlsruhe13C6D‐Glucose CambridgeIsotopes,AndoverL‐Glutathionereduced(GSH) Roth,KarlsruheGlycerol Roth,KarlsruheGlycine AppliChem,DarmstadtHellmanexII™ HellmaAnalytics,MüllheimHoechst33258 Lifetechnologies,DarmstadtHydrochloricacid(HCl) AppliChem,Darmstadt4‐(2‐Hydroxyethyl)piperazine‐1‐ethanesulfonicacid(HEPES) Roth,KarlsruheImidazole Fluka,BuchsImmersionoil Roth,KarlsruheIron(III)citrate Sigma‐Aldrich,SteinheimIsopropanol Caldic, DüsseldorfIsopropyl‐β‐D‐thiogalactopyranoside(IPTG) Applichem,DarmstadtKanamycinsulfate(Kan) Roth,KarlsruheMagnesiumchloride(MgCl2) Roth,KarlsruheMagnesiumsulfate(MgSO4) Fluka,BuchsManganese(II)chloride(MnCl2) Fluka,BuchsMeatextract Fluka,Buchs2‐Mercaptoethanol Roth,KarlsruheMidoriGreen™ NipponGenetics,DürenMinimumEssentialMedium(MEM;1000x) PAN‐Biotech,AidenbachNickel(II)chloride(NiCl2) Roth,KarlsruheNicotineamideadeninedinucleotide(NADH) Roth,KarlsruheOptimem Lifetechnologies,DarmstadtOrangeG Sigma‐Aldrich,SteinheimParaformaldehyde(PFA) Sigma‐Aldrich,SteinheimPCRwater Lifetechnologies,DarmstadtPenicillin‐streptomycin Lifetechnologies,DarmstadtPeptone Fluka,BuchsPhenylmethylsulfonylfluoride(PMSF) Serva,HeidelbergPhosphoenolpyruvicacid(PEP) Sigma‐Aldrich,SteinheimPotassiumchloride(KCl) Roth,KarlsruhePotassiumdihydrogenphosphate(KH2PO4) Roth,KarlsruheReactionbufferT4ligase Metabion,MartinsriedRotiphoreseGel30 Roth,KarlsruheRubidiumchloride(RbCl2) Sigma‐Aldrich,SteinheimSodiumchloride(NaCl) Roth,KarlsruheSodiumdihydrogenphosphate(NaH2PO4) Roth,KarlsruheSodiumdodecylsulfate(SDS) Roth,KarlsruheSodiumhydroxide(NaOH) Roth,KarlsruheSodiummolybdate(Na2MoO4) Roth,KarlsruheSodiumselenite(Na2SeO3) Fluka,BuchsD‐Sucrose Roth,KarlsruheN,N,N',N'‐Tetramethylethylenediamine(TEMED) Roth,Karlsruhe


32

Name(abbreviation) SupplierThiaminehydrochloride(vitaminB1) Merck,DarmstadtTricine Roth,KarlsruheTrifluoroaceticacid(TFA) Caldic,DüsseldorfTris(hydroxymethyl)aminomethane(Tris) AppliChem,DarmstadtTritonX‐100 AppliChem,DarmstadtTrypsin‐EDTA PAN‐Biotech,AidenbachTryptone Roth,KarlsruheTween20 Roth,KarlsruheXtremeGene™ Roche,BaselYeastextract Roth,KarlsruheZINCdatabasecompounds ChemDiv,SanDiegoZincsulfate(ZnSO4) Roth,Karlsruhe

2.1.2 Purchasedpeptides

Purchasedpeptidesusedforthisthesisarelistedintable2.3.Allpeptidesweresynthesizedby

themanufacturersusingsolidphasepeptidesynthesisandpurifiedto>95%purityusinghigh‐

pressureliquidchromatography(HPLC).

Table2.2Peptidesused.

Sequence* Description Supplier[Mca]-GKGQVDGADDK-[DNP] CS1Taspasesubstrate ChinaPeptides,Shanghai[Abz]-KISQLDGVDDK-[DNP] CS2Taspasesubstrate CASLOApS,Lyngby[Abz]-KISQLAAVDDK-[DNP] MutatedTaspasesubstrate CASLOApS,Lyngby[Succ]-AAPF-[pNA] Chymotrypsinsubstrate Bachem,BubendorfAPPNIMTTRFSLAAFKRNKRKLELAERV

DTDFMQLKKRRQSS ShortTaspaseloop CASLOApS,Lyngby

*[Mca]:7‐amino‐4‐methylcoumarin;[Dnp]:dinitrophenol;[Abz]:anthranilicacid;[Succ]:succinate;[pNA]:p‐nitroaniline

2.1.3 Buffersandmedia

Buffersandmediausedforthisthesisare listedintable2.3.pHvalueswereadjustedatroom

temperature.Mediawereautoclavedforsterilization.

Table2.3Buffersandmediaused.

Solution CompositionBradfordreagent 5 %ethanol,10 %H3PO4,0.01 %CoomassiebrilliantblueG250,

ddH2O4xLaemmlistackinggelbuffer 500 mMTris/HClpH6.8,ddH2O4xLaemmliseparatinggelbuffer 3 MTris/HClpH8.8,ddH2O2xLaemmlisamplebuffer 125 mM Tris/HCl pH 6.7, 50 % glycerin, 10 % SDS, 10 mM

2‐mercaptoethanol,0.01%bromophenoleblue,ddH2O


33

Solution CompositionLaemmlibuffer 0.1 %SDS,0.303 %Tris,1.441 %glycine,ddH2OCoomassiestainingsolution 1.25 gCoomassiebrilliantblueG250,227mlethanol,46ml

aceticacid,227mlddH2OCoomassiedestainingsolution 50 mlaceticacid,75 mlethanol,875 mlddH2OPBS 137 mMNaCl,2.7 mMKCl,10 mMNaH2PO4,2mMKH2PO4,

ddH2OTBSTbuffer 50 mMTris‐HCl,150 mMNaCl,1 %Tween‐20,ddH2O,pH7.5TAEbuffer 40 mMTrispH 8.5,1 mMEDTA,20 mMaceticacid,ddH2OLysisbuffer 25 mMTrispH7.4,150 mMKCl,5 mMMgCl2,5%glycerol,1 %

TritonX‐100,2mM2‐mercaptoethanol,1mMPMSF,1tabletcOmplete,ddH2O

GSTwashbuffer 50 mMNaH2PO4,450 mMNaCl,1 mMDTT,1%TritonX‐100,pH7.3

GSTelutionbuffer 50 mMNaH2PO4,450 mMNaCl,1 mMDTT,20mMglutathione,pH8.0

GSTequilibrationbuffer 50 mMNaH2PO4,450 mMNaCl,1 mMDTT,1%TritonX‐100,pH7.3

NiNTAequilibrationbuffer 50 mMNaH2PO4, 450 mMNaCl,10 mMimidazole,pH8.0NiNTAwashbuffer 50 mMNaH2PO4,450 mMNaCl,20 mMimidazole,pH8.0NiNTAelutionbuffer 50 mMNaH2PO4,450 mMNaCl,250 mMimidazole,pH8.0Gelfiltrationbuffer 50 mMNaH2PO4,450 mMNaCl,1 mMDTT,pH7.9Activityassaybuffer 50 mMNaH2PO4,10 %sucrose,1 mMDTT,pH7.950xTS2 2.4 mMZnSO4,0.15 mMMnCl2,4.9 mMH3BO3,0.84mMCoCl2,

0.84mMNiCl2,58µMCuCl2,4.3mMNa2MoO4,0.12mMNa2SeO4,ddH2O

LBmedium 1 %tryptone,0.5 %yeastextract,1 %NaCl,ddH2O,pH7.0LBagarplates 500 mlLBmedium,7.5 gagarM9medium 0.68 %Na2HPO4,0.3 %KH2PO4,0.05 %NaCl,0.1%NH4Cl,0.2 %

1MMgSO4,0.2%50xTS2,0.1%10mMiron(III)citrate,0.4%glucose,0.01%CaCl2,0.01%0.5%vitaminB1,ddH2O

Nutrientmedium 0.5 %peptone,0.3 %meatextract,ddH2OSOCmedium 2 %tryptone,0.5 %yeastextract,10 mMNaCl,2.5mMKCl,

10mMMgCl2,10mMMgSO4,20mMglucose,ddH2OHelamedium 10 %FCS,1 %MEM,1 %penicillin‐streptomycin,DMEMHCT116/HEK293Tmedium 10 %FCS,1 %penicillin‐streptomycin,DMEM

2.1.4 Kits

Allkitsusedarelistedintable2.4.

Table2.4Kitsused.

Kit SupplierLightningQCsite‐directedmutagenesiskit Agilent,SantaClaraNucleoSpinPlasmid MachereyNagel,DürenNucleoSpinExtractIIkit MachereyNagel,DürenPierceECLpluskit ThermoScientific,Waltham


34

2.1.5 Instruments

Instrumentsanddevicesusedarelistedintable2.5.Standardlaboratoryequipmentisnotlisted.

Table2.5Instrumentsanddevicesused.

Description Instrument ManufacturerAgarosegelanalyzer BioDocAnalyze Biometra,GöttingenCasyTT Cellcounter OMNILifeScience,BremenCDspectrometer J‐710CDspectrometer Jasco,Gross‐UmstadtCentrifuge Centrifuge5415R Eppendorf,HamburgCleanbench HERAsafe ThermoScientific,LangenselboldCuvettes Quartzcuvettes(104‐QS,108F‐

QS,105.251‐QS,110‐QS)HellmaAnalytics,Müllheim

Developerforfilms Cawomat2000 Cawo,SchrobenhausenDNAgelsystem MiniSubcellGT BioRad,BerkeleyElectroporationcuvette Electroporationcuvette,0.2 cm BioRad,BerkeleyElectroporationdevice MicroPulser BioRad,BerkeleyFluorescencemicroscope BX61fluorescencemicroscope Olympus,HamburgFluorescencespectrometer CaryEclipse AgilentTechnologies,WaldbronnFPLCsystem BioRadDuoFlow BioRad,BerkeleyFPLCsystem ÄktaFPLCP920 GEHealthcare,SolingenIncubator BD53 Binder,USAIncubator ThermomixerHT Infors,BottmingenLC‐MSESI‐TOFsystem LCQFleetESIspectrometer ThermoScientific,DarmstadtMALDI‐MSsystem Autoflexspeed BrukerDaltonics,BremenMicrowave SharpMSTdevice MonolithNT.115 NanoTemper,MunichMSTdevice(labelfree) MonolithNT.LabelFree NanoTemper,MunichNanodrop NanoDropND1000 Peqlab,ErlangenNMRspectrometer Ultrashield700NMR BrukerBioSpin,FällandenPCRcycler T3000thermocycler Biometra,GöttingenpHmeter SevenEasy Mettler‐Toledo,GießenPhotometer BioPhotometer Eppendorf,HamburgPowersupply PowerPacbasic BioRad,BerkeleyScanner Scanjet8200 HewlettPackard,ChinaSDSgelshaker Duomax1030 Heidolph,SchwalbachSDSgelsystem XCellMiniProtean3 BioRad,BerkeleyThermomixer Thermomixercomfort Eppendorf,HamburgTriggeredFviewIIFWcamera

Fluorescencemicroscopycamera Olympus,Hamburg

Ultracentrifuge OptimaL‐90K Beckmann,PaloAltoUltrasounddevice Sonopuls Bandelin,HamburgUV/Visspectrophotometer Cary100 AgilentTechnologies,WaldbronnWaterfilterunit Milli‐QBiocel Millipore,Schwalbach


35

2.1.6 Consumables

Consumablesusedarelistedintable2.6.Standardlaboratoryequipmentisnotlisted.

Table2.6Consumablesused.

Description Name SupplierCellculturedishes 100x20 mm GreinerBioOne,SolingenCentricon AmiconUltraconcentrator Millipore,SchwalbachCentricon Concentrator5301 Eppendorf,HamburgCoverslips 15x15mm Menzel,BrunswickDesaltingtips SupelTipsC18 Sigma‐Aldrich,SeelzeGelfiltrationcolumn HiLoad16/600Superdex200pg GEHealthcare,SolingenGelfiltrationcolumn HiLoad26/600Superdex200pg GEHealthcare,SolingenGelfiltrationcolumn 10/300GLSuperdex200pg GEHealthcare,SolingenGSTcolumn GSTrapFF16/10,20 ml GEHealthcare,SolingenNiNTAcolumn ProtinoNiNTAAgarose MachereyNagel,DürenMALDItargetframe MTP384groundsteel BrukerDaltonics,BremenNAPcolumn PD‐10gravityflowcolumn GEHealthcare,SolingenNitrocellulosemembrane Protran Whatman,DasslNMRtubes AP5‐800‐7NMRtubes SchottEconomics,MainzObjectslides SuperfrostPlus ThermoScientific,DarmstadtSterilefilter 0.22µm Millipore,TullagreenX‐Rayfilms CL‐Xposurefilms ThermoScientific,Darmstadt

2.1.7 Software

Softwareusedislistedintable2.7

Table2.7Softwareused.

Software Application ReferenceCcpNmrAnalysis2.3 Visualization,editing andevaluationof

NMRspectra[111]

ACD/ChemSketch12.00

Visualizationofchemicalstructures,calculationoflogP

AdvancedChemistryDevelopment,Toronto

Chimera Visualizationofproteinstructures,preparationofstructuresfordocking

[112]

ClustalO Sequencealignments [113]DOCK6 Virtualdocking [114]FlexControl MALDImassspectra BrukerDaltonics,BremenGraphPadPrism5.04 Statisticaldataevaluation [115]ImageJ1.47v Evaluationofmicroscopyimages [116]NanoTemperanalysissoftware

NanoTemperanalysissoftware NanoTemper,Munich

OligoCalc Calculationofprimerproperties [117]CellP Florescencemicroscopyimaging Olympus,HamburgOPTIMIZER E.colioptimizationofDNAsequences [118]


36

Software Application ReferenceProtparam Calculationofproteinproperties [119]PyMOL1.4 Visualizationofproteinstructures [120]SIAS Calculationofsequencehomology [121]SnapGeneViewer2.2.2

Visualizationofplasmidmaps GSLBiotech,Chicago

SpectraManager2.0 EvaluationofCDspectra Jasco,Gross‐UmstadtSwiss‐Model Homologymodeling [122]Topspin3.0 RecordingandprocessingofNMRspectra BrukerBiospin,RheinstettenYASARATwinset11.12.31/WHATIF

Visualizationofproteinstructures,MDsimulations,homologymodeling

YASARABiosciences,Vienna

2.1.8 Enzymesandantibodies

Enzymesandantibodiesusedarelistedintable2.8andwerestoredat‐20°C.

Table2.8Enzymesandantibodiesused.

Enzyme SupplierApaI Fermentas,St.Leon‐RotChymotrypsin Sigma‐Aldrich,SteinheimDreamTaqGreenDNApolymerase ThermoScientific,DarmstadtLacticdehydrogenase Sigma‐Aldrich,SteinheimLysozyme Fluka,BuchsParvalbumin madein‐housePreScissionprotease madein‐houseProteinaseK NewEnglandBiolabsT4ligase Metabion,MartinsriedTrypsin Sigma‐Aldrich,SteinheimXhoI Fermentas,St.Leon‐Rot

Actin(ACTB/ACTC)polyclonalantibody(N‐term) Abgent,HeidelbergTASP1monoclonalantibody Origene,Rockville

2.1.9 Bacterialstrains

Bacterialstrainsusedforplasmidamplificationandproteinexpressionaregivenintable2.9.

Table2.9Bacterialstrainsused.

Strain Genotype SupplierE.coliElectroSHOXcompetentcells

F‐mcrAΔ(mrr‐hsdRMS‐mcrBC)ɸ80lacZΔM15ΔlacX74recA1endA1araΔ139Δ(ara,leu)7697galUgalK‐rpsL(StrR)nupGλ‐

Bioline,Luckenwalde

E.coliBL21(DE3)‐T1R F‐ompThsdSB(rB‐ mB‐)galdcmλ(DE3)tonA Sigma‐Aldrich,SteinheimE.coliC43pLysDE3 F–ompThsdSB(rB‐ mB‐)galdcm(DE3)

pLysS(CmR)Lucigen,Heidelberg

E.coliRosetta2(DE3) F‐ompThsdSB(rB‐ mB‐)galdcm (DE3)pRARE2(CamR)

Merck,Darmstadt


37

2.1.10 Plasmids

Vectormapsof theplasmidsused in thisworkcanbe found in thesupplement(supplemental

figure 7.1 and figure 7.2). The plasmidwith human Taspase cDNA and Taspase plasmids for

eukaryotictransfectionwerecourtesyoftheKnauergroupandthepRARE2plasmidwascourtesy

of the Ehrmann group. For cloning with GST tag, a modified pET‐41b(+) vector (Merck,

Darmstadt)wasused.Inthisvector,aPreScissioncleavagesiteandanApaIcleavagesitewere

insertedinthemultiplecloningsiteaftertheGSTtagbetweentheBamHIandSacIIcleavagesites.

TheoriginalApaIcleavagesite in the lacI genewasmutated(figure2.1).With thisvector, the

targetproteinhasonlyanN‐terminalGly‐ProadditionafterPreScissioncleavage.

Figure2.1VectormapofthemodifiedpET‐41bvector.

TheApaIrestrictionsite(red)wasmutatedandaPreScissioncleavagesitewasinsertedbetweentheSacII(green)andBamHI(blue)restrictionsite.Theinsertedsequenceisshownatthebottom.

MATERIALANDMETHODS MOLECULARBIOLOGICALMETHODS

38

2.1.11 ProteinandDNAmarker

PageRulerUnstainedBroadRangeProteinLadder(ThermoScientific;figure2.2a)andPageRuler

PlusPrestainedProteinLadder(Metabion;figure2.2b)wereusedasproteinmarkers.mi‐100bp+

DNAmarkerGo(Metabion;figure2.2c)andmi‐1kb(‐)DNAMarkerGo(Metabion;figure2.2)were

usedasDNAmarker.

Figure2.2ProteinandDNAmarker.

aPageRulerUnstainedBroadRangeProteinLadder forSDS‐PAGE.bPageRulerPlusPrestainedProteinLadderforSDS‐PAGE.cmi‐100bp+DNAmarkerGoforagarosegelelectrophoresis.dmi‐1kb(‐)DNAMarkerGoforagarosegelelectrophoresis.

2.2 Molecularbiologicalmethods

2.2.1 CloningofactiveTaspaseandtheinactivemutant

ThenucleotidesequenceofthebicistronicconstructreferredtoasactiveTaspasewasdesigned

insilicobasedontheconstructdescribedbyKhanetal.[56].Theexpressionconstructcomprises

the α‐subunit with an N‐terminal hexahistidine tag, followed by a stop codon and a second

ribosomalbindingsitefortheβ‐subunitfollowingdownstream(figure2.3).Theunstructuredloop

attheC‐terminusoftheα‐subunitwasshortenedby27aminoacids,endingatAla206.Thecodon

usageofthewholesequencewasoptimizedforexpressioninE.coliusingOPTIMIZER[118].The

genewassynthesizedandclonedintoamodifiedpET‐41bvectordescribedinfigure2.1(GeneArt,

Regensburg).


39

Figure2.3ConstructforactiveTaspase.

TheconstructcomprisesanN‐terminalHis‐tag,followedbytheshortenedα‐subunit(aminoacids1‐206)andastopcodon.Afterashortlinkersequence,theribosomalbindingsite(RBS)enablesthetranscriptionoftheβ‐subunit(aminoacids234‐420).ThisconstructwasclonedinamodifiedpET‐41bvector.

TheinactivemutantisbasedonthewildtypeconstructofhumanTaspasecDNAinapET‐22b

vectorwithN‐terminalhexahistidinetaggeneratedbyJensRabenstein[82].Twomutationsinthe

active site (D233A and T234A) prevent autocatalytic activation of the proenzyme and hence

reduce catalytic activity. These mutations were inserted with the help of a QuikChange kit

(Agilent, Santa Clara) using wild type Taspase as template. Primers were designed with

NEBaseChanger (NEB, Ipswich; table 2.10), synthesized (Metabion,Martinsried) and used for

QuikChangeaccordingtothemanufacturer’sprotocol(table2.11)withaPCRprogramdescribed

intable2.12.

Table2.10PrimersusedtogeneratetheinactiveTaspasemutant.

Substitutedbasesarehighlightedinyellow.

Primer SequenceForward 5’-CTCAGGCACTTTGGCGGCGGTAGGCGCTGTGG-3’

Reverse 5’-CCACAGCGCCTACCGCCGCCAAAGTGCCTGAG-3’

Table2.11PCRcomponentstogeneratetheinactiveTaspasemutant.

Component Amount[µl]10xbuffer 2.5WildtypeTaspase(template) 1(=100pg)Forwardprimer 0.6(=125ng)Reverseprimer 0.6(=125ng)dNTPmix 0.5(=0.5mM)QuikChangereagent 0.75DNApolymerase 0.5PCRwater 19.2

Table2.12PCRprogramtogeneratetheinactiveTaspasemutant.

Temperature[°C] Time[min] Numberofcycles95 2:00 195 0:20

2568 0:1068 4:0068 5:00 137 ∞


40

ThePCRproductwastreatedwith1µlDpnIfor20minutesat37°CtodigesttemplateDNAand

thentransformedintocompetentE.colishoxcells(section2.3.2).Threecloneswereselectedfor

DNAisolationusingtheMiniPrepkit(MachereyNagel,Düren)accordingtothemanufacturer’s

instructionsandtheconstructwassequenced(GATCBiotech,Cologne)toverifythecorrectDNA

sequence(supplementalfigure7.1b).Permanentplasmidstocksinshoxcellswereobtainedby

mixing600µlofanovernightculturewith400µlofsterile86%glycerol.Thestocksolutionwas

shockfrozenandstoredat‐80°C.

2.2.2 CloningoftheTaspaseloop

The construct for the GST‐tagged Taspase loop comprising 56 amino acids (G178‐D233)was

generated by amplification of the loop sequence using wild type Taspase as template and

subsequentlyclonedintoamodifiedpET‐41vectorwithN‐terminalGSTtag(supplementalfigure

7.3andfigure2.1).

PCRamplificationoftheinsert:

ForPCR, theprimerpair listed in table2.13wasusedwith theplasmid forwild typeTaspase

(supplementalfigure7.1aandfigure2.1)astemplate.ThepipettingschemeandPCRprogramare

listedintable2.16andtable2.17,respectively.Afteramplification,a1.3%agarosegelwasused

toverifythecorrectsizeof the fragmentandtheDNAwasextractedusingagelextractionkit

(MachereyNagel,Düren)accordingtothemanufacturer’sinstructions.

Table2.13PCRprimersusedtogeneratetheTaspaseloop.

Primer Sequence Tm[°C]Forward 5’-CACACAGGGCCCGGAATACCGTCTTGCCCTCC-3’ 59.5

Reverse 5’-GGTTGGCTCGAGCTTAGTCCAAAGTGCCTGAGTCGTTC-3’ 61.2

Restrictionandligation:

BoththeamplifiedTaspaseloopandthetargetvectorpET‐41‐PreScweredigestedwithApaIand

XhoIfor30minutesatroomtemperature(table2.14).

Table2.14ApaI/XhoIrestrictionassay.

Component Vector[µl] Insert[µl]pET‐41‐PreScDNA 2(=300 ng) ‐LoopDNA ‐ 8(=2 µg)ApaI 1(=1 FDU) 1(=1 FDU)XhoI 1(=1 FDU) 1(=1 FDU)10xbuffer 3 5PCRwater 23 35

AfterDNAcleanup(MachereyNagel,Düren),vectorandinsertweremixedinamolarratioof1:3

andligatedfor1hat16°C(table2.15).


41

Table2.15LigationoftheTaspaseloopandpET‐41‐PreSc.

Component Vector[µl]pET‐41bDNA 1.26(=50 ng)LoopDNA 0.95(=4.33 ng)T4ligase 0.5(=125 U)10xbuffer 2PCRwater 15.3

Subsequently,1µloftheligatedDNAwasusedfortransformationintoelectrocompetentE.coli

shoxcellsandthecellswereplatedonLBmediumwithkanamycin.

Table2.16PCRcomponentstoamplifytheTaspaseloop.

Component Amount[µl]WildtypeTaspase(template) 0.5(=10ng)10xbuffer 5Forwardprimer 5(=50pg)Reverseprimer 5(=50pg)dNTPmix 2.5(=0.5 mM)DMSO 0.5(=1 %)DNApolymerase 1(=5 U)PCRwater 30.5

ColonyPCR:

TencoloniesweretestedforloopDNAemployingacolonyPCR(table2.18)withthesamePCR

programusedfortheamplificationoftheloop(table2.17).

Table2.17PCRprogramusedtoamplifytheTaspaseloop.

Temperature[°C] Time[min] Numberofcycles95 0:30 195 0:3056 0:30 372 0:3095 0:3058 0:30 3272 0:3072 10:00 14 ∞

MATERIALANDMETHODS MICROBIOLOGICALMETHODS

42

Table2.18ColonyPCRcomponentstoamplifytheTaspaseloop.

Component Amount[µl]Cellsuspension(template) 1010xbuffer 4Forwardprimer 2(=20 pg)Reverseprimer 2(=20 pg)dNTPmix 1(10 mM)DNApolymerase 0.5(2.5 U)PCRwater 20

Theplasmidsof cloneswith correct sizewere isolatedusing aMiniPrep kit (MachereyNagel,

Düren)andverifiedbysequencing.

2.3 Microbiologicalmethods

2.3.1 Preparationofcompetentcells

Competentcellswerecourtesyofothergroupmembers.Forthepreparationofelectrocompetent

cells,500mlnutrientmediumwere inoculatedwith1%of anovernight cultureof shoxcells

(Bioline,Luckenwalde)andcultivatedtoanOD600of0.8at37°Cand160rpm.Thecellswere

cooleddownonicefor15minutesandpelletedat4000gfor15minutesat4°C.Thepelletwas

resuspendedin500mlcoldddH2Oandcentrifugedagain.Thepelletwaswashedoncemoreby

resuspensionin250mlcoldddH2Oandcentrifugation,afterwardsresuspendedin20ml10%

glycerol,centrifugedandfinallyresuspendedin0.5ml10%glycerol.50µlaliquotswereshock

frozeninliquidnitrogenandstoredat‐80°C.

Chemically competent BL21(DE3)T1r cells were generated starting from 200ml LBmedium

inoculatedwith2mlofanantibiotic‐freeovernightculture(160rpm,37°C).AtanOD600of0.6,

cellswerepelletedfor5minutesat3000rpm,4°C(slowstartandstop)andresuspendedin45ml

cold0.5MCaCl2solution.Afterincubationonicefor30minutes,cellswerepelletedagainand

resuspendedin2ml0.5MCaCl2solution.50µlaliquotswereshockfrozenandstoredat‐80°C.

2.3.2 Transformationofcompetentcells

Electrocompetentcellswereincubatedonicewith1µlofthePCRproductandelectroporated

subsequentlywith therespectivebacterialprograminaMicroPulser(BioRad,Berkeley).After

incubationfor60minutesin300µlSOCmediumat37°Cand400rpm,cellswereplatedonLB

agardishescontainingrespectiveantibioticsandincubatedovernightat37°C.

Chemically competent cells were incubatedwith 1 µl of plasmid DNA for 20minutes on ice,

followedbyaheatshockat42°Cfor45secondsandanotherincubationsteponicefor2minutes.

Afterthat,thecellsweretransferredto300µlofSOCmediumandincubatedfor60minutesat

37 °C and 400 rpm. Subsequently, the cells were plated on LB agar dishes with respective

antibioticsandincubatedat37°Covernight.


43

2.3.3 Expressiontest

TofindoptimalconditionsfortheexpressionandsolubilityofthedifferentTaspaseconstructs,

theirexpressionwasassayedaftervarious timesandatvarious temperatures.Theexpression

plasmidwastransformedintochemicallycompetentE.coliBL21(DE3)T1rcellsandgrownover

night.ForwildtypeTaspase,theE.colistrainRosetta2,whichisdesignedtoexpresseukaryotic

proteinswithcodonsrarelyusedinE.coli,wastestedinadditiontotheBL21(DE3)T1rstrain.

1.5%ofanLBovernight cultureof transformedbacteriawasused to inoculate100mlofTB

medium.AtanOD600of0.8,200µMIPTGwasaddedandthecellsuspensionwassplituptoallow

expressionatdifferenttemperatures.Beforeinductionandatdefinedtimesafterinduction,10ml

samplesweretakenfromthecellsuspensions,pelleted(4000g,4°C,20min)andfrozen.The

pellets were resuspended in 50 mM NaH2PO4, 450 mM NaCl, 10 mM imidazole, pH 8.0

(supplementedwith1%TritonX‐100 for theTaspase loop), lysedby sonification (3 cycles à

30secondsat30%intensity)andinsolublecell fragmentswereremovedat14000rpm,4°C,

90min.Thesupernatantwascollectedandthepelletswereresuspendedin500µlbuffer(50mM

NaH2PO4,450mMNaCl,10mMimidazole,pH8.0).ForSDS‐PAGE,20µlofsupernatantorpellet

solutionwereaddedto20µlof2xSDSsamplebuffer.

2.3.4 ExpressionofTaspaseconstructs

TheplasmidscontainingthesequenceforTaspase(sections2.2.1and2.2.2)weretransformed

intocompetentexpressionstrainsofE.coli.ForactiveTaspase,thepET‐41bplasmidencodingthe

activeTaspasewastransformedintotheC43(DE3)pLysSstrainasdescribedabove(section2.3.2).

CellswereplatedonLBmediumwith60µg/mlkanamycinand50µg/mlchloramphenicol.After

incubationovernight at 37 °C, a single clonewasused to inoculate anovernight culture (see

below).

ThesequencescodingforwildtypeTaspase,inactiveTaspaseandtheTaspaseloopwerebased

onhumancDNAandthereforenotoptimizedforexpressioninE.coli.Toovercomethis,1µlofa

pRARE2plasmid(Merck,Darmstadt)containingthetRNAforcodonsrarelyusedbyE.coliwas

co‐transformed additionally to the Taspase construct (in pET‐22b vector) intoBL21(DE3)T1r

chemicallycompetentcells(section2.3.2).PlatingonLBmediumwith140µg/mlampicillinand

50 µg/ml chloramphenicol allowed for selection of cells that had taken up both the Taspase

plasmid (with ampicillin resistance gene) and the pRARE2 plasmid (with chloramphenicol

resistancegene).

OvernightculturesweregrowninLBmediumwithrespectiveantibiotics.4 lTBmediumwith

antibioticswereinoculatedwith15mlovernightcultureperliterofmediumandbacteriawere

grownat37°Cand160rpm.Growthofbacteriawasmonitoredbymeasuringtheopticaldensity

at600nmwavelength.

ForNMRspectroscopy,15Nor15N/13Clabeledproteinwasused.Toachieveexpressionofisotopic

labeledprotein, first,1 lofLBmediumwas inoculatedwith15mlofanovernightcultureand


44

grownat37°C,160rpmuntiltheOD600reached0.7.Cellswerepelletedgently(20°C,3000rpm,

15 min) and resuspended in 4 l of M9 minimal medium supplemented with 15N ammonium

chlorideand/or13Cglucose.Bacterialgrowthwasallowedtocontinueat37°C,160rpm.

For both labeled and unlabeled Taspase protein expression, at OD600 of 0.8, expression was

inducedbyadditionof200µMIPTGandthetemperaturewasreducedto30°C.6hafterinduction,

cellswerepelletedat4°C,6300gfor20minutesandresuspendedin50mMNaH2PO4,450mM

NaCl, 10mM imidazole, pH8.0. TheTaspase loopwas expressed over night at 25 °C and the

bacteriapelletwasresuspendedin50mMNaH2PO4,450mMNaCl,1mMDTT,1%TritonX‐100,

pH 7.0. Cell pellets were shock frozen in liquid nitrogen and stored at ‐20 °C until used for

purification.

2.3.5 Cultivationofeukaryoticcelllines

Cervical cancer cells (Hela)were cultivated, cryopreserved and passaged as suggested by the

AmericanTypeCultureCollection(ATCC):Cellswerecultivatedin10cmdisheswith10mlcell

culturemediumandsplitat70to90%confluency.Forthis,cellswerewashedwithPBS,before

1ml of trypsin‐EDTAwas added and the cellswere incubated for 5minutes to detach. 9ml

prewarmedmediumwasaddedandcellsweredispersed,priortoseedinginasplitratioof1:5to

1:25infreshmedium.Mediumwaschangedasneeded.Forcryopreservation,2x106cells/mlwere

centrifuged gently (3min, 300 g, 25 °C) and resuspended in freezemedium containing 10%

DMSO.Aliquotsof1mlwereplacedinacontrolled‐ratefreezerprecooledto4°Candcooleddown

to‐70°Cwitharateof‐1°Cperminute.Forlong‐termstorage,cellsweretransferredtoa‐130°C

freezer.Recoveryofcryopreservedcellswasperformedbyrapidthawingto37°C,followedby

additionof1mlfreshmediumandgentlecentrifugation(3min,300g,25°C).Thesupernatant

wasdiscarded,cellswereresuspendedin10mloftheirrespectivemediumandtransferredinto

10cmdishes.Mediumwaschangedthenextday.

2.3.6 Transfectionofeukaryoticcellsandpreparationformicroscopy

Transfectionwasperformedin12wellplateswithcoverslipsonedayafterseeding2x105cells

perwell.CellswerecountedinaCasyTT(OMNILifeScience,Bremen)calibratedindividuallyfor

HelaandHEK293Tcells.Trypsin‐EDTAwasusedtodetachthecells,beforetheyweredispersed

in 9ml culturemedium. 50 µl of cellswere diluted in 10ml of CasyTon solution (OMNI Life

Science)andonlycountswithmorethan3000eventswereregardedasvalid.1µgDNAwasmixed

with3µlXtremeGeneTransfectionreagentin100µlOptimem(Lifetechnologies)mediumand

incubatedfor15minutesatroomtemperature.Thissolutionwasaddedtothecellswith900µl

oftheirmedium.Ifneeded,nanoparticleswereadded3hoursaftertransfection.

Onedayaftertransfection,cellswerewashedwithPBSand500µlof3.7%PFAwith0.1%Tween

wasaddedforfixation.Afterincubationfor20minutesat4°C,cellswerewashedtwicewithPBS

and1µg/mlHoechst33258inPBSwasaddedfor5minutestostainthenuclei.Subsequently,the

MATERIALANDMETHODS BIOCHEMICALMETHODS

45

cellswerewashed firstwithPBS and thenwithwater. 10 µlAntifade gold (Life technologies,

Darmstadt)wasspottedonamicroscopeslide,onwhichthecoverslipwasplaced.

2.3.7 Fluorescencemicroscopyandimageevaluation

AftertheAntifadegoldsolutionhaddriedovernight,fluorescencemicroscopyimagesweretaken

withanOlympusBX61fluorescencemicroscopeequippedwithatriggeredFviewIIFWcamera.

The microscope was operated by the Olympus CellP software. Exposure times for Hoechst,

Taspase‐mCherry,Taspase‐GFP,Substrate‐mCherryandSubstrate‐GFParelistedintable2.19.

Table2.19Channelsettingsusedforfluorescencemicroscopy.

Channel Excitationwavelength

Emissionwavelength

Exposuretime

Hoechst 345nm 455 nm 25 msGFP 480nm 510 nm 500 msmCherry 545nm 610 nm 500 ms

ImageswereprocessedwithImageJ[116]andcomprisedthefollowingsteps.First,cellborders

wereidentifiedwiththehelpofbright‐fieldandautofluorescence.Second,nucleiweredefined

using the DAPI channel. The area of the cell borders minus the nucleic area was defined as

cytoplasm.Theaverageintensitiesofthenucleicandcytoplasmicareaswerecalculated,andthe

ratio of nucleus to cytoplasm (N/C) was computed. Statistical evaluation of these ratios was

executedinGraphPadPrism5.03[115].Anunpairedtwo‐tailedstudent’st‐testwasapplied.p‐

valuesbelow0.05wereregardedassignificant,valuesbelow0.001ashighlysignificant.

2.3.8 Cellpenetrationassayswithpeptidicinhibitor

Uptake of rhodamine‐labeled peptidic inhibitor was assayed with Hela cells. 5 µM or 50 µM

fluorescentinhibitorwereaddedto2x105cellsin12wellplates.Cellswereeithercultivatedin

medium with or without serum and fixated 1 or 5 hours after addition of the inhibitor.

Fluorescencemicroscopy imageswere takenasdescribedabove (section2.3.7)using theRFP

channel and the fluorescence intensity of the cells was compared to the autofluorescence of

untreatedcellsusingImageJ.Forstatisticalevaluation,anunpairedtwo‐tailedstudent’st‐testwas

appliedusingGraphPadPrism.

2.4 Biochemicalmethods

2.4.1 Bacterialcelllysis

Bacterialpelletswerethawedrapidlyinawaterbathand1mMoftheproteaseinhibitorPMSF

wasadded,aswellasaspatulatipoflysozymeand0.001%antifoam.Afterstirringfor60minutes

at 4 °C, cells were lysed by sonification. For this, the cell suspension underwent 3 pulses of


46

sonification(30secondseach,60%intensity)onicewith1minuteintervalstoavoidexcessive

heat.Cellfragmentsandinsolubleproteinswerepelletedbyultracentrifugationfor70minutesat

142000g,4°Candthesupernatantwascollected.Afterfiltrationthrougha0.22µmfilter, the

volumewasadjustedto150mlandloadedintoaSuperloop(GEHealthcare,Solingen).

2.4.2 ProteinpurificationofHis‐taggedTaspase

Affinity chromatographywas performedwith anÄkta‐FPLC system (GEHealthcare, Solingen)

using a HisTrap HP 5ml NiNTA column (GE Healthcare) equilibrated with 50mM NaH2PO4,

450mMNaCl,10mMimidazole,pH8.0.Thedetailedbuffers,volumesandflowratesusedfor

affinitypurificationarelistedintable2.20.Inshort,theHis‐taggedproteinwasallowedtobindto

thecolumnmatrix,beforenon‐targetproteinswerewashedaway.Forelution,theconcentration

ofimidazolewasgraduallyincreasedtoamaximumof250mMimidazole.

Table2.20ProgramusedforNiNTAaffinitypurificationwithaHisTrapHP5mlNiNTAcolumn.

Step Buffer Volume Flowrate Sizeofcollectedfractions

Equilibration NiNTAequilibrationbuffer

25 ml 1.5 ml/ml ‐

Applicationofsupernatant

Supernatant(inNiNTAequilibrationbuffer)

150 ml 1 ml/min ‐

Wash NiNTAwashbuffer 50 ml 1 ml/min ‐Elutiongradient

GradientfromNiNTAwashbuffertoNiNTAelutionbuffer

100 ml 1 ml/min 3ml

Finalelution NiNTAelutionbuffer 20 ml 1 ml/min 5ml

Elutionfractionscontainingproteinwereidentifiedviaabsorbanceat280nmandthecorrectsize

oftheproteinwascheckedbySDS‐PAGE.Forthispurpose,20µlofanelutionfractionwasadded

to20µlof2xSDSsamplebufferandseparatedina15%SDSgel.AllfractionscontainingTaspase

wereconcentratedinaCentriconwith30kDacut‐offtoavolumeof2mlandusedforgelfiltration.

PreparativegelfiltrationswereperformedonanÄktaFPLCsystemwithaSuperdex200HiLoad

16/600gelfiltrationcolumn(GEHealthcare)toremovesolubleaggregatesandimidazole.Thegel

filtrationbuffercontained50mMNaH2PO4,450mMNaCl,1mMDTTatpH7.7.Flowratesand

fractionsizesusedintheprogramarelistedintable2.21.Theelutedproteinwasconcentratedin

Centriconswith30kDacut‐off,aliquotswereshockfrozeninliquidnitrogenandstoredat‐20°C.

Table2.21ProgramusedforgelfiltrationwithaSuperdex20016/600gelfiltrationcolumn.


Equilibration Gelfiltrationbuffer 130 ml 1.5 ml/ml ‐


47

Sampleinjection

ConcentratedTaspase(in50mMNaH2PO4450mMNaCl~125mMimidazolepH8.0)

5 ml 1 ml/min ‐

Elution Gelfiltrationbuffer 130 ml 1 ml/min 3ml

2.4.3 ProteinpurificationoftheGST‐taggedTaspaseloop

A20mlGSTrapcolumn(GEHealthcare,Solingen)wasequilibratedwithanÄkta‐FPLCsystem(GE

Healthcare)with50mMNaH2PO4,450mMNaCl,1mMDTT,1%TritonX‐100,pH7.0,beforethe

GST‐taggedTaspase loopwasbound,washedandelutedwithglutathionebuffer.Thedetailed

buffers,volumesandflowratesarelistedintable2.22.

Table2.22ProgramusedforGSTaffinitypurificationwitha20mlGSTrapcolumn.


Equilibration GSTequilibrationbuffer

100 ml 1.5 ml/ml ‐

Application ofsupernatant

Supernatant(inGSTequilibrationbuffer)

150 ml 1 ml/min ‐

Wash GSTwashbuffer 100 ml 1 ml/min ‐Elution GSTelutionbuffer 100 ml 1 ml/min 3 ml

AlthoughtheTaspaseloopitselfdoesnotabsorblightat280nmduetoitslackoftyrosinesand

tryptophans, theabsorbanceof theGST tag canbeused to identifyprotein containingelution

fractions.ThesefractionswereconcentratedinaCentriconwith10kDacut‐offtoafinalvolume

of5ml.ProteinconcentrationwasdeterminedusingUV/Visspectroscopy,assuminganextinction

coefficient for the GST‐loop fusion protein of 42860 M‐1 cm‐1. The PreScission cleavage site

betweentheGSTtagandtheTaspaseloopallowedspecificremovalofthetagbyadditionof20µg

PreScissionproteaseper6mgfusionproteinandincubationovernightat4°C.Thisrendersthe

cleavedTaspaseloopwithanN‐terminalGly‐ProextensionfromtheGSTtag.

FiltrationthroughaCentriconwith10kDacut‐offresultedinaretentionofthe28kDaGSTtag,

whilethe6.5kDaTaspaseloopcouldpassthroughthemembrane.Hence,theflow‐throughwas

collectedandconcentratedinanotherCentriconwith3kDacut‐off.Proteinconcentrationwas

determinedusingtheBradfordassay,aliquotswereshockfrozeninliquidnitrogenandstoredat

‐20°C.

2.4.4 Quantificationofprotein,peptideandDNAconcentrations

ProteinconcentrationsofpurifiedproteinweredeterminedviaUV/Visspectroscopyusingthe

Lambert‐Beer law. Extinction coefficients at 280 nm were calculated using the online tool


48

Protparam[119]andarelistedintable2.23.AbsorbancespectrawererecordedwithaNanoDrop

ND1000.

Protein concentrationsof eukaryotic cell lysatesweredetermined inaBradfordassay.1µlof

samplewasincubatedwith999µlBradfordreagentfor5minutesandabsorbancewasmeasured

at 595nmandablankvaluewas subtracted. Calibrationwith a bovine serumalbumin (BSA)

standardcurveallowedanestimationoftheproteinconcentrationinthesample.

ForthequantificationofinhibitorscontaininganFmocmoiety,theextinctioncoefficientofFmoc

wasdeterminedexperimentally.Forthis,theabsorbanceofaserialdilutionofFmoc‐valinewas

measuredat300nmwavelength.Linearregressionofthestandardcurveyieldedanextinction

coefficientof4985M‐1cm‐1.Thisvaluewasusedtoconverttheabsorbanceat300nmtoinhibitor

concentrations.

TheconcentrationofoverexpressedGFP‐Taspaseineukaryoticcelllysateswasobtainedusinga

combination of absorbance and fluorescence spectroscopy. This was necessary, because the

concentrationswere too low to be detected by absorbance spectroscopy. In a first step, cells

transfectedwithonlyGFPwerelysedandtheGFPconcentrationinthe lysatewasdetermined

with a Nanodrop ND 1000 applying an extinction coefficient of 9500 M‐1 cm‐1 at 475 nm.

Transfection and expression is much more efficient for GFP than for GFP‐Taspase and a

concentrationof15.8µMGFPinthecelllysatewascalculated.Fromthislysate,astandardcurve

for fluorescence emission at 509 nm was created (excitation wavelength: 395 nm;

excitation/emissionslits:10/10nm;PMT:600V).Inasecondstep,thefluorescenceemissionof

alysatewithoverexpressedGFP‐Taspasewasmeasured.Withthehelpofthestandardcurve,this

valuecouldbeconvertedtoaGFP‐Taspaseconcentrationof370nM.

Concentrations of the peptidic substrate used in the kinetic assays were determined via the

absorbance of the anthranilic acid. An extinction coefficient of 14650 M‐1 cm‐1 at 355 nm

wavelength was used. The concentration of the rhodamine‐labeled inhibitor was determined

likewise,assuminganextinctioncoefficientof80000M‐1cm‐1at555nm[123].

DNA concentrations were determined with a Nanodrop ND 1000, assuming an extinction

coefficientat360nmof0.02µgml‐1cm‐1.


49

Table2.23CalculatedparametersofTaspaseproteins.

Protein Extinction coefficientat280nm(ε280)

Molecularweight Aminoacids

wtTaspase 26930M‐1cm‐1 45278Da 420InactiveTaspase 26930M‐1cm‐1 45204 Da 420ActiveTaspase 26930M‐1cm‐1 42082 Da 399Taspaseloop 0M‐1cm‐1 6 611 Da 58

2.4.5 SDS‐PAGE

The composition of polyacrylamide gels for sodium dodecyl sulfate polyacrylamide gel

electrophoresis(SDS‐PAGE)isgivenintable2.24.Proteinswereseparatedforabout90minutes

at 150 V and stained with Coomassie. For this, the gel was immersed in staining solution

(section2.1.3),microwavedbrieflyandshakenfor10minutes.Afterthis,thegelwastransferred

todestainingsolution,microwavedbrieflyandshakenforanother10minutes.

Table2.24CompositionofpolyacrylamidegelsusedforSDS‐PAGE.

Component Stackinggel(5x) Separatinggel(4x)0.5MTris,pH6.8 500 µl ‐1.5MTris,pH8.0 ‐ 4.4 ml10%SDS 50 µl 200 µlAcrylamide/Bis‐acrylamide 850 µl 10 ml10%Ammoniumperoxydisulfate(APS) 50 µl 200 µlTEMED 5 µl 8 µlddH2O 3.5 ml 5.2 ml

2.4.6 Analyticalgelfiltration

Analytical gel filtrations were performed with a BioRad DuoFlow FPLC system on a

Superdex 200 10/300 gel filtration column (GE Healthcare, Solingen). The column was

equilibratedwith30mlof50mMNaH2PO4,450mMNaCl,1mMDTT,pH8.180µgTaspasewas

loaded into a 140 µl loop and separated with 0.5 ml/min flow rate. Due to the low protein

concentration,theabsorbanceofthepeptidebondat214nmwasusedtoevaluatetheretention

times.Forcalibration, ferritin (450kDa),aldolase(161kDa), conalbumin(75kDa),α‐amylase

(54kDa)andribonucleaseA(13.7kDa)wereused.Theretentiontimeswereplottedagainstthe

logarithmicmolecularweightforlinearregression.

2.4.7 LabelingofTaspasewithAtto594‐NHS

ThewildtypeproteinofTaspaseandtheinactivemutantwereindependentlylabeledattheC‐

terminuswithAtto594.1mgofTaspasewasdilutedin1mlof50mMNaH2PO4,450mMNaCl,

1mMDTT,pH8.Afterwards,a1.2‐foldmolarexcessofAtto594(dissolvedinDMSO)wasadded

andthesolutionwasincubatedfor60minutesatroomtemperatureinthedark.Unbounddyewas


50

removedwiththehelpofaPD‐10columnandtheconjugatewasconcentratedinaCentricon.To

determinetheproteinconcentration(c)anddegreeoflabeling(DOL),theabsorbanceat280nm

and601nmwasmeasuredusingaNanoDropND1000andconvertedasfollows:

∙ 26930 ∙ 0.51 ∙ 120000

∙ 0.51

1 ∙ 26930

ThesecalculationsyieldedforwildtypeTaspaseaproteinconcentrationof551µMandaDOLof

1.0 and for the inactivemutant a concentration of 87 µMwith aDOL of 1.01. Aliquots of the

conjugateswereshockfrozenandstoredat‐20°C.

2.4.8 LabelingofTaspasewithAtto488‐maleimide

WildtypeTaspasewaslabeledwiththecysteinereactivedyeAtto488‐maleimide.Thebufferwas

changed for labeling to 50 mM NaH2PO4, 450 mM NaCl, pH 7.4 by repeated dilution and

concentration in a 30 kDa cut‐off Centricon. A 2.5‐foldmolar excess of Atto488was used for

labeling. Labeling and removal of unbounddyewasperformed asdescribed for labelingwith

Atto594‐NHS(section2.4.7).DOLandproteinconcentrationweredeterminedusingthefollowing

formula:

∙ 26930 ∙ 0.1 ∙ 90000

∙ 0.1

1 ∙ 26930

After labelingwas completed, the bufferwas changed to the previous storage buffer (50mM

NaH2PO4,450mMNaCl,1mMDTT,pH8.0).Thisprocedureresultedinafinalconcentrationof

84µMconjugatewithaDOLof2.45.Aliquotsoftheconjugatewereshockfrozenandstoredat‐

20°C.

2.4.9 AssayforautocatalyticactivationofTaspase

Spontaneouscleavageoffull‐lengthTaspasebetweenD233andT234intoα‐andβ‐subunitwas

visualizedwithSDS‐PAGE.First,9µMpurifiedTaspasewasincubatedindifferentbuffersat30°C.

Samplesweretakenafteracertainperiodoftime,preparedforSDS‐PAGEandstoredat‐20°Cto

stop the cleavageprocess. The two subunitswere separated via SDS‐PAGE in gelswith 15%

acrylamide.

MATERIALANDMETHODS SPECTROSCOPICMETHODS

51

2.4.10 LimitedproteolysisassaywithproteinaseK

Limited proteolysis with proteinase K was performed in gel filtration buffer (section 2.1.3).

5µg/mlproteinaseKwasusedtodigest6.4µgTaspaseinatotalvolumeof20µlfor10minutes

at21°C,before5mMPMSFwasaddedtostopthereaction.SamplesweremixedwithSDSsample

bufferandseparatedbySDS‐PAGE.

2.5 Spectroscopicmethods

2.5.1 Taspaseenzymeactivityassay

ForthedeterminationofcatalyticactivityofTaspase,afluorogenicassaywasdesigned[56].The

substratecontainstheCS2cleavagesequenceasitoccursintheTaspasetargetproteinMLL,with

an additional anthranilic acid coupled to the N‐terminus and dinitrophenol‐conjugated lysine

coupledtotheC‐terminusofthepeptide(CASLOApS,Lyngby).Thefluorescentdyeanthranilic

acidcanbequenchedbydinitrophenol.Sinceduringthequenchingprocesstheexcitedelectron

transfers its excess energy to an electron of the quenchermolecule, the quenching efficiency

depends on the distance between dye and quencher. The intact peptide is quenched very

effectively,i.e.90%oftheenergyabsorbedbyanthranilicacidistransferredtodinitrophenoland

dissipated as heat (figure 2.4a). Cleavage by Taspase separates dye and quencher and the

quenchedfluorescenceofthedyeisrecoveredinaratedirectlyrelatedtothecatalyticactivityof

Taspase(figure2.4b).

Figure2.4PrincipleofthefluorogenicTaspaseactivityassay.aThemodelsubstratecontainsthecleavagesiteforTaspase,anN‐terminalanthranilicacid(Abz)andaC‐terminaldinitrophenolmoiety(DNP).AfterexcitationoftheAbzdye,theenergyistransferredtotheDNPandthefluorescenceisquenched.bUponcleavagebyTaspase,Abzdyeandquencherareseparated.Hence,Abzisnolongerquenchedandcanemitphotons.

Aslongasthesubstrateconcentrationisatsaturatinglevels,theincreaseinfluorescenceislinear

andmirrorstheinitialrate.Hence,thechangeinfluorescenceduringthefirst180secondswas

fittedwith linearregressionandconverted toenzymeactivity.As fluorescence ismeasured in

arbitrary values, a quantification of the kinetics with absolute values was not trivial. One


52

prerequisite for this was that all measurements were taken with the same parameters with

respect to temperature,pH,excitationslitwidth,emissionslitwidth,PMTvoltageandsample

purity.Furthermore,thesubstrateconcentrationhadtobelowtoavoidinnerfiltereffectsand

allow the assumption of a linear correlationof intensity change and substrate cleavage. 8 µM

provedtobeasuitablesubstrateconcentrationwithanabsorbancewellbelow0.1,whilestillhigh

enoughtoyieldanadequatetimetomeasurethereactionbeforesubstratedepletionbegan.

Unless specified otherwise, all samplesweremeasured in 3mm cuvetteswith a Varian Cary

Eclipsespectrofluorimeterequippedwithapeltierelement.Foreachassay,300nMTaspasewas

addedto8µMsubstrateinatotalvolumeof60µlwithbuffercontaining100mMHEPES,10%

sucrose,10mMDTT,pH7.9.Thetemperaturewasadjustedto37°Candmeasurementswere

takenfor10minuteswiththesettingsprovidedintable2.25.

Table2.25Parametersusedforkineticmeasurementswiththefluorogenicassay.

Parameter ValueExcitationwavelength 320nmEmissionwavelengths

420nm

Excitationslitwidth 20nmEmissionslitwidth 20nmPMTvoltage 480VAveragetime 5sTemperature 37°C

Fluorescence intensities were recorded for 8 µM of the intact and the completely cleaved

substrate. The measured intensities (I) in arbitrary units were converted to substrate

concentrations(c)accordingtothefollowingequation

8μ

wherecistheconcentrationofcleavedsubstrate,Iisthemeasuredfluorescenceintensity,Iintactis

theintensityoftheintactsubstrateandIcleavedrepresentstheintensityofthecleavedsubstrate.

Thechangeinconcentrationovertime(Δc/Δd)correspondstotheenzymaticactivityofTaspase.

Hence,thespecificcatalyticactivitycouldbecalculatedusingtheformulabelow.

∆

min ∙

μmin ∙

aspecrepresentsthespecificcatalyticactivityinµmolcleavageperminuteandmgTaspase,Δcis

theconcentrationofsubstrateconsumedin1minuteinµMand[Taspase]istheconcentrationof

Taspaseinmg/ml.

For thedeterminationof theMichaelisconstantKM, the initialvelocities (v)weremeasuredat

different substrate concentrations [S]. Nonlinear regression using least squares fitting to the


53

Michaelis‐MentenequationgivenbelowusingPrismsoftware [115]yieldedmaximumvelocity

vmaxandMichaelisconstantKm.

∙

Taspase requires high salt concentrations (>300mM) for stability. However, salt inhibits the

catalyticactivity.Areasonablecompromisebetweenstabilityandactivityisabuffercontaining

sucrose,whichwasusedforkineticassays.Inabuffercontaining10%sucrose,thehalf‐lifeof

Taspasewaslongenoughtoperformkineticmeasurementswithoutlosingactivityduetoprotein

denaturation.Despitethestabilizingeffectofsucrose,thehalf‐lifewastooshortforacomplete

testseries,suchasatitrationcurve.Tocircumventthisproblemandmaximizereproducibility,

aliquots of 300 nM Taspase in sucrose were flash frozen and stored at ‐80 °C. For each

measurement,analiquotwasthawedrapidlyinathermomixerat37°Candincubatedforexactly

5minutesat500rpmtoensureahomogeneoustemperature.Aftertheincubationtime,preheated

Taspase was added to the equally preheated substrate mixed with inhibitor or buffer and

immediately filled into cuvettes preheated to 37 °C. The first data point of each trace was

measured10secondsafterthereactionhadstarted.

With this assay, several compounds were tested for inhibitory effects on Taspase activity.

Togetherwiththesubstrate,thecompoundswerebrieflyincubatedat37°Casdescribedabove

before Taspase was added. Some of the inhibitors tested were insoluble in water at higher

concentrations and required dissolving in DMSO. After dilution in sucrose buffer, the DMSO

concentrationwasbelow3%andallmeasurementswerereferencedtothebuffercontrolswith

respectiveDMSOconcentrations.

Inhibitionconstant(Ki)andIC50valueswereobtainedbyapplyingnonlinearleastsquaresfitting.

For IC50 fits, enzymatic activity of Taspase was plotted against logarithmic inhibitor

concentrationsandfittedtothefollowingmodel.

1 10

Bottom and Top represent the inhibition at very low and very high inhibitor concentrations,

respectivelyandIC50istheconcentrationatwhich50%oftheeffectisobserved.

For the experimental estimation of Ki values, the enzyme activity was measured at several

combinations of substrate and inhibitor concentration and data was fitted to the following

equation.


54

1 ∙

∙

∙ 1

1 ∙

Wherevistheobservedvelocity,vmaxisthemaximumenzymevelocitywithoutinhibitor,Kmis

theMichaelisconstant,Kiistheinhibitionconstant,[S]isthesubstrateconcentration,[I]isthe

inhibitorconcentrationandαdescribesthemechanismofinhibition.Verysmallαvaluesindicate

uncompetitiveinhibition,verylargeαvaluesindicatecompetitiveinhibitionandavalueof1is

seen for noncompetitive inhibition. If Km and substrate concentration [S] is known, it is also

possibletoderivetheKivaluefromIC50valuesusingtheCheng‐Prusoffequation[124].

1

Thevaluesobtained fromexperimentaldeterminationand thevaluesgainedbymathematical

conversion from IC50 valueswere similar. For this reason,most of the Ki values reported are

convertedfromIC50values,whichareeasiertomeasurepractically.

Asnegativecontrol,theaminoacidsatthecleavagesiteofthesubstratewereexchangedagainst

alanine,yielding[Abz]‐KISQLAAVDDK‐[DNP](Abz:anthranilicacid;DNP:dinitrophenol;CASLO

ApS, Lyngby). Another negative control was performedwith inactive Taspase using identical

parameters. For positive controls, a peptide comprising the CS1 cleavage site of MLL was

synthesized with 7‐amino‐4‐methylcoumarin (MCA) as N‐terminal fluorescent dye

([MCA]‐GKGQVDGADDK‐[DNP];ChinaPeptides,Shanghai).Concentrationsofthispeptidewere

determinedusinganextinctioncoefficientof15000M‐1cm‐1at350nm.Measurementsweretaken

withexcitationandemissionwavelengthof320and400nm,respectively.

The activity of eukaryotic cell lysates was measured under slightly different conditions.

Measurementsweretaken in lysisbuffer(25mMTrispH7.4,150mMKCl,5mMMgCl2,5%

glycerol,1%tritonX‐100,2mM2‐mercaptoethanol,1mMPMSF,1cOmpletetablet)with8µM

substrateadded.ThecOmpleteProteaseInhibitorCocktailprecipitatesat37°C.Hence,theassay

temperaturewasloweredto30°Ctominimizeprecipitation.

2.5.2 Bindingstudieswithdetectionoffluorescenceanisotropy

The principle of fluorescence anisotropy is based on photoselection. This means that after

excitationwithlinearpolarizedlight,fluorophoreswithanexcitationdipolemomentparallelto

the electric vector of the incident light, are preferentially excited. In the excited state, the

fluorophorerotatesduetoBrownianmotion.Thedegreeofrotationbeforeemissiondependsof

rotation correlation time and fluorescence lifetime. Emission occurs with an electric vector


55

parallel to the emission dipole of the fluorophore. Hence, the emitted light of a fast rotating

fluorophore is less polarized and themolecule has a lower anisotropy than the emission of a

slowlyrotatingfluorophore.Forglobularproteins,therotationcorrelationtimeθdependsonthe

molecularweightasdescribedbythePerrinequation.

∙

Whereθistherotationcorrelationtime,ηistheviscosityofthesample(0.01poiseforwater,

0.011poisefor10%sucrose),Misthemolecularweight,Risthegasconstant,Trepresentsthe

temperature,visthespecificvolume(typicallynear0.73ml/g)andhisthehydration(typically

near0.23mgH2Opermgofprotein).Thetheoryofanisotropyexplainedaboveshowsthatbinding

ofasmallfluorescentligandtoaproteincausesanincreaseinanisotropy.

Measurements were taken with a Varian Cary Eclipse equipped with polarizer wheels and a

peltierelement.Unlessspecifiedotherwise,thesamplevolumewas60µlin3mmcuvettesand

readingswereaveragedover5minutesat20°Cwiththeparametersstatedintable2.26.

Table2.26Parametersusedforanisotropymeasurements.

Parameter Atto488 Atto594 Rhodamine Tryptophan AnthranilicacidExcitationwavelength 501nm 601 nm 555 nm 280 nm 320nmEmissionwavelengths 523nm 627 nm 582 nm 340 nm 420nmExcitationslitwidth 10nm 20 nm 5 nm 10 nm 20nmEmissionslitwidth 20nm 20 nm 10 nm 20 nm 20nmPMTvoltage 400V 400 V 500 V 600 V 600VAveragetime 5s 5 s 5 s 5 s 5sTemperature 20°C 20 °C 20 °C 20 °C 20°CGfactor 1.6724 2.1338 1.9587 0.6537 1.1373Concentration 1µM 1 µM 500 nM 10 µM 1µM

G factors were applied to compensate different transmission efficiency of the polarizers for

horizontally and vertically polarized light. To obtain these values, the emission of every

fluorophoreusedwasrecordedconsecutivelywithpolarizersorientedhorizontallyandvertically

and the ratio of the intensities yielded theG factor. The sampleswere excitedwith vertically

polarizedlightandemissionintensitiesofhorizontally(IH)andvertically(IV)polarizedlightwere

recordedseparately.Withthesevaluestheanisotropy(r)wascalculated.

∙2 ∙

Theincreaseinanisotropywasplottedagainstligandconcentrationandbindingconstantswere

derivedfromnonlinearcurvefittingusingPrismsoftware[115]withanequationforsaturation

binding.


56

Δ ∙

WhereΔr is the increase in anisotropy, rmax is the increase in anisotropy at saturating ligand

concentrationsandKdisthedissociationconstant.

2.5.3 Bindingdetectedwithmicroscalethermophoresis

Thermophoresis is thedirectedmovementofparticlesduetoa temperaturegradient. Inmost

cases,thismovementoccursfromhottocoldanddependsonthehydrationshell,chargeandsize

ofthemolecules.Ifbindingofanothermoleculechangesanyofthesethreeparameters,thiscan

be detected by microscale thermophoresis (MST; NanoTemper, Munich). The temperature

gradient is generated by an infrared laser and themovement of themolecules is followedby

changeinlocalfluorescenceintensityinsideacapillary.

Binding of Taspase to nanoparticles was assessed in standard glass capillaries with a

MonolithNT.115during aMSTdemonstration course. Taspasewas labeledwithAtto488 and

50 nM labeled protein was added to a 16‐step serial dilution of nanoparticles, ranging from

200µg/mlto12ng/ml.ToavoidadsorptionofTaspasetothecapillarywalls,0.1%Tweenwas

addedtoabuffercontaining50mMNaH2PO4,450mMNaCl,1mMDTT,pH8.0inafinalvolume

of 40 µl. Binding of the peptidic Taspase inhibitor Comp9 was assessed under label‐free

conditions, using intrinsic tryptophan fluorescence for detection with a Monolith

NT.115LabelFree.1µMofTaspasewasaddedtoaserialdilutionof400µMto12nMComp9in

50mMNaH2PO4,450mMNaCl,1mMDTT,0.1%Tween,pH8.0.Dataanalysisanddetermination

ofKdvalueswasperformedwiththesoftwaretools implemented intheNanoTemperAnalysis

software.

2.5.4 Determinationofinhibitorstabilitywithfluorescenceanisotropy

Fluorescence anisotropy is sensitive to changes in the hydrodynamic radius, as explicated in

section2.5.2.Therefore,itcanbeusedasameasureforthecleavageoffluorescentsubstrates.The

stabilityofthepeptidicrhodamine‐labeledinhibitorwasassayedinthepresenceof1µMTaspase

and compared to the regular fluorescent substrate (section 2.5.1) as positive control.

Measurementsweretakenin200µlof50mMNaH2PO4,450mMNaCl,1mMDTT,pH8.0in3mm

cells.Inhibitororsubstrateconcentrationsamountedto8µMandreadingswereperformedwith

aVarianCaryEclipsefluorescencespectrometerusingthesettingslistedintable2.27.

Table2.27Parametersusedtomeasureinhibitorstabilitywithanisotropy.


57

Parameter Substrate InhibitorExcitationwavelength 320nm 495 nmEmissionwavelength 420nm 582 nmExcitationslitwidth 20nm 20nmEmissionslitwidth 20nm 20nmPMTvoltage 600V 600 VAveragetime 5s 5 sInterval 10min 10 minTemperature 20°C 20 °CGfactor 1.1373 1.9640

Curveswerefittedtoamodelformono‐exponentialdecaywithGraphPadPrism:

∙

whererrepresentstheanisotropy,rintactistheanisotropyvalueoftheuncleavedstate,rcleavedis

theanisotropyvalueinthecleavedstate,tisthetimeandKtherateconstant.Thehalf‐lifet0.5can

becalculatedas

.2

2.5.5 CDspectroscopy

CDspectraandmeltingcurveswererecordedwithaJascoJ‐710CDspectropolarimeter(Jasco,

Gross‐Umstadt).Far‐UVspectrawererecordedin0.1cmcuvettes(HellmaAnalytics,Müllheim).

ForTaspaseandTaspasemutantsthebuffercontained50mMpotassiumphosphatebufferatpH

6.5,whilethesyntheticlooppeptidewasdissolvedin50mMpotassiumphosphate,pH6.0.The

protein concentration was determined precisely and 200 µl of a 0.2 mg/ml solution were

measuredwiththeparametersgivenintable2.28.

Tocorrectthey‐driftofthedevice,allspectrawereshiftedsothattheellipticitybetween250and

260nmwaszero.Fromeachproteinspectrumabufferspectrumwassubtractedandthemdeg

unitswereconvertedtomolarellipticityusingpathlengthandproteinconcentration.

∙ /

whereψspecisthespecificellipticityindegcm3dm‐1pergramofaminoacid,ψistheobserved

ellipticityinmdeg,listhepathlengthincmandcistheconcentrationinmg/ml.

Thesespectrashowifthepurifiedproteinisfoldedandallowanestimationofhelix,sheet,turn

andrandomcoilcontent.Fortheestimationofsecondarystructurecontentbydeconvolution,the

CDSSTRalgorithm[125]wasused.

Melting curves were recorded in a 1 cm cuvettes sealed with a stopper (Hellma Analytics,

Müllheim)with0.05mg/mlproteinin50mMpotassiumphosphatebufferatpH6.5.Thesetof


58

parametersused toobtain themeltingcurves is listed in table2.28.Nonlinearcurve fitting to

obtain melting temperatures was performed with the Jasco Spectra Manager 2.0 software

package.

Table2.28Parametersusedtorecordfar‐UVCDspectrawithaJascoJ‐710.

Far‐UVspectra MeltingcurvesParameter Value Parameter valueStart 260 Wavelength 225 nmEnd 200 Temperatureslope 1 °C/minScanningmode Continuous Delaytime 180 sResponsetime 0.5s Responsetime 8 sBandwidth 2 nm Bandwidth 1 nmDatapitch 0.2nm Datapitch 0.1 °CTemperature 21°C Temperature 20‐85 °CAccumulations 20Scanningspeed 100nm/min

2.5.6 Fluorescencemeltingcurves

ForfluorescencemeltingcurvesofTaspase,thesolvatochromismoftryptophanisexploited.The

emission maximum of tryptophan depends on the polarity of the environment, because two

differentelectrontransitionscanoccurwhenthefluorescentindoleringabsorbsaphoton.These

twoperpendicularelectronictransitionsarereferredtoas1Laand1Lbanddifferintheirenergy

levels.Foratryptophanintheproteincore,1Lbhaslowerenergythan1Laandisthepredominant

transition.Sincethe1Latransitionisdirectedthroughthering‐NHgroup,itcanhavedipole‐dipole

interactionswithpolarsolventmoleculesandisthereforethelowestenergytransitionforsurface

exposedtryptophans.Asemissionfrom1Lboccursfromahigherenergylevelthanemissionfrom1La,thetryptophanemissionundergoesabathochromicshiftduringproteinunfolding.

Accuratedeterminationoftheemissionmaximumtoonenanometerischallenging,becauseofthe

broademissionspectrumoftryptophan.Inordertoincreasetheaccuracyofthemeasurement,

theintensityattwowavelengthswithconsiderablechangesinintensityduringunfoldingwere

observed during unfolding. This procedure also allowed a mathematical elimination of the

thermalquenchingoccurringduringthemeasurementviadivisionbythesumoftheintensities.

Measurementsweretakenin3mmquartzcuvetteswithaVarianCaryEclipsespectrofluorimeter

equippedwithapeltierelement.10µMofTaspasein200µlwasassayedtodeterminetheeffect

of buffer composition and nanoparticle binding. After excitation at 280 nm, the fluorescence

emissionintensitiesat334and376nmwererecorded,whilethetemperaturewasincreasedfrom

20to95°C.Thedetailedsettingsforthisassaycanbefoundintable2.29.Table2.29Parametersusedforrecordingoffluorescencemeltingcurves.

Parameter ValueExcitationwavelength 280nmEmissionwavelengths 334nmand376 nmExcitationslitwidth 20nm


59

Emissionslitwidth 10nmPMTvoltage 480VAveragetime 5sHoldtime 180sTemperatureslope 0.5°C/minDatainterval 0.1°CTemperaturerange 20‐95°C

Theintensitiesmeasuredat334nm(I334)and376nm(I376)wereconvertedtoarbitraryfolding

units(F)usingthefollowingformula.

Thesevalueswerenormalizedonascalefrom0to1usingthevaluesofthefoldedandunfolded

statesasreferences.Thefractionofunfoldedprotein(U)atatemperature(T)isdefinedas

1

whereF(T)representsthevalueofthefoldingvalue(asdefinedabove)attemperatureT,while

FunfoldedandFfoldedrepresentfoldingvalueoftheunfoldedandnativeprotein,respectively.

2.5.7 NMRspectroscopy

ForNMRspectraofTaspase,thebufferwaschangedto50mMNaH2PO4,300mMNaCl,1mMDTT,

pH7.9byrepeateddilutionandconcentrationinaCentriconandmeasurementsweretakenat

30°C,unlessspecifiedotherwise.Sampleconcentrationsrangedbetween400and600µMandto

allsamples2%(v/v)deuteriumoxidewasaddedasfrequencylock,and50µMDSS(2,2‐dimethyl‐

2‐silaptane‐5‐sulfonate)toallowacalibrationof1Hchemicalshifts.Thefinalsamplevolumeof

600µlwastransferredintoa5mmthin‐wallNMRtube.

NMRspectrawererecordedwithaBrukerUltrashield700NMRspectrometerwithinversetriple

resonancecryoprobe(BrukerBiospin,Rheinstetten)withconstanttransmitterfrequencies(1H:

700,22MHz;15N:70.952676MHz;13C:176,070459MHz).RecordingandprocessingofNMR

spectrawasexecutedwiththeTopspin3.0softwarepackage(BrukerBiospin,Rheinstetten)using

theparametersspecifiedinsupplementaltable7.1,table7.2,andtable7.3.Allpulseprograms

usedweretakenfromtheBrukerstandardlibrary.

2.5.8 RecordingandprocessingofNMRspectra

FreeinductiondecayswereFouriertransformedineverydimension(command“xfb”for2Dor

“tf3;tf2;tf1”for3D)andπ/4shiftedsinefilterfunctionswereusedforapodization.Afifthdegree

polynomialwassubtractedforbaselinecorrectionineverydimension(commands“abs2.water;


60

abs1”for2Dor“tabs3;tabs2;tabs1”for3D).Spectrawerephasecorrected,withimaginarydata

addedto2Dplanesof3DspectrabyHilberttransformation(command“xht2;xht1”)toallowa

phasecorrectionof3Dspectra.Fortheevaluationandpeakassignment,CCPNsuite2.3wasused.

2.5.9 Assignmenttheory

The principle of backbone atom assignment and identification of sequential amino acids is

explainedhereusingsimplifiedexamplespectra(figure2.5).Theassignmentwasbasedona1H‐15NHSQC,whichcanberegardedas the fingerprintofaprotein.ThisspectrumshowsallH‐N

correlations,which aremainly backbone amide groups (figure 2.5b), but also side‐chain H‐N

groups of tryptophans, asparagines and glutamines are visible. Essentially, every amino acid

(except for theN‐terminaloneandprolines) causesonepeak in this spectrum.Withonly the

Figure2.5SchematicspectraandinformationforNMRbackboneassignment.

a1H‐15NHSQCspectrumshowingthreeaminoacidslabeled“a”,“b”and“c”.bNamingconventionforaminoacidatoms,shownforalanine inapolypeptidechain.Heavyatomsaredisplayed inblack,hydrogens inwhite.c3‐dimensionalHNCaCbspectrumshowingthecarbonresonancesoftheaminoacids“a”,“b”and“c”. Alpha carbon atoms are displayed in blue, beta carbon atoms in red.d 3 dimensional CbCaCONHspectrumshowingthecarbonresonancesoftheaminoacidsprecedingaminoacids“a”,“b”and“c”.Alphaandbetacarbonatomsare indistinguishable in thisspectrum.eReferencechemicalshifts for thealpha(blue)andbeta(red)carbonatomsofall20aminoacids.ReferencechemicalshiftswereextractedfromCCPNandarebasedontheRefDBdatabase[126].H‐Nchemicalshifts,itisnotpossibletoidentifywhichtypeofaminoacidthesignalbelongsto.

Hence,thesepeaksarearbitrarilynumbered.Thisisshownforthreeaminoacidsinfigure2.5a.

Toobtaintheinformationneededforanassignmentoftheaminoacidtype,informationfrom3D

spectrawererecorded.InHSQC‐based3dimensionalspectra,athirdaxiswithinformationabout

thechemicalshiftofcarbonatomsisadded.Inthespectrumshowninfigure2.5c,thechemical

shiftofthealphaandbetacarbonatomsisrecordedandcanbevisualizedincolumnsabovethe


61

HSQCspectrum,becausethechemicalshiftoftherespectiveH‐Ngroupsisalsoincludedinthis

spectrum.Alphaandbetacarbonatomscanbedistinguisheddueto theiroppositephase.The

chemicalshiftsofthecarbonalphaandbetaatomsdependstronglyontheaminoacidtype.Since

thecombinationofCαandCβshiftvaluesisuniqueforsomeaminoacidtypes(figure2.5e),this

allowsassigningtheresiduetype.Intheexampleshowninfigure2.5,residue“a”hasaCβwitha

chemical shift higher than the Cα. As seen in figure 2.5e, this is only the case for serine and

threonine.Consequently,residue“a”isprobablyaserineorthreonine.Residue“b”doesnothave

aCβ,butonlyaCα.AsallaminoacidsbutglycinescontainaCβatom,residue“b”mustbeaglycine.

TheCβshiftofresidue“c”iswith20ppmremarkablylow.Thisistypicalforalanines,indicating

thatresidue“c”ismostlikelyanalanine.

Proteinsusuallycontainanaminoacidtypemorethanonce,whichpreventsanassignmentofthe

positioninthepolypeptidechainwiththeinformationofHSQCandHNCaCbspectra.Thismeans,

it ispossible todetermine that theHSQCpeak labeledas residue “b” is aglycine,but it isnot

possibletodeterminewhichglycineintheproteincausedthispeak.Furthermore,theorderofthe

threeaminoacidscannotbederivedfromthesespectra.

Information about the sequential arrangement of the amino acids canbe obtainedwith pulse

programs,inwhichthemagnetizationistransferredfromthecarbonatomstotheH‐Ngroupof

thesucceedingaminoacid(i+1).Figure2.5dshowssuchaspectrum(calledCbCaCONH),inwhich

theCαandCβchemicalshiftinformationoftheprecedingaminoacid(i‐1)isrecordedwiththeN‐

Hshiftsof theaminoacidconsidered.Thismeansthat thecolumnsabovetheHSQCspectrum

(figure2.5d)nowrepresentthechemicalshiftsoftheprecedingaminoacid.

LookingattheCbCaCONHspectrum(figure2.5d)ofresidue“b”,itisnoticeablethatthechemical

shiftvaluesinCdimensionareidenticaltothoseofresidue“a”intheHNCaCbspectrum(figure

2.5c).Asexplainedabove,thiscanonlybethecase,ifresidue“a”precedesresidue“b”.Residue“a”

hasonlyonesinglepeakintheCbCaCONHspectrum(figure2.5d),whichhasthesamechemical

carbonshiftas theCαof residue“c” (figure2.5c).Hence, residue“c”precedesresidue“a”.The

CbCaCONHpeaksofresidue“c”indicatethecarbonshiftsoftheaminoacidprecedingresidue“c”.

Withthesethreespectra(HSQC,HNCaCb,CbCaCONH)theorderofthethreeresiduescouldbe

determinedasGly(residue“c”)–Ser/Thr(residue“a”)–Ala(residue“b”).Thisprocedureiscalled

chaintracing.

Inmanycases,itispossibletolocalizethisshortfragmentintheproteinsequence.Iftheprotein

has bothGly‐Ser‐Ala andGly‐Thr‐Ala in its sequence, or themotif occursmultiple times, it is

necessary to assign longer fragments, before it canbeunambiguously assigned to theprotein

sequence.Other spectra reveal information about the chemical shifts of side chain hydrogens

and/orhydrogensof thepreceding amino acids.These values arehelpful both to identify the

aminoacidtypeandsequentialorder.

Itshouldbementionedthatspectraareneverasperfectasintheexampleexplainedabove.Inthe

HNCaCbspectrum,thecarbonshiftsoftheprecedingaminoacidarefrequentlyvisibleinaddition

tothecarbonshiftsoftheaminoaciditself.Furthermore,usuallynotallpeaksareclearlyvisible

duetooverlap,insufficientsignal‐to‐noiseratio,toohighortoolowcorrelationtimes,orterminal

degradation.Additionally,smallshiftdifferencesbetweentwospectraarecommonlyobserved,


62

becauseofslightconformationalchangesorminordeviationsinsamplepreparation.Iftwostable

conformations exist,which is often observed as cis/trans conformations for amino acids near

prolines, both conformations are visible with their signal intensity corresponding to the

abundanceoftherespectiveconformation.

2.5.10 Lactatedehydrogenaseactivityassay

Asacontrolfornanoparticleinhibition,lactatedehydrogenase(LDH)wasused.Theconversion

of NADH to NAD+ during the reduction of pyruvate to lactate was followed by absorbance

spectroscopy.Thereactionwasfollowedin1cmcuvettesat340nmwavelengthinaVarianCary

100spectrophotometerequippedwithapeltierelement.Measurementswereperformedwitha

spectralbandwidthof1nmand5secondsaveragetimeat20°C.Priortoeachmeasurement,the

blankrateof250µMNADHand1mMpyruvatein100mMsodiumphosphatebufferpH7.4was

measuredfor2minutes.Tostartthereaction,100ng/mlLDHwereaddedandthelineardecrease

ofabsorbancewasfollowedoverseveralminutes.WithamolarextinctioncoefficientforNADHof

6220 M‐1 cm‐1, the NADH consumption per minute was calculated and the blank rate was

subtracted. To test for inhibition by nanoparticles, Amsil8 and Amsil125 nanoparticles were

addedtothefinalreactionvolumeof300µlandincubatedbriefly,beforetheactivitywasassayed.

2.5.11 Chymotrypsinactivityassay

Anotherproteasetestedforinhibitionbynanoparticleswastheserineproteasechymotrypsin.As

substrate,thepeptideSucc‐Ala‐Ala‐Pro‐Phe‐pNAwasused,whereSuccissuccinicacidandpNA

isp‐nitroaniline.ChymotrypsincleavesthesubstratebetweenPheandpNA.FreepNAexhibitsa

strongabsorbance,whichcanbefollowedbyUV/Visspectroscopy.Measurementsweretakenat

20°Cwith300µlof50mMNaH2PO4,10%sucrose,450mMNaCl,1mMDTT,pH8.0in1cm

cuvetteswith a Varian Cary 100. In this assay, the blank rate of 150µM substrate peptide in

presenceorabsenceofnanoparticleswasfollowedat390nmfor2minutes,beforethereaction

wasstartedbyadditionof350nMchymotrypsin.Afterbaselinesubtraction,thecomparisonof

the increase in absorbance with and without nanoparticles allowed a determination of the

inhibitoryeffect.

2.5.12 MALDI‐TOFmassspectrometry

MassspectraofproteinswererecordedwithanAutoflexspeedMALDI‐TOFmassspectrometer

(Bruker Daltonics, Bremen). The matrix solution was created by dissolving 7.6 mg of 2’,5’‐

dihydroxyacetophenone in 375 µl analytical ethanol and adding it to 125 µl of an 18mg/ml

aqueoussolutionofammoniumhydrogencitrate.ProteinsweredesaltedwiththehelpofSupel‐

TipsC18(Sigma‐Aldrich,Seelze)andelutedin2µlofa50:50‐mixture(v/v)ofacetonitrileand

0.1%trifluoroaceticacid.2µlof2%TFAand2µlofmatrixsolutionwereaddedtotheeluted

protein. 0.5 µl of thismixturewas spotted on anMTP 384 ground steel target plate (Bruker

MATERIALANDMETHODS BIOINFORMATICMETHODS

63

Daltonics,Bremen)andallowedtodry.Samplesweremountedontoatargetframeandanalyzed

withthestandardmethodsLP_20‐50kDaorRP_5‐20kDafromtheBrukerlibrary.

2.5.13 LC‐MSESI‐TOFmassspectrometry

Liquid chromatography‐mass spectrometry electrospray ionization time‐of‐flight mass

spectrometry(LC‐MSESI‐TOFMS)measurementsweretakenwithanLCQFleetESIspectrometer

(ThermoScientific,Darmstadt)equippedwithanEclipseXDB‐C18column(Agilent,Böblingen).

Chromatography was performed with 5 mM ammonium acetate and linear gradient from

acetonitriletowaterwasappliedover10ml.Theflowratewassetto1ml/minandabsorbance

wasdetectedat210nm.

2.6 Bioinformaticmethods

2.6.1 GenerationofTaspasemodelsandpreparationfordocking

Missing parts of the Taspase crystal structure were added by homology modeling using the

availablestructure(PDB2a8j)astemplate.ForhomologymodelingwithYASARA,sequenceand

structureweresubjectedtotheimplementedalgorithmwiththeparametersgivenintable2.30.

Table2.30ParametersforhomologymodelinginYASARA.

Parameter ValuePsiBLASTs 1EValueMax 0.5TemplatesTotal 5TemplatesSameSeq 1OligoState 4Alignments 5LoopSamples 50TermExtensions 90Speed slow

ForhomologymodelscreatedwithSwiss‐Model[122],theautomatedmodelingmodewasused

withdefaultparametersand2a8iastemplate.

AsYASARAandSwiss‐Modeldonotsupportheterodimermodeling,themodelofcleavedTaspase

was generated in a stepwise approach. First, homologymodels of the α‐ and β‐subunit were

computed independently. Subsequently, bothmodelswere aligned to the full‐length structure

2a8i,toreconstitutetheactiveαβ‐dimer.Thisdimerwasrefinedinanenergyminimizationand

1nsmoleculardynamics(MD)simulationusingYASARA.

ThestructureswerepreparedfordockingusingUCSFchimera[112]andprogramsoftheDOCK6

suite [114]. Inchimera,solventmoleculesweredeleted,foralternativeatomlocationsonlythe

highest occupancy was kept and charges were calculated. Next, cavities on the surface were


64

identified using sphgen [127] and the active sitewas selected as target area for docking. The

overlappingspheresareusedlaterinthedockingprocess.Asthirdstep,asquareboxwith50Å

edgelengthwasbuiltaroundtheactivesite.Inthisbox,twoenergygridswerecomputedusing

grid [128]with the inputparameters listed in table2.31.Thebumpgrid contains information

necessarytoidentifystericoverlapofligandandreceptor.Theenergygridisusedforscoringand

holdsinformationaboutthevanderWaalsandelectrostaticinteractions.

Table2.31Parameterstocalculatetheenergygridfordocking.

Parameter Valuegrid_spacing 0.3Åenergy_cutoff_distance 9999Åatom_model a(allatoms,includingH)attractive_exponent 6repulsive_exponent 12dielectric_factor 4bump_overlap 0.75

2.6.2 Compounddatabaseforvirtualdocking

CompoundsfortheinsilicovirtualscreenwereobtainedfromtheZINCdatabase[129].Toreduce

the amount of structures and still cover amaximum of different features, compoundswith a

similar structurewere discarded. For this, the structureswere converted to chemical hashed

fingerprintswithChemAxon [130], the fingerprintsof the structureswere comparedandonly

compoundsthatdifferfromallothersbymorethan60%wereselected.Forthis,thecompounds

werecomparedpairwise.Bothcompoundswerecheckedforthepresenceorabsenceofcertain

structuralfeaturesandtheportionofsharedfeatureswascalculated.Thisisadditionallydonefor

longer fragments, branching points and cyclic patterns. The structures of this subset were

downloadedfromtheZINCdatabaseinmol2fileformat.

2.6.3 VirtualDocking

Twodockingmethodswereusedconsecutively.First,thecompoundswerescreenedinaflexible

dockingprocess,andthencompoundssimilartothebestfittingcompoundswereobtainedand

dockedusinganAMBERforcefield[131].

Flexible docking was performed with the Anchor‐and‐Grow method on 4 processors. This

algorithmpartitionsthecompound,whichistobedocked,intorigidsegments.Thesesegments

containonlybondswithdoublebondsorpartialdoublebondcharacter.Thelargestrigidfragment

isusedasanchorandorientedintheactivesite.Thisisdonebymatchingtheheavyatomsofthe

fragmentwiththespheresgeneratedwithsphgen(seesection2.6.1)andsubsequentenergetic

optimization. This process is repeated 500 times and 100 orientationswith the best docking

scoresareusedforgrowth.Inthegrowthstage,theflexiblepartsofthecompoundsareaddedone

byoneandonlytheconformationswiththebestbindingenergiesareusedforthenextgrowth


65

step.Afterthelastcycle,allthefragmentsarecombinedandtheoriginalcompoundisdockedin

theactivesite.Afinalenergycalculationisperformedtoallowarankingofallthecompoundsthat

aredocked.Parametersusedforthissteparelistedintable2.32.

Allstructureswithadockingscorebelowanarbitrarilydefinedcut‐off(‐40kJ/mol)wereselected

forthefollowingstep.Fortheselectedstructures,similarstructuresweresearchedandusedfor

thenextdockinground.Thisseconddockingroundwasperformedwith thesameparameters

usedforthefirstround(table2.32)on8processors.Thedockedconformationswereprocessed

withtheprepare_amberscriptoftheDOCKsuite,whichaddshydrogens,AMBERforcefieldatom

types,charges,topologyandcoordinates.31structurescontainingcyclobutanemoietiescouldnot

beparameterizedcorrectlyandthuswereexcludedfromtheset.Theamberinputfileswereused

forthefinaldockingroundwithAMBERscoring.Inthisdockingprocess,all26aminoacidswithin

5Åaroundtheactivesiteweretreatedasflexible.Theligand‐receptor‐complexwassubjectedan

energy minimization with 200 conjugate gradient steps, followed by a 10 picoseconds MD

simulationandanotherenergyminimization.Thesecalculationswereperformedon8processors

withtheinputvariablesgivenintable2.33.

Table2.32ParametersusedforflexibleAnchor‐and‐Growdocking.

Parameter Valueautomated_matching yesmax_orientations 500min_anchor_size 40pruning_max_orients 100pruning_clustering_cutoff 100simplex_anchor_max_iterations 500simplex_growth_max_iterations 500atom_model a(allatoms)

Table2.33ParametersusedforAMBERscoredocking.

Parameter Valueamber_score_movable_distance_cutoff 5 Åamber_score_before_md_minimization_cycles 200amber_score_after_md_minimization_cycles 200amber_score_md_steps 10000amber_score_temperature 300 Kamber_score_nonbonded_cutoff 18 Å

2.6.4 GenerationofTaspaseloopmodels

Twomodels of the isolated Taspase loopwere generated. The firstmodelwas obtainedwith

conventionalhomologymodeling,asdescribedforthefull‐lengthprotein(section2.6.1).Starting

pointforthesecondmodelwasalinearpeptidechain.Afterenergyminimizationand100nsMD

simulation,thestructurereachedanenergeticminimum,whichwasconsideredasstablestate.

RESULTS CLONINGANDPURIFICATIONOFTASPASEPROTEIN

66

3 Results

3.1 CloningandpurificationofTaspaseproteinTheDNAforTaspaseproteinswasclonedinexpressionvectorsandtransformedincompetent

bacterialexpressionstrains.Theconditionsforproteinexpressionwereindividuallyoptimized

forwildtypeTaspase,theactivemutantandthelooppeptide.Affinitytagsallowedapurification

of heterologously expressed protein in themilligram scale. The pure proteins were used for

biochemicalcharacterizationininvitroassays.

3.1.1 CloningofTaspaseanditsmutants

IsolationofthecDNAofhumanTaspasefromSEMcellsandcloningintoapET‐22bvectorwithC‐

terminal His‐tag was performed by Jens Rabenstein [82]. The bicistronic construct for active

Taspase with shortened loop (figure 2.3) was designed in silico. The DNA was synthesized

(GeneArt,Regensburg)andsubsequentlyclonedviaNdeI/XhoIrestrictionsitesintoamodified

pET‐41b vector [132,132]. The catalytically inactive mutant D233A/T234A was created by

QuikChangeaccordingtomanufacturer’sinstructionsusingtheconstructforwildtypeTaspase

as template. Vectormaps of all bacterial expression constructs can be found in supplemental

figure7.1.

Figure3.1AgarosegelafterPCRamplificationoftheTaspaseloop.

PCRwasperformedwithwildtypeTaspaseastemplate.ThevisiblebandexhibitstheexpectedsizeoftheTaspaseloopof168bp.

ForcloningoftheTaspaseloop,thefulllengthTaspaseconstructwasusedastemplatewiththe

primer pair given in table 2.13. After PCR amplification, the PCR productwas analyzed in an

agarosegel(figure3.1).Abandatapproximately150bpcanbedetected,whichcorrespondsto

theexpectedsizeof168bpfortheTaspaseloop.ThisbandwascutfromthegelandtheDNAwas

isolated.Subsequently,both the insert (PCRproduct)and thevector (pET‐41b)weredigested

separatelywithApaIandXhoI.Insertandvectorwerepurifiedagain,beforetheywereligated(3

to 1 ratio) using T4 ligase. After transformation of the ligation product into shox cells and

incubationovernight,tenclonesweretestedinacolonyPCRandtheresultwasagaincontrolled

inanagarosegel(figure3.2).


67

Figure3.2AgarosegelaftercolonyPCRforverificationofpositiveclones.

TencloneswereselectedafterligationofTaspaseloopandmodifiedpET‐41bvector.ComparisontoaPCRwithwildtypeTaspaseplasmid(Ctrl)showsthatallclonesexceptforclone6exhibitabandatthesizeoftheTaspaseloop(168bp).

Allclonesexceptforclone6showedabandatthesizeofthepositivecontrol(168bp).Clones1

and3wereselectedarbitrarilyforDNAisolationandsequencing.Bothclonestestedpossessed

the correct sequence and the DNA of clone 1 was transformed into the expression strain

BL21(DE3)T1r.

3.1.2 ExpressiontestofwildtypeTaspase

Figure3.3SDS‐PAGEgelsofexpressiontestforwtTaspase.

ExpressionofwildtypeTaspasewastestedinE.coliBL21(DE3)T1rcells(BL21)andRosetta2cells(Ros)at18°C,24°Cand30°Cafter3h,6handovernight(o.n.).Supernatant(S)andpelletfractions(P)werecompared to a sample before IPTG‐induction (‐IPTG). The expected sizes of full‐length Taspase (fl), α‐subunit(α)andβ‐subunit(β)areindicatedwitharrows.OptimumexpressionconditionsareachievedinRosetta2cellsafter6hat30°C.


68

HeterologousexpressionoftheHis‐taggedwildtypeTaspaseproteininE.coliwastestedintwo

differentstrains(BL21(DE3)T1randRosetta2)atthreedifferenttemperatures(18°C,24°Cand

30°C)andforthreedifferentexpressiontimes(3h,6h,overnight)inTBmedium.Theamountof

overexpressedproteinwascomparedtotheproteincontentbeforeadditionofIPTG.Afterlysis

andultracentrifugation,pelletandsupernatantproteinswereinvestigatedwithSDS‐PAGE(figure

3.3).

Undermostconditions,themajorfractionofTaspasecanbefoundinthepellet,whileonlyasmall

fraction ispresent inthesupernatant.The largestamountofsolubleTaspase isachievedafter

expressionfor6hat30°CinRosetta2cells.Consequently, theseconditionswereusedforthe

expressionofwild typeTaspase, aswell as for the expressionof the inactiveTaspasemutant

(D233A/T234Amutant).

3.1.3 PurificationofwildtypeTaspaseandinactiveTaspase

Infigure3.4,thechromatogramsandgelsforthepurificationofwildtypeTaspasearedisplayed.

Thecorrespondingfiguresfortheinactivemutantcanbefoundinsupplementalfigure7.4.

Afterexpressionin2lunderoptimizedconditions(section3.1.2),bacteriawereresuspendedin

phosphatebufferandlysedbyincubationwithlysozyme,followedbysonification.Aftercelldebris

andinsolubleproteinshadbeenremovedbyultracentrifugationandsterile filtration,aNiNTA

affinitychromatographywasperformedandthepurityoftheelutedfractionswascontrolledby

SDS‐PAGE(figure3.4).

Figure3.4NiNTAaffinitypurificationofwildtypeTaspase.

The upper panel shows the chromatogram of the NiNTA affinity purification including the gradient ofelutionbuffer.SamplesforSDS‐PAGEweretakenfromthesupernatantbeforeNiNTApurification(S)andfrom elution fractions indicated with arrows. The corresponding gel (lower panel) reveals fractionscontainingfull‐lengthTaspase(fl),α‐subunit(α)andβ‐subunit(β).Fractionshighlightedinbluewerepureenoughforgelfiltration,whiletheimpurefractionshighlightedingreenwereappliedtoanotherNiNTAaffinitychromatography.


69

Thefirstelutionpeak(greeninfigure3.4)emergesatlowimidazoleconcentrations(~35mM)

andcontainssignificantamountsofTaspaseinadditiontosevereimpuritiesbyE.coliproteins.

Thefractionsofthispeakweredilutedto100ml,reducingtheimidazoleconcentrationto10mM,

andappliedanothertimetoaNiNTAaffinitycolumn.Thechromatogramofthispurificationand

therespectivegelfiltrationcanbefoundinsupplementalfigure7.5.

Thesecondelutionpeakat125mMimidazole(blueinfigure3.4)containsTaspasewithonlyfew

impuritiesataround35kDa.Thispeakwaspooledandconcentratedto2ml,beforetheprotein

wassubjectedtogelfiltrationonaSuperdex200columnforfurtherpurificationandremovalof

imidazole.Twopeaksarevisible,whichwereanalyzedbySDS‐PAGE(figure3.5).

Figure3.5PreparativegelfiltrationofwildtypeTaspase.

aTheupperpaneldisplaysthegelfiltrationchromatogram.ThetwopeakswereanalyzedbySDS‐PAGE(lowerpanel),showingthatthesecondpeak(highlightedinblue)containspurefull‐lengthTaspase(fl),α‐subunit(α)andβ‐subunit(β).bThecorrespondinggelfortheinactiveTaspasemutant(D233A/T234A)showsonlyonebandnear45kDa,correspondingtothefull‐lengthenzyme.The first peak at 50 ml corresponds to the void volume and contains large complexes or

aggregates.ForwildtypeTaspase,themajorpeak(blueinfigure3.5a)containsalmostexclusively

theproenzymeoffulllengthTaspaseortheactiveformconsistingofα‐andβ‐subunits.Forthe

inactivemutant, only the proenzymewas found (figure 3.5b). This peakwas concentrated to

18mg/mlandstoredinaliquotsat‐20°C.Theoverallproteinyieldamounted25mgperliterof

bacteriaculture.


70

3.1.4 PurificationofactiveTaspasemutant

To express a cleaved Taspase variant, the two subunits were co‐expressed in E. coli with

shortenedTaspaseloop.ThisvariantissubsequentlyreferredtoasactiveTaspase.Thebicistronic

constructforactiveTaspasecomprisestheaminoacids1to206(His‐taggedα‐subunit)andamino

acids234to420(β‐subunit;figure2.3).Expressiontestsindicatedmaximumproteinexpression

after6hoursat37°C(seesupplementalsection7.6).Thelowamountofoverexpressedprotein

couldpartiallybecompensatedbyincreasingtheexpressionculturesizeto6liter.Celllysisand

NiNTAaffinitypurificationwereperformedasdescribedforthewildtypeTaspase(section3.1.3).

Althoughonlytheα‐subunitisHis‐tagged,theβ‐subunitcanbeco‐purifiedduetodimerizationof

thesubunits.Thechromatogramsof theNiNTAaffinitychromatogramandthesubsequentgel

filtrationaredisplayedinfigure3.6.

Figure3.6PurificationofactiveTaspasebyaffinitychromatographyandgelfiltration.

aNiNTAaffinitypurificationofactiveTaspase.TheHis‐taggedα‐subunitandtheuntaggedβ‐subunitwereco‐purified.Thetoppanelshowsthechromatogramincludingtheelutiongradient.TheSDSgelofproteincontainingelutionfractionsisshowninthelowerpanel.Thesecondpeak(highlightedinblue)containsα‐subunit(α)andβ‐subunit(β)withfewimpuritiesandwasusedforsubsequentgelfiltration.bGelfiltrationofactiveTaspase.Theupperpaneldisplays thechromatogram.TheSDS‐PAGEgelofproteincontainingfractionsisshowninthelowerpanel.Thethirdpeak(highlightedinblue)containspureactiveTaspase.

The first elution peak of the NiNTA column (figure 3.6a) between 275 and 300 ml contains

Taspase,butalsoasignificantamountofimpurities.Hence,onlythesecondelutionpeak(bluein

figure3.6a)wasconcentratedandusedforgelfiltration.Thegelfiltrationchromatogram(figure

3.6b)exhibitsthreeprominentpeaks.Theleftpeakcorrespondstothevoidvolumeandcontains

aggregatedprotein.ThemiddlepeakcontainsTaspasewithimpuritiesoflargerproteinsandthe

right peak (blue in figure 3.6b) contains Taspasewith some impurities below 20 kDa. These

proteinswereprobablydegradationproductsandcouldbesuccessfullyremovedbyusingafilter

with 30 kDa cut‐off for final concentration to 15mg/ml. The overall yields of active Taspase

rangedfrom1to2mgperliterofbacteriaculture.

RESULTS CHARACTERIZATIONOFTASPASEANDITSMUTANTS

71

3.2 CharacterizationofTaspaseanditsmutantsWithpurifiedwildtype,activeandinactiveTaspaseavailable,adetailedbiochemicalcomparison

andcharacterizationwasaccomplished.Dimerizationwasassessedbyanalytical gel filtration,

while CD as well as fluorescence spectroscopy gave insight into the stabilization by sodium

chloride. Possible posttranslational modifications could be excluded with the help of mass

spectrometry, and the autocatalytic processing was followed by SDS‐PAGE and NMR

spectroscopy. Furthermore, first structural information about the loop region, which is not

resolvedinthecrystalstructures,wasobtainedbyNMRandCDspectroscopy.

3.2.1 CDspectroscopicstudies

In order to analyze the secondary structure of the Taspasemutants, far‐UV CD spectrawere

recorded(figure3.7a).

Figure3.7Far‐UVCDspectraofTaspase.

aFar‐UVCDspectraofwildtypeTaspase(wt),activeTaspase(withshortenedloop)andinactiveTaspase(D233A/T234A)indicateasimilarsecondarystructurecompositionofallthreeproteins.Theshapeofthespectrum,especiallytheminimaat222nmand208nmhintatahelix‐richproteinwithbeta‐standportions.b Secondarystructuredeconvolutionof thewtCDspectrumusing theCDSSTRalgorithmconfirm55%helix,23%sheet,11%turnand10%randomcoilelements.Numbersindicatepercentvalues.

AllthreeCDspectraforwildtype,activeandinactiveTaspasearesimilarandexhibitminimaat

220nmand208nmwiththe220nmminimumexhibitingthelowestellipticity.Comparisonwith

areferencedatasetwiththeCDSSTRalgorithmyielded55%helixand23%sheetcontent(figure

3.7b).

3.2.2 StabilityofTaspase

Secondarystructureislost,whenproteinsareunfolded.Hence,CDspectroscopycanbeusedto

monitor thermal unfolding of proteins (figure 3.8a). To find suitable buffer conditions for

subsequentassays,Taspasestabilitywasmeasured indifferentbuffers.AsCDspectroscopy is

limitedwithrespecttochlorideandsucroseconcentration,tryptophanunfoldingviafluorescence

spectroscopywasemployedasorthogonalmethod(figure3.8b).


72

Figure3.8MeltingcurvesofTaspase.

aCDmeltingcurvesofwildtypeTaspase(wtTaspase;black;Tm=59°C),activeTaspase(blue;Tm=60°C)and inactiveTaspase (red;Tm=63 °C) in50mMphosphatebuffer.bTryptophan fluorescencemeltingcurvesshowthatwildtypeTaspasein50mMphosphatebuffer(black;Tm=56°C)canbestabilizedbyadditionof10%sucrose(blue;Tm=73°C)or450mMNaCl(red;Tm=77°C).

CDspectroscopyinphosphatebufferrevealsameltingtemperature(Tm)of59°Cforwildtype

Taspase,whichconsistsofα‐andβ‐subunitandaTmof60°CfortheactiveTaspasewithshortened

loop. In contrast, the inactive mutant is made up of a continuous polypeptide chain and is

somewhatmorestablewithameltingtemperatureof63°C.Fluorescencespectroscopyyieldeda

meltingtemperatureof56°CforwildtypeTaspase.Khanetal.[56]usedchlorideandsucrosein

theirbuffers,whichwastestedhereforastabilizingeffect.Additionof10%sucrose(292mM)

elevates themelting temperatureby17 °C toaTmof73 °C.A similar effect couldbe seen for

450mMNaCl,whichstabilizedTaspaseby21°Cto77°C.

3.2.3 MALDI‐MSanalysis

Massspectrometryallowsaprecisedeterminationofthemolecularweightofproteins.Spectra

wererecordedformassesbetween15and50kDa,whichcoversthemassoftheproenzyme,as

wellasthemassesofthesubunits(figure3.9).

Forthewildtypeenzyme(blackinfigure3.9),fourmajorpeaksarevisible.Theycorrespondto

theproenzyme(T2‐H428)withasodiumandpotassiumionasadduct(Δm/z=0.01%),theα‐

subunit(T2‐D233)withapotassiumionasadduct(Δm/z=‐0.01%)andtheβ‐subunit(T234‐

H428)withasodiumionasadduct(Δm/z=‐0.05%).Thedoublechargedproenzymecausesthe

fourthpeakatexactlyhalfthemassoftheproenzyme(22722Da).

The active Taspase (blue in figure 3.9) exhibits twomain peaks in themass spectrum. They

correspondtotheα‐subunit(H1‐A212)withapotassiumionasadduct(Δm/z=‐0.05%)andthe

β‐subunit(T234‐E420)withasodiumionasadduct(Δm/z=0.01%).

The inactive mutant (red in figure 3.9) exhibits three major peaks, corresponding to the

proenzymewithpotassiumandsodiumionsasadduct(Δm/z=‐0.1%),aswellasthedoubleand

triplechargedproenzyme.


73

Figure3.9MALDI‐TOFmassspectraofTaspase.

Mass spectra ofwild type Taspase (top panel), active Taspase (middle panel) and 15N‐labeled inactiveTaspase (lower panel)were recorded in the range of 15 to 50 kDa. The arbitrary intensity unitswererescaled.Arrowsindicatemassesoftherespectivefull‐lengthprotein(fl),α‐subunit(α),β‐subunit(β),andshortenedα‐subunit(aminoacids1‐206or1‐195).Forimprovedclarity,doubleandtriplechargedmassesarenotlabeled.

3.2.4 Analyticalgelfiltration

AnalyticalgelfiltrationwasperformedtotestforTaspasemultimerization.Thecalibrationrange

coversthesubunits(22kDa),thedimer(45kDa)aswellasthetetramer(90kDa)andcanbe

foundinsupplementalfigure7.7.Gelfiltrationrunsof1to400µMTaspaseandinactiveTaspase

(D233A/T234A) in thepresenceof450mMNaClonaSuperdex200columnaredisplayed in

figure3.10.

BothwildtypeTaspaseandtheinactivemutanteluteintherangeofatetramer(90‐110kDa;gray

area in figure3.10) independentof theprotein concentration.Theexpected range for theαβ‐

dimer(45‐55kDa)ishighlightedingreen.Nopeakcanbedetectedinthisareaforthewildtype

or the inactivemutant. Consistentwith this, no peak for the separate α‐ or β‐subunit can be

detectedat19mlretentionvolume.

Theresultisessentiallythesameatlowersaltconcentrations(100mMNaCl):WildtypeTaspase,

activeandinactiveTaspaseeluteastetramerwithapparentmolecularweightsaround100kDa

(supplementalfigure7.8).


74

Figure3.10AnalyticalgelfiltrationofTaspase.

AnalyticalgelfiltrationchromatogramsofacalibratedSuperdex200columnrevealthatwildtypeTaspase(a) and inactive Taspase (b) elute in the range of 90‐110 kDa (labeled T), corresponding to a hetero‐tetramer.Attheexpecteddimersizeof45‐55kDa(labeledD),nopeakisvisible.

3.2.5 AutocatalyticprocessingoftheTaspaseproenzyme

TheTaspaseproenzymeundergoesspontaneousintramolecularautocatalyticprocessing.Inthis

process,thepolypeptidechainopensbetweenAsp233andThr234,yieldingthecatalyticactive

Thr234with free hydroxyl group. This process is time dependent and can be visualized by a

separationofthesubunitsbySDS‐PAGE(figure3.11a).

Figure3.11AutocatalyticprocessingofTaspase.

a10µMwildtypeTaspasewasincubatedingelfiltrationbufferat37°C.Atindicatedtimepoints,samplesweretakenforSDS‐PAGE.Theexpectedsizesoffull‐lengthTaspase(fl),α‐subunit(α)andβ‐subunit(β)aremarkedwitharrowsinthe15%SDS‐PAGEgel.bNoautocatalyticprocessingwasobservedfortheinactiveTaspasemutantafter7days.

Sixhoursincubationat37°Ccausedanattenuationofthebandforthefull‐lengthenzyme,while

thebandsforα‐andβ‐subunitgainedintensity.Afteroneday,thebandforthefull‐lengthenzyme

is diminished, while both subunit bands became stronger. Theweak band for the full‐length


75

enzymeisvisibleevenafteroneweek.Incontrast,incubationoftheinactivemutantforoneweek

resultedinasinglebandforthefull‐lengthenzyme(figure3.11b).

3.2.6 NMRspectroscopicanalysisofTaspase

Forthefirsttime,TaspasewasanalyzedbyNMRspectroscopyinthisprojecttogaininformation

abouttheflexibleparts,whicharenotresolvedinthecrystalstructures.A1H‐15NHSQCspectrum

of15N‐labeledwildtypeTaspasewasrecordedat30°C(figure3.12a).

Figure3.121H‐15NHSQCspectrumassignment.

1H‐15NHSQCspectrumofwildtypeTaspaseafteroneweekat30°Cingelfiltrationbuffer.Thecrowdedregion in the center of the spectrum (dashed lines) is magnified. For amino acids labeledwith letterswithoutnumbers,onlytheresiduetypecouldbedeterminedunambiguously.

TheHis‐taggedproteincomprises428aminoacids,including15prolines,whichresultsin412

backbone peaks theoretically visible in a 1H‐15NHSQC spectrum.However, only 56 backbone

peaksarevisibleinthe1H‐15NHSQCspectrum.

These peaks were assigned to the corresponding amino acids using backbone (HNCaCb,

CbCaCONH, HNCOCa) and side chain spectra (15N‐TOCSY‐HSQC, 15N‐COSY‐HSQC, HbHaNH,

HbHaCONH,HCCH‐TOCSY)recordedwith15Nand13Clabeledprotein(supplementarytable7.4).

Thisallowedanassignmentof40residuetypes(71%),ofwhich32(55%)couldbeassignedto

specificaminoacidpositions.Moreover,aspectrumrecordedatpH6.5(datanotshown)showed


76

onlymarginalshiftsandthesamenumberofpeakscomparedtothespectrumatpH7.9shownin

figure3.12a.

AnotherstrikingobservationwasatimedependentappearanceofpeaksintheHSQCspectrum.

The spectrum of a Taspase sample recorded immediately after purification exhibits only 22

backboneNH peaks (supplemental figure 7.9). The other 34 peaks become visible by and by

duringtheincubationforfivetosevendays.Aspectrumofasampleincubatedfor30days(data

not shown) still shows the 56 peaks seen after seven days and only few signs of terminal

degradation, indicating that the underlyingprocess is completed after sevendays. The amino

acidsinitiallyvisibleintheHSQCspectrumcouldbeassignedtotheN‐terminus(5residues),the

flexibleloop(3residues)andaloopneartheactivesite(1residue).

Theinactivemutantlackingtheabilityofautocatalyticactivationduetoadoublemutationofthe

active site (D233A/T234A)was 15N isotopic labeled and 1H‐15NHSQC spectrawere recorded

(supplementalfigure7.10)underthesameconditionsusedforwildtypeprotein.Bothproteins

wereincubatedfor7daysat30°Cbeforespectrawererecorded.Comparedwiththewildtype

spectrum,9backbonepeaksofaminoacidslocatedintheflexibleloop(S222‐D233)andthepeak

ofGlu34intheN‐terminusaremissing.

3.2.7 CharacterizationoftheTaspaseloop

Insilicostudies

TheTaspaseloopisreleasedattheC‐terminalendbyanautocatalyticevent,whichcreatesthe

catalyticallyactiveThr234attheN‐terminusoftheβ‐subunit.Suitablealgorithms,suchasJNet,

CFSSPortheYASARAtool,canpredictthesecondarystructureelementsoftheloopbasedonthe

aminoacidsequence(figure3.13a).

All three algorithms consistently predict a helical content around 50 % and the rest as

unstructured.BothN‐andC‐terminusarepredictedasrandomcoilregions,whilethecentralpart

(N185‐S223)ispredictedtoformhelices.

Toinvestigateiftheloopconformationchangesaftertheautocatalyticcleavage,releasingtheloop

attheC‐terminus,homologymodelsoftheproenzymeandactiveTaspaseweregenerated(figure

3.13candd).Inbothmodels,theloopregioncomprisestwointeractinghelices.Theconformation

oftheloop,however,differsbetweentheproenzymeandactiveTaspase.Whileintheproenzyme

theloopisattachedtotheactivesiteandthereforeblockingthecatalyticcenter,inthemodelof

activeTaspasetheloopislocatedoutsidetheactivesite,renderingthecatalyticThr234accessible

forsubstrates.


77

Figure3.13SecondaryandtertiarystructurepredictionoftheTaspaseloop.

aPrimarysequencebasedsecondarystructurepredictionoftheloopregionG178‐D233,whichismissinginthecrystalstructureofactiveTaspase(PDB2a8j).ThealgorithmspredictastretchofhelicesbetweenN185andS223.For theYASARAmodel and the JNetprediction, cylinders representhelicalareas, linesrepresent random coil and loops are represented as curves. Confidence (conf) values range from 0(uncertain)to9(confident).Residuespredictedasburiedbysol25andsol0arelabeled(B).NotethattheN‐terminalGPoriginates fromtheGSTtag.b Intheproenzyme, the loop(redandorange) iscovalentlyattachedtotheactivesite(green).Themodelisbasedonthecrystalstructureoftheproenzyme(PDB2a8i).Theloopaminoacidsthatarestructuredinthecrystalstructureoftheproenzyme,butnotintheprocessedprotein are colored in orange. Newly modeled amino acids are displayed in red. c After autocatalyticactivation,theloop(red)canleavetheactivesite(green).ThemodelisbasedonthestructureofactiveTaspase(PDB2a8j).

Invitroexperiments

Testingthisinsilicomodelinvitrorequirespurifiedlooppeptide.Forthis,the56aminoacidloop

fragmentwasobtainedbyrecombinantexpressioninE.coli.Inaddition,ashorterloopfragment

of 42 amino acids covering only the two helices was synthesized (CASLO ApS, Lyngby).

Recombinant expression of the short fragment was considered ineffective, as short peptide

fragmentsaretypicallypronetodegradation,eveniflongtags,suchasGST,areused.

Expression tests of the long loop in E. coli (supplemental figure 7.11) indicated optimum

expressionconditionsof25°CovernightinLBmediumandadditionof1%TritonX‐100tothe

bufferinwhichthepelletwasresuspended.TheGSTfusionproteinwaspurifiedbyGSHaffinity

chromatography(figure3.14a).


78

Figure3.14PurificationoftheTaspaseloop.

aGSHaffinitychromatogram(toppanel)andtheSDS‐PAGEgelofthefractionsindicatedbyarrows.TheoverexpressedGST‐loop fusionprotein visible in the supernatant before affinity chromatography (S) isboundtothecolumnandtherebyremovedfromtheflow‐through(firstlane).Afterwashing,theGST‐loopiselutedandconcentrated(‐PreSc).PreScission (+PreSc) removes theGST tagandremovalofGST ina10kDacut‐offCentriconwithsubsequentconcentrationoftheflow‐throughyieldedpureloopprotein.bThecorrectsizeoftheloopprotein(6621Da)wasconfirmedbyMALDI‐MS.

ThesupernatantafterultracentrifugationwiththeoverexpressedGST‐loop(Sinfigure3.14a)was

appliedtoaGSHcolumn.Efficientbindingtothecolumnisconfirmedbytheabsenceofthetarget

proteinintheflow‐through.TheelutionfractionscontainGST‐loop,whichisstillimpurifiedby

additional proteins. The GST‐tagwas removed by PreScission protease treatment (+ PreSc in

figure3.14a)andfiltrationthrougha10kDamembraneyieldedpurelooppeptide(Cinfigure

3.14a). TheGST tag andother impuritieswere retained.The correctmassof thepeptidewas

confirmedbyMALDI‐MS(figure3.14b).Withthetwovariantsofthelooppeptideavailable,the

secondarystructurewasanalyzedbyCDspectroscopy(figure3.15a).

Theshortandlongloopdifferintheirsecondarystructure,indicatingastronginfluenceofthe16

additionalaminoacidspresentinthelongerloop(5N‐terminal;11C‐terminal).Thespectrumof

theshort looppeptideexhibitsunambiguous featuresofanunstructuredrandomcoilprotein,

such as positive ellipticity at 220 nm and a minimum at 200 nm. In contrast, the long loop

possessesaminimumat222nmand increasingellipticitybelow200nm,which is typical for

helices.Nevertheless,theminimumat200nmisaclearsignforrandomcoilregionsinthelong

loop.DeconvolutionoftheCDspectra(figure3.15b)yieldsahelixcontentof58%forthelong

loopandonly19%fortheshortloop.

RESULTS CATALYTICACTIVITYOFTASPASE

79

Figure3.15SpectroscopicanalysisandMDsimulationoftheTaspaseloop.

aCDspectraoftheshort(P183‐S222;black)andlongTaspaseloop(G178‐D233;red)showmorehelicalstructure for the long loop.bCDspectrumbasedsecondarystructuredeconvolutionusingCDSSTRandCONTINconfirmahigherhelicalcontent(blue)inthelongloopcomparedwiththeshortloop.Numbersaregiveninpercent.c7nsmoleculardynamicssimulationstartingwithalinearloopindicatesformationoftransienthelices,especiallyaroundArg190.

Toinvestigatehowtheshortlooppeptidebehavesinsolution,a7nanosecondsMDsimulation

was performed (figure 3.15c). Analysis of the adopted conformations over time reveals the

formationanddisorganizationofshorthelicalareas.Mostproneforhelixformationisthestretch

around Arg190, while in the terminal regions and the central part around Lys202 no helix

formationcouldbeobservedduringthesimulation.

3.3 CatalyticactivityofTaspaseQualitativeassaysofTaspaseactivityhavebeenpublishedbefore [56,63,69]. In this thesis,an

improvedassayispresented.Itallowedquantificationofthespecificactivityforthefirsttimeand

revealed different catalytic rates with respect to the two MLL cleavage sites. For this, the

oppositionalroleofsodiumchloride,whichisrequiredforstabilitybutinhibitsTaspaseactivity

[56],hadtobeovercome.Furthermore,amodificationoftheassayfacilitatedadeterminationof

specificTaspaseactivityincelllysates.


80

3.3.1 QuantitativeTaspaseactivityassay

ThecatalyticactivityofTaspasewasassayedwiththehelpofamodelsubstratecomprisingthe

Taspase target sequence. In the intact peptide, anN‐terminal fluorophor is quenched by a C‐

terminaldinitrophenolmoiety,whichisremoveduponcleavage.Afterthereactionisstartedby

additionofTaspase,theintensityincreaseslinearlyaslongassaturatingsubstrateconcentrations

arepresent(figure3.16).Withdecreasingsubstrateconcentrations,thereactionratedropsuntil

allsubstrateisconsumedandaplateauisreached.Ifthetwoaminoacidsadjacenttothecleavage

siteofthesubstratepeptidearesubstitutedbyalanines,noincreaseinfluorescenceintensityis

observedafteradditionofTaspase.

Figure3.16Taspaseactivityassay.

Emission of the fluorogenic substrate increases over time in presence of Taspase. The initial rate andplateauatsubstratedepletionincreasewithsubstrateconcentration.Substratewithmutatedcleavagesite(red)showsnoincreaseinfluorescence.

Asexplicatedinthesectionsabove,sodiumchloridestabilizesTaspaseandpreventsprecipitation.

However,chlorideinhibitsTaspaseactivitywithanIC50valueof68±7mM(supplementalfigure

7.12).Hence,chloridewasomittedfromthebufferandallactivitymeasurementsweretakenin

10%sucrose.

Wild typeTaspasewas incubated inabufferwith10%sucroseat37 °Cand theactivitywas

measuredafterdifferentincubationtimestotestforapotentiallossofactivity(figure3.17).

Figure3.17LossofTaspaseactivityovertime.

IncubationofwildtypeTaspaseat37°Cinmeasurementbuffercontaining10%sucroseshowsalossofactivityovertime.Nonlinearfityieldsahalf‐lifearound2.5hours.


81

After 24 hours, the residual activity of Taspase is less than 10 % compared to the activity

immediatelyafterthawing.Thecalculatedhalf‐lifeofTaspaseactivityisnear2.5hours.Although

sucroseandsaltshowasimilarincreaseofthemeltingtemperature(figure3.8b),incubationfor

24hoursinabufferwith450mMNaClresultedinnodetectablelossofactivity(datanotshown).

Toovercometheinstabilityinmeasurementbuffer,freshTaspasealiquotswerethawedforeach

assay.ThisapproachenabledareproduciblequantitativemeasurementofTaspaseactivity.

Figure3.18ActivityoftheTaspasevariants.

Specific activity of wild type Taspase (wt, black, 0.086 ± 0.0009 µmol * min‐1 * mg‐1), active Taspase(shortened loop, blue, 0.063 ± 0.002 µmol * min‐1 * mg‐1), inactive Taspase (D233A/T234A mutant,0.00003µmol*min‐1*mg‐1)andbuffercontrol(no,0.00006µmol*min‐1*mg‐1)inthepresenceof8µMsubstrate. For the inactivemutant and the buffer control no activitywas observed. Error bars indicatestandarddeviations.

WildtypeTaspase,activeTaspaseandtheinactivemutantwereusedinequalconcentrationsof

300nMforactivityassayswith8µMsubstrate(figure3.18).Theactivemutantexhibits75%of

the activitymeasured forwild type Taspase. In contrast, the inactive Taspasemutant has no

catalyticactivity,whichalsoappliestothebuffercontrolwithoutTaspase.

3.3.2 CleavagesequencespecificityofTaspase

Anatural substrateofTaspase is theMLLprotein.Taspase cleavesMLLat twocleavagesites,

calledCS1andCS2.50%oftheaminoacidsareidenticalforthetwosequences(table3.1).Here,

kineticconstantsforCS1(GKGQVDGADDK)andCS2(KISQLDGVDDK)peptidesarereportedfor

thefirsttime(figure3.19).


82

Figure3.19SubstratespecificityofTaspase.

Michaelis‐Mentenplotsat37°CforthetwoTaspasepeptideswiththecleavagesitesCS1(bluesquares)andCS2(blackcircles)oftheMLLprotein.Catalyticparametersobtainedbynonlinearfittingcanbefoundintable3.1.

Taspasehasanaffinityinthelow‐micromolarrangeforbothtargetsequences,withasevenfoldhigheraffinityfortheCS2peptide(18µM),comparedtotheCS1peptide(2.7µM).Furthermore,themaximumreactionvelocity(vmax)and,hence,alsotheturnovernumber(kcat)differsbetweenthetwosubstrates.Bothvaluesareabout3timeshigherforCS2thanforCS1(CS1:vmax=0.03±0.003µmolmin‐1mg‐1,kcat=0.0274s‐1;CS2:vmax=0.11±0.002µmolmin‐1mg‐1,kcat=0.083s‐1).Asignificant difference can be seen if the specificity constants (kcat/Km) are compared. Here,cleavageoftheCS2peptideismorethan20‐foldmoreefficientthanCS1cleavage.Table3.1CatalyticparametersofTaspasetargetsequencesat37°C.

CS1[Mca]‐GKGQVDGADDK‐[DNP]

CS2[Abz]‐KISQLDGVDDK‐[DNP]

Km(µM) 18.4±2.8 2.7 ±0.1

vmax(µmolmin‐1mg‐1) 0.03±0.003 0.11±0.002

kcat(s‐1) 0.0274 0.083

kcat/Km(lmol‐1s‐1) 1487 31190

[Mca]:7‐amino‐4‐methylcoumarin;[DNP]:dinitrophenol;[Abz]:anthranilicacid

3.3.3 Taspaseactivityinhumancelllysates

TaspaselevelsineukaryoticHela,HEKandHCT116cellsarelowandthereforebelowdetection

levels the western blots performed in this project (supplemental figure 7.13). Consequently,

Taspase levelshadtobe increasedbyoverexpression.Toallowasubsequentquantificationof

Taspaseconcentration,aGFPtagwasadded,whichdoesnotimpairthecatalyticactivity[70].This

approachyieldedTaspaseconcentrationsof370nMinHelacelllysates.Additionally,theassay

temperaturehadtobeloweredfrom37°Cto30°Ctopreventprecipitationofthelysisbufferafter

supplementedwithPMSF.TocomparetheactivityofHela lysateswithoverexpressedTaspase

andTaspaseisolatedfromE.coli,measurementsweretakenunderidenticalconditions(figure

3.20).


83

Figure3.20SpecificactivityofeukaryoticTaspase.

Specificactivityat30°Cinthepresenceof8µMsubstrateofHelalysateswithoverexpressedTaspase‐GFPfusionprotein(transfected)andpurifiedTaspaseafterheterologousbacterialexpression(bacterial).BothTaspasespeciesshowasimilarspecificactivity.

Hela lysates andHela controls transfectedwith an empty PC3 vector showed nomeasurable

Taspaseactivity(datanotshown).Incontrast,thelysatewithoverexpressedGFP‐Taspaseisable

tocleavethesubstratepeptide.ComparedwithbacteriallyexpressedandpurifiedTaspase,the

specificactivityofoverexpressedTaspaseinHelalysatesisnearlyidentical.

3.3.4 TaspaseinhibitiontestusingNSC48300

Withthisassayinhands,theonlyavailableTaspaseinhibitorNSC48300describedbyChenetal.

[63],whichhasbeencontroversiallydiscussed[133],couldbetested(figure3.21).

Figure3.21InhibitiontestwiththereportedTaspaseinhibitorNSC48300.

aThereportedTaspaseinhibitorNSC48300exhibitsnoinhibitoryactivityatconcentrationsupto333µM,comparedtoDMSOcontrols.NotethatthereportedKiis4.22µM.bStructureofNSC48300.

In the in vitro activity assay, the inhibitory effect observed by Chen et al. [63] could not be

confirmed.NSC48300yieldednodecreaseincatalyticactivity,evenatconcentrationsof333µM,

RESULTS EXPLORATIONOFTASPASEINHIBITORSBYVIRTUALDOCKING

84

which exceeds the reportedKimore than70‐fold. Thepresence andpurity ofNSC48300was

confirmedbyESI‐LC‐MSanalysis(supplementalfigure7.14).

3.4 ExplorationofTaspaseinhibitorsbyvirtualdockingInthisproject,differentapproacheswerefollowedtofindnovelTaspaseinhibitors.Oneofthese

approacheswas a virtual docking of a large compound library. Subsequently, the top binding

candidateswerepurchasedandtestedforinhibitioninvitro.Alltestedcandidatesfailedtoexhibit

aninhibitoryeffect.

3.4.1 Modelpreparation

Virtualdockingbasicallyrequirestwomolecules,termedligandandreceptor.Asreceptor,two

differentstatesofTaspasewereused:Ontheonehandtheproenzyme,wherebindingofasmall

moleculecouldinterferewiththeautocatalyticactivationprocessand,thereby,keepTaspasein

the inactiveproenzymestateandontheotherhandtheactiveTaspaseprotein,whereasmall

moleculecouldinhibitthecatalyticactivitybyblockingtheactivesite.

Inafirststep,themissingloopareaswereaddedtothemonomersofthecrystalstructures(figure

3.22a)toincludetheN‐terminus(M1‐R40),theTaspaseloop(A206‐S229)andashortloopinthe

β‐subunit(S352‐Q362).Thecorrespondingmodelqualitystatisticscanbefoundinsupplemental

figure7.15.

Figure3.22PreparationofTaspasemodelsforvirtualdocking.

aModelofwildtypeTaspaseproenzymebasedonPDB2a8iwithmissingloops(red)modeled.YASARAhomologymodelingwasusedtoaddtheN‐terminus(M1‐R40),theTaspaseloop(A206‐229)andashortloop in the β‐subunit (S352‐Q362). These residueswere not defined in the crystal structure.b Spherecluster(blue)generatedbysphgen.TheclustercoversthecatalyticThr234(red).cTheactivesite(green)wassurroundedwithacube(yellow),inwhichtheenergygridswereprecomputed.Inthisarea,thespacewasperformed.

Partial charges were calculated with the help of Chimera [112] and suitable target areas for

dockingwereidentifiedbythesphgenalgorithm(figure3.22b).Theareacoveringthesitewas

selected as ligandbinding site and in a cubewith30Å edge length theenergygrids for rigid

dockingwereprecomputed(figure3.22c).


85

Fortheliganddatabase,alargevarietyofcompoundsandavailabilityofthecompoundsfor in

vitrousewasrequired.TheZINCdatabase[129]metbothcriteria,with12millionpurchasable

compoundsavailable.Toreducethenumberofcompoundstoareasonableamountfordocking

and still cover a maximum of different conformations, chemically different molecules were

selectedwiththehelpoftheSUBSETalgorithm[130].Onlymoleculeswhichdifferfromallother

by at least 60%were used for docking.With this approach, the number of compoundswas

reducedfrom12millionto15865.

3.4.2 VirtualdockingofZINCdatabasecompounds

Asafirststep,thecompoundsoftheZINCdatabaseweretreatedasflexibleanddockedwithan

anchor‐and‐growalgorithm(section2.6.3).Taspaseresidueswithin5Åoftheactivesitewere

regardedasflexible,aswell.Thisallowedafastandyetsufficientlyaccurateexplorationofseveral

hundreddifferentbindingconformationsforeachcompound.Thefreebindingenergyforthetop

poseofeachligandwascomputedwithaprecomputedgrid,whichincludesvanderWaalsand

electrostaticcomponents.Adistributionofthebindingenergiesisshowninfigure3.23.

Figure3.23Energydistributionofthedockedcompounds.

a+b15865compoundsoftheZINCdatabaseweredockedontheproenzyme(a)oractiveTaspase(b)withananchor‐and‐growalgorithm.ArankingofthefreebindingenergiesfollowsaGaussiandistribution(blacklines).Forcompoundswithbindingenergybelow40kcal/mol(red),similarcompoundsweresearched.c+dThesesimilarcompoundsweredockedinasecondroundontheproenzyme(c)oractiveTaspase(d)andanAMBERscoringfunctionwasusedforrankingafterashortMDsimulation.


86

The sample size of 15865 compoundswas large enough to obtain a Gaussian distribution of

bindingscores(figure3.23aandb)withameanbindingenergyof26kcal/molfortheproenzyme

and 30 kcal/mol for the active Taspase. The best binding compounds were used for the

subsequentdockinground.Asthreshold,twicethestandarddeviationwasused,whichresulted

inacut‐offbindingenergyof40kcal/mol.Thiscriterionwasmetby1%oftheproenzymedataset

and2%oftheactiveTaspasedataset.

Sinceafiltereddatasetwasused,inwhichonlycompoundsthatdifferfromeachotherbyatleast

60%werepresent,thedatasetofthetopcompoundswasexpandedbytheinclusionofsimilar

molecules.Forthis,theZINCdatabasewassearchedforcompoundswithatleast70%similarity

to the topcompounds.Thisapproach increasedthenumberofcompoundsto1925(from363

compoundsfortheproenzyme)and7313(from1817compoundsforactiveTaspase)andensured

thatderivativesofthetopcompoundswereincludedinthenextdockinground.

Toincreasethereliabilityofthebindingscore,anenergyminimizationandshortMDsimulation

wasperformedforeachcompoundafterdocking,beforethebindingenergywascalculatedusing

anAMBERforcefieldscoring.ThescoringdistributionsfortheproenzymeandactiveTaspaseare

shown in Figure 3.23c and d, respectively. Again, a Gaussian distribution is almost obtained

indicatingasufficientsamplesize.

3.4.3 Inhibitiontestswithcompoundsobtainedbyvirtualdocking

For both datasets (proenzyme and active Taspase), the top poses of the modeled Taspase‐

inhibitor‐complexeswereinspected, therespectivecompoundswerereviewedfortoxicityand

availability.Figure3.24aandbshowsthedockedconformationsoftwotopcandidatesfoundfor

theproenzyme,referredtoasZINC52513265andZINC19323748,respectively.Bothcompounds

block the catalytic Thr234 (cyan in figure 3.24a and b) and have contact to the surrounding

residues(paleorangeinfigure3.24aandb).


87

Figure3.24Dockedconformationandinhibitiontestofputativeproenzymeinhibitors.

a+bDockedposesofZINC52513265(a)andZINC19323748(b)withtheproenzyme.ThecatalyticThr234ishighlightedincyanandinteractingTaspaseresiduesinpaleorange.c+dAdditionofZINC52513265(c)or ZINC19323748 (d) shows no significant inhibition of Taspase activity. The structures of the twocompoundsareinserted.

Asharedfeatureofthesetwocompoundsisthesulfategroup,whichinteractswiththeThr234.

Anoverlayofthedockedposesofthetoptencompoundsrevealsthatthisinteractionofasulfate

groupwiththeThr234isasharedfeatureamongthesecompounds(figure3.25).

Figure3.25Sharedfeaturesamongtopcandidates.

aAnoverlayofthetencompoundswithbestbindingscoresrevealscommonfeatures.bAdditionallytoacentralsulfatemoiety,twoaromaticrings(5‐and6‐membered)arepresentinmoststructures.


88

Thiscommonframeworkisdiversifiedbydifferentmoietiesatthearomaticringsanddifferent

linkersfromthesulfatemoietytotherings.Forthisreason,twocompoundswereselected,which

containacentralsulfategroupbutdifferinmostoftheothermoieties.Thesecompoundswere

purchased,incubatedataconcentrationof200µMwithTaspasefor6htoallowaninhibitionof

theautocatalyticactivation,andweretestedinthefluorogenicassay.However,theproteolytic

Taspase activity is not inhibited in presence of up to 1mM ZINC52513265 (figure 3.24c) or

600µMZINC19323748(figure3.24d).

Evaluation of the best binding compounds in the dataset for active Taspase showed a more

heterogeneousgroupofinhibitorsthantheproenzymedataset.Asulfategroupispresentinsome

inhibitors,aswell,butanoverlayofthetopinhibitorsrevealednostrikingcommonframework.

Hence, three compounds (ZINC63428205, ZINC21096526, and ZINC21096535)were selected

randomlyforinvitroanalysis(figure3.26a‐c).

Figure3.26DockedconformationandinhibitionofputativeactiveTaspaseinhibitors.

a‐c Structures of ZINC63428205 (a), ZINC21096526 (b) and ZINC21096535 (c). d‐f Docked poses ofZINC63428205(d),ZINC21096526(e)andZINC21096535(f)withactiveTaspase.ThecatalyticThr234ishighlightedincyan,interactingTaspaseresiduesinpaleorangeandthechlorideioningreen.gAllthreeZINCcompoundsexhibitnoinhibitoryeffectintheTaspaseactivityassay.

Allthreecompoundscagethechlorideionintheactivesiteanddonotdirectlyinteractwiththe

catalyticThr234(figure3.26d‐f).Interestingly,thecompoundsshowninfigure3.26eandfcage

the chloride ion at different positions, despite their close structural similarity. In theTaspase

activityassay,however,all threecompoundsexertno inhibitoryeffect inconcentrationsupto

600µM(figure3.26g).

RESULTS INHIBITIONOFTASPASEBYNANOPARTICLES

89

3.5 InhibitionofTaspasebynanoparticlesRecent studies disclosed binding of serum proteins to nanoparticles [99]. Furthermore, silica

nanoparticlescaninhibittheproteasome,theonlyotherN‐terminalnucleophile(Ntn)threonine

protease (Shirley Knauer, personal communication). Hence, binding of Taspase to silica

nanoparticles and their effect on the proteolytic activity was tested. Three different sizes of

nanoparticles inhibitedTaspaseat lownanomolarconcentrations,while theactivitiesof three

controlenzymeswerenotimpaired.

3.5.1 NanoparticlesbindandinhibitTaspaseinvitro

BindingofTaspase tosilicananoparticleswasobserved inapreliminaryexperimentduringa

microscale thermophoresis (MST) demonstration course. Here, Taspase was added to

nanoparticleswith8or125nmdiameter,andbindingwasobservedlabel‐freeinsolution(figure

3.27a).

Figure3.27NanoparticlesbindTaspase.

aMicroscalethermophoresis(MST)titrationsrevealbindingofTaspasetonanoparticles.For8nm(Amsil8)and 125nm (Amsil125) silica nanoparticles EC50 values of 35 nM± 0.5 nM and10 pM± 0.4 pMwereobtained.bFluorescenceanisotropytitrationswith1µMAtto488‐labeledTaspaseyieldastoichiometryofsome 10 µmol/mg nanoparticles. This corresponds to roughly 5000 Taspase molecules per Amsil125nanoparticleand6TaspasemoleculesperAmsil8nanoparticle.

Both sizes of nanoparticles were able to bind Taspase with high affinity. The obtained EC50

concentrationswereinthenanomolarrangeforAmsil8(35nM±0.5nM)andinthepicomolar

rangeforAmsil125(10pM±0.4pM).ActivesitefluorescenceanisotropytitrationswithAtto488‐

labeledTaspaseconfirmedtherangesofbindingaffinitiesdetectedbyMSTanddisplayedahigher

change in anisotropyuponbinding forAmsil125 than forAmsil8due to thehighermolecular

weight. Based on these data, a stoichiometry of roughly 10 nmol per mg nanoparticles was

calculatedforbothAmsil8andAmsil125(figure3.27b).Thiscorrespondstoapproximatelysix

TaspasemoleculesboundperAmsil8nanoparticleand5000TaspasemoleculesperAmsil125

nanoparticle.Toexplorepossible consequencesofnanoparticles forTaspaseactivity, thiswas

testedinaninvitroassay(figure3.28).


90

Figure3.28NanoparticlesinhibitTaspaseactivitynoncompetitively.

a Silica nanoparticles with 8 nm (Amsil8) 20 nm (Amsil20) and 125 nm (Amsil125) diameter inhibitTaspaseactivity.EC50valuesof41±9nM,1.4±0.3nMand119±21pMwerecalculatedforAmsil8,Amsil20and Amsil125, respectively. b Fluorescence anisotropy of Atto488‐labeled changes in the presence ofAmsil125, while the anisotropy of the fluorogenic substrate and the Atto488‐labeled control proteinparvalbuminaltersonlymarginally.cThechangeinanisotropyuponadditionofAmsil125nanoparticlesisthesameforwildtypeTaspase(wt)andthe inactiveTaspasemutant(red), indicatingthattheybindtonanoparticlesinthesamemanner.WildtypeandinactiveTaspaseproteinswerelabeledwithAtto488.dTitrations with Amsil125 at different substrate concentrations pinpoint a noncompetitive inhibitionmechanism(R2=0.98).AKivalueof380±55pMwascalculated.

Allthreesizesofsilicananoparticlestested(8,20and125nmdiameter)inhibitedtheTaspase

activity already at low concentrations. As observed for binding to nanoparticles, the EC50

concentrationsarenegativelycorrelatedwiththenanoparticlediameter.Hence,Amsil125hasthe

highest affinity with 119 ± 21 pM, while Amsil20 and Amsil8 bind Taspase at nanomolar

concentrations(EC50=1.4±0.3nMand41±9nM,respectively).

However, nanoparticles would artificially lower Taspase activity, if they bind the fluorescent

substrateandrenderitunavailableforTaspasecleavage.Tothisend,bindingofthesubstrateto

Amsil125 nanoparticles was analyzed by fluorescence anisotropy and compared to Taspase

binding (figure 3.28b). Since only the anisotropy of Taspase changed in the presence of

nanoparticles,whiletheanisotropyofthesubstrateremainedvirtuallyconstant,bindingofthe

substratetonanoparticlescanbeexcludedasanexplanationfortheinhibitionofTaspaseactivity.

Interestingly, theproenzymebinds toAmsil125comparably to thewild typeenzyme.For this

experiment, thetitrationwasperformedwithAtto488‐labeled inactiveTaspasemutant(figure

3.28c).Furthermore,theconstantanisotropyoftheAtto488‐labeledcontrolproteinparvalbumin


91

excludesanAtto488‐dependentbindingtonanoparticles.Onlyathighconcentrations,abinding

ofparvalbumintonanoparticlescouldbeobserved(supplementary figure7.16).However, the

EC50 value of 3 ± 0.8 µM determined for this interaction is 300000‐fold higher compared to

Taspase.

To gain a first insight into the inhibitorymechanism, the inhibition typewas determined for

Amsil125 by a variation of both substrate and nanoparticle concentrations (figure 3.28d).

NonlinearfittingrevealsareversiblenoncompetitiveinhibitionwithaKivalueof380±55pM.

One possible explanation for the observed inhibition is a destabilization of Taspase by

nanoparticles, which causes unfolding and thereby reduces the activity. Hence, the impact of

nanoparticlesonthesecondaryandtertiarystructureofTaspasewasanalyzed.Tothisend,CD

spectrawererecordedtoassessthesecondarystructure(figure3.29a).

Figure3.29Nanoparticlesdonotaltersecondarystructureorstability.

aCDspectrainthepresenceofnanoparticleswith8nm(Amsil8)20nm(Amsil20)and125nm(Amsil125)diametershownochangeinsecondarystructurecontentofTaspase.bTryptophanfluorescencemeltingcurvesrevealnoalterationofTaspasestability.MeltingpointsinphosphatebufferwithoutNaCl(solidlines)arenear56°Candwith450mMNaClnear77°C(dashedlines),bothinthepresenceandabsenceof100µg/mlnanoparticleswith8nm(Amsil8)20nm(Amsil20)and125nm(Amsil125)diameter.The obtained spectra in the presence and absence of nanoparticles are indistinguishable,

indicating that neither nanoparticle type caused a change in secondary structure content at

inhibitory concentrations of 40 µg/ml. However, a change in the tertiary structure, such as a

movementortorsionofawholesubunit,cannotbeexcludedbythesedata.

Consequently, fluorescence tryptophan melting curves were recorded to get insight into the

tertiarystructure,aswellastheoverallfoldingstateandstabilityofTaspase(figure3.29b).As

explicated in section3.2.2,Taspasehas amelting temperaturenear56 °Cand77 °C inbuffer

without andwith 450mMNaCl, respectively. These values do not change in the presence of

100µg/mlnanoparticles(=266nMAmsil8;16nMAmsil20;204pMAmsil125).

Takentogether,thesedataargueagainstchangesinsecondaryortertiarystructureandindicate

thattheinhibitionbynanoparticlesisnotcausedbyadestabilizationofTaspase.


92

3.5.2 Enzymeinhibitionisnogeneralfeatureofnanoparticles

So far, inhibition of enzyme activity could be not only restricted to Taspase, but instead be a

general feature of nanoparticles. To test the hypothesis of a general inhibitory effect of

nanoparticles on enzyme activity, three control enzymes were studied in activity assays,

measuringeitherlactateformationbylactatedehydrogenase(LDH)orproteolyticactivityofthe

serineendopeptidasechymotrypsinandoftheserineendo‐andexopeptidaseproteinaseK(figure

3.30).

Figure3.30LDH,chymotrypsinandproteinaseKareweaklyinhibitedbynanoparticles.

a Lactate dehydrogenase (LDH) activitywas assayed in the presence of silica nanoparticleswith 8 nm(Amsil8) 20 nm (Amsil20) and 125 nm (Amsil125) diameter. Amsil8 shows weak inhibition at highconcentrations(3mg/ml).Amsil20inhibitstheactivitybyupto25%withananomolarEC50.Amsil125reducestheactivitybyupto15%.bChymotrypsinactivity isnot inhibitedbynanoparticleswith8nm(Amsil8)20nm(Amsil20)and125nm(Amsil125)diameter.cProteinaseKactivityisnotinfluencedinthepresenceof100µg/mlAmsil8.Forconcentrationsbetween0.05and50µg/mlproteinaseK,thedigestionpatterndoesnotchangeinthepresenceofnanoparticles.fl,αandβrepresentthesizeoffull‐lengthTaspase,α‐andβ‐subunit,respectively.

Amsil125 inhibitsLDHactivitybyup to15%at lownanomolar concentrations (figure3.30a;

blue).HigherconcentrationscausenofurtherreductionofLDHactivity.ForAmsil20,inhibitionis

notseeninthepicomolarrange,whereTaspaseisalreadysignificantlyinhibited.However,higher

concentrationscanreduceLDHactivitybyupto25%.Similarly,micromolarconcentrationsof


93

Amsil8arerequiredtoinhibitLDHactivityby10%,whichistwoordersofmagnitudemorethan

forTaspasewithanEC50valueof41nM.

Concerning the second control enzyme chymotrypsin, the enzymatic activity is not alteredby

nanoparticles(figure3.30b).Neitherofthenanoparticlescouldreducetheproteolyticactivity,

eveniftheconcentrationsexceededtheEC50ofTaspasebytwoordersofmagnitude.

ToassaytheactivityofthethirdcontrolenzymeproteinaseK,theTaspasedigestionpatternwas

analyzed by SDS‐PAGE after limited proteolysis by proteinase K (figure 3.30c). The digestion

bandsbelowboththefull‐lengthenzymeandthetwosubunitsindicatethathighconcentrations

ofproteinaseKcandegradeallthreeTaspasespecies.Acomparisonofthedegradationpatternin

the presence and absence of 100 µg/ml Amsi8 reveal no inhibition for all 11 proteinase K

concentrationstested.

3.5.3 NanoparticlesinhibitTaspaseincelllysatesandincells

Studieswiththreecontrolenzymesindicatedthatinhibitionofenzymeactivityseemedtobeno

general feature of nanoparticles. Hence, the selective inhibition of Taspase activity by

nanoparticlescouldalsobepossibleincelllysates.ThishypothesiswastestedinHelacelllysates

withoverexpressedGFP‐Taspase(figure3.31).

Figure3.31NanoparticlesinhibitTaspaseactivityincelllysates.

TaspaseactivitywasassayedinHelacelllysatesinthepresenceofnanoparticleswith8nm(Amsil8)20nm(Amsil20)and125nm(Amsil125)diameter.NonlinearfittingyieldedEC50valuesof80±34µg/ml(=213±90nM)forAmsil8,251±24µg/ml(=40±4nM)forAmsil20and42±15µg/ml(=86±31pM)forAmsil125.

OnestrikingfindingisthatnanoparticlesreduceTaspaseactivityonlybyupto50%inlysates.

Moreover, the EC50 values (40‐250 µg/ml) are somewhat higher than for pure Taspase (8‐

60µg/ml),whichadditionallyreducestheinhibitoryeffectincelllysates.

However,theinhibitoryconcentrationsinlysateswerestillinthenanomolarandsub‐nanomolar

range.Sincenanoparticlescanbetakenupbycells[99],theconsequentexperimentwastotest

the effect of nanoparticles on cellular Taspase activity. This endeavor required a cell‐based

activityassay,inwhichafluorescentreportersubstrate,calledBioTasp[71],withnuclearimport

andexportsequenceswasused(figure3.32a).


94

Figure3.32CellularTaspaseactivityassay.

aTheBioTaspreportersubstratecomprisesanN‐terminalnuclearlocalizationsignal(NLS)eGFP(green),glutathioneS‐transferase(GST),Taspasecleavagesites(red;scissors)andareversenuclearexportsignal(NES).bBioTasp(green;secondcolumn)isfoundequallydistributedinnucleusandcytoplasminHelacellswithout Taspase (first row) and Hela cells transfected with inactive Taspase (second row) due to thepresenceofbothnuclearlocalizationandexportsignal(NLSandNES,respectively).NucleiwerestainedwithHoechst(blue;firstcolumn).InthepresenceofwildtypeTaspase(wtTaspase;thirdrow),thenucleusexportsignal(NES)iscleavedandthenuclearlocalizationsignal(NLS)causesanaccumulationofBioTaspinthenucleus.cStatisticalanalysisshowsnosignificant(ns)differenceinthenucleustocytoplasmratio(N/C)ofBioTaspbetweenuntransfectedcellsandcellswithinactiveTaspase.Incontrast,theN/CratioissignificantlyhigherincellswithwildtypeTaspase.50cellsfrom3independentexperimentswerecounted.***indicatep‐values<0.001.

BioTaspcontainsagreenfluorescentprotein(GFP)fusedtoaGSTtopreventpassivediffusion

throughthenuclearpores.ThesensorshuttlesbetweennucleusandcytoplasmbyitsNLSand

NES.Duetoawell‐consideredcombinationofsignalstrength,thesteadystateofthesensorisan

even distribution throughout the cell (figure 3.32b). The localization of BioTasp changes, if

Taspaseisoverexpressed.TaspasecleavesBioTaspattheintegratedCS2cleavagesitefromMLL

andremovestheNES,whichleadstoanaccumulationofBioTaspinthenucleus.Controlswith

overexpressedinactiveTaspaseshowadistributionsimilartocellswithoutTaspase.Statistical

evaluationconfirmedasignificantlyincreasednuclearlocalizationincellswithwildtypeTaspase


95

comparedtocellswithinactiveornoTaspase(figure3.32c;N/CratiosfornoTaspase:1.25±0.03,

inactiveTaspase:1.03±0.03,wildtypeTaspase:2.81±0.1).

Withthisassay,thesilicananoparticlesof8,20and125nmdiameterweretested.Forthis,cells

were challenged with 200 µg/ml nanoparticles 3 hours after transfection. The presence of

nanoparticles had no visible effect on the cell viability ormorphology (supplementary figure

7.17).TheBioTasplocalizationwasanalyzedafter18hoursincubation(figure3.33).

Figure3.33NanoparticlesinhibitTaspaseactivityincells.

aHelacellstransfectedwithTaspase(red;mCherry)andtheBioTaspreportersubstrate(green;GFP)werestainedwithHoechst(blue)aftertreatmentwith200µg/mlAmsil8(firstrow),Amsil20(secondrwo)orAmsil125(thirdrow)for18h.TheequaldistributionofthegreenBioTaspreportersubstrateindicateslowTaspaseactivityinpresenceofnanoparticles.Hoechst,GFPandmCherrychannelsaredisplayedseparatelyincolumns1‐3,respectively,andmergedincolumn4.bStatisticalanalysisofthenucleartocytoplasmratio(N/C) shows a significantly reducedN/C ratio in thepresence of nanoparticles compared to untreatedcontrols.50cellsfrom3independentexperimentswerecounted.***indicatep‐values<0.001.

Figure3.33ashowsanevendistributionofBioTaspthroughoutthewholecellafternanoparticle

challenge.TheredfluorescenceconfirmsthesuccessfuloverexpressionofwildtypeTaspaseand

its perfect overlap with the Hoechst staining indicates an exclusively nuclear localization.

Statisticalevaluationrevealsthatallthreetypesofnanoparticlessignificantlyreducethenuclear

localizationofBioTasptoalevelmidwaybetweencellswithwildtypeandinactiveTaspase(figure

3.33b).SincethecontrolcellswithwildtypeTaspasecanberegardedas100%activityandthe

cellswithinactiveTaspaseas0%activity,nanoparticlesreducetheTaspaseactivityincellsby

roughly50%at200µg/ml.Furthermore,theinhibitoryeffectofAmsil125seemstobesomewhat

lowercomparedtoAmsil8andAmsil20.

In summary, inhibition of Taspase activity in cell lysates and cellswas observed for all three

nanoparticlesatnanomolarconcentrations.

RESULTS INHIBITIONOFTASPASEBYPEPTIDYL‐SUCCINIMIDYLPEPTIDES

96

3.6 InhibitionofTaspasebypeptidyl‐succinimidylpeptidesBesides nanoparticles and compounds obtained by virtual docking, substrate analog Taspase

inhibitors designed by Prof.MarkusKaiserwere tested in this project. Starting point for this

endeavor was the proteolytic cleavage mechanism proposed by the Bogyo group [69]. This

mechanism involves the formation of a succinimide‐hydrate intermediate (figure 3.34),which

facilitatesthepeptidebondtransferof theproteinbackbonetotheaspartatesidechainof the

substrate.

Figure3.34Designofthepeptidyl‐succinimidylpeptide.

Intheproposedcleavagemechanism[69],theaspartatesidechainattacksthefollowingglycine(redarrow).Thisreactionyieldsasuccinimide‐hydrateintermediate(toppanel;succAsphydrate),whichwasusedasstartingpointforsubstrateanalogTaspaseinhibitors.Thepotentialinhibitormotifpeptidyl‐succinimidyl(succAsp)isshowninthetoprightpanel.InthetestcompoundComp2(bottompanel),theLeuatP2wassubstituted by Val to reduce the polarity. The Fmoc protection group was left for detection by UVabsorption.Thisfigurewastakenfrom[134].

Thenovel classof substrateanalog inhibitorsdeveloped for this thesis featuresa succinimide

moiety (figure 3.34). Optimization of the initial inhibitors yielded the most potent Taspase

inhibitorsofarwithaKiof0.4±0.1µMinvitro.Moreover,thiscompoundwasactiveincelllysates,

albeittheaffinityis30‐foldattenuated.Furthermore,thedevelopedinhibitorscanenterthecell,

however,theachievedintracellularconcentrationsweretoolowforameasurableinhibitionof

Taspase activity. This project was accomplished in cooperation with the Kaiser group, who

designedandsynthesizedthecompounds.

3.6.1 Peptidyl‐succinimidylpeptidesinhibitTaspaseinvitro.

Thepeptidyl‐succinimidylpeptidesweresynthesizedbytheKaisergroupusingFmoc/tBusolid

phasepeptidesynthesis.Leaving theN‐terminal fluorenylmethyloxycarbonyl (Fmoc)groupon


97

thefinalproductallowedanexactspectroscopicdeterminationoftheconcentration.However,

thisadditionalaromaticmoietycouldchangetheinhibitorypropertiesofthecompound.Torule

outaninhibitionbytheFmocgroup,theinhibitoryeffectofFmoc‐valine(Comp1)wastestedin

the in vitro assay (supplemental figure 7.18a). This titration resulted in a negligible effect on

Taspaseactivity,evenathighFmoc‐Valconcentrations.

Inthenextstep,thedesignedinhibitor(figure3.34)aswellaseightderivativesweretitratedto

determineIC50andKivalues(table3.2andsupplementalfigure7.18).

Table3.2Inhibitorypotentialofthesuccinimidylpeptidesandtheirderivatives.Deviationsfromthecommoninhibitorscaffoldarehighlightedinyellow.Compound Sequence* IC50 [µM] Ki[µM]Comp1 Fmoc‐Val >1000 n.d.Comp2 Fmoc‐Gln‐Val‐succAsp‐Gly‐Val‐Asn‐Asp‐NH2 38.7±12.1 4.3±1.3Comp3 Fmoc‐Gln‐Val‐piperidineAsp‐Gly‐Val‐Asn‐Asp‐NH2 30.1±10.6 3.3±1.2Comp4 Fmoc‐Gln‐Val‐allocAsp‐Gly‐Val‐Asn‐Asp‐NH2 143±39 15.9±4.3Comp5 Fmoc‐Gln‐Val‐succAsp‐βAla‐Val‐Asn‐Asp‐NH2 61±9.5 6.8±1.1Comp6 Fmoc‐Gln‐Val‐succGlu‐Gly‐Val‐Asn‐Asp‐NH2 14.4±2.8 1.6±0.3Comp7 Fmoc‐Gln‐Val‐Asp‐Gly‐Val‐Asn‐Asp‐NH2 >1000 n.d.Comp8 Fmoc‐Gln‐Val‐isoAsp‐Gly‐Val‐Asn‐Asp‐NH2 424±58 47±6.4Comp9 Fmoc‐Prg‐Bpa‐Gln‐Val‐succAsp‐Gly‐Val‐Asn‐Asp‐NH2 3.6±0.8 0.4±0.09

Comp10 Rhd‐Gln‐Val‐succAsp‐Gly‐Val‐Asn‐Asp‐NH2 373±23 41±2.6*Fmoc:fluorenylmethyloxycarbonyl;succAsp:succinimide;piperidineAsp:Aspwithpiperidineprotectiongroup; allocAsp: Asp with alloc protection group; βAla: β‐alanine; succGlu: glutarimide; isoAsp:isoaspartate;Prg:propargylglycine;Bpa:benzophenonealanine;Rhd:rhodamineB

Comp2exhibitedapromisinginhibitoryeffectinthelowmicromolarrange(Ki=4.3±1.3µM).To

ensurethattheinhibitorwasnotsimplyusedascompetitivesubstrateandcleavedbyTaspase,

thesuccinimidemoietywasomitted,yieldingapeptidicTaspasesubstrate(Comp7intable3.2).

Incontrast,thiscompoundwasdegradedwithoutareductionofTaspaseactivity.Thisindicates

thatComp7isdegradedrapidly,whilethesuccinimidemoietyinComp2trulyactsasaninhibitor

andcannotbedegradedbyTaspase.

Duringsynthesis,asideproductwithanLC‐MSmassequivalenttoapiperidineadditionproduct

could be isolated (Comp3). Although not further analyzed, this compound was presumably

generatedbyanadditionofpiperidinetothesuccinimideresidue.TheKiandIC50valuesofComp3

aresimilartothoseofComp2(Ki=3.3±1.2µM).Anothersideproductoccurringduringsynthesis

displayed an LC‐MS mass corresponding to a compound with alloc‐protected Asp (Comp4).

However,comparedwithComp3,theinhibitoryeffectofComp4wasfivefoldlower(Ki=15.9±

4.3µM).

Intheproposedcleavagemechanism[69],theproteinbackboneofthesubstrateistransferredto

theaspartatesidechain,forminganisoaspartateintermediate(figure3.35).


98

Figure3.35Isoaspartateintermediateduringproteolysis.

Intheproposedcatalyticmechanism,thesuccinimide‐hydrateintermediate(succAsphydrate)isusedtotransfer theproteinbackbone to theaspartatesidechain.Thisyieldsan isoaspartate (isoAsp),which isneededforTaspasecleavage.

TotestifisoaspartatecantrulybecleavedbyTaspaseorinhibitsthereaction,asubstratewith

isoaspartateinsteadofaspartatewassynthesized(Comp8).Thiscompoundexhibitedatenfold

weakerinhibitionofTaspasethanthesuccinimidecompoundComp2(IC50=424±58µM),which

excludedisoaspartateasinhibitorymoietyforTaspase.

For further improvementof thebindingaffinity, theroleof theS’ siteofTaspase for inhibitor

bindingwasexplored.Tothisend,theP’residueglycinewasreplacedbyβ‐alanine(Comp5).This

smallstructuralchangecausedaminorreductionofthebindingaffinity(Ki=6.8±1.1µM).

Foranextensionofthe5‐memberedsuccinimideringtoa6‐memberedring,thesuccinimidewas

replacedby a glutarimide (Comp6).This change slightly improved the affinity to aKi of 1.6±

0.3µM.

The most promising compound was the photoreactive inhibitor Comp9, which possesses an

additional N‐terminal benzophenone as well as a propargyl glycine group. With a Ki in the

nanomolarrange(0.4±0.09µM)thiscompoundshowedatenfoldhigherinhibitorypotentialthan

the best published Taspase inhibitor. Binding of Comp9 was confirmed by microscale

thermophoresis(MST)asalternativemethod(figure3.36a).

Figure3.36Bindingofthepeptidyl‐succinimidylpeptideinhibitors.

BindingofComp9wasverifiedbymicroscalethermophoresis(MST)titrations.AKdvalueof22.1±0.6µMwasobtained.


99

In contrast to the activity assay, MST allows the determination of binding constants in the

presenceofNaCl.Hence,theMSTmeasurementswereperformedinthepresenceof450mMNaCl

tomaintainTaspasestability.Thisprocedureallowsinspectingiftheinhibitorbindsintheactive

siteandcompeteswiththechlorideion,whichwasfoundintheactivesite[56].

AcomparisonoftheaffinityofComp9inthepresence(22.1±0.6µM;MST)andabsenceofNaCl

(0.4±0.09µM;activityassay)reveals50‐foldweakerbindinginthepresenceofNaCl.Thisfinding

supports that binding is competitive with respect to chloride. This was also found for the

substrate,indicatingasimilarbindingmodeofsubstrateandinhibitor.

3.6.2 Peptidyl‐succinimidepeptidesareresistantagainstTaspasecleavage

To verify the competitive type of inhibition, Taspase inhibitionbyComp2was assayed in the

presenceofdifferentsubstrateconcentrations(figure3.37a).Nonlinearfittingofthedatawitha

model for competitive inhibition confirms that Comp2 truly acts as a reversible competitive

inhibitor.

Figure3.37Stabilityofcompetitivepeptidyl‐succinimidylpeptides.

aCompetitiveinhibitionofpeptidyl‐succinimidylpeptideswasconfirmedwithComp2.AKiof2.5±2µMwasobtainedbynonlinearfitting.bRhodamine‐labeledinhibitorwasincubatedwith1µMTaspasetoassaystability.Theinhibitorshowsnodecreaseinanisotropy,whiletheanisotropyofthesubstratecontroldropsduetocleavage.cTaspasewasincubatedwithandwithout63µMComp2,beforeactivitywasassayed.Thecalculated inhibitionreflectstheratiooftheactivities inthepresenceandabsenceofthe inhibitoraftercertainincubationtimes.Constantinhibitionbyroughly60%indicatesasufficientlyhighstabilityoftheinhibitoragainstdegradation.Barsindicatestandarderrors.

Competitiveinhibitionofaprotease,however,carriestheriskofproteolyticdegradationofthe

inhibitorduetoitssubstrateanalognature.Hence,thestabilityofthesuccinimidemotifagainst

Taspasemediatedproteolysiswastested.Thisrequiredthesynthesisofafluorescentderivative

oftheinhibitor(Comp10).AlthoughtheincorporationofabulkyrhodaminedyereducetheKi

twentyfold (Ki = 41 ± 2.6 µM), the affinity was still sufficiently high for stability tests. The

fluorescentdyeallowedfollowingthedegradationovertimebyfluorescenceanisotropy(figure


100

3.37b).TheanisotropyobtainedforComp10inthepresenceofTaspasewasconstantover5days.

Incontrast,theanisotropyofthesubstratedropsrapidlyduetoproteolyticcleavage.

ToexcludeaninfluenceofthefluorescentdyeattachedtoComp10,Taspasewasincubatedwith

63µMComp2,andtheactivitywasassayedafterdefinedtimes(figure3.37c).Theinitialinhibition

by roughly 60% remained constant even after incubation for 24 hours.However, the overall

TaspaseactivitywasreducedovertimeduetotheinstabilityofTaspaseintheassaybuffer(see

section3.2.2).

Altogether, these experiments demonstrate the stability of peptidyl‐succinimide inhibitors

againstproteolyticcleavagebyTaspase.

3.6.3 Peptidyl‐succinimidepeptidesbindandinhibitTaspaseincelllysates

Thenextsteptowardsinvivoexperimentswastotesttheinhibitorypotentialofthedeveloped

peptidyl‐succinimidepeptidesincelllysates.Tothisend,Helacelllysatewassupplementedwith

recombinantTaspaseandincubatedwiththephotoreactivecompoundComp9.Irradiationwith

UVlightcross‐linkedtheinhibitorwiththeboundprotein.Asubsequentadditionofarhodamine

azideviaclickreactionallowedvisualizationbyfluorescenceimaging(figure3.38).

Figure3.38Photolabelingandlysateinhibitionwithpeptidyl‐succinimidepeptides.

aTaspase(5µM)wasaddedtoHelacelllysatesandsubsequentlycross‐linkedwith5µMoftheUV‐reactiveComp9inhibitor.Aclickreactionwithrhodamineamidealloweddetectionviafluorescence.Theα‐subunitandthebandforthefull‐length(fl)Taspasearevisibleinthegel,butnottheβ‐subunit.Cross‐linking,clickreactionandfluorescencegelimagingwasperformedbySusanneZweerink,Kaisergroup.bComp9inhibitstheactivityofoverexpressedTaspaseinHelalysateswithanIC50of111±30µM.

Thefluorescentbandsat45kDaand28kDacorrespondtofull‐lengthTaspaseandtheα‐subunit,

respectively. This indicates that the inhibitor binds to the α‐subunit of Taspase, while no

interactionwith the β‐subunit is visible. Furthermore, no interaction above backgroundwith

other proteins in the lysatewas observed. However, the comparatively high concentration of

recombinantTaspase(5µM)addedtothelysatesshouldbeconsideredintheinterpretationof

thisfinding.

TheinhibitorypotentialwasassayedinHelalysatestransfectedwithGFP‐Taspase.Theassaywas

performed in lysis buffer containing 150 mM NaCl. Under these conditions, the inhibition is

roughly30‐timesweaker(IC50=111±30µM)thanforrecombinantTaspaseintheinvitroassay.


101

3.6.4 Peptidyl‐succinimidylpeptidesenterthecell

Althoughthemolecularweightofthepeptidyl‐succinimidepeptidesofnearly1kDaisratherlargetoenter the cell, their ability to cross the cellmembranewas tested. For this,Hela cellswere

treatedwiththefluorescentcompoundComp10.Afterdefinedincubationtimes,thecellswere

washed and the intracellular fluorescence was used as a measure for cellular uptake of the

compound(figure3.39a‐c).

Figure3.39ThepeptidicinhibitorsenterthecellbutexhibitnoTaspaseinhibition.

aThefluorescencelabeledpeptidicinhibitorcanenterHelacells.Theintracellularfluorescenceintensityisincreasedafter1h(middlepanel)and5h(rightpanel)comparedtothebackgroundintensity(leftpanel).bStatisticalanalysisconfirmselevatedfluorescenceduetoinhibitorentryafter1and5hours.Theinhibitoruptakeisincreasedbyincubationintheabsenceofserum.cHelacellstreatedwith50µMComp9for20hshownosignificant(ns)inhibitionofTaspaseactivityincomparisontountreatedcells(wtTaspase).InthiscellularTaspaseactivityassay,thenucleustocytoplasmratio(N/C)correspondstotheTaspaseactivity.20(c)or50(e)cellsfrom3independentexperimentswereevaluated.***indicatep‐values<0.001.

UntreatedHelacellsshowaverylowbackgroundfluorescenceintheRFPchannel(figure3.39a).

Aftertreatment for5hourswiththepeptide, thecellsshowadistinctredfluorescenceevenly

distributedthroughoutthecell,indicatingacellularuptakeofthefluorescentinhibitor.Statistical

evaluationof thesamplesreveals that the increase in fluorescence ishighlysignificant (figure

3.39b).Evenafterincubationfor1hour,thecellsdisplayanincreasedfluorescence,whichisnot


102

visiblebyeye(figure3.39a).Notably,theincubationinserum‐freemediumincreasedtheuptake

of the inhibitor more than twofold in the samples treated for 5 hours. With fluorescence

microscopy,adeterminationoftheintracellularconcentrationoftheinhibitorwasnotpossible.

Therefore,theuptakecouldonlybeconfirmedqualitatively.

Nevertheless,theinhibitoryeffectofthepeptidyl‐succinimideinhibitorwastestedinHelacells

usingacell‐basedactivityassay(section3.5.3).Tothisend,Helacellsweretreatedwith50µM

Comp9for18hours,beforecellswerefixatedandthelocalizationofthereportersubstratewas

evaluated(figure3.39c).However,cellstreatedwithComp9displayedthesameTaspaseactivity

asuntreatedcells.

Insummary,thepeptidyl‐succinimidepeptidessuccessfullyinhibitedrecombinantTaspaseand

celllysates,butshowednoinhibitoryeffectinthecellularactivityassaydespitecellularuptakeof

thecompounds.

DISCUSSION TASPASEISAUNIQUEPROTEASEANDREQUIRESUNIQUEINHIBITORS

103

4 DiscussionProteolyticactivationisacommonregulatorymechanismforvariouscellularprocesses.Inthe

caseof a t(4;11) translocation,Taspase is thekeyenzyme for theproteolyticactivationof the

oncogenic MLL protein. Consequently, inhibition of Taspase could prevent the malignant

transformationofcellswitht(4;11)translocationandprobablyalsosolidtumors.

Forthedevelopmentofnewinhibitors,adetailedunderstandingofthebiochemicalproperties,

the catalytic mechanism and the structure of the target enzyme are of tremendous value.

Therefore, Taspase was subjected to extensive biochemical characterization with respect to

structure, stability, autocatalytic activation, posttranslational modification and substrate

specificity.Withthisknowledge,novelTaspaseinhibitorswereevaluatedinvitro,incelllysates

andincells.ThisapproachyieldedtwonovelclassesofTaspaseinhibitors:Silicananoparticles

andsubstrateanalogsuccinimidepeptides.

4.1 TaspaseisauniqueproteaseandrequiresuniqueinhibitorsTaspasebelongstothetypeIIasparaginasefamilyandsharesthetypicalαββαasparaginasefold,

anautocatalyticactivationprocess,aswellasanN‐terminalnucleophile(Ntn)attheβ‐subunit.

However,despitethesesimilarities,TaspasediffersfromothertypeIIasparaginasesinsomekey

properties,whichwereanalyzedindetail.

4.1.1 CatalyticconstantsofTaspase

TheMLLproteinpossessestwosimilarTaspasecleavagesites(CS1andCS2;table3.1)withCS2

beingmoreconservedandoptimalforcleavage[57].Withaquantitativeactivityassayinhand,

thiscleavagesitespecificitycouldbequantifiedforthefirsttime(figure3.19).AlthoughtheCS1

andCS2modelsubstratesusedfortheinvitroassaydifferslightlyinthefluorescentdyesattached

attheN‐terminus,theeffectonthecatalyticactivityisexpectedtobenegligibleduetosufficient

distancetothecleavagesite.Theresultingcatalyticconstantsrevealedlowmicromolaraffinities

oftheCS1andCS2substrates(Km=18and3µM,respectively),whichisingoodagreementwith

thepreviouslypublishedaffinityforCS1(12µM;[56]).TheturnovernumbersforCS1andCS2

cleavagewerequantified for the first time(0.03and0.08s‐1, respectively)andarerather low

compared to other asparaginases, for example from the enterobacterial Pectobactrium

carotovorum (120 s‐1; [135]), from Yersinia pseudotuberculosis (0.22 s‐1; [136]), or from the

hyperthermophilicPyrococcusfuriosus(888s‐1at80°C;[137]).Theresultingspecificityconstant

confirmsatwentyfoldhighercleavagepreferenceofCS2overCS1.

Post‐translationalmodificationsareacommonmechanismfortheregulationofenzymeactivity

andhavealsobeenspeculatedforTaspase[71].ForTaspaseproteinexpressedheterologouslyin

E.coli,post‐translationalmodificationscouldnotbedetectedbyMALDI‐MS(figure3.9).While

methylation and acetylation have been described for E. coli [138], enzymatic regulation by

phosphorylationiscommonineukaryotes,butrarelyobservedinE.coli.However,thefactthat


104

Taspaseexpressedinbacteriaandeukaryoticcellsexhibitthesamespecificactivity(figure3.20),

suggests that Taspase activity does not need activation by post‐translationalmodifications in

eukaryoticcells.However,modificationsunderstressconditionssuchasphosphorylationcannot

beexcludedwiththesedata.NotethattheGFP‐tagneededforthisexperimentisreportedtohave

noinfluenceontheactivity[71].

Moreover,thesedataruleoutthehypothesisthatbacteriallyexpressedTaspasedisplaysonlyan

attenuatedactivity[71].Additionally,theenzyme/substrateratiorequiredforefficientsubstrate

cleavage is below 1:500, which is more than two orders of magnitude more efficient than

published [71]. One plausible explanation that the catalytic activity in vitro is higher than

published before refers to the instability of Taspase in the assay buffer. In this regard, the

requirement of salt for stability was pinpointed (figure 3.8) and NaCl concentrations below

450 mM caused precipitation of Taspase (data not shown). To avoid inhibition by chloride,

Taspasewasstabilizedbysucroseforactivitymeasurementsanditwasdemonstratedforthefirst

timethatalbeitsucrosestabilizesTaspase,thehalf‐lifeisstillonlynear2.5hours.Hence,Taspase

requiresfastpurificationwithhighsaltconcentrationsandshortincubationtimestomaintainits

catalyticactivity.Insufficientconsiderationofanyoftheserequirementswasfoundtoresultina

dramaticlossofactivity(figure3.17andfigure3.8).

4.1.2 AutocatalyticprocessingofTaspasetakesseveraldays

While the general principle of autocatalytic processing of a proenzyme is shared among

asparaginases, the kinetics of the activation rates varies significantly between different

asparaginases(table4.1).Forinstance,activationofE.coliasparaginasetakeslessthanoneday

[62],contrarytothefivefoldsloweractivationofFlavobacteriummeningosepticumasparaginase

[139]. The slowest activation rate was reported for human asparaginase with 90 days. For

Taspase,theonlyreportedvalueisthatittakeslongerthan6hours[63].

Table4.1PropertiesofTaspase‐homologproteins.

ProteinHomology

sequence structureOligomericstate

Xtal solution AutocatalysisTaspase 100% 0Å dimer dimer 3daysAsparaginase 36% 0.9Å dimer dimer+monomer 90daysF.m.Glycoamidase 27% 1.1Å dimer monomer <5daysE.c.Asparaginase 21% 1.5Å dimer dimer <1daysProteasome 14% 4.7Å heptamer heptamer α‐SU‐dependent

Adetailedstudyof theautocatalyticprocessprovided for the first time informationabout the

Taspase activation rate. Separation of the active subunits from the proenzyme by SDS‐PAGE

indicatesthattheactivationprocesstakesapproximatelyoneday(figure3.11).Importantly,the

activationofTaspasedoesnotbeginatthestartoftheassay,butalreadyduringthepurification

processoftheprotein.ThisbecomesclearbycomparisonoftheSDSgelsofthebacteriallysate


105

(figure 3.3; almost only proenzyme), after affinity purification (figure 3.4; predominantly

proenzyme)andaftergelfiltration(figure3.5;ca.50%proenzyme).Hence,thetwodaysneeded

forproteinpurificationhavetobeaddedtotheactivationtimedeterminedintheassay,yielding

atotalactivationtimeofthreedays.Moreover,thehighamountofproenzymeinthebacterial

lysatesafterexpressionovernightcouldhintatalowerprocessingincellscomparedtobuffer

conditions.

Interestingly,thebandfortheproenzymelosesintensityovertime,butdoesnotdisappeareven

aftersevendays.Twoexplanationsarepossibleforthisphenomenon.EitherafractionofTaspase

isstillintheproenzymeform,orthebandiscausedbyanSDS‐resistantdimerofthetwosubunits.

Since the autocatalytic processing is an irreversible spontaneous event and no equilibrium

reaction,astagnationofthisprocessisunlikely.Additionally,thepresenceofanSDS‐resistantαβ‐

dimerisinagreementwiththehighaffinityofthesubunitsdiscussedbelow.

Moreover,Nommeandcolleaguesreportedanaccelerationofthehumanasparaginaseactivation

processbyadditionofglycine[66].However,Taspaseactivationwasindependentofglycine,salt,

sucroseormalate(datanotshown).

Togainfurtherinsightintothisactivationprocess,time‐dependentNMRstudieswereapplied.

With this technique, the activation can be followedmore detailed thanwith SDS‐PAGE at the

atomic level. The NMR peak intensities of the amino acids in the fragment released by the

autocatalyticactivationevent(referredtoasTaspaseloop)increasedovertime(supplemental

figure 7.9). Since the assigned peaks in the 1H‐15N‐HSQC spectrum of Taspase correspond to

flexible loop regions and disordered regions (see section 4.1.4 for detailed explanation), the

appearanceofpeakscorrelateswithan increased flexibilityof thecorrespondingaminoacids.

Therefore, the increasing flexibilityof the loop residues reflects theautocatalyticactivationof

Taspase.

Thepeakintensitiesoftheloopresiduesreachtheirmaximumafterapproximatelysevendays,

indicatingthattheautocatalyticactivationiscompletedafterthisperiod.Hence,comparedtothe

SDS‐PAGE‐based analysis of the autocatalytic activation, the NMR experiments indicate a

somewhat slower activation rate. This difference can presumably be attributed to the lower

incubation temperature during the NMR experiment (30 °C) compared to the SDS‐PAGE

experiment(37°C),astheprobabilityforanucleophilicattack,whichisneededfortheactivation

process,isusuallytemperaturedependentasdescribedfortheactivationofotherNtn‐hydrolases

[139].Additionally,bothexperimentsareonlyqualitative.

Insummary,thesefirstdataoftheTaspasematurationprocessrevealthatTaspaseactivationin

vitro takes approximately three days,which is significantly shorter than its closest structural

homolog,humanasparaginase(90days).

4.1.3 Taspaseisahetero‐tetramerwithhighaffinity

After autoproteolysis, the subunitsof asparaginase type II familyproteins assemble to anαβ‐

heterodimer. In Taspase, the αβ‐heterodimer displays a remarkably high stability. This high

affinity of the subunits allowed a co‐purificationwith aHis‐tag only present at the α‐subunit


106

(section3.1.4).Furthermore,thetwosubunitswereretainedbya30kDafiltermembraneduring

purification,althougheachsubunitissmallenoughforpassage(25and20kDa).Accordingly,a

FRET‐basedassay,inwhichAtto488‐labeledTaspasewasmixedwithAtto594‐labeledTaspase,

failed todetect an exchangeof the subunits indicatingnoor slowsubunit exchange (datanot

shown).Alookatthecrystalstructurerevealsthatthehighaffinityofthesubunitsisbasedona

largehydrophobicinterfacewithapositivelychargedpatchoftheα‐subunitburiedinanegatively

chargedgrooveintheβ‐subunit(figure4.1).

Figure4.1InterfaceofthetwoTaspasesubunits.aTheα‐subunitpossessesapositivelychargedpatch(blue)attheinterfacetotheβ‐subunit(yellow).bConversely, theβ‐subunit showsanegatively chargedpatch (red),whichallows interactionwith theα‐subunit(yellow).ChargesweremappedonamodelbasedonPDB2a8j.

Whiletheαβ‐dimerisgenerallyacceptedforTaspase,thehetero‐tetramericoligomerofTaspase

seen in all asparaginase‐like crystal structures [56,66,139,140] is, however, discussed

controversially.Ontheonehand,Taspaseunambiguouslyadoptsahetero‐tetramericstructure

uponcrystallizationandthelargehydrophobicinterfacebetweenthetwoαβ‐dimerssuggestsa

highstabilityofthecomplex[56].Ontheotherhand,pull‐downexperimentsandgelfiltrations

indicated only a low affinity of the two αβ‐heterodimers in solution [68]. Nevertheless, these

contradictory experiments can be brought in line by a consideration of the Taspase

concentrations.Athighconcentrationsused for crystallography,Taspase is ahetero‐tetramer,

whereasthelowconcentrationsusedinthebiochemicalassaysfavortheαβ‐dimer.Accordingly,

alsothehomologF.meningosepticumglycosylasparaginaseisadimerinsolution,despiteshowing

aheterotetramerinthecrystalstructure(table4.1and[140]).

Forthefirsttime,concentrationdependentgelfiltrationswereperformedwithTaspasetoshed

lightontheoligomericstateofTaspaseinsolution.Forconcentrationsbetween1and400µM

Taspasewasexclusivelyfoundasαββα‐heterotetramer(section3.2.4).Cellularconcentrationsof

Taspaseareassumedtobemuchlower,butcouldnotbeassessedbygelfiltration,astheyfall

below the detection limit of theUV detector [68]. However, the fact that the 90 kDa Taspase

hetero‐tetramercanpassa50kDafiltermembrane(supplementalfigure7.19)indicatesthatthe

complexisnotasstableasproposedbyKhanandcolleaguesbasedonthedimerinterfaceareain

thecrystalstructure[56].


107

4.1.4 StructureoftheTaspaseloop

Intheprevioussections,theautocatalyticmaturationoftheTaspaseproenzymewasdiscussed.

ThisprocessyieldsaC‐terminalstretchattheα‐subunitofallNtn‐hydrolases,whichismissingin

allavailablecrystalstructuresduetoalackofelectrondensityforthisregion(figure4.3b).Inthe

caseofTaspase,thisregioncomprises56aminoacids(G178‐D233)andisreferredtoasTaspase

loop.Ascrystallographyfailedtoprovideinformationaboutthispartoftheprotein,NMRandCD

spectroscopywereusedascomplementarymethodstogainfirstinformationabouttheflexible

Taspaseloop.

Thefirststrikingfindingwasthatthe1H‐15NHSQCspectrumoffull‐lengthTaspaseshowslow1H

signaldispersionof0.8ppm(between7.8and8.6ppm),whichistypicalforunstructuredregions

[141–143].Tomaptheseresidues,therespectiveareasinthemodelwereaddedbyhomology

modeling(figure4.2).

Figure4.2Mappingofthe1H‐15NHSQCsignals.Mapping of the assigned residues on a model of wild‐type Taspase based on the structure PDB 2a8j.Assignedresiduesarehighlightedinred;theactivesiteismarkedingreen.MostoftheassignedresiduescanbefoundintheflexibleN‐terminusandtheflexibleTaspaseloop.

Indeed,theassignedsignalsinthespectrumcorrespondtoflexibleregionsintheN‐terminus(15

residues) or the Taspase loop (12 residues), while the rigid parts with a defined secondary

structure remain invisible.Hence, themissing regions of theTaspase crystal structure canbe

selectivelyinvestigatedbyNMRspectroscopy.AnimportantfactorforthesignalintensityinNMR

spectroscopyistherotationcorrelationtimeoftheprotein,whichcorrelateswiththeproteinsize.

This rotational correlation time is related to both the longitudinal (T1) and transversal (T2)

relaxationtimes.LargeproteinslikeTaspaseexhibitlongT1andshortT2times,whichresultsin

line‐broadeningintheresultantNMRspectrumduetoalossofmagnetizationandrendersthe

peaks invisible. However, the large loop regions of Taspase obviously behave like separate

disordereddomainswith independentrotationalcorrelation times.Thus, therigidcoreregion

losesitsmagnetizationquickly,whiletheflexibleregionspossesslowerrotationcorrelationtimes

andthesignalscanberecordedbeforethemagnetizationislost.Conversely,theinactivemutant

doesnotcleaveitself,sothattheloopretainstherotationalcorrelationtimeofthecoreandis

thereforeinvisibleintheNMRspectraofinactiveTaspase(figure7.10).


108

Thus,accordingtoNMRspectroscopicstudiestheTaspaseloopseemstobeflexible.Furthermore,

thesedataarein linewiththehelix‐turn‐helixmodelgeneratedbyhomologymodeling(figure

3.13).Essentially,theaminoacidsprecedingSer222werenotvisibleintheNMRspectra,whereas

the region C‐terminal of Ser222 was seen in the NMR spectra. This strongly supports the

occurrenceofamorerigidsecondarystructureendingatGlu221.

Basedonthesedata,thefirststructuralmodelwasgeneratedfortheloopregionmissinginthe

crystalstructureofTaspaseandallotherasparaginases.Themodelfulfilsthecriteriamentioned

above,namelyacentralregionofdiscontinuoushelices,aswellasadefinedstructureC‐terminal

ofSer222.

Figure4.3ProposedstructureoftheTaspaseloop.

a The proposed structure is based on NMR and CD spectroscopy data as well as secondary structureprediction.Themodelshowsahelix‐turn‐helixmotif.NLSaminoacidsarehighlightedinorange,criticalNLSresiduesaredisplayed inred.Theoverlayof twopossibleconformations indicates flexibilityof thetermini.bLocationoftheloopintheTaspaseprotein(PDB2a8j).

The 3D model generated by YASARA displays the proposed helix‐turn‐helix structure of the

Taspaseloopwithflexibletermini(figure4.3a).Interestingly,thehelix‐turn‐helixmotifendsat

Glu221andallowstheinteractionofthetwohelices,sothatthebipartitebasicNLS[74]formsa

continuouspatch(redinfigure4.3a).Furthermore,theinteractionofthetwohelicesagreeswith

theinterfacevisibleinthewheelprojections(datanotshown).

Asmentionedbefore,allothertypeIIasparaginasescompriseaflexibleloopregionN‐terminalof

the catalytic site that is disordered in the crystal structures of human glycosylasparaginase,

human asparaginase, plant asparaginase, E. coli asparaginase, and F. meningosepticum

glycosylasparaginase.Sequence‐basedmodelsoftheloopfortheseTaspasehomologs(figure4.4)

revealthatonlytheTaspaseloopispredictedtobepredominantlyhelical.Incontrast,theother

loops are predicted as pure random coil (human asparaginase, E. c. asparaginase, and F.m.

glycosylasparaginase) or random coil interspersed with short helices (human

glycosylasparaginaseandplantasparaginase).


109

Figure4.4AsparaginasetypeIIloopmodels.

Modeledstructuresof theTaspase loophomologousregions inother type IIasparaginases.The loopofhumanTaspase(templatePDB2a8i; loop length:25aa) ispredictedaspredominantlyhelical,while forhumanglycosylasparaginase(templatePDB1apy;looplength:20aa),humanasparaginase(templatePDB4pvs; loop length: 14 aa), plant asparaginase (template PDB 2gez; loop length: 33 aa),Escherichia coliasparaginase (template PDB 2zal; loop length: 17 aa), and Flavobacterium meningosepticumglycosylasparaginase(templatePDB1ayy;looplength:10aa)mainlyrandomcoilelementsarepredicted.Themodeledloopregionsarehighlightedinred;theactivesiteThr234ishighlightedingreen.

Although loops shorter than the 25 amino acids Taspase loop are unstructured (F. m.

glycosylasparaginase: 10 amino acids; human asparaginase: 14 amino acids; human

glycosylasparaginase:20aminoacids),alsothelongerplantasparaginaseloop(33aminoacids)

possessespredominantlyrandomcoilelements.

FurtheranalyseswiththeisolatedloopproteinconfirmedthehelicalstructureoftheTaspaseloop

and revealed that the terminal loop region are required for helix formation (figure 3.15a).

Moreover,MDsimulationsoftheloopwithouttheterminiindicatethatalthoughshorthelicescan

form,theyareonlytransientandcannotestablishadefinedtertiarystructure(figure3.15c).

4.1.5 InhibitionstrategiesforTaspase

Dueto thekeyroleofTaspase in theactivationof theoncogenicMLLproteinassociatedwith

t(4;11)leukemias,Taspaseinhibitorsareofhighdemand.However,althoughseveralstrategies

havebeenfollowedtofindinhibitors,abreakthroughisstillawaitedandnovelTaspaseinhibitors

areurgentlyrequired.


110

Incontrast, inhibitionof thehomologenzymeasparaginaseisundesiredformedicalpurposes.

Rather,highasparaginaseactivityisdesiredforcancertreatment,sinceasparagineisessentialfor

leukemic cells. Consequently, depletion of asparagine by asparaginase leads to inhibition of

protein synthesis and subsequent cell death [144]. Thus, asparaginase is successfully used as

chemotherapeutic and was released as E. coli asparaginase (Elspar™), PEGylated E. coli

asparaginase(Oncaspar™),orErwiniachrysanthemiasparaginase(Erwinase™)forthetreatment

ofchildhoodALL[145].

Asexplicatedintheprevioussections,TaspasediffersfromallothertypeIIasparaginasesinits

substrate specificity and the catalytic mechanism. This is assumed to be the reason, why

commercial protease inhibitors, suchas the serineprotease inhibitorPMSFor the cOmplete™

inhibitor cocktail, but also the threonine protease inhibitor bortezomib [63], do not inhibit

Taspase.However,thisuniquepositionofTaspaseoffersvariouswaysfornewinhibitors,some

ofwhichhavealreadybeenmentionedintheintroduction(section1.3.4):

Compoundscreeningyieldedanallostericarsenicinhibitor(NSC48300)withaKiof4.2µM[63].

However, this compound is not specific for Taspase and was initially identified as an anti‐

angiogenic drug. Besides the toxicity of the arsenic moiety, this compound has been under

discussion, since inhibition could not be reproduced in a cellular assay [133]. This lack of

inhibitionbyNSC48300washereconfirmedinvitro(figure3.21),whichquestionsthepotential

ofthecompoundevenmore.

Virtualdocking byKnauerandcolleagues resulted in twocompounds (CHC‐A4andDHC‐C1)

withweakinhibitorypotential[71].TheauthorsdiscussedthemissingpartsintheTaspasecrystal

structureandlimitedcellentryofthecompoundsaspossiblecausesforthepoorscreeningoutput.

Furthermore,thecomparativelysmallsizeofthecompounddatabasemightalsohavelowered

thenumberofhits.

Forthesereasons,thevirtualdockingwasrepeatedwithanimprovedTaspasemodelincluding

themissingloopregionsandafiftyfoldlargercompoundlibrary(section3.4.2).Additionally,the

hurdleofcrossingthecellmembranewaseliminatedbyusinganinvitroassayforthereadout.

Nevertheless,theresultingcompoundsdisplayednoinhibitionofTaspaseactivity.Thissupports

thetheorythatTaspaseundergoesastructuralchangeuponsubstratebinding,whichwouldlimit

thevalueofthecrystalstructureforvirtualdockingapproaches.

SubstrateanalogswerefirsttestedbyLeeandcolleagueswithaseriesofsubstrate‐derivedvinyl

sulfoneandvinylketonederivatives [69].Despite themoderate inhibitorypotential (IC50near

30µMforthebestcompoundyzm18),thesestudiesprovedhelpfulforthedesignofnewsubstrate

analog compounds. These succinimide peptides tested in this thesis are discussed separately

below(section4.3).

EnforceddimerizationbythefusionofJun‐andFos‐tagstotheC‐terminusofTaspaseresulted

in a loss of catalytic activity [68]. This finding seems tobe in contrast to the catalytic activity

obtainedwithTaspase at concentrationwhere the dimer is the predominant oligomeric state

(figure3.37b).Asimpleexplanationforthesecontradictoryfindingsisthattheenforceddimer

adoptsadifferentconformationthanthenaturaldimer.Particularly, theC‐terminallyattached

dimertagspresumablybringtheC‐terminiincloseproximityintheenforceddimer.However,the

DISCUSSION NANOPARTICLESASNOVELINHIBITORS

111

C‐termini of the natural dimer aremore than 60Å apart on the opposite ends of the hetero‐

tetramer(figure7.20).Intheenforceddimer,asterichindrancecouldthereforepreventsubstrate

binding and thus abolish catalytic activity. Furthermore, the artificial generation of higher

oligomericTaspasespeciesbyJun/Fosmediateddimerizationwasnotexcluded,whichmightbe

anotherreasonforthemissingactivity.

Molecularplugs are a comparatively newmechanism to occlude the active site by artificial

ligands. The Schmuck group achieved an inhibition of the serine protease β‐tryptase with

nanomolarKiusingafour‐armedligandtoblocktheactivesitenoncompetitively[146,147].This

approach could also be applicable to Taspase, if the compounds sterically prevent substrate

binding.Consequently,41ofthesemolecularplugderivativeswereobtainedfromtheSchmuck

group and tested for Taspase inhibition. Although these compounds covered combinations of

different charges, lengths and hydrophobicity, none of them was able to inhibit Taspase

(supplementalfigure7.21andfigure7.22).

4.2 NanoparticlesasnovelinhibitorsOwing to theirhigh surface tovolumeratio,nanoparticlespossessahigh free surfaceenergy,

which allows the adsorbance of othermolecules [148]. The interaction of nanoparticles with

biomoleculesisalreadyusedforabroadrangeofdiagnosticandtherapeuticmethodsinmedicine

[149].Recentstudiesrevealedthattheproteasome,theonlyotherNtn‐typethreonineprotease,

bindstosilicananoparticles(ShirleyKnauer,unpublisheddata).Forthisreason,Taspasebinding

tosilicabasednanoparticleswasinvestigated.

4.2.1 Taspasebindingdependsonnanoparticlesize

Indeed,microscalethermophoresisandfluorescenceanisotropyexperimentsshowedbindingof

Taspasetosilicananoparticlesatlownanomolarconcentrations(section3.5.1).Furthermore,in

contrasttoTaspase,thecontrolproteinparvalbuminshoweda300000‐foldweakeraffinity.To

understand why some proteins bind to nanoparticles, while others are rejected, the protein‐

nanoparticle interface needs to be examined in more detail. However, the interactions of

nanoparticleswithbiomoleculesarestillnotfullyelucidated,despiteextensiveresearchinrecent

years.Nevertheless, someconcepts emerged from theseefforts,whichhelp tounderstand the

influenceof nanoparticlesonproteins [92]. Someof thepossible interactionsaredisplayed in

figure4.5.


112

Figure4.5Interactionsofproteinswithnanoparticles.

Bindingtonanoparticles(red)isdependentonprotein‐specificon/offrates,aswellashydrophobicity,sizeandchargeoftheparticlesthemselves.Thegreencirclesrepresentvariousproteins.Modifiedaccordingto[92].

Obviousfactorsthatruletheinteractionsarechargeandhydrophobicityoftheparticleandthe

protein.Surprisingly,arecentstudyrevealedthatnegativelychargedproteinsshowincreased

bindingtonanoparticles,irrespectiveoftheirnetcharge[99].Otherimportantdeterminantsare

thesizeandcurvatureofthenanoparticlesandstericfactors.Theproteinlayersurroundingthe

nanoparticlesiscalledcorona.Duetotheparametersoutlinedinfigure4.5,thiscoronaisspecific

foreach typeofnanoparticleandestablishes rapidly.Hence,nanoparticles interact selectively

withonlysomeproteinsanddonotbindeveryprotein[99,150].Thisisinagreementwiththe

bindingofTaspasetosilicananoparticles,whileparvalbuminisrejected(section3.5.1).

For the interactionofTaspasewithnegativelychargedamorphoussilicaparticles (Amsil), the

surfaceareaandcurvature showedagreat influenceonTaspasebindingaffinityand capacity

(figure 3.27). TheEC50 of nanoparticleswith125nmdiameter (Amsil125) is 3500‐fold lower

compared with nanoparticles of 8 nm diameter (Amsil8). This is also reflected in the

stoichiometryofroughly5000and6TaspasemoleculeperAmsil125andAmsil8nanoparticle,

respectively,whichisingoodagreementwiththestoichiometryof620:1reportedforHSAwith

70nmNIPAM/BAM‐nanoparticles[151].Notably,anAmsil8particlebindsroughly800timesless

TaspasemoleculesthananAmsil125particle,althoughthesurfaceareaisonlyroughly250times

smallerforAmsil8.ThisisprobablythecasebecauseTaspaseisabletobindwithhigherdensity

toasurfacewithlowcurvature(Amsil125)comparedtohighcurvaturesurfaces(Amsil8).

4.2.2 Possibleinhibitorymechanisms

Here,forthefirsttimesilicananoparticleswerefoundnotonlytobind,butalsotoinhibitTaspase

(figure3.28).Notably,bindingandinhibitionoccuratsimilarnanoparticleconcentrations(figure

3.27and figure3.28), indicating thateveryTaspasemoleculebound tonanoparticles loses its

activity.However,theinhibitionofenzymesbynanoparticlesispoorlyinvestigatedandscarcely

described.Oneexampleistheinhibitionofchymotrypsinmulti‐walledcarbonnanotubes,which

were specifically imprinted to enable chymotrypsin binding [152], whichwas also applied to


113

trypsinbyanindependentgroup[153].Anotherstudyinvestigatedtheinfluenceofsingle‐walled

carbon nanotubes on soybean peroxidase and chymotrypsin activity [109]. They found an

irreversible inhibition of chymotrypsin activity and more than 90 % inhibition of soybean

peroxidase activity. For the inhibition mechanism of chymotrypsin, the authors suggested a

denaturationof theprotein,while the inhibitorymechanismforsoybeanperoxidaseremained

unresolved. Furtherpossiblemodels for enzyme inhibitionbynanoparticles aredisplayed for

Taspaseinfigure4.5[92].

Figure4.6Structuralchangesofproteinsboundonnanoparticles.

Interactionwithnanoparticlescanchangetheproteinstructureandfunction.ThegreencirclesrepresentTaspasesubunits.ModelsforTaspaseinhibitionbynanoparticlesincludestericocclusionoftheactivesiteafter binding as hetero‐tetramer or dimer, separation of the subunits, as well as structural orconformationalchanges,crowdingeffects,fibrillationandoxidation.Modifiedaccordingto[92].

Structural changes have been described for the human β2‐microglubulin upon binding to

copolymerparticles,ceriumoxideparticles,quantumdotsandnanotubes[154].Inthecaseofβ2‐

microglubulin,nanoparticlesincreasetheprobabilityofnucleationofproteinfibrilsandthereby

induceamyloidformation.Thisfibrillationprocessisaccompaniedbyatransitionfromahelixto

asheetandthussignificantchangesintheproteinstructure.However,thesecondarystructureof

Taspaseisnotaltereduponbindingtonanoparticles(figure3.29a),whichrulesoutfibrillationas

explanationforinhibition.Inaddition,thereversibilityoftheinhibitiondoesnotagreewithan

irreversiblefibrillationandaggregationprocess.

Moreover,thestabilityofenzymesboundtonanoparticlescandifferfromtheenzymeinsolution

[109,155].Interestingly,thiseffectcaneitherbestabilizingordestabilizing,dependingonboth

thenanoparticle and theprotein investigated. For instance, soybeanperoxidaseand subtilisin

Carlsberg were stabilized by spherical nanoparticles such as C60 fullerenes, silica or gold

nanoparticles [155], while chymotrypsin unfolds upon binding to carbon nanotubes [109].

Additionally,thestoragestabilityofglucoseoxidasewassignificantlyincreasedafterbindingto

silica‐encapsulated magnetic nanoparticles [156]. Nevertheless, both stabilization and


114

destabilizationcanbeexcludedasthecauseforTaspaseinhibition,sincethestabilityisnotaltered

bynanoparticles(figure3.29).

Anothermechanismistheoxidationofbiomoleculescatalyzedbynanoparticles,whichhasbeen

demonstratedforDNAinthepresenceofTiO2nanoparticles[157],butcanalsoaffectproteinsand

otherbiomolecules[93,158].Indeed,Taspasepossessesatleastsixcysteinsandtwomethionines

withsufficientsolventaccessibilityforoxidationandespeciallymethionineoxidationwasfound

tobe important for inhibition forexample in thecaseofcatalase [159].However,oxidationof

Taspaseisunlikelyduetothehighlyreducingmilieuinthebuffercontaining10mMDTT.

Moreover, binding of proteins to nanoparticles can greatly increase the local protein

concentrationinthenanoparticlecoronaandcauselayeringorcrowdingeffects,whichareknown

topromoteaggregation[160,161].Indeed,thestoichiometryof6and5000Taspasemoleculesper

Amsil8 and Amsil125 nanoparticle (figure 3.27b), respectively, indicate high local Taspase

concentrations. However, the CD spectra in the presence of nanoparticles contradict the

aggregationofTaspase.Moreover,assuminganinterfacesizeof35x42ÅperTaspasemonomer

andperfectstericpacking,thesurfacecoverageofAmsil8isbelow50%androughly150%for

Amsil125.ThissuggeststhataheavilycurvedsurfacecanbindlessTaspasemoleculescompared

to flatter surfaces (figure 4.7). However, a burying of Taspase by multiple layers in the

nanoparticlecoronaisprobablynotthecauseforthereducedactivity.

Figure4.7True‐to‐scalemodelofTaspaseboundtonanoparticles.

a Scaled model of Taspase bound to Amsil8 nanoparticles (red) as αβ‐heterodimer (black and blue,respectively).Theactivesite(greencircle)isoccludedintheboundstate.Thecalculatedstoichiometryof6:1isdisplayed.bSizecomparisonofTaspaseandAmsil125.

Hence,Taspaseislikelytobindtonanoparticlesinasinglelayerwithoutmajorconformational

changes.ForthisitisimportanttoconsidertheoligomericstateofTaspase.IfTaspasebindsas

hetero‐tetramer, the two active sites are located at opposite ends and cannot be covered

simultaneouslybyonenanoparticle.Asinactivationofoneprotomerhasnoeffectontheother


115

protomer[85],oneprotomerislikelytobeactiveifTaspasebindsashetero‐tetramer,unlesssmall

conformationalchangesundetectablebyCDandfluorescencespectroscopyoccur.

Withthecurrentdata,themostplausibleexplanationforTaspaseinhibitionbynanoparticlesis

bindingasanαβ‐dimerinasinglelayerwithoutconformationalchangeasshowninfigure4.7a.

This either results in noncompetitive inhibition (figure 3.28) that is either allostery‐like or

occludestheactivesite.However,thelattermechanismismorelikely,sincethepositivelycharged

active site is probably part of the binding interface for the strongly negatively charged

nanoparticles.Asaresult,theactivesiteisreversiblyoccludedandinaccessibleforthesubstrate.

Giventhefactthattheinactivemutantcanstillbindnanoparticles,thepositivelychargedTaspase

loop(pI=9.8)probablysupportsbinding,whilethelocationdoesprobablynotinfluencebinding

strength.

Finally,bindingofonlyonesubunitandseparationofthehighlyaffineαβ‐dimercanberuledout,

asthiswouldresultinfreeα‐orβ‐subunits,whichagaincanbeexcludedsinceneithersubunitis

foundafterfiltrationthrougha30kDamembrane(datanotshown).However,aseparationofthe

twosubunitsandbindingofbothsubunitstonanoparticlesisanotherplausibleexplanationfor

theinhibitoryeffect.

Importantly, control experimentswith lactatedehydrogenase, chymotrypsin andproteinaseK

showedno or low inhibition of the respective enzyme activity at nanoparticle concentrations

needed for Taspase inhibition (figure 3.30). In summary, the nanoparticle concentrations

requiredforLDHinhibitionare1‐2ordersofmagnitudehigherthanforTaspase,andnoinhibitory

effectwasfoundforchymotrypsinandproteinaseK.Moreover,nanoparticlescannotreduceLDH

activitybymorethan25%,incontrasttothe90%inhibitionseenwithTaspase.Thispinpoints

thatsilicananoparticlesdonotinhibitenzymesingeneral.Whilesilicananoparticlesdisplayedan

inhibitoryeffectonthethreonineproteasesTaspaseandtheproteasome,theyfailedtoinhibitthe

serine proteases chymotrypsin and proteinase K. This is in agreement with the finding that

nanoparticlesbindaselectivesetofproteins[99].However,forapredictionwhetheranenzyme

isinhibitedbynanoparticles,furtherstudiesarerequiredtogainadeeperunderstandingofthese

interactions.

4.2.3 InvivoeffectsofnanoparticlesonTaspase

Incontrasttoinvitroassayswithpurifiedprotein,nanoparticlesfacealargevarietyofdifferent

proteinsinvivo.Hence,thenanoparticlecoronawillconsistofadiversemixtureofproteinsthat

compete for binding. To test Taspase binding under such conditions, Hela cell lysates with

overexpressed GFP‐Taspase were tested for Taspase activity in the presence of silica

nanoparticles.

WhilenanoparticleswereabletoinhibitTaspasebyupto90%invitro,themaximuminhibitory

effect is limited to50% incell lysates (figure3.31).Consequently, thenanoparticle surface is

probablypartlyoccupiedbyothercellularproteins,whichreducetheinhibitorypotentialwith

respect to Taspase. Nevertheless, the nanomolar affinitywas still good enough to investigate

Taspaseinhibitionincells.


116

Owing to their large size, nanoparticles cannot enter the cell passively, butmust be takenup

activelybythecell.However,theefficiencyofthiscellularuptakeagaindependsonsizeandshape

of the nanoparticle [162]. The current model for the uptake mechanism involves energy‐

dependentwrappingofthenanoparticlewiththehelpofclathrinandthecytoskeleton[92].Ina

recentstudy,tissue‐specifictargetingofcompoundswasachievedbycoatingnanoparticleswith

alung‐targetingpeptide,whichresultedinanenrichmentofthesenanoparticlesinthelungafter

venousinjectioninmice[163].Eventargetingofcellularcompartmentswassuccessfullyshown

in living cells with nanoparticles conjugated to a small signal peptide. With this approach,

fluorescentnanocrystalquantumdotsaccumulatedselectivelyinthemitochondriaorthenucleus

within15minutes[164].

Somenanoparticlesposearisktocellsduetotheirtoxicpotential.Forinstance,ZnOnanoparticles

were tested for cytotoxicity against human immune cells and monocytes were found to be

particularly susceptible to the reactive oxygen species induced by the nanoparticles [165].

Especially crystalline silica (SiO2) is known to induce chronic inflammation, fibrosis and lung

cancer upon inhalation [166,167]. In stark contrast to the high toxicity of crystalline silica,

amorphous silica nanoparticles are regarded to be safewithout displaying chronic effects, as

studied in rats [168]. This is in agreement with the good viability of the cells after Amsil

nanoparticlechallengewithnormalproliferation,cellsizeandmorphology(figure3.33).

Using fluorescently labeled nanoparticles, Tenzer and colleagues cold confirm that the silica

nanoparticlesusedinthisprojectarereadilytakenupwithinroughlyonehour[99].Furthermore,

the cellular uptake rate was increased in the presence of serum, compared to serum‐free

incubation.

Taspasedoesnotresideinthecytoplasm,butislocatedmainlyinsidethenucleus[74].Notably,

thisnuclearlocalizationisnotstatic,butTaspaseshuttlesbetweennucleusandcytoplasminan

importin‐α/nucleophosmin‐dependent manner [74]. Importantly, the nuclear localization is

retained in the presence of nanoparticles,which are unlikely to be shuttled actively between

nucleusandcytoplasm(figure3.33a).ThisfindingexcludesthatnanoparticlesbindTaspaseinthe

cytoplasmandpreventthenuclearimport,asthiswouldbereflectedinadistinctaccumulationof

Taspaseatthenuclearenvelopeandinthecytoplasm.Hence,thissuggeststhatnanoparticleshave

accesstothenucleus,whereTaspasebindingandinhibitionoccurs.Thediameterofthecentral

channel of the nuclear pore complex (NPC) is with 40 nm large enough for the passage of

nanoparticles[169,170].Indeed,nanoparticlescanenterthenucleus[164],whichwasalsoshown

forsilicananoparticles[99].Sinceproteinswithamolecularweightabove60kDacannotpassthe

porebydiffusion,nanoparticlesareprobablyactivelyimported.

Whiletheobservedinhibitionto50%oftheinitialactivityisconsistentwiththevaluesobserved

incelllysatesforAmsil8andAmsil20,theinhibitoryeffectofthelargerAmsil125wassignificantly

lower incells (figure3.33a).Thisoutcome is inagreementwith the finding that thewrapping

processneededforcellularuptakeisoptimalfornanoparticleswith15‐30nmdiameter[92,162].

Hence,thelargeAmsil125nanoparticlesaretakenupatalowerrateandexcludedfromnuclear

importduetotheirsize,whichresultsinalowerinhibitorypotential.

DISCUSSION SUCCINIMIDEPEPTIDESINHIBITTASPASECOMPETITIVELY

117

Altogether, the inhibition of Taspase by silica nanoparticles at nanomolar concentrations

highlightsforthefirsttimethatnanoparticlesareabletoqualitativelyinhibitenzymesinvivo,and

thereby contributes to the comprehension of nanoparticle interactions in cells. Although the

natureofsilicananoparticlesdoesnotsuggestaselective inhibition, thiseffectdoesnotaffect

proteasesingeneral.However,possibleinhibitoryeffectsofnanoparticlesonenzymesneedtobe

consideredinfuturestudiesandforpossiblemedicalapplications.

4.3 SuccinimidepeptidesinhibitTaspasecompetitivelyThespecificconsensussequenceofTaspase(Q‐[F,I,L,V]‐D|G‐X‐X‐DorQ‐[F,I,L,V]‐D|G‐X‐D‐X)

facilitatestargetingtheactivesitebysubstrate‐derivedinhibitorsandhasalreadybeenattempted

bytheBogyogroupin2009[69].However,thesevinylketoneandvinylsulfonederivativesonly

targeted the S‐site, but not the S’‐site anddisplayed onlymoderate inhibition (IC50 = 30µM).

Consequently, a new class of substrate analog inhibitors targeting both the S‐ and S’‐sitewas

designedandtestedincooperationwiththeKaisergroupforchemicalbiology.

4.3.1 Implicationsforsubstratecleavageandbindingmode

Forthedesignofnovelinhibitors,thecleavagemechanismproposedforTaspasewasanalyzed

(figure 1.9; [69]). Concerning the high structural similarity of Taspase with other type II

asparaginases[66,139],thecatalyticmechanismislikelytoshowsomesimilarities.Nevertheless,

asparaginaseshydrolyzetheasparaginesidechain,whileTaspaseisanendopeptidase.However,

if the substrate undergoes a rearrangement via a succinimide‐hydrate intermediate, Taspase

couldcleavethe isopeptidebondat theasparaginesidechain(figure1.9),which isagainvery

similar to the asparaginase cleavage mechanism [66]. The fact that the compound with

isoaspartate(Comp8)displayedonlyalowinhibitorypotential,indicatesthatTaspasecanindeed

cleaveisopeptidebonds.Incontrast,compoundswithasuccinimidemoiety,(Comp2,5,9,10)could

effectivelyinhibitTaspase.

Forthefirsttime,conclusionscanbedrawnaboutthecleavagemechanismofTaspasebasedon

these experimental data. Indeed, these data strongly support the previously hypothesized

substrate rearrangement via a succinimide‐hydrate to form an isopeptide bond between

aspartateandglycine.Inotherwords,theproteinbackboneistransferredtotheaspartateside

chainbeforethefinalcleavagesteptakesplace.Hence,Taspasedoesnotdirectlyhydrolyzethe

peptidebond,butthenewlyformedisopeptidebond(figure1.9).Furthermore,theformationof

asuccinimide‐hydrateintermediateisinlinewiththemandatorypresenceofthesmallresidue

glycineaftertheaspartatetoallowthenucleophilicattackontheaspartatesidechain.

Moreover,aphotoreactivederivativeofthesuccinimideinhibitorwasusedtoshedlightonthe

bindingmodeofthesubstrate,whichisstillcompletelyenigmaticduetomissingexperimental

dataandalackofhomologousproteases.Forthiscriticalexperiment,thephotoreactivecross‐

linkerwaslocatedattheP4positionofthesubstrate.Aftercross‐linking,theinhibitorwasfound

attachedtotheα‐subunit,butnottheβ‐subunit(figure3.38).Asbothsubunitscontributetothe


118

activesitelocatedbetweenthem,thecross‐linkingexperimentallowsamappingoftheputative

substrateorientationontheTaspasestructure(figure4.8).

Figure4.8Proposedsubstrateorientation.

Thesuggestedbindingmodederivedbythephotolabelingexperiment(seemaintext)wasmodeledonthedimerofTaspase(PDB2a8j).aThesubstrate(yellow)isproposedtobindtotheactivesite(red)withitsP‐site(green)orientedtotheα‐subunit(gray)andtheP’‐site(green)totheβ‐subunit(blue).N‐andC‐terminiofthesubstratearelabeled,aswellasthesubstratecleavagesiteandtheTaspasesubunits.bThesideviewshowsthecatalyticcleftformedbybothsubunits,inwhichthesubstratecanbind.

Thus,theP‐siteofthesubstrateislikelytobeorientedtowardstheα‐subunit,whichinturnmeans

thattheP’‐siteisprobablypointingtowardstheβ‐subunit.Assumingthatthisbindingmodeis

correct,thegrooveformedbybothα‐andβ‐subunitcouldserveasbindingcleftforthesubstrate

as shown in figure 4.8b. Although these experiments provide valuable first data about the

substratebinding,detailedinformationaboutthesubstratebindingmodecannotbederivedfrom

thesepreliminaryexperiments.Nevertheless,thedevelopedcompoundspresentapowerfultool

forfurtherstudies.Possibleexperimentsarediscussedintheoutlooksection.

4.3.2 Scopeandlimitsofthesuccinimideinhibitors

Invitroapplications

The inhibition type of the newly designed peptidyl‐succinimide inhibitors is expected to be

competitive,asthisisusuallyseenforsubstrateanalogcompounds.Indeed,thesubstratecould

replacetheinhibitor(figure3.36b).Furthermore,bindingoftheinhibitortoTaspaseinabsence

of the substrate was verified by microscale thermophoresis (figure 3.36a) and excludes

uncompetitive inhibition (i.e. bindingof the enzyme‐substrate complex).Moreover, prolonged

incubation times of enzyme and inhibitor did not result in increased inhibition, hinting at a

reversible nature of the inhibition. However, unlike allosteric inhibitors, substrate analog

protease inhibitorsareprone tocleavagedue to theirhighstructural similarity to thenatural

substrate.InthecaseofTaspase,adegradationofthesuccinimidemotifcouldberuledoutand

the inhibitory potential remained constant over time (figure 3.37). Moreover, an artificial

influence of the N‐terminal Fmoc protection group, which was used for photometric


119

determinationoftheconcentration,wasexcluded,sinceFmoc‐valinedidnotinfluenceTaspase

cleavage.

SinceTaspaseinhibitorsaddressingonlytheS‐siteofTaspasedisplayedonlyweakinhibition[69],

theroleof theS’‐sitewasexplored.Tothisend, theP’‐positionwaschangedfromglycinetoa

structurallyrelatedisoaspartate.Thissmallstructuralchange indeedcausedatwofoldweaker

inhibition (Ki=6.8±1.1µM),whichpinpoints the importanceof theS’‐site for tight inhibitor

binding.

Moreover,derivatizationoftheinhibitorysuccinimidemotifshowedthatanextensionofthe5‐

membered succinimidemotif (Ki = 4.3 ±1.3µM) to a 6‐membered glutarimide ring yielded a

compoundwithsimilarinhibitorypotential(Ki=1.6±0.3µM).However,itshouldbenoted,that

althoughglutarimideisthemostprobableconformationofthecompound,anucleophilereaction

oftheglutamatewiththeprecedingNHbackbonegroupcannotbeexcludedbytheLC‐MSspectra.

Surprisingly, the most potent inhibitor was a succinimide derivative with a photoreactive

benzophenone(Comp9;Ki=0.4±0.09µM).ThiscompoundrepresentsthemostpotentTaspase

inhibitordescribedsofar.Comparedwiththenextbestpublishedinhibitor,thenoncompetitive

bisarsonicacidNSC48300[63],itsKivalueistenfoldbetteranditdoesnotrelyonthetoxicarsenic

acidmoiety.

Despite its promising inhibitory effect in vitro, the inhibitory effect in cell lysates with

overexpressed GFP‐Taspase was 30‐fold lower (IC50 = 111 µM). Regarding this, the chloride

concentrationof150mMNaClinthelysisbufferistobeconsidered,whichwasnotpresentinthe

in vitro assay. Chloride is known to bind to the active site and therefore inhibit Taspase

competitively[56].Forthisreason,thepresenceofachlorideionintheactivesiteisalsoexpected

tocompetewithsubstrateanaloginhibitors.Thishypothesisissupportedbytheloweraffinityof

Comp9inthepresenceof450mMNaCl(22.1µM),comparedtothesaltfreebindingaffinity(0.4

±0.09µM).Consequently,sophisticatedequationsareneededtodescribethissystemwiththree

ligands(substrate,inhibitor,chloride)bindingcompetitivelytoonebindingsite.However,ifthe

substrateconcentration iskeptconstant, theCheng‐Prusoffequation [124] canbeapplied ina

modifiedformtoestimatethesaltinfluenceoninhibitorbindingusingtheknownchlorideaffinity

(68±7mM;supplementalfigure7.12).

_

Thisshowsthat30%ofthe30‐foldlowerinhibitoryeffectinlysatescanbeattributedtohigher

saltconcentrations.However,toexplaintheremaining70%,correspondingtoa10‐foldreduction

oftheinhibitoryeffect,otherfactorsneedtobetakenintoaccount.Inparticular,thesixpeptide

bondsoftheinhibitormightbecleavedbyotherproteases,sincetheproteaseinhibitorspresent

inthelysisbuffercannotguaranteecompleteinhibition.Moreover,unspecificbindingtoother

proteinscanlowertheefficiencyoftheinhibitor.Nevertheless,thesestudiesconfirmedthatthe

succinimideinhibitorsinhibiteukaryoticTaspaseincelllysates,althoughtheinhibitorypotential

isratherweak.


120

Anotherapplicationofthephotoreactiveinhibitorderivative(Comp9)isactivitybasedprotein

profiling (ABPP). For this, the compound was added to Hela cell lysate spiked with 5 µM

recombinantTaspase.Subsequentcross‐linkingoftheinhibitorandattachmentofafluorescent

dyealloweddetectionofallproteinsinteractingwiththeinhibitorbyfluorescenceimaging(figure

3.38).TheSDS‐PAGEgelrevealsonlytwobandscorrespondingtothesizesoffull‐lengthTaspase

andtheα‐subunit.Here,thebandforthefull‐lengthenzymecaneitherrepresentaninteraction

with the proenzyme, or more likely an interaction with the SDS‐resistant αβ‐dimer (see

section 4.1.3). Furthermore, the absence of other bands in the lysate could indicate that the

inhibitorbindsselectivelytoTaspase.However, thecomparativelyhighconcentrationof5µM

recombinantTaspaseaddedtothelysateprecludesanunequivocalexclusionofbindingtoless

abundantoff‐targetproteinsbelowthedetectionbackground.

Invivoapplications

Forcellularapplications,aninhibitorneedstocrossthehydrophobiccellmembrane.Forthis,a

molecule must generally possess lipophilic groups. For small molecules, the octanol/water

partitioncoefficient(logP)isoftenusedtopredictthelikelinessofmembranepassage.Forthe

developedsuccinimideinhibitors,a logPvalueof0.21±0.91wascalculated,whichindicatesa

goodpassageofthecellmembrane[171].However,otherproperties,suchassize,thenumberof

hydrogenbonddonorsandacceptors,andmolarrefractivity,arealsoimportantfactorsforcell

penetration. Table 4.2 shows that the succinimide compounds fail to fulfil these criteria, also

knownasrule‐of‐five[172],mainlybecauseoftheirhighmolecularweight.

Table4.2Parametersfordruglikelinessofthepeptidyl‐succinimideinhibitors.

Parameter Qualifyingrangea Comp9logP ‐0.4to5.6a 2.83 ±0.97c

Mw 160to480Daa 1296 DaMolarrefractivity 40to130a 333c

Numberofatoms 20to70a 168H‐bonddonors ≤ 5b 9H‐bondacceptors ≤10b 15

aaccordingto[171];baccordingto[173];cpredictedwithChemSketch

Nevertheless,eukaryoticcellswereassayedfortheabilitytotakeupafluorescentderivativeof

the succinimide compounds. Unexpectedly, the compound was indeed able to cross the cell

membraneandwasfoundevenlydistributedinthecellsafter5hours(figure3.39c).Importantly,

the uptakewasmore than twofold higher if the cells were incubated in serum‐freemedium.

Hence,a fractionoftheinhibitor isobviouslyeitherdegradedorboundtoserumproteins like

albumin.However,theassayisonlyqualitative,sothattheintracellularinhibitorconcentration

couldnotbedetermined.

UsingacellularTaspaseactivityassay,theinhibitorypotentialofthesuccinimidecompoundswas

testedinvivo.However,theseinhibitorsfailedtoinhibitTaspaseincellsafter18hourschallenge

with50µMComp9(figure3.39e).Themostlikelyexplanationfortheinhibitorstobeinactivein


121

vivodespitetheirpromisinginvitroactivityisthattheirsizepreventssufficientinternalization.

Consequently,theintracellularconcentrationsareprobablytoolowtocauseaninhibitoryeffect.

Furthermore,thecellularchlorideconcentrationsofroughly50mM[174–176]furtherreducethe

affinity of the inhibitor, as explicated above. Moreover, degradation or modification of the

inhibitor by cellular proteases and transferases, respectively, as well as binding to off‐target

proteinscanlimititseffectiveness.

Altogether,thepeptidyl‐succinimide inhibitorsdisplayanovelclassofTaspaseinhibitorswith

promisinginvitroinhibitionpotential.Inparticular,thenewlydiscoveredsuccinimidescaffoldis

expectedtobeusefulforrationaldrugdesigninsubsequentstudies.Suggestionstoimprovethe

currentlylimitedinvivobioactivityofthecompoundsarediscussedintheoutlooksectionbelow.

OUTLOOK SUCCINIMIDEPEPTIDESINHIBITTASPASECOMPETITIVELY

122

5 OutlookInthisproject,threedifferentstrategieswereappliedtofindnovelTaspaseinhibitors.Whilethe

bioinformatics approach did not yield active compounds, silica nanoparticles were found to

inhibit Taspase noncompetitively. However, the most promising compounds were substrate

analogpeptides containing a succinimidemoiety, because it is not expected that selectivity is

achievable for uncoated nanoparticles by rational design with the current knowledge. These

compoundsexhibitedgoodinhibitorypotentialinvitro,whiletheinvivopotentialwasratherlow.

Hence,thesecompoundsmightbeimprovedtoachievehigherstabilityandbetterbioavailability.

Primarily, better bioavailability requires a reduction of the molecular weight and increased

hydrophobicity[172,177].Tothisend,themolecularweightcanbereducedbyroughly15%by

removing theN‐terminal Fmoc group,whichwas shown to be irrelevant for binding. Cellular

experimentstoelucidatetheTaspaseconsensussequencerevealedthatthepresenceofonlyone

ofthetwoaspartatesatP3’andP4’issufficientforcleavage[70].Consequently,thepeptidepart

of the inhibitor couldbe shortenedby theC‐terminal aspartateatP4’, if theP3’ asparagine is

replacedby aspartate. Furthermore, a substitution of the aspartic acid by benzoic acidwould

contributetohigherhydrophobicityandthereforeincreasecellularuptake.However,increased

hydrophobicityalsoincreasestheriskofunspecificbindingtootherproteinsandshouldtherefore

betestedthoroughly.

Moreover,stabilityandresistanceagainstdegradationbycellularproteasescanbetestedincell

lysateswiththefluorescentderivative.If fragmentationisobservedafterHPLCanalysis,either

thesidechainsorthepeptidebackbonecanbealtered.Withregardtothelatter,asubstitutionof

the peptide bond by N‐methyl, ether, ketomethylene or retro‐inverso groups is a common

approach(figure5.1;[178]).

Figure5.1Substitutionsforamidebonds.

Possible substitutions for amide bonds in peptidomimetics includeN‐methyl, ketomethylene and ethergroups.Thesemodificationsyieldhigherstabilityandresistanceagainstcellularproteases.Accordingto[178].

Furthermore,sidechainmodificationscanbothstabilizethecompoundandimprovetheaffinity.

EspeciallymodificationsoftheCβatomareknowntoreducethesusceptibilitytochymotryptic

cleavage[178],butalsoothersidechainmodificationsareaproventooltoenhancethepotential

of peptidic ligands. For example, methylation, methoxy or benzyl ether modifications greatly

enhanced the affinity of analogs for the neuropeptide substance P [179]. Altogether, critical

evaluationofthesemodificationsmighthelptoenhancethe invivopotencyofthesuccinimide

inhibitors.

OUTLOOK SUCCINIMIDEPEPTIDESINHIBITTASPASECOMPETITIVELY

123

Finally, co‐crystallizationofTaspaseand the inhibitor could reveal theactive conformationof

Taspase.Thisfacilitatesbioinformaticsapproachesandcouldultimatelyyieldpotentinhibitors

byvirtualdocking.

ABSTRACT SUCCINIMIDEPEPTIDESINHIBITTASPASECOMPETITIVELY

124

6 AbstractThethreonineproteaseTaspase1isofhighinterestasnewtargetforcancertherapeutics,dueto

its implication in leukemiasandsolid tumors.AlthoughTaspase isnotanoncogene itself, it is

crucially required for the proteolytic activation of the MLL oncogene as well as MLL fusion

proteinsgeneratedbychromosomaltranslocations.Hence,MLL‐dependentcellscansuccessfully

be targeted by inhibition of Taspase activity.However, Taspase is insensitive to inhibition by

general serine, cysteine or metalloprotease inhibitors. Furthermore, the limited number of

availableTaspase inhibitorsdisplay either low inhibitorypotential or broadbioactivity. Thus,

novelTaspaseinhibitorsarestillurgentlyrequired.

Inafirststep,detailedcharacterizationrevealedforthefirsttimethatrecombinantTaspaseand

protein overexpressed in eukaryotic cells are comparable with respect to their catalytic

parameters.

Moreover,thespecificcatalyticactivity(0.03U/mg)wasdetermined,aswellasa20‐foldhigher

efficiency regarding cleavage of the CS2 cleavage site of MLL over CS1. Regarding stability,

Taspase rapidly loses its catalytic activity (half‐life < 3 h), unless it is stabilized by high

concentrationsofNaCl(≥450mM).Additionally,theautocatalyticprocessingeventrequiredfor

catalytic competencewas found tooccurona time scaleof2‐3days.The flexible loop region

betweenglycine178andaspartate233generatedbythisactivationeventisnotresolvedinthe

crystalstructure.Hence,NMRandCDspectroscopywereappliedandtheresultssuggestahelix‐

turn‐helixstructure.Furthermore,analyticalgelfiltrationpinpointedthatTaspaseisactiveasa

hetero‐tetramerinsolution.Withthesedata,afluorogenicinvitroassaywasoptimizedforinitial

inhibitortests.

WhilethecontroversiallydiscussedbisarsonicinhibitorNSC48300andcompoundsderivedfrom

virtualdockingexhibitednoinhibitoryeffect,bothsilicananoparticlesandsuccinimidepeptides

were discovered as novel Taspase inhibitors. In fact, silica nanoparticles are known to bind

proteins, albeit nanoparticle‐mediated inhibition of enzyme activity is scarcely described.

Surprisingly, IC50 values in the nanomolar and picomolar range indicate high affinity binding,

whichwasobservedinvitroandincells.Inaddition,comparisonofnanoparticleswithdifferent

diameters (8‐125nm)suggestsapositivecorrelationofsizeandaffinity.Subsequentanalyses

characterizedtheinhibitiontypeasnoncompetitiveandpinpointedthatTaspasedoesnotchange

itsstructureuponbinding.Importantly,inhibitionbynanoparticlesisnogeneralfeature,asthree

otherenzymes(lactatedehydrogenase,chymotrypsin,andproteinaseK)showednooronlyweak

inhibitionbysilicananoparticles.

Another approach for Taspase inhibition exploited the cleavage sequence and the proposed

mechanismtodesignsubstrateanalogpeptideswithasuccinimidemoietyatthecleavagesite.

ThemostpromisingcompoundexhibitedaKivalueof400±88nManddisplaysthusatenfold

betterinhibitorypotentialthanthebestreportedTaspaseinhibitor.Moreover,thecompetitive

inhibitiontypeandtheresistanceagainstTaspasecleavagewereverified.However,degradation

by other proteases and/or limited cellular uptake as well as competition with chloride ions

preventstheinvivoinhibitionofTaspase.


125

Altogether,thesefindingsdemonstratethepotentialofsilicananoparticlesaswellassuccinimide

peptidesasTaspaseinhibitorsandtoolsforsubsequentstudies.

ZusammenfassungDie Threonin‐Protease Taspase1 stellt aufgrund ihrer Beteiligung an Leukämien und soliden

TumoreneinenwichtigenneuenAnsatzpunktfürKrebsmedikamentedar.ObwohlTaspaseselbst

keinOnkogenist,istsieunabdingbarfürdieproteolytischeAktivierungdesOnkogensMLLundder durch Translokationen entstandenen MLL‐Fusionsproteine. Deshalb reagieren MLL‐

abhängigeZellenbesonderssensitivaufTaspase‐Inhibition.Allerdingssindsämtlicheallgemei‐

nenSerin‐, Cystein‐ undMetalloprotease‐InhibitorenunwirksamgegenTaspase, und auchdiewenigen verfügbaren Taspase‐Inhibitoren besitzen entweder nur geringes Inhibitorpotential

oder einbreitesWirkspektrum.Daher sindneueTaspase‐Inhibitorennachwievorvonhoher

Relevanz.Ein Vergleich rekombinanter Taspasemit in eukaryotischen Zellen überexprimiertemProtein

zeigteeinevergleichbarespezifischekatalytischeAktivität(0.03U/mg).Weiterhinkonnteeine

20‐fachhöhereEffizienzhinsichtlichderSpaltungderCS2‐SchnittstellevonMLLgegenüberderCS1‐Schnittstellefestgestelltwerden.BeiBetrachtungderProteinstabilitätfielauf,dassTaspase

sehr schnell an Aktivität verliert (Halbwertszeit < 3 h), sofern das Protein nicht durch hohe

Salzkonzentrationen(≥450mMNaCl)stabilisiertwird.Zudemwurde fürdieautokatalytischeProzessierungdesProenzyms,welchewichtigzurErlangungkatalytischerKompetenz ist,eine

Zeitspannevon2‐3Tagenermittelt.DerdabeientstehendeflexibleBereichzwischenGlycin178

und Aspartat 233 ist in der Kristallstruktur ungeordnet, konnte aber dank NMR‐ und CD‐Spektroskopie als potentielles Helix‐Turn‐Helix‐Motiv identifiziert werden. Des Weiteren

belegtenanalytischeGelfiltrationendieAktivitätvonTaspasealsHetero‐Tetramer.Mithilfedieser

InformationenwurdeeininvitroAssayoptimiertundfürInhibitortestsverwendet.Während der kontrovers diskutierte Inhibitor NSC48300 und die mittels virtuellen Dockings

ermittelten Substanzen keine Inhibition zeigten, konnten Silikatnanopartikel und Succinimid‐

Peptide als neue Taspase‐Inhibitoren identifiziert werden. Obwohl Proteinbindung anNanopartikel bekannt ist, ist einedurchdiese vermittelteEnzyminhibitionkaumbeschrieben.

UmsoüberraschendersinddieIC50‐Werteimnano‐undpicomolarenBereich,diesowohlinvitro

alsauchinZellenbeobachtetwurden.WeiterhinkonntemitverschiedenenNanopartikelgrößeneine positive Korrelation von Durchmesser und Affinität belegt werden. Weitere Analysen

ergaben eine nichtkompetitive Inhibition und keine Strukturänderung bei Bindung.

Bemerkenswerterweise stellt Enzyminhibition keine generelle Eigenschaft von Nanopartikelndar,dadreiKontrollenzyme(Lactatdehydrogenase,ChymotrypsinundProteinaseK)kaumdurch

Nanopartikelinhibiertwurden.

IneinemanderenAnsatzwurdedieTaspase‐SchnittsequenzunddervermuteteMechanismusfürdasDesignpeptidischerSubstratanalogamitSuccinimid‐MotivanderSchnittstellegenutzt.Die

beste Verbindung zeigte einenKi‐Wert von 400 ± 88nMund ist damit ein zehnfachbesserer

Hemmstoff als der beste zuvor beschriebene Taspase‐Inhibitor. Des Weiteren konnte diekompetitive Hemmung und die Stabilität gegenüber Abbau durch Taspase gezeigt werden.


126

AllerdingsverhindernschlechteZellaufnahmeund/oderAbbaudurchandereProteasensowieKompetitionmitChlorid‐IoneneineInhibitioninvivo.

Insgesamt belegen diese Ergebnisse das Potential von Silikatnanopartikel sowie Succinimid‐

PeptidenalsTaspase‐InhibitorenundWerkzeugefürnachfolgendeStudien.

SUPPLEMENT RECORDINGANDPROCESSINGOFNMRSPECTRA

127

7 Supplement

7.1 RecordingandprocessingofNMRspectraTable7.1PulsesusedtorecordNMRspectra.

Pulse Powerlevel[W]

Length Description

p1 7.8239 20.86µs 1H90°highpowerpulsep3 40.617 18µs 13C90°highpowerpulsep21 89.492 35.5µs 15N90°highpowerpulsepcpd1 0.17596 50µs 1HDIPSI‐2decouplingpcpd2 3.6535 60µs 13Cdecoupling

pcpd3 4.352 180µs 15NGARPdecoupling

sp1 0.00039212 1H90°shapedpulseforH2Oonresonance

sp2 53.544 13C90°shapedpulseforCali onresonance

sp3 47.381 13C180°shapedpulseforCali onresonance

sp5 47.381 13C180°shapedpulseforC=Ooffresonance

sp8 53.544 1H90°shapedpulsefortimereversedpulse

d1 2s relaxationdelayd9 90µs TOCSYmixingtime

Table7.2RecordingparametersforNMRspectra.

Spectrum F1 F2 F3pulseprogram

nucSWH[ppm]

O1[ppm] TD

AQ‐mode* nuc

SWH[ppm]

O1[ppm] TD

AQ‐mode* nuc

SWH[ppm]

O1[ppm] TD

AQ‐mode*

1H‐15NHSQChsqcetf3gpsi2

15N 40 115 256 DQD 1H 16.08 4.695 1024 E‐AE ‐ ‐ ‐ ‐ ‐

sofast‐HMQCsfhmqcf3gpph

15N 34 115 200 DQD 1H 15 4.695 1024 STPPI ‐ ‐ ‐ ‐ ‐

15N‐TOCSY‐HSQCdipsihsqc3gpsi3d

1H 15.94 4.695 128 STPPI 15N 39 119 48 STPPI 1H 16 4.696 2048 DQD

HNCACBhncacbgpwg3d

13C 80 39 128 STPPI 15N 34 119 40 STPPI 1H 15 4.696 2048 DQD

CBCACONHcbcaconhgpwg3d

13C 75 39 128 STPPI 15N 34 119 36 STPPI 1H 15 4.696 2048 DQD

HBHANHhbhanhgpwg3d

1H 15 4.695 128 STPPI 15N 34 119 40 STPPI 1H 8 4.695 2048 DQD

HBHACONHhbhaconhgpwg3d

1H 15 4.695 128 STPPI 15N 34 119 40 STPPI 1H 8 4.695 2048 DQD

HCCH‐TOCSYhcchdigp3d

1H 8 4.695 128 STPPI 13C 75 39 64 STPPI 1H 16 4.695 2048 DQD

HNCAhncagpwg3d

13C 32 54 64 STPPI 15N 30 117 40 STPPI 1H 8.4 4.695 2048 DQD

*E‐AE:Echo‐Antiecho;STPPI:States‐TPPI

SUPPLEMENT RECORDINGANDPROCESSINGOFNMRSPECTRA

128

Table7.3ProcessingparametersforNMRspectra.

F1 F2 F3Spectrum

nucSI[ppm] MC2* REV WDW nuc

SI[ppm] MC2* REV WDW nuc

SI[ppm] MC2* REV WDW

15N‐HSQC 15N 2048 E‐AE False Qsine 1H 4096 ‐ False Qsine ‐ ‐ ‐ ‐ ‐sofast‐HMQC 15N 1024 STPPI False Qsine 1H 2048 ‐ False Qsine ‐ ‐ ‐ ‐ ‐15N‐TOCSY‐HSQC 1H 256 STPPI False Qsine 15N 128 E‐AE False Qsine 1H 1024 ‐ False QsineHNCACB 13C 256 STPPI False Qsine 15N 128 STPPI True Qsine 1H 2048 ‐ False QsineHBHACONH 13C 256 STPPI False Qsine 15N 64 STPPI True Qsine 1H 2048 ‐ False QsineHBHANH 1H 256 STPPI False Qsine 15N 128 STPPI True Qsine 1H 2048 ‐ False QsineHBHACONH 1H 128 STPPI False Qsine 15N 64 STPPI True Qsine 1H 2048 ‐ False QsineHCCH‐TOCSY 1H 512 STPPI False Qsine 13C 256 STPPI False Qsine 1H 2048 ‐ False QsineHNCA 13C 128 STPPI False Qsine 15N 64 STPPI True Qsine 1H 1024 ‐ False Qsine*E‐AE:Echo‐Antiecho;STPPI:States‐TPPI

SUPPLEMENT BACTERIALVECTORMAPS

129

7.2 Bacterialvectormaps

Figure7.1Mapsofbacterialexpressionvectors.

aMapofthemodifiedpET‐22bvectorcontainingwildtypeTaspasewithC‐terminalHis‐tag.Primerbindingsitesfortheloopcloningareindicated.bMapofthemodifiedpET‐41bvectorcontainingactiveTaspasewithN‐terminalHis‐tag. The ribosomal binding site (RBS) between the two subunits has the sequence“AGGAGG”.cMapofthemodifiedpET‐22bvectorcontaininginactiveTaspasemutantwithC‐terminalHis‐tag.D233andT234attheactivesiteweremutatedtoA233andA234,respectively.dMapofthemodifiedpET‐41bvectorcontainingtheTaspaseloop(G178‐D233)withN‐terminalGST‐tag.AllvectorsareIPTG‐inducibleandresistancegenesaswellasrestrictionsitesusedforcloningareindicated.

SUPPLEMENT EUKARYOTICVECTORMAPS

130

7.3 Eukaryoticvectormaps

Figure7.2Mapsofeukaryoticvectorsusedfortransfection.

aMap of the pC3 vector containingwild type Taspasewith C‐terminal eGFP.bMap of the pC3 vectorcontaininginactiveTaspase(T234V)withC‐terminalBFP.cMapofthepC3vectorcontainingthereportersubstrate (BioTasp)used for thecellularTaspaseactivityassay.Thecodedproteincomprisesanuclearlocalization signal (NLS), eGFP, glutathione S‐transferase (GST), CS1 Taspase cleavage site (CS1), CS2Taspasecleavagesite(CS2)andareversenuclearexportsignal(NES).Resistancegenesandrestrictionsitesareindicated.

SUPPLEMENT TASPASELOOPCLONINGSCHEME

131

7.4 Taspaseloopcloningscheme

Figure7.3SchematicforcloningoftheTaspaseloop.

TheTaspase loopwas amplified from the plasmid forwild typeTaspase (top left panel) by PCR usingprimerswithbindingsitesindicatedinpurple.Thetargetvector(toprightpanel)wasdigestedwithApaIandXhoI.Respectiverestrictionsitesare indicated.Vectorand insertwere ligated,yieldingtheTaspaseloopwithN‐terminalGSTtag(lowerpanel).

SUPPLEMENT PURIFICATIONOFTASPASE

132

7.5 PurificationofTaspase

Figure7.4PurificationoftheinactiveTaspasemutant.

Purification of inactive Taspase was performed as described for wild type Taspase (section 3.1.3).aChromatogramoftheNiNTAcolumnelution.Thefractionshighlightedinbluewereusedforgelfiltration,while the fractions highlighted in green were purified again by NiNTA affinity chromatography. bChromatogram of the NiNTA affinity chromatography of the peak highlighted green in (a). Fractionshighlighted inbluewereused forgel filtration.cTheSDS‐PAGEgelshows inactiveTaspasewithminorimpuritiesafterNiNTAaffinitychromatography.dandeshowgelfiltrationchromatogramsafterNiNTAaffinitychromatographyrunsdisplayedin(a)and(b),respectively.Thefractionshighlightedinbluewereconcentrated,shockfrozenandstoredat‐20°C.fTheSDS‐PAGEgelconfirmspureinactiveTaspaseaftergelfiltration.

Figure7.5PurificationofthefirstelutionpeakafterNiNTAaffinitychromatography.

As the firstelutionpeakafterNiNTAaffinitychromatographycontainedasignificantamountof impureTaspase(section3.1.3),thepeakfractionsweredilutedtoreducetheimidazoleconcentrationto10mMandagainappliedtoaNiNTAcolumn.Allparametersweretakenfromthefirstpurification,describedinsection 3.1.3. a Chromatogram of the NiNTA column elution. The fractions highlighted in blue wereconcentrated and gel filtrated as described in the main text (section 3.1.3). b Chromatogram of thesubsequentgelfiltration.Thefractionshighlightedinbluewereconcentratedto18mg/mlandstoredat‐20°C.

SUPPLEMENT EXPRESSIONTESTOFACTIVETASPASE

133

7.6 ExpressiontestofactiveTaspaseThebicistronicconstructforactiveTaspasecomprisestheaminoacids1to206(His‐taggedα‐

subunit)andaminoacids234to420(β‐subunit;figure2.3).Initialexpressiontestsshowedpoor

solubilityofbothsubunits,aswellasdegradation,whenexpressedlongerthan3hours,plusan

increasednumberofdeadcells.Thishintsatapotentialtoxicityoftheexpressedprotein.Hence,

expressionwas optimized in the C43(DE3)pLysS strain,which is effective in expressing toxic

proteins.

Additionallytothethreetemperatures(20°C,30°C,37°C)andthethreeexpressiontimes(3h,

6h,overnight),thecombinationofhighcelldensity(OD600=1.5)andshortexpressiontimewas

tested.TherespectiveSDSgelsofpelletandsupernatantfractionsafterultracentrifugationare

showninfigure7.6.

Figure7.6SDS‐PAGEgelsofexpressiontestforactiveTaspase.

ExpressionofactiveTaspasewastestedinE.coliBL21(DE3)T1rcellsat20°C,30°Cand37°Cafter3h,6handovernight(o.n.).Additionally,expressionlevelsat37°Cwereanalyzed1h,2hand5hafterinductionofcellswithanOD600of1.5(37°COD600=1.5).Supernatant(S)andpelletfractions(P)werecomparedtoasamplebeforeinductionbyIPTG(‐IPTG)andindicatedoptimumexpressionconditionsat37°Cafter6horovernight.Theexpectedsizesofα‐subunit(α)andβ‐subunit(β)areindicatedbyarrows.

Thehighest expression levelsof solubleproteinare achievedat37 °Cafter6horovernight.

Nevertheless,itshouldbenotedthatthetotalamountofactiveTaspaseissignificantlylowerthan

forwildtypeprotein.

SUPPLEMENT CALIBRATIONFORANALYTICALGELFILTRATION

134

7.7 Calibrationforanalyticalgelfiltration

Figure7.7CalibrationplotforSuperdex20010/300.

Calibrationwasperformedwithgelfiltrationbuffer(section2.1.3)containing100or450mMNaCl(greenandblue,respectively).Ferritin(450kDa),aldolase(161kDa),conalbumin(75kDa),α‐amylase(54kDa)and ribonuclease A (13.7 kDa) were used as references. Elution volumes were plotted against thelogarithmicmolecularweightandlinearregressionwasperformed.Thedashedlinesindicatetheelutionvolume(14.5ml)andcorrespondinglog(Mw)(2.01=102kDa).

7.8 Analyticalgelfiltrationwithlowersaltconditions

Figure7.8AnalyticalgelfiltrationofTaspasewith100mMNaCl.

Analyticalgelfiltrationswithwildtype(wt),inactiveandactiveTaspasewereperformedasdescribedinthemaintext(section3.2.4).WildtypeandinactiveTaspaseshowanapparentmolecularweightofaround100kDa,whiletheshorteractiveTaspaseelutesataround90kDa.

SUPPLEMENT NMRSHIFTSOFTASPASE

135

7.9 NMRshiftsofTaspaseTable7.4NMRshiftsforwildtypeTaspase.

Shiftvaluesarepresentedinppm.aa Res HN N Cα Cβ Hα Hβ Hγ Hδ Hε9 Ser 8.379±0.004 117.112±0.142 58.552±0.096 64.041±0.199 4.499±0.005 3.894±0.029 ‐ ‐ ‐

10 Gly 8.382±0.006 110.706±0.097 45.332±0.078 ‐ 3.95±0.017 ‐ ‐ ‐ ‐

11 Glu 8.223±0.005 120.5±0.049 56.763±0.055 30.194±0.015 4.259±0.007 1.997±0.017 2.197±0.042

12 Gly 8.386±0.003 109.791±0.086 45.18±0.011 ‐ 3.914±0.02 ‐ ‐ ‐ ‐

13 Leu 7.994±0.005 122.667±0.032 53.124±0.028 41.668 4.599±0.011 1.594±0.009 1.533 0.896±0.002 ‐

23 Gly 8.263±0.005 107.765±0.036 45.327±0.005 ‐ 3.912±0.011 ‐ ‐ ‐ ‐

24 Lys 8.02±0.004 120.824±0.055 56.274±0.054 33.037±0.083 4.305 1.72±0.01 1.382 ‐ 3.006

25 Ile 8.15±0.006 122.453±0.08 61.219±0.057 38.712±0.114 4.206±0.002 1.831±0.027 1.258/1.463 ‐ ‐

26 Thr 8.2±0.004 118.662±0.061 61.838±0.079 70.13±0.084 4.32±0.012 ‐ ‐ 1.164 ‐

27 Ala 8.287±0.005 126.192±0.059 52.94±0.091 19.053±0.094 4.238±0.027 1.36±0.019 ‐ ‐ ‐

28 Lys 8.187±0.003 120.311±0.069 ‐ ‐ 4.21±0.015 1.716±0.022 ‐ ‐ ‐

32 Thr 8.06±0.003 115.009±0.064 62.212±0.055 69.791±0.074 4.265±0.008 1.091 ‐ ‐

33 Lys 8.219±0.004 123.407±0.104 56.576±0.081 32.539 4.259 1.741±0.007 ‐ ‐

34 Gln 8.331±0.003 121.675±0.047 56.822±0.08 30.11±0.13 4.214±0.007 1.978±0.019 2.202 ‐ ‐

221 Gln 8.281±0.002 122.044±0.021 56.506±0.128 29.762 4.299±0.028 2.02±0.013 ‐ ‐ ‐

222 Ser 8.213±0.006 116.916±0.11 58.492±0.012 63.854 ‐ 3.817 ‐ ‐ ‐

223 Ser 8.322±0.009 117.672±0.154 58.731±0.131 63.948 ‐ 3.899±0.02 ‐ ‐ ‐

224 Glu 8.37±0.007 122.597 ±0.119 56.784±0.054 30.032±0.025 4.25±0.012 1.79±0.009 ‐ ‐ ‐

225 Lys 8.181±0.003 121.261±0.157 56.585±0.113 32.936 4.262±0.016 1.735±0.011 ‐ ‐ ‐

226 Glu 8.282±0.003 121.292±0.072 57.157±0.007 29.807 4.202±0.004 2.02±0.021 2.223 ‐ ‐

227 Asn 8.368±0.004 119.275±0.056 53.357±0.041 39.106±0.031 4.159 2.772±0.011 ‐ ‐ ‐

228 Asp 8.302±0.008 121.214±0.057 54.354±0.096 41.452±0.072 ‐ 2.675±0.013 ‐ ‐ ‐

229 Ser 8.298±0.002 116.311±0.091 58.872±0.131 63.95±0.134 4.433±0.006 3.88±0.013 ‐ ‐ ‐

230 Gly 8.431±0.004 110.723±0.151 45.496±0.068 ‐ 3.965±0.021 ‐ ‐ ‐ ‐

231 Thr 7.959±0.006 113.633±0.051 61.926±0.078 70.004±0.124 4.323±0.023 4.03±0.005 1.171±0.012 ‐ ‐

232 Leu 8.287±0.003 124.764±0.116 55.2±0.062 42.366±0.151 4.388±0.014 1.616±0.014 1.585 0.84/0.889 ‐

233 Asp 7.837±0.004 126.462±0.032 55.802±0.053 42.484±0.072 4.337 2.529/2.589 ‐ ‐ ‐

251 Ser 8.284±0.004 119.094±0.067 58.483±0.095 63.81±0.002 4.39±0.022 3.84±0.015 ‐ ‐ ‐

428 His 8.014±0.056 125.365±0.085 ‐ ‐ ‐ 3.095±0.026 ‐ ‐ ‐

SUPPLEMENT 1H‐15NHSQCSPECTRAOFTASPASE

136

7.10 1H‐15NHSQCspectraofTaspase

Figure7.9TimedependenceoftheTaspaseNMRspectra.

1H‐15NHSQCspectraofwildtypeTaspasewererecordedoffreshlythawedprotein(black)andafteroneweek(red).Peaksofthefreshlythawedproteinarelabeled;alabelingofallassignedpeakscanbefoundinfigure3.12a.

Figure7.10ComparisonofwildtypeandmutantTaspase.

1H‐15NHSQCspectraofwildtypeTaspase(red)andinactiveTaspase(blue)after7daysincubationat30°C.Allbackbonepeaksoftheinactivemutantcanbefoundinthespectrumofthewildtypeprotein.The10peaksvisibleonly inthewildtypeproteinare labeled.9of thesepeaksare located inthe loopwhich isreleasedbyautocatalyticactivation.Alabelingofallassignedpeakscanbefoundinfigure3.12a.

SUPPLEMENT EXPRESSIONTESTOFTHEGST‐TAGGEDTASPASELOOP

137

7.11 ExpressiontestoftheGST‐taggedTaspaseloop

Figure7.11ExpressiontestfortheTaspaseloop.

ExpressionoftheGST‐taggedTaspase‐loop(GST‐loop)wastestedinLBmediumat25°C,30°Cand37°Cafter 3 h, 6 h and over night (o.n.). After centrifugation, supernatant (S) and pellet (P) fractionswerecomparedtoasamplebeforeinductionbyIPTG(‐IPTG).OptimumexpressionconditionswithmaximumGST‐loopcontentwasfoundforexpressionat25°Covernight.

7.12 NaClinhibitsTaspaseactivity

Figure7.12InhibitionofTaspaseactivitybyNaCl.ActivitymeasurementsofwildtypeTaspaseinthepresenceofdifferentNaClconcentrationsconfirmedaninhibitoryeffectofNaCl.NonlinearfittingyieldedIC50=68±7mM.

7.13 WesternblotanalysisofTaspaselevelsinhumancelllysates50µgcelllysatesand0.5µgrecombinantwildtypeTaspasewereseparatedina12.5%SDS‐PAGE

gel.Westernblottingonnitrocellulosemembrane(Whatman)wasperformedfor45minutesat

40mApergel(80mAfor2gels).Membraneswereblockedwith5%milkpowderinTBSTfor

SUPPLEMENT WESTERNBLOTANALYSISOFTASPASELEVELSINHUMANCELLLYSATES

138

2hoursatroomtemperature,beforetheprimaryantibodieswereaddedin5%BSAovernightat

4°C.TheTaspaseantibody(Origene;goatantimouse)wasdiluted1:500andtheβ‐actinantibody

(Abgent;sheepantirabbit)wasusedina1:250dilutioninTBST+5%milkpowder.Subsequently,

themembranewaswashedthreetimeswithTBSTandtherespectivesecondaryantibody(anti

mouseforTaspase;antirabbitforβ‐actin)wasappliedina1:10000dilutionin5%milkpowder

for2hoursatroomtemperature.ThemembranewaswashedagainthreetimeswithTBSTandan

ECL kit (Thermo Scientific) was applied according to the manufacturer’s instructions for

visualization. Films (Thermo Scientific) were exposed for 2 minutes and developed using a

Cawomat2000.

Figure7.13WesternblotofTaspaseincelllysates.50µgeukaryoticcelllysatesofHela,humanembryonickidney293T(HEK),humancoloncarcinoma116(HCT)and0.5µgofrecombinantTaspaseaspositivecontrolwereseparatedbySDS‐PAGE,blottedonanitrocellulosemembraneandTaspasewasdetectedwitha specificmonoclonalantibody(upperpanel).Taspaseconcentrationswerebelowthedetectionlimitinallthreecelllinestested.RecombinantTaspaseshowstwobands,correspondingtothefull‐lengthenzymeandtheα‐subunit.β‐actin(lowerpanel)wasusedasloadingcontrol.

SUPPLEMENT ESI‐LC‐MSOFNSC48300

139

7.14 ESI‐LC‐MSofNSC48300

Figure7.14ESI‐LC‐MSoftheTaspaseinhibitorNSC48300.aChromatogramoftheLC‐MSanalysisinnegativemode.Thepeakat1.27mincontainstheNSC48300.bFull‐rangenegativemodemassspectrum.Peaksarelabeledwiththemostprobablemoleculesandtheirstructure.TheexpectedmassofNSC48300is416Da.

7.15 Modelqualitystatistics

Figure7.15Modelqualityofthefull‐lengthTaspasehomologymodel.Ahomologymodeloffull‐lengthTaspasewasgeneratedbasedonthecrystalstructurePDB2a8i.75aminoacidswereaddedtothetemplate.Modelquality(andZscore)areasfollows.Dihedrals:optimal(1.317);packing1D:satisfactory(‐1.450);packing3D:good(‐0.718);overall:good(‐0.708).

SUPPLEMENT BINDINGCONTROLSWITHNANOPARTICLES

140

7.16 Bindingcontrolswithnanoparticles

Figure7.16Nanoparticlesbindparvalbuminwithlowaffinity.AnisotropytitrationsofAtto488‐labeledparvalbuminwithAmsil8nanoparticlesyieldanEC50valueof3±0.8µM,whichisa300000‐foldweakerinteractioncomparedtoTaspase.

7.17 Cellmorphologyinthepresenceofnanoparticles

Figure7.17Nanoparticlesdonotaltercellmorphology.CellmorphologyofHelacellsincubatedwith200µg/mlsilicananoparticlesfor18hours(lowerpanel)aresimilartountreatedcells(toppanel;ctrl).GFPfluorescenceisdisplayedforcellshape.

SUPPLEMENT TITRATIONCURVESOFSUCCINIMIDEPEPTIDES

141

7.18 Titrationcurvesofsuccinimidepeptides

Figure7.18Inhibitioncurvesofthepeptidyl‐succinimidepeptides.Taspaseactivitywastestedinafluorogenicinvitroassayatdifferentinhibitorconcentrations.CurveswerenormalizedandnonlinearfittingwasappliedtoobtainIC50values.AlistoftheIC50valuesandthestructuresofthecompoundscanbefoundintable3.2.

SUPPLEMENT PASSAGETHROUGH50KDAMEMBRANE

142

7.19 Passagethrough50kDamembrane

Figure7.19Taspasecanpassa50kDamembrane.Taspasewasdilutedingelfiltrationbufferand10‐foldconcentratedagainina50kDaCentricon.This10‐folddilutionfollowedby10‐foldconcentrationwasrepeatedtwomoretimes.Thefractionretainedbythe50kDamembranewasanalyzedbySDS‐PAGE(firstlane).Theflow‐throughwascollectedandconcentratedina30kDaCentricon(secondlane).Subsequently,theproteinwasdiluted10‐foldandconcentratedagaintwotimes.Theflow‐throughwasconcentratedina10kDaCentriconbeforeanalysis(thirdlane).Asaresult,Taspasecanpassa50kDamembrane.

7.20 StructureoftheTaspasehetero‐tetramer

Figure7.20LocationoftheC‐terminiinthehetero‐tetramerofTaspase.TheC‐termini(red)ofTaspasearelocated60Åapartattheoppositeendsofthehetero‐tetramer.EnforceddimerizationbyC‐terminaltags,asperformedin[68],isthereforelikelytoresultinadimerwithdifferentconformation.

SUPPLEMENT MOLECULARPLUGSTESTEDASTASPASEINHIBITORS

143

7.21 MolecularplugstestedasTaspaseinhibitors

Figure7.21MolecularplugsfailtoinhibitTaspase.aSeveralcompoundsdesignedbytheSchmuckgroupasmolecularplugsforavarietyofotherenzymeswereprescreened forTaspase inhibition invitro.Measurementswere taken in3%DMSOtoguaranteesolubilityofthecompoundsandreferencedtoaDMSOcontrol.Errorbarsindicatestandarddeviationsfromtwo independent experiments. The compounds are listed in supplemental figure 7.22 below. b Twocompounds with promising inhibitory values in the prescreen and sufficient water solubility weredeliberatelyselectedfortitrations.Bothcompoundsrevealednoinhibitoryeffect.


144


145

Figure7.22Structureofthemolecularplugcompounds.StructuresofthemolecularplugcompoundstestedforTaspaseinhibition.AllcompoundsarecourtesyoftheSchmuckgroup.

REFERENCES

146

8 References 1. HowladerN,NooneAM,KrapchoM,Garshell J,MillerD,Altekruse SF,KosaryCL,YuM,Ruhl J,

TatalovichZ,MariottoA, LewisDR, ChenHS, FeuerEJ, CroninKA (2014) SEERCancerStatisticsReview,1975‐2011.

2. BoyleP,FerlayJ(2005)CancerincidenceandmortalityinEurope,2004.AnnOncol16:481‐488.mdi098[pii];10.1093/annonc/mdi098[doi].

3. Perez‐DiezA,MorgunA,ShulzhenkoN(2007)Microarraysforcancerdiagnosisandclassification.AdvExpMedBiol593:74‐85.10.1007/978‐0‐387‐39978‐2_8[doi].

4. HanahanD,WeinbergRA(2000)Thehallmarksofcancer.Cell100:57‐70.S0092‐8674(00)81683‐9[pii].

5. PokharelM(2012)Leukemia:AReviewArticle.IJARPB2:397‐407. 6. BufflerPA,KwanML,ReynoldsP,UrayamaKY(2005)Environmentalandgeneticriskfactorsfor

childhoodleukemia:appraisingtheevidence.CancerInvest23:60‐75. 7. BrownCM,LarsenSR,IlandHJ,JoshuaDE,GibsonJ(2012)Leukaemiasintothe21stcentury:part

1:theacuteleukaemias.InternMedJ42:1179‐1186.10.1111/j.1445‐5994.2012.02938.x[doi].

8. GibsonJ,IlandHJ,LarsenSR,BrownCM,JoshuaDE(2013)Leukaemiasintothe21stcentury.Part2:thechronicleukaemias.InternMedJ43:484‐494.10.1111/imj.12135[doi].

9. SlanyRK(2009)Themolecularbiologyofmixed lineage leukemia.Haematologica94:984‐993.haematol.2008.002436[pii];10.3324/haematol.2008.002436[doi].

10. Belson M, Kingsley B, Holmes A (2007) Risk factors for acute leukemia in children: a review.EnvironHealthPerspect115:138‐145.

11. FreedmanDM,StewartP,KleinermanRA,WacholderS,HatchEE,TaroneRE,RobisonLL,LinetMS(2001)Householdsolventexposuresandchildhoodacutelymphoblasticleukemia.AmJPublicHealth91:564‐567.

12. RinskyRA,YoungRJ,SmithAB(1981)Leukemiainbenzeneworkers.AmJIndMed2:217‐245. 13. IlhanG,KarakusS,AndicN(2006)Riskfactorsandprimarypreventionofacuteleukemia.Asian

PacJCancerPrev7:515‐517. 14. MitelmanF,JohanssonB,MertensF(2007)Theimpactoftranslocationsandgenefusionsoncancer

causation.NatRevCancer7:233‐245.nrc2091[pii];10.1038/nrc2091[doi]. 15. RabbittsTH,StocksMR(2003)Chromosomaltranslocationproductsengendernewintracellular

therapeutictechnologies.NatMed9:383‐386.10.1038/nm0403‐383[doi];nm0403‐383[pii].

16. Aplan PD (2006) Chromosomal translocations involving theMLL gene:molecularmechanisms.DNA Repair (Amst) 5: 1265‐1272. S1568‐7864(06)00174‐1[pii];10.1016/j.dnarep.2006.05.034[doi].

17. MeloJV(1996)Themolecularbiologyofchronicmyeloidleukaemia.Leukemia10:751‐756. 18. LovettBD,LoNL,RappaportEF,BlairIA,OsheroffN,ZhengN,MegonigalMD,WilliamsWR,Nowell

PC,FelixCA(2001)Near‐precise interchromosomal recombinationand functionalDNAtopoisomeraseIIcleavagesitesatMLLandAF‐4genomicbreakpointsintreatment‐relatedacute lymphoblastic leukemiawith t(4;11) translocation. ProcNatl Acad Sci U S A 98:9802‐9807.10.1073/pnas.171309898[doi];171309898[pii].

19. KrivtsovAV,ArmstrongSA(2007)MLLtranslocations,histonemodificationsandleukaemiastem‐celldevelopment.NatRevCancer7:823‐833.nrc2253[pii];10.1038/nrc2253[doi].

20. ReichelM,GillertE,AngermullerS,HenselJP,HeidelF,LodeM,LeisT,BiondiA,HaasOA,StrehlS,Panzer‐GrumayerER,GriesingerF,BeckJD,GreilJ,FeyGH,UckunFM,MarschalekR(2001)BiaseddistributionofchromosomalbreakpointsinvolvingtheMLLgeneininfantsversuschildrenandadultswitht(4;11)ALL.Oncogene20:2900‐2907.10.1038/sj.onc.1204401[doi].

21. Daser A, Rabbitts TH (2005) The versatile mixed lineage leukaemia gene MLL and its manyassociationsinleukaemogenesis.SeminCancerBiol15:175‐188.S1044‐579X(05)00008‐8[pii];10.1016/j.semcancer.2005.01.007[doi].

22. SoCW,KarsunkyH,PassegueE,CozzioA,WeissmanIL,ClearyML(2003)MLL‐GAS7transformsmultipotent hematopoietic progenitors and induces mixed lineage leukemias in mice.CancerCell3:161‐171.S1535610803000199[pii].

23. OnoR,IharaM,NakajimaH,OzakiK,Kataoka‐FujiwaraY,TakiT,NagataK,InagakiM,YoshidaN,KitamuraT,HayashiY,KinoshitaM,NosakaT(2005)DisruptionofSept6,afusionpartner

REFERENCES

147

gene of MLL, does not affect ontogeny, leukemogenesis induced by MLL‐SEPT6, orphenotype induced by the loss of Sept4.Mol Cell Biol 25: 10965‐10978. 25/24/10965[pii];10.1128/MCB.25.24.10965‐10978.2005[doi].

24. GuY,NakamuraT,AlderH,PrasadR,CanaaniO,CiminoG,CroceCM,CanaaniE(1992)Thet(4;11)chromosome translocation of human acute leukemias fuses the ALL‐1 gene, related toDrosophilatrithorax,totheAF‐4gene.Cell71:701‐708.0092‐8674(92)90603‐A[pii].

25. HuretJL,DessenP,BernheimA(2001)Anatlasofchromosomesinhematologicalmalignancies.Example:11q23andMLLpartners.Leukemia15:987‐989.

26. HildenJM,DinndorfPA,MeerbaumSO,SatherH,VillalunaD,HeeremaNA,McGlennenR,SmithFO,WoodsWG,SalzerWL,JohnstoneHS,DreyerZ,ReamanGH(2006)Analysisofprognosticfactorsofacutelymphoblasticleukemiaininfants:reportonCCG1953fromtheChildren'sOncologyGroup.Blood108:441‐451.2005‐07‐3011[pii];10.1182/blood‐2005‐07‐3011[doi].

27. PuiCH,SandlundJT,PeiD,CampanaD,RiveraGK,RibeiroRC,RubnitzJE,RazzoukBI,HowardSC,HudsonMM,ChengC,KunLE,RaimondiSC,BehmFG,DowningJR,RellingMV,EvansWE(2004)Improvedoutcomeforchildrenwithacutelymphoblasticleukemia:resultsofTotalTherapy Study XIIIB at St Jude Children's Research Hospital. Blood 104: 2690‐2696.10.1182/blood‐2004‐04‐1616[doi];2004‐04‐1616[pii].

28. CorralJ,LavenirI,ImpeyH,WarrenAJ,ForsterA,LarsonTA,BellS,McKenzieAN,KingG,RabbittsTH (1996) An Mll‐AF9 fusion gene made by homologous recombination causes acuteleukemiainchimericmice:amethodtocreatefusiononcogenes.Cell85:853‐861.S0092‐8674(00)81269‐6[pii].

29. BursenA,SchwabeK,RusterB,HenschlerR,RuthardtM,DingermannT,MarschalekR(2010)TheAF4.MLL fusion protein is capable of inducing ALL in mice without requirement ofMLL.AF4. Blood 115: 3570‐3579. blood‐2009‐06‐229542 [pii];10.1182/blood‐2009‐06‐229542[doi].

30. GaussmannA,WengerT,EberleI,BursenA,BracharzS,HerrI,DingermannT,MarschalekR(2007)Combined effects of the two reciprocal t(4;11) fusion proteins MLL.AF4 and AF4.MLLconferresistancetoapoptosis,cellcyclingcapacityandgrowthtransformation.Oncogene26:3352‐3363.1210125[pii];10.1038/sj.onc.1210125[doi].

31. Zeleznik‐LeNJ,HardenAM,RowleyJD(1994)11q23translocationssplitthe"AT‐hook"cruciformDNA‐bindingregionandthetranscriptionalrepressiondomainfromtheactivationdomainofthemixed‐lineageleukemia(MLL)gene.ProcNatlAcadSciUSA91:10610‐10614.

32. XiaZB,AndersonM,DiazMO,Zeleznik‐LeNJ(2003)MLLrepressiondomaininteractswithhistonedeacetylases, the polycomb group proteins HPC2 and BMI‐1, and the corepressor C‐terminal‐binding protein. Proc Natl Acad Sci U S A 100: 8342‐8347.10.1073/pnas.1436338100[doi];1436338100[pii].

33. FairK,AndersonM,BulanovaE,MiH,TropschugM,DiazMO(2001)ProteininteractionsoftheMLLPHDfingersmodulateMLLtargetgeneregulationinhumancells.MolCellBiol21:3589‐3597.10.1128/MCB.21.10.3589‐3597.2001[doi].

34. YuBD,HessJL,HorningSE,BrownGA,KorsmeyerSJ(1995)AlteredHoxexpressionandsegmentalidentityinMll‐mutantmice.Nature378:505‐508.10.1038/378505a0[doi].

35. HessJL(2004)MLL:ahistonemethyltransferasedisruptedinleukemia.TrendsMolMed10:500‐507.S1471‐4914(04)00213‐8[pii];10.1016/j.molmed.2004.08.005[doi].

36. HessJL(2004)MechanismsoftransformationbyMLL.CritRevEukaryotGeneExpr14:235‐254.22d69e8b2d47b5ea,0c09f91b76652c8a[pii].

37. Terranova R, Agherbi H, Boned A, Meresse S, Djabali M (2006) Histone and DNAmethylationdefectsatHoxgenes inmiceexpressingaSETdomain‐truncatedformofMll.ProcNatlAcadSciUSA103:6629‐6634.0507425103[pii];10.1073/pnas.0507425103[doi].

38. Kramer OH, Stauber RH, Bug G, Hartkamp J, Knauer SK (2013) SIAH proteins: critical roles inleukemogenesis.Leukemia27:792‐802.leu2012284[pii];10.1038/leu.2012.284[doi].

39. NakamuraT,MoriT,TadaS,KrajewskiW,RozovskaiaT,WassellR,DuboisG,MazoA,CroceCM,CanaaniE(2002)ALL‐1isahistonemethyltransferasethatassemblesasupercomplexofproteins involved in transcriptional regulation. Mol Cell 10: 1119‐1128.S1097276502007402[pii].

40. HsiehJJ,ErnstP,Erdjument‐BromageH,TempstP,KorsmeyerSJ(2003)ProteolyticcleavageofMLLgeneratesacomplexofN‐andC‐terminalfragmentsthatconfersproteinstabilityandsubnuclearlocalization.MolCellBiol23:186‐194.

REFERENCES

148

41. TurkB(2006)Targetingproteases:successes,failuresandfutureprospects.NatRevDrugDiscov5:785‐799.nrd2092[pii];10.1038/nrd2092[doi].

42. VillamilMA,LiangQ,ChenJ,ChoiYS,HouS,LeeKH,ZhuangZ(2012)Serinephosphorylationiscriticalfortheactivationofubiquitin‐specificprotease1anditsinteractionwithWD40‐repeatproteinUAF1.Biochemistry51:9112‐9123.10.1021/bi300845s[doi].

43. CardoneMH,RoyN,StennickeHR,SalvesenGS,FrankeTF,StanbridgeE,FrischS,ReedJC(1998)Regulationofcelldeathproteasecaspase‐9byphosphorylation.Science282:1318‐1321.

44. NicholsonDW(1999)Caspasestructure,proteolyticsubstrates,andfunctionduringapoptoticcelldeath.CellDeathDiffer6:1028‐1042.10.1038/sj.cdd.4400598[doi].

45. NeurathH,WalshKA(1976)Roleofproteolyticenzymesinbiologicalregulation(areview).ProcNatlAcadSciUSA73:3825‐3832.

46. Vergnolle N (2000) Review article: proteinase‐activated receptors ‐ novel signals forgastrointestinalpathophysiology.AlimentPharmacolTher14:257‐266.apt690[pii].

47. StauberRH,KnauerSK(2013)Taspase1—LizenzzumSchneiden.BIOspektrum19:134‐136. 48. DeuE,VerdoesM,BogyoM(2012)Newapproachesfordissectingproteasefunctionstoimprove

probe development and drug discovery. Nat Struct Mol Biol 19: 9‐16. nsmb.2203[pii];10.1038/nsmb.2203[doi].

49. RakashandaS,RanaF,RafiqS,MasoodA,AminS(2012)Roleofproteases incancer:Areview.BiotechnologyandMolecularBiologyReview7:90‐101.

50. Lopez‐OtinC,HunterT(2010)Theregulatorycrosstalkbetweenkinasesandproteasesincancer.NatRevCancer10:278‐292.nrc2823[pii];10.1038/nrc2823[doi].

51. SoreideK,JanssenEA,KornerH,BaakJP(2006)Trypsinincolorectalcancer:molecularbiologicalmechanisms of proliferation, invasion, and metastasis. J Pathol 209: 147‐156.10.1002/path.1999[doi].

52. Kawasaki G, Kato Y, Mizuno A (2002) Cathepsin expression in oral squamous cell carcinoma:relationshipwith clinicopathologic factors. Oral SurgOralMedOral Pathol Oral RadiolEndod93:446‐454.S1079210402889028[pii].

53. TeitzT,WeiT,ValentineMB,VaninEF,GrenetJ,ValentineVA,BehmFG,LookAT,LahtiJM,KiddVJ(2000)Caspase8isdeletedorsilencedpreferentiallyinchildhoodneuroblastomaswithamplificationofMYCN.NatMed6:529‐535.10.1038/75007[doi].

54. DuffyMJ(1996)Proteasesasprognosticmarkersincancer.ClinCancerRes2:613‐618. 55. HoellerD,HeckerCM,DikicI(2006)Ubiquitinandubiquitin‐likeproteinsincancerpathogenesis.

NatRevCancer6:776‐788.nrc1994[pii];10.1038/nrc1994[doi]. 56. Khan JA, Dunn BM, Tong L (2005) Crystal structure of human Taspase1, a crucial protease

regulating the function of MLL. Structure 13: 1443‐1452. S0969‐2126(05)00265‐0[pii];10.1016/j.str.2005.07.006[doi].

57. HsiehJJ,ChengEH,KorsmeyerSJ(2003)Taspase1:athreonineaspartaserequiredforcleavageofMLLandproperHOXgeneexpression.Cell115:293‐303.S009286740300816X[pii].

58. HsuehWA,BaxterJD(1991)Humanprorenin.Hypertension17:469‐477. 59. NagelW,RobelKP,WilligF(1965)ÜberdieAktivierungproteolytischerProenzymedesPankreas.

KlinischeWochenschrift43:173‐174. 60. Arendt CS, Hochstrasser M (1999) Eukaryotic 20S proteasome catalytic subunit propeptides

preventactivesiteinactivationbyN‐terminalacetylationandpromoteparticleassembly.EMBOJ18:3575‐3585.10.1093/emboj/18.13.3575[doi].

61. SaarelaJ,LaineM,TikkanenR,OinonenC,JalankoA,RouvinenJ,PeltonenL(1998)Activationandoligomerizationofaspartylglucosaminidase.JBiolChem273:25320‐25328.

62. MichalskaK,Hernandez‐SantoyoA,JaskolskiM(2008)Themechanismofautocatalyticactivationof plant‐type L‐asparaginases. J Biol Chem 283: 13388‐13397. M800746200[pii];10.1074/jbc.M800746200[doi].

63. Chen DY, Lee Y, Van Tine BA, Searleman AC,Westergard TD, LiuH, Tu HC, Takeda S, Dong Y,Piwnica‐WormsDR,OhKJ,KorsmeyerSJ,HermoneA,GussioR,ShoemakerRH,ChengEH,Hsieh JJ (2012)Apharmacologic inhibitorof theproteaseTaspase1effectively inhibitsbreast and brain tumor growth. Cancer Res 72: 736‐746. 0008‐5472.CAN‐11‐2584[pii];10.1158/0008‐5472.CAN‐11‐2584[doi].

64. OinonenC,RouvinenJ(2000)StructuralcomparisonofNtn‐hydrolases.ProteinSci9:2329‐2337.10.1110/ps.9.12.2329[doi].

65. Oinonen C, Tikkanen R, Rouvinen J, Peltonen L (1995) Three‐dimensional structure of humanlysosomalaspartylglucosaminidase.NatStructBiol2:1102‐1108.

REFERENCES

149

66. Nomme J, Su Y, Konrad M, Lavie A (2012) Structures of apo and product‐bound human L‐asparaginase: insights into themechanismof autoproteolysis and substratehydrolysis.Biochemistry51:6816‐6826.10.1021/bi300870g[doi].

67. Groll M, Ditzel L, Lowe J, Stock D, BochtlerM, Bartunik HD, Huber R (1997) Structure of 20Sproteasomefromyeastat2.4Aresolution.Nature386:463‐471.10.1038/386463a0[doi].

68. BierC,KnauerSK,WunschD,KunstL,ScheidingS,KaiserM,OttmannC,KramerOH,StauberRH(2012) Allosteric inhibition of Taspase1's pathobiological activity by enforceddimerization in vivo. FASEB J 26: 3421‐3429. fj.11‐202432 [pii];10.1096/fj.11‐202432[doi].

69. LeeJT,ChenDY,YangZ,RamosAD,HsiehJJ,BogyoM(2009)Design,syntheses,andevaluationofTaspase1 inhibitors. Bioorg Med Chem Lett 19: 5086‐5090. S0960‐894X(09)00989‐5[pii];10.1016/j.bmcl.2009.07.045[doi].

70. BierC,KnauerSK,KlapthorA,SchweitzerA,RekikA,KramerOH,MarschalekR,StauberRH(2011)Cell‐basedanalysisofstructure‐functionactivityofthreonineaspartase1.JBiolChem286:3007‐3017.M110.161646[pii];10.1074/jbc.M110.161646[doi].

71. KnauerSK,FetzV,RabensteinJ,FriedlS,HofmannB,SabianiS,SchroderE,KunstL,ProschakE,ThinesE,KindlerT,SchneiderG,MarschalekR,StauberRH,BierC(2011)BioassaystomonitorTaspase1functionfortheidentificationofpharmacogeneticinhibitors.PLoSOne6:e18253.10.1371/journal.pone.0018253[doi];PONE‐D‐10‐05577[pii].

72. TakedaS,ChenDY,WestergardTD,FisherJK,RubensJA,SasagawaS,KanJT,KorsmeyerSJ,ChengEH, Hsieh JJ (2006) Proteolysis of MLL family proteins is essential for taspase1‐orchestrated cell cycle progression. Genes Dev 20: 2397‐2409. 20/17/2397[pii];10.1101/gad.1449406[doi].

73. ZhouH,SpicugliaS,HsiehJJ,MitsiouDJ,HoibyT,VeenstraGJ,KorsmeyerSJ,StunnenbergHG(2006)UncleavedTFIIAisasubstratefortaspase1andactiveintranscription.MolCellBiol26:2728‐2735.26/7/2728[pii];10.1128/MCB.26.7.2728‐2735.2006[doi].

74. Bier C, Knauer SK, Docter D, Schneider G, Kramer OH, Stauber RH (2011) The importin‐alpha/nucleophosmin switch controls taspase1 protease function. Traffic 12: 703‐714.10.1111/j.1600‐0854.2011.01191.x[doi].

75. Luo J, SoliminiNL,ElledgeSJ (2009)Principlesof cancer therapy:oncogeneandnon‐oncogeneaddiction. Cell 136: 823‐837. S0092‐8674(09)00200‐1 [pii];10.1016/j.cell.2009.02.024[doi].

76. ChenDY,LiuH,TakedaS,TuHC,SasagawaS,VanTineBA,LuD,ChengEH,HsiehJJ(2010)Taspase1functionsasanon‐oncogeneaddictionproteasethatcoordinatescancercellproliferationandapoptosis.CancerRes70:5358‐5367.0008‐5472.CAN‐10‐0027[pii];10.1158/0008‐5472.CAN‐10‐0027[doi].

77. Schulte TW, Akinaga S, Soga S, SullivanW, Stensgard B, Toft D, Neckers LM (1998) AntibioticradicicolbindstotheN‐terminaldomainofHsp90andsharesimportantbiologicactivitieswithgeldanamycin.CellStressChaperones3:100‐108.

78. LuoJ,EmanueleMJ,LiD,CreightonCJ,SchlabachMR,WestbrookTF,WongKK,ElledgeSJ(2009)Agenome‐wideRNAi screen identifiesmultiple synthetic lethal interactionswith theRasoncogene. Cell 137: 835‐848. S0092‐8674(09)00529‐7 [pii];10.1016/j.cell.2009.05.006[doi].

79. Adams J, KauffmanM (2004) Development of the proteasome inhibitor Velcade (Bortezomib).CancerInvest22:304‐311.

80. SarubbiE,NolliML,AndronicoF,StellaS,SaddlerG,SelvaE,SiccardiA,DenaroM(1991)Ahighthroughput assay for inhibitors of HIV‐1 protease. Screening ofmicrobialmetabolites.FEBSLett279:265‐269.0014‐5793(91)80164‐X[pii].

81. Yang W, Wang L, Paschen W (2013) Development of a high‐throughput screening assay forinhibitors of small ubiquitin‐like modifier proteases. J Biomol Screen 18: 621‐628.1087057113479971[pii];10.1177/1087057113479971[doi].

82. Rabenstein, Jens (2008) Inhibierung des pathomolekularen Mechanismuseiner t(4;11)‐assoziiertenLeukämie[dissertation].Goethe‐UniversitätFrankfurtamMain.

83. WunschD,FetzV,HeiderD,TenzerS,BierC,KunstL,KnauerS,StauberR(2012)Chemico‐geneticstrategiestoinhibittheleukemicpotentialofthreonineaspartase‐1.BloodCancerJ2:e77.10.1038/bcj.2012.22[doi].

84. ThielP,KaiserM,OttmannC(2012)Small‐moleculestabilizationofprotein‐proteininteractions:anunderestimatedconceptindrugdiscovery?AngewChemIntEdEngl51:2012‐2018.10.1002/anie.201107616[doi].

REFERENCES

150

85. BierC,HechtR,KunstL,ScheidingS,WunschD,GoessweinD,SchneiderG,KramerOH,KnauerSK,Stauber RH (2012) Overexpression of the catalytically impaired Taspase1 T234V orTaspase1D233AvariantsdoesnothaveadominantnegativeeffectinT(4;11)leukemiacells.PLoSOne7:e34142.10.1371/journal.pone.0034142[doi];PONE‐D‐11‐20159[pii].

86. DavisME,ChenZG,ShinDM(2008)Nanoparticletherapeutics:anemergingtreatmentmodalityforcancer.NatRevDrugDiscov7:771‐782.nrd2614[pii];10.1038/nrd2614[doi].

87. YahCS, SimateGS, Iyuke SE (2012)Nanoparticles toxicity and their routes of exposures. Pak JPharmSci25:477‐491.

88. MohanrajVJ,ChenY(2006)Nanoparticles‐Areview.TropicalJournalofPharmaceuticalResearch5:561‐573.

89. BurlesonDJ,DriessenMD,PennRL(2004)Onthecharacterizationofenvironmentalnanoparticles.JEnvironSciHealthAToxHazardSubstEnvironEng39:2707‐2753.

90. Elsaesser A, Howard CV (2012) Toxicology of nanoparticles. Adv Drug Deliv Rev 64: 129‐137.S0169‐409X(11)00232‐8[pii];10.1016/j.addr.2011.09.001[doi].

91. Oberdorster G, Oberdorster E, Oberdorster J (2005) Nanotoxicology: an emerging disciplineevolvingfromstudiesofultrafineparticles.EnvironHealthPerspect113:823‐839.

92. NelAE,MadlerL,VelegolD,XiaT,HoekEM,SomasundaranP,KlaessigF,CastranovaV,ThompsonM(2009)Understandingbiophysicochemical interactionsat thenano‐bio interface.NatMater8:543‐557.nmat2442[pii];10.1038/nmat2442[doi].

93. NelA,XiaT,MadlerL,LiN(2006)Toxicpotentialofmaterialsatthenanolevel.Science311:622‐627.311/5761/622[pii];10.1126/science.1114397[doi].

94. LeroueilPR,BerrySA,DuthieK,HanG,RotelloVM,McNernyDQ,BakerJR, Jr.,OrrBG,HollMM(2008)Widevarietiesofcationicnanoparticlesinducedefectsinsupportedlipidbilayers.NanoLett8:420‐424.10.1021/nl0722929[doi].

95. ChenM,vonMA(2005)Formationofnucleoplasmicproteinaggregatesimpairsnuclearfunctionin response to SiO2 nanoparticles. Exp Cell Res 305: 51‐62. S0014‐4827(04)00733‐5[pii];10.1016/j.yexcr.2004.12.021[doi].

96. Marano F, Hussain S, Rodrigues‐Lima F, Baeza‐Squiban A, Boland S (2011) Nanoparticles:molecular targets and cell signalling. Arch Toxicol 85: 733‐741. 10.1007/s00204‐010‐0546‐4[doi].

97. ShangW,NufferJH,DordickJS,SiegelRW(2007)UnfoldingofribonucleaseAonsilicananoparticlesurfaces.NanoLett7:1991‐1995.10.1021/nl070777r[doi].

98. MehrabiM,WilsonR(2007)IntercalatinggoldnanoparticlesasuniversallabelsforDNAdetection.Small3:1491‐1495.10.1002/smll.200700230[doi].

99. TenzerS,DocterD,KuharevJ,MusyanovychA,FetzV,HechtR,SchlenkF,FischerD,KiouptsiK,Reinhardt C, Landfester K, Schild H, Maskos M, Knauer SK, Stauber RH (2013) Rapidformationofplasmaprotein corona critically affectsnanoparticlepathophysiology.NatNanotechnol8:772‐781.nnano.2013.181[pii];10.1038/nnano.2013.181[doi].

100. BagreAP,JainK,JainNK(2013)Alginatecoatedchitosancoreshellnanoparticlesfororaldeliveryof enoxaparin: in vitro and in vivo assessment. Int J Pharm 456: 31‐40. S0378‐5173(13)00770‐9[pii];10.1016/j.ijpharm.2013.08.037[doi].

101. Garcia‐BennettAE,KozhevnikovaM,KonigN,ZhouC,LeaoR,KnopfelT,PankratovaS,TrolleC,BerezinV,BockE,AldskogiusH,KozlovaEN(2013)Deliveryofdifferentiationfactorsbymesoporous silica particles assists advanced differentiation of transplanted murineembryonic stem cells. Stem Cells Transl Med 2: 906‐915. sctm.2013‐0072[pii];10.5966/sctm.2013‐0072[doi].

102. SoniS,TyagiH,TaylorRA,KumarA(2013)Roleofopticalcoefficientsandhealthytissue‐sparingcharacteristics ingoldnanorod‐assistedthermaltherapy. Int JHyperthermia29:87‐97.10.3109/02656736.2012.753162[doi].

103. Matsumura Y, Maeda H (1986) A new concept for macromolecular therapeutics in cancerchemotherapy:mechanismof tumoritropicaccumulationofproteinsand theantitumoragentsmancs.CancerRes46:6387‐6392.

104. CondeJ,delaFuenteJM,BaptistaPV(2013)Nanomaterialsforreversionofmultidrugresistanceincancer:anewhopeforanoldidea?FrontPharmacol4:134.10.3389/fphar.2013.00134[doi].

105. Teodor E, Litescu SC, Lazar V, Somoghi R (2009) Hydrogel‐magnetic nanoparticles withimmobilizedL‐asparaginaseforbiomedicalapplications.JMaterSciMaterMed20:1307‐1314.10.1007/s10856‐008‐3684‐y[doi].

REFERENCES

151

106. Ortac I,RuffL,YehY,EsenerS,MessmerBT(2013)NanoparticleEncapsulatedL‐Asparaginase.Blood122:2669.

107. Wang C, Liu D, Wang Z (2012) Gold nanoparticle based dot‐blot immunoassay for sensitivelydetectingAlzheimer'sdiseaserelatedbeta‐amyloidpeptide.ChemCommun(Camb)48:8392‐8394.10.1039/c2cc33568a[doi].

108. PengG,TischU,AdamsO,HakimM,ShehadaN,BrozaYY,BillanS,Abdah‐BortnyakR,KutenA,HaickH(2009)Diagnosing lungcancer inexhaledbreathusinggoldnanoparticles.NatNanotechnol4:669‐673.nnano.2009.235[pii];10.1038/nnano.2009.235[doi].

109. KarajanagiSS,VertegelAA,KaneRS,DordickJS(2004)Structureandfunctionofenzymesadsorbedonto single‐walled carbonnanotubes. Langmuir 20: 11594‐11599. 10.1021/la047994h[doi].

110. LinetMS,RiesLA,SmithMA,TaroneRE,DevesaSS(1999)Cancersurveillanceseries:recenttrendsinchildhoodcancer incidenceandmortality in theUnitedStates. JNatlCancer Inst91:1051‐1058.

111. VrankenWF,BoucherW,StevensTJ,FoghRH,PajonA,LlinasM,UlrichEL,MarkleyJL,IonidesJ,LaueED(2005)TheCCPNdatamodelforNMRspectroscopy:developmentofasoftwarepipeline.Proteins59:687‐696.10.1002/prot.20449[doi].

112. PettersenEF,GoddardTD,HuangCC,CouchGS,GreenblattDM,MengEC,FerrinTE(2004)UCSFChimera‐‐avisualizationsystemforexploratoryresearchandanalysis.JComputChem25:1605‐1612.10.1002/jcc.20084[doi].

113. SieversF,WilmA,DineenD,GibsonTJ,KarplusK,LiW,LopezR,McWilliamH,RemmertM,SodingJ, Thompson JD, Higgins DG (2011) Fast, scalable generation of high‐quality proteinmultiple sequence alignments using Clustal Omega. Mol Syst Biol 7: 539. msb201175[pii];10.1038/msb.2011.75[doi].

114. Ewing TJ, Makino S, Skillman AG, Kuntz ID (2001) DOCK 4.0: search strategies for automatedmoleculardockingofflexiblemoleculedatabases.JComputAidedMolDes15:411‐428.

115. GraphPadSoftwareSDCU(2009)GraphPadPrism,version5.03[computerprogram]. 116. SchneiderCA,RasbandWS,EliceiriKW(2012)NIHImagetoImageJ:25yearsofimageanalysis.

NatMethods9:671‐675. 117. KibbeWA(2007)OligoCalc:anonlineoligonucleotidepropertiescalculator.NucleicAcidsRes35:

W43‐W46.gkm234[pii];10.1093/nar/gkm234[doi]. 118. PuigboP,GuzmanE,RomeuA,Garcia‐VallveS(2007)OPTIMIZER:awebserverforoptimizingthe

codon usage of DNA sequences. Nucleic Acids Res 35: W126‐W131. gkm219[pii];10.1093/nar/gkm219[doi].

119. GasteigerE,GattikerA,HooglandC,IvanyiI,AppelRD,BairochA(2003)ExPASy:Theproteomicsserverforin‐depthproteinknowledgeandanalysis.NucleicAcidsRes31:3784‐3788.

120. SchrödingerL(2010)ThePyMOLMolecularGraphicsSystem,version1.5[computerprogram]. 121. [Anonymous](2014)SIAS‐sequenceidentityandsimilarity. 122. Arnold K, Bordoli L, Kopp J, Schwede T (2006) The SWISS‐MODEL workspace: a web‐based

environment for protein structure homology modelling. Bioinformatics 22: 195‐201.bti770[pii];10.1093/bioinformatics/bti770[doi].

123. BartasunP,CieslinskiH,BujaczA,Wierzbicka‐WosA,KurJ(2013)AstudyontheinteractionofrhodamineBwithmethylthioadenosinephosphorylaseproteinsourcedfromanAntarcticsoil metagenomic library. PLoS One 8: e55697. 10.1371/journal.pone.0055697[doi];PONE‐D‐12‐29171[pii].

124. Cheng Y, Prusoff WH (1973) Relationship between the inhibition constant (K1) and theconcentration of inhibitor which causes 50 per cent inhibition (I50) of an enzymaticreaction.BiochemPharmacol22:3099‐3108.

125. SreeramaN,WoodyRW(2000)Estimationofproteinsecondarystructurefromcirculardichroismspectra: comparison of CONTIN, SELCON, and CDSSTR methods with an expandedreference set. Anal Biochem 287: 252‐260. 10.1006/abio.2000.4880 [doi];S0003‐2697(00)94880‐2[pii].

126. ZhangH,NealS,WishartDS(2003)RefDB:adatabaseofuniformlyreferencedproteinchemicalshifts.JBiomolNMR25:173‐195.5118302[pii].

127. Kuntz ID, Blaney JM, Oatley SJ, Langridge R, Ferrin TE (1982) A geometric approach tomacromolecule‐ligandinteractions.JMolBiol161:269‐288.0022‐2836(82)90153‐X[pii].

128. MengEC,ShoichetBK,KuntzID(1992)Automateddockingwithgrid‐basedenergyevaluation.JCompChem13:505‐524.

REFERENCES

152

129. Irwin JJ, SterlingT,MysingerMM,BolstadES,ColemanRG (2012)ZINC:a free tool todiscoverchemistryforbiology.JChemInfModel52:1757‐1768.10.1021/ci3001277[doi].

130. VoigtJH,BienfaitB,WangS,NicklausMC(2001)ComparisonoftheNCIopendatabasewithsevenlargechemicalstructuraldatabases.JChemInfComputSci41:702‐712.ci000150t[pii].

131. CaseDA,CheathamTE,III,DardenT,GohlkeH,LuoR,MerzKM,Jr.,OnufrievA,SimmerlingC,WangB,WoodsRJ(2005)TheAmberbiomolecularsimulationprograms. JComputChem26:1668‐1688.10.1002/jcc.20290[doi].

132. Neumann, D. (2006) Strukturelle und biochemische Charakterisierungder humanen PAPS‐Synthetase2[dissertation].Universität Duisburg-Essen.

133. StauberRH,BierC,KnauerSK(2012)TargetingTaspase1forcancertherapy‐‐letter.CancerRes72:2912.0008‐5472.CAN‐12‐0150[pii];10.1158/0008‐5472.CAN‐12‐0150[doi].

134. vandenBoomJ,MamicM,BacchelliereD,ZweeringS,KaschaniF,KnauerS,BayerP,KaiserM(2014) Peptidyl succinimidyl peptides as potent Taspase1 inhibitors; manuscriptsubmitted.

135. Krasotkina J, Borisova AA, Gervaziev YV, Sokolov NN (2004) One‐step purification and kineticpropertiesoftherecombinantL‐asparaginasefromErwiniacarotovora.BiotechnolApplBiochem39:215‐221.10.1042/BA20030138[doi];BA20030138[pii].

136. PokrovskayaMV,AleksandrovaSS, PokrovskyVS,OmeljanjukNM,BorisovaAA,AnisimovaNY,SokolovNN(2012)Cloning,expressionandcharacterizationoftherecombinantYersiniapseudotuberculosis L‐asparaginase. Protein Expr Purif 82: 150‐154. S1046‐5928(11)00351‐2[pii];10.1016/j.pep.2011.12.005[doi].

137. Bansal S, Srivastava A, Mukherjee G, Pandey R, Verma AK, Mishra P, Kundu B (2012)Hyperthermophilic asparaginase mutants with enhanced substrate affinity andantineoplasticactivity:structuralinsightsontheirmechanismofaction.FASEBJ26:1161‐1171.fj.11‐191254[pii];10.1096/fj.11‐191254[doi].

138. NesterchukMV, Sergiev PV, Dontsova OA (2011) PosttranslationalModifications of RibosomalProteinsinEscherichiacoli.ActaNaturae3:22‐33.

139. BorekD,JaskolskiM(2000)CrystallizationandpreliminarycrystallographicstudiesofanewL‐asparaginaseencodedbytheEscherichiacoligenome.ActaCrystallogrDBiolCrystallogr56:1505‐1507.S0907444900010076[pii].

140. XuanJ,TarentinoAL,GrimwoodBG,PlummerTH,Jr.,CuiT,GuanC,VanRP(1998)CrystalstructureofglycosylasparaginasefromFlavobacteriummeningosepticum.ProteinSci7:774‐781.10.1002/pro.5560070327[doi].

141. YaoJ,DysonHJ,WrightPE(1997)Chemicalshiftdispersionandsecondarystructurepredictioninunfolded andpartly foldedproteins. FEBSLett 419: 285‐289. S0014‐5793(97)01474‐9[pii].

142. Neri D, Wider G, Wuthrich K (1992) Complete 15N and 1H NMR assignments for the amino‐terminaldomainofthephage434repressorintheurea‐unfoldedform.ProcNatlAcadSciUSA89:4397‐4401.

143. VrankenWF,RiepingW(2009)Relationshipbetweenchemicalshiftvalueandaccessiblesurfaceareaforallaminoacidatoms.BMCStructBiol9:20.1472‐6807‐9‐20[pii];10.1186/1472‐6807‐9‐20[doi].

144. Fu CH, Sakamoto KM (2007) PEG‐asparaginase. Expert Opin Pharmacother 8: 1977‐1984.10.1517/14656566.8.12.1977[doi].

145. MullerHJ,BoosJ(1998)UseofL‐asparaginaseinchildhoodALL.CritRevOncolHematol28:97‐113.S1040‐8428(98)00015‐8[pii].

146. JiangQQ,BartschL,SickingW,WichPR,HeiderD,HoffmannD,SchmuckC(2013)Anewapproachtoinhibithumanbeta‐tryptasebyproteinsurfacebindingoffour‐armedpeptideligandswithtwodifferentsetsofarms.OrgBiomolChem11:1631‐1639.10.1039/c3ob27302d[doi].

147. EastmanJ,WilsonEJ,CervenanskyC,RosenberryTL(1995)Fasciculin2bindstotheperipheralsiteonacetylcholinesteraseandinhibitssubstratehydrolysisbyslowingastepinvolvingprotontransferduringenzymeacylation.JBiolChem270:19694‐19701.

148. XiongS,QiW,ChengY,HuangB,WangM,LiY(2011)Modelingsizeeffectsonthesurface freeenergy of metallic nanoparticles and nanocavities. Phys Chem Chem Phys 13: 10648‐10651.10.1039/c0cp02102d[doi].

149. NiemeyerCM(2001)Nanoparticles,Proteins,andNucleicAcids:BiotechnologyMeetsMaterialsScience.AngewandteChemie40:4128‐4158.

REFERENCES

153

150. TenzerS,DocterD,RosfaS,WlodarskiA,KuharevJ,RekikA,KnauerSK,BantzC,NawrothT,BierC,SirirattanapanJ,MannW,TreuelL,ZellnerR,MaskosM,SchildH,StauberRH(2011)Nanoparticle size is a criticalphysicochemicaldeterminantof thehumanbloodplasmacorona: a comprehensive quantitative proteomic analysis. ACS Nano 5: 7155‐7167.10.1021/nn201950e[doi].

151. Cedervall T, Lynch I, Lindman S, Berggard T, Thulin E, Nilsson H, Dawson KA, Linse S (2007)Understandingthenanoparticle‐proteincoronausingmethodstoquantifyexchangeratesand affinities of proteins for nanoparticles. Proc Natl Acad Sci U S A 104: 2050‐2055.0608582104[pii];10.1073/pnas.0608582104[doi].

152. ZhengC,ZhangXL,LiuW,LiuB,YangHH,LinZA,ChenGN(2013)Aselectiveartificialenzymeinhibitorbasedonnanoparticle‐enzymeinteractionsandmolecularimprinting.AdvMater25:5922‐5927.10.1002/adma.201302064[doi].

153. GuerreiroA,PomaA,KarimK,MoczkoE,TakaradaJ,deVargas‐SansalvadorIP,TurnerN,PiletskaE,deMagalhaesCS,GlazovaN,SerkovaA,OmelianovaA,PiletskyS(2014)InfluenceofSurface‐Imprinted Nanoparticles on Trypsin Activity. Adv Healthc Mater .10.1002/adhm.201300634[doi].

154. LinseS,Cabaleiro‐LagoC,XueWF,LynchI,LindmanS,ThulinE,RadfordSE,DawsonKA(2007)Nucleationofprotein fibrillationbynanoparticles.ProcNatlAcadSciUSA104:8691‐8696.0701250104[pii];10.1073/pnas.0701250104[doi].

155. Asuri P,Karajanagi SS,VertegelAA,Dordick JS,KaneRS (2007)Enhanced stability of enzymesadsorbedontonanoparticles.JNanosciNanotechnol7:1675‐1678.

156. AshtariK,KhajehK,Fasihi J,AshtariP,RamazaniA,ValiH(2012)Silica‐encapsulatedmagneticnanoparticles:enzymeimmobilizationandcytotoxicstudy.IntJBiolMacromol50:1063‐1069.S0141‐8130(11)00486‐7[pii];10.1016/j.ijbiomac.2011.12.025[doi].

157. LeeWA,PernodetN,LiB,LinCH,HatchwellE,RafailovichMH(2007)Multicomponentpolymercoating to block photocatalytic activity of TiO2 nanoparticles. Chem Commun (Camb )4815‐4817.10.1039/b709449c[doi].

158. AuffanM,AchouakW,RoseJ,RoncatoMA,ChaneacC,WaiteDT,MasionA,WoicikJC,WiesnerMR,BotteroJY(2008)Relationbetweentheredoxstateofiron‐basednanoparticlesandtheircytotoxicitytowardEscherichiacoli.EnvironSciTechnol42:6730‐6735.

159. Trueb RM (2009) Oxidative stress in ageing of hair. Int J Trichology 1: 6‐14. 10.4103/0974‐7753.51923[doi].

160. EllisRJ,MintonAP(2006)Proteinaggregationincrowdedenvironments.BiolChem387:485‐497.10.1515/BC.2006.064[doi].

161. vandenBergB,EllisRJ,DobsonCM(1999)Effectsofmacromolecularcrowdingonproteinfoldingandaggregation.EMBOJ18:6927‐6933.10.1093/emboj/18.24.6927[doi].

162. DecuzziP,FerrariM(2007)Theroleofspecificandnon‐specificinteractionsinreceptor‐mediatedendocytosis of nanoparticles. Biomaterials 28: 2915‐2922. S0142‐9612(07)00153‐6[pii];10.1016/j.biomaterials.2007.02.013[doi].

163. AkermanME,ChanWC,LaakkonenP,BhatiaSN,RuoslahtiE(2002)Nanocrystaltargetinginvivo.ProcNatlAcadSciUSA99:12617‐12621.10.1073/pnas.152463399 [doi];152463399[pii].

164. HoshinoA,FujiokaK,OkuT,NakamuraS,SugaM,YamaguchiY,SuzukiK,YasuharaM,YamamotoK (2004) Quantum dots targeted to the assigned organelle in living cells. MicrobiolImmunol48:985‐994.JST.JSTAGE/mandi/48.985[pii].

165. HanleyC,ThurberA,HannaC,PunnooseA,ZhangJ,WingettDG(2009)TheInfluencesofCellTypeandZnONanoparticleSizeonImmuneCellCytotoxicityandCytokineInduction.NanoscaleResLett4:1409‐1420.1556‐276X‐4‐1409[pii];10.1007/s11671‐009‐9413‐8[doi].

166. MossmanBT,ChurgA (1998)Mechanisms in thepathogenesisofasbestosis and silicosis.Am JRespirCritCareMed157:1666‐1680.10.1164/ajrccm.157.5.9707141[doi].

167. OvrevikJ,LagM,SchwarzeP,RefsnesM(2004)p38andSrc‐ERK1/2pathwaysregulatecrystallinesilica‐inducedchemokinerelease inpulmonaryepithelialcells.ToxicolSci81:480‐490.10.1093/toxsci/kfh214[doi];kfh214[pii].

168. WarheitDB,McHughTA,HartskyMA (1995)Differential pulmonary responses in rats inhalingcrystalline,colloidaloramorphoussilicadusts.ScandJWorkEnvironHealth21Suppl2:19‐21.76[pii].

169. CooperGM(2000)TheNuclearEnvelopeandTrafficbetweentheNucleusandCytoplasm.In:TheCell‐AMolecularApproach.BostonUniversity.

REFERENCES

154

170. PanteN,KannM(2002)Nuclearporecomplexisabletotransportmacromoleculeswithdiametersofabout39nm.MolBiolCell13:425‐434.10.1091/mbc.01‐06‐0308[doi].

171. Ghose AK, Viswanadhan VN, Wendoloski JJ (1999) A knowledge‐based approach in designingcombinatorial ormedicinal chemistry libraries for drug discovery. 1. A qualitative andquantitativecharacterizationofknowndrugdatabases.JCombChem1:55‐68.

172. LipinskiCA(2003)ChrisLipinskidiscusseslifeandchemistryaftertheRuleofFive.DrugDiscovToday8:12‐16.S1359644602025564[pii].

173. Moriguchi I, Hirono S, Liu Q, Nakagome I, Matsushita Y (1992) Simple method of calculatingOctanol/WaterPartitionCoefficient.ChemPharmBull40:127‐130.

174. Faundez V,Hartzell HC (2004) Intracellular chloride channels: determinants of function in theendosomal pathway. Sci STKE 2004: re8. 10.1126/stke.2332004re8[doi];stke.2332004re8[pii].

175. Pilas B, Durack G (1997) A flow cytometricmethod formeasurement of intracellular chlorideconcentration in lymphocytes using the halide‐specific probe 6‐methoxy‐N‐(3‐sulfopropyl) quinolinium (SPQ). Cytometry 28: 316‐322. 10.1002/(SICI)1097‐0320(19970801)28:4<316::AID‐CYTO7>3.0.CO;2‐9[pii].

176. HuangZM,PrasadC,BrittonFC,YeLL,HattonWJ,DuanD(2009)FunctionalroleofCLC‐2chlorideinwardrectifierchannelsincardiacsinoatrialnodalpacemakercells.JMolCellCardiol47:121‐132.S0022‐2828(09)00158‐8[pii];10.1016/j.yjmcc.2009.04.008[doi].

177. CongreveM,CarrR,MurrayC,JhotiH(2003)A'ruleofthree'forfragment‐basedleaddiscovery?DrugDiscovToday8:876‐877.S1359644603028319[pii].

178. KlebeG(2009)Peptidomimetika.In:Wirkstoffdesign.SpektrumVerlag.pp.137‐150. 179. MacLeodAM,MerchantKJ, CascieriMA, Sadowski S,BerE, SwainCJ,BakerR (1993)N‐acyl‐L‐

tryptophanbenzylesters:potentsubstancePreceptorantagonists.JMedChem36:2044‐2045.

ACKNOWLEDGEMENT

155

9 AcknowledgementAnersterStellemöchteichmichbeimeinemDoktorvaterProf.PeterBayerfürseinevorbildliche

Betreuungbedanken.VorallemdieFreiheiten,diedumirbeiExperimentengelassenhastunddie

Offenheit,mitderduneuenIdeengegenübertrittst,weißichsehrzuschätzen.Ebensowaressehr

angenehm,dassduauchabseitsderArbeitimmerbemühtbist,eineunkomplizierteLösungfür

unsereProblemezufinden.

BeiProf.ShirleyKnauermöchteichmichfürdieHilfebeiderProjektsuchebedankenundfürdievielenhilfreichenGesprächewährendmeinerDoktorarbeit.DieKooperationhatmirnichtzuletztwegendeinesEinsatzesfürmichunddeinerHilfesovielFreudegemacht.

WeiterhingehtmeinDankanProf.MarkusKaiserfürdieInhibitoren,ohnediedieseArbeitnurdie Hälfte wert wäre. Deine unverblümte Direktheit hat die Gespräche sehr unkompliziertgemacht.

BesondererDankgehtanFranzi,dankefürsgemeinsameLaufen,dieGassi‐RundenmitSamundvieleGrillabende.DuhastmirmitdeinerehrlichenArtnichtnurwissenschaftlichweitergeholfen.

MuchasgraciasaAnjaporserunabuenaoyente,porresistirmisataquesyporreírsesiempredemischistes.

HvalaanEdisa fürdieVersorgungmitSüßigkeiten,dankeanLukas,derauchmeine flachstenWitzenochunterbietenkann,unddankeanAndréfürdiespannendenDSA‐Runden.

Außerdemmöchte ichallenMitgliedernundEhemaligenderAGBayer–Tina,Daniel,Cristina,Peter,Jenny,Irina,Julia,Dana,Mina,Gila,AlmaundChristoph–fürdievielenlustigenMomentewährendderArbeitdanken.VielenDankanDanieleundMarijafürdieVersorgungmitInhibitorenundanLena fürDNA‐Nachschub.Dankean JuliaundManu fürdiekreativen„Accessoires“ fürmeinenArbeitsplatz.

MeinenElterndanke ich fürdieUnterstützungwährenddesgesamtenStudiums.UndmeinemBruderdankeich,dassermirregelmäßigmeinePizzenwegisst–zumGlücknurdievegetarischen.

EinganzgroßesDankeschöngehtanClaudia,dieimmerzumirgehaltenhat,auchwennesnichtimmerleichtwar.DafürliebeichdichundohnedichhätteichimLebenhöchstenshalbsovielFreude.

156

10 CurriculumVitae

Forreasonsofdataprotection,thecurriculumvitaeisnotincludedintheonlineversion.

158

11 Declarations

Erklärung:

Hiermit erkläre ich, gem. § 6 Abs. (2) f) der Promotionsordnung der Fakultäten für Biologie,

Chemie undMathematik zur Erlangung der Dr. rer. nat., dass ich das Arbeitsgebiet, dem das

Thema „Novel inhibitors for the protease Taspase1“ zuzuordnen ist, in Forschung und Lehre

vertreteunddenAntragvonJohannesvandenBoombefürworteunddieBetreuungauchimFalle

einesWeggangs,wennnichtwichtigeGründedementgegenstehen,weiterführenwerde.

Essen,den_________________________________________________________

UnterschrifteinesMitgliedsderUniversitätDuisburg‐Essen

Erklärung:

Hiermit erkläre ich, gem. §7Abs. (2) c) + e) derPromotionsordnungFakultäten fürBiologie,

ChemieundMathematik zurErlangungdesDr. rer. nat., dass ichdie vorliegendeDissertation

selbständigverfasstundmichkeineranderenalsderangegebenenHilfsmittelbedienthabe.

Essen,den________________________________________________________

Unterschriftdes/rDoktoranden/in

Erklärung:

Hiermiterkläreich,gem.§7Abs.(2)d)+f)derPromotionsordnungderFakultätenfürBiologie,

ChemieundMathematikzurErlangungdesDr.rer.nat.,dassichkeineanderenPromotionenbzw.

PromotionsversucheinderVergangenheitdurchgeführthabeunddassdieseArbeitvonkeiner

anderenFakultät/Fachbereichabgelehntwordenist.

Essen,den________________________________________________________

UnterschriftdesDoktoranden

novel inhibitors for the protease taspase1 - core.ac.uk · m molar, mol/l maldi matrix‐assisted...

Documents