New Insights to Resistance of a Novel Drug Bedaquiline using in-vitro Mutants of ATP Synthase in Mycobacterium Tuberculosis

Ameeruddin Nusrath Unissa1*; Appisetty Ramya Lakshmi1; Swathi Sukumar 1; Luke Elizabeth

Hanna2

1 Centre for Biomedical Informatics, National Institute for Research in Tuberculosis, India2 Division of Clinical Research, National Institute for Research in Tuberculosis, India

*Corresponding to :A. Nusrath Unissa, Department of Clinical Research, National Institute for Research in

Tuberculosis (NIRT), Indian Council of Medical Research (ICMR), No. 1, Mayor Sathyamoorthy Road, Chet-

put, Chennai 600 031, Tamil Nadu, India.

Phone: 91-044 -2836 9597; Fax: +91-(044)-2836 2528; E-mail: nusrathunissa@gmail.com

Chapter 2

Diagnosis and Management of Tuberculosis

Abstract Bedaquiline(BDQ)isthenewfirst-in-classanti-tuberculosis(TB)compoundbelonging to theclassofdiarylquinolonewithactivityagainstdrug-sensitiveanddrug-resistantMycobacterium tuberculosis (M. tuberculosis).ThisnoveldrughastheimmensepotentialtoshortenTBtreatmentdurationandhasbeenadvocatedformulti-drugresistant(MDR)-TBtreatment.Therefore,BDQresistancecanbecon-sideredasamajorpublichealthproblemandmolecularinvestigationofthesameistheutmostneedofthehour.ThetargetbasedconceptofresistancetoBDQiscausedbymutationinc-ringofadenosinetriphosphate(ATP)synthase,acriticalenzymeinthesynthesisofATPinM.tuberculosiscodedbyatpEgene.BDQinhibitstheprotonpumpofmycobacterialATPsynthase.TounderstandthemolecularbasisofBDQre-sistanceusingmutants(MTs)astheemergenceofstrainsresistanttoBDQmayposeapotentialthreattotheTBcontrolprogram,weundertookainitiativetostudysevenin vitromutantsofAtpEviz.,Asp28Gly,Asp28Ala,Asp28Pro,Asp28Val,Glu61-Asp,Ala63ProandIle66Met(D28G,D28A,D28PD28V,E61D,A63P,andI66M)which involves in resistanceagainstBDQ.Molecularmodelinganddockingwasperformed tounderstand the interactingbehaviourofmutant (MT)enzymeswithBDQincomparisontowild-type(WT).TheseresultsindicatethatthesubstitutionsinAtpEexceptD28GshowedhighaffinitytowardsthedrugincomparisontotheWT.ThiscouldbeduetothefavourableinteractionsinmutantscomparedtoWT.

DiagnosisandManagementofTuberculosis

2

www.openaccessebooks.comUnissaA

N

Itcanbeinferredfromthisconciseanalysisthatthemutants(D28A,D28PD28V,E61D,A63P,andI66M)duetohighaffinitybindswiththeBDQtightlyleadingtoslowornoreleaseoftheBDQwhicheventuallyresultsinhighlevelBDQresis-tance,comparedwithD28Gthatcouldleadtolow-levelresistance.

key words:Mycobacteriumtuberculosis;BDQresistance;ATPsynthase;mutants

1. Introduction

Despitebeingacontrollable,preventableandcurabledisease,tuberculosis(TB)stillre-mainasamajorpublichealthprobleminmanypartsoftheworld.Theincreaseinmulti-drugresistant(MDR-TB)isdefinedasresistancetothetwomosteffectivefirstlineTBdrugs:ri-fampicin(RIF)andisoniazid(INH)hasintensifiedthemagnitudeofthesituation.Extensivelydrug-resistantTB(XDR-TB) isemergingasanevenmoreominous threat, that is resistanttoanyfluoroquinolone,andatleastoneofthreeinjectablesecond-linedrugs(capreomycin,kanamycin,andamikacin),inadditiontoINHandRIF.DrugresistanceinTBisessentiallyapotentialthreattotheTBcontrolprograms.Themostrecentdrug-resistancesurveillancedataissuedbytheWorldhealthorganization(WHO)estimatesthatanaverageofroughly9%ofMDR-TBcasesareXDR-TB[1].

1.2. Properties of Bedaquiline (BDQ)

Bedaquiline(BDQ)isalsoknownasTMC207orR207910,isadrugbelongingtotheclassofdiarylquinolineandfoundtobeeffectiveagainstbothdrugsusceptibleanddrugresis-tanceTB[2].ItisaverypromisingandrelativelynewcandidatedrugfortreatmentofTBandmycobacterial infections[2].TheactivityspectrumofTMC207includes themycobacterialspecies thatarepathogenic tohumans, suchasMycobacterium tuberculosis (M. tuberculo-sis),butalsoatypicalpathogenicspecies,suchasM. avium complex, M kansasii,andthefastgrowersM. fortuitum and M. abscessus[2].Itisfoundtobeactivewithinmacrophages,andanimportantagentinshorteningthedurationofanti-TBtreatment[3].

1.3. BDQ in MDR treatment

PreclinicalstudieshaveshowntheefficacyofBDQintermsofreductioninbacterialloadandtreatmentduration.BasedonthepositiveinfluenceofBDQfortreatmentofMDR-TB[4],in2012,foodanddrugadministration(FDA)approvedBDQfortreatmentofMDR-TBandXDR-TB,followedthismanyclinicaltrialswereundertakenandwasalsoevaluatedinamulti-centricstudyfortreatmentofMDR-TBandXDR-TB[5].PhaseIIclinicalstudieshaveestablishedthesafety, tolerabilityandearliersputumconversiontimeinpatientswithMDR-TB[6].ItiscurrentlyinphaseIIIclinicaltrialsforpatientswithMDR-TB.Recently,astudyfromFrancebasedonlargeproportionofpatientsshowedthatBDQ-containingregi-mensachieved favourableoutcomesandprolonged treatmentwasoverallwell tolerated intheircohort[7].

DiagnosisandManagementofTuberculosis

3

1.4. Mechanism of action of BDQ

BDQ exhibits a novelmechanism of actionwhich efficiently inhibits the adenosine5-triphosphate(ATP)synthaseofmycobacteriasuchasM. tuberculosis[2].IttargetssubunitcoftheATPsynthaseofM. tuberculosis,leadingtoinhibitionoftheactivityofprotonpumpintheenzyme[8].ThestructureofATPsynthasewasdescribedintermsoftwosectors,amem-branousF0(ab2c10–15)andamembrane-extrinsicF1(α3β3γδε)[9].BindingofBDQatthelevelof theproton-bindingsite totheoligomericandproteolipicsubunitc inF0domainofATPsynthaseblocksrotationofdiscsofATPsynthase,whichculminatesintheinhibitionofATPsynthesisandeventuallytothedeathofmycobacteria[10,11].

1.5. Mechanism of resistance of BDQ

Targetbasedresistancewasfoundtooccurinresistantisolateswithmutationsinsubunitcatpositions28(Asp→Gly/Ala/Val/Pro),61(Glu→Asp),63(Ala→Pro),and66(Ile→Met)basedontheM. tuberculosisaminoacidnumberingsystem[12,13].Theassumptionwassup-portedbythefactthatthemycobacterialspeciesnaturallyresistanttoBDQ,i.e.,M. xenopi, M. novacastrense, and M. shimoidei,displayaMetatposition63insubunitcinplaceofaconservedAlainthespeciessusceptibletothedrug[12,13].

Recentelucidationofcrystal structureofATPsynthase fromM. Phlei (4VIF,4VIG)[14],enablesthemechanismofactionandresistanceofBDQinM. tuberculosis inamoreprecisemannercomparedtoearlierstudies[12,15].Inthelightoftheaboveinthisstudy,tounderstandthemolecularbasisofBDQresistance,sevenmutants(MT)proteinsofATPsyn-thasewithimportantresistantmutationsatcodonposition28suchasAsp→Gly/Ala//Pro/Val(D28G,D28A,D28PandD28V)andtheremainingthreemutantsatcodon61,63and66withAsp,Pro,MetinplaceofGlu,AlaandIle(E61D,A63P,andI66M)weremodeledanddockedwithBDQusingin silicoapproaches.

2. Materials and Methods

2.1. Proteins: Template selection and model building

Inthepresentstudy,ATPsynthasewasmodeledusingMODELLER9.14[16],becauseitscrystalstructureissofarnotdeduced.ATPsynthaseiscodedbyatpEgene;thetargetAtpEprotein (Rv1305) sequence ofM. tuberculosis obtained from theTuberculist databasewassubmittedtoBLASTp[17]programandsearchedagainstproteindatabank(PDB).TheWTproteinfromM. phlei(PDBcode-4V1F)[14]wasconsideredastemplateproteindisplayingmaximumidentitywith theWTprotein. Inaddition, theminimuminhibitoryconcentration(MIC)reportedforM.phleiisverylow(0.05mg/ml)andbasicallyidenticaltothatreportedforM.tuberculosis(0.06mg/ml)[18].InthePDBfileoftemplate4V1F,heteroatomssuch

DiagnosisandManagementofTuberculosis

4

aswaterandotherchainswereremoved;retainingchainAandcommandlineoptionswereprovidedformodelbuildinginMODELLER9v14.ThesoftwarealignstheFASTAsequenceoftemplateandtarget,followingwhichittakesupthePDBtemplatefileandgeneratesthemodelbasedontheconceptofsequencepredictstructure.SequencealignmentbetweenWTandtemplateproteinwasperformedwiththecommandlineoptions,andaseriesofcommandswereprovidedformodelbuildingusingthesoftwareMODELLER9v14.Residuesatpositions28,61,63,and66ofWTAtpEproteinweresubstitutedforgeneratingsevendifferentMTproteins,followingtheaboveprocedure.ThesamesetofprotocolwasfollowedforchainBgenerationfortheWTandsevenmutants.

2.2. Chain combination

Inthisstudy,two(AandB)chainsweremodeledseparatelyfollowingtheaboveproce-dure;chainswerethencombinedusingAmbersoftware(Version-12)[19]toenabledockingefficiently,since,ATPsynthaseisacomplexoligomericproteincomprisesof11AtpEsub-units,forproperbindingofBDQatleasttwochainsarerequired.

2.3. Model evaluation

Validationofthemodelswasdonebyramachandranplot[20].Furtherthedeviationbe-tweentheWTandthetemplate4V1FuponstructuralsuperimpositionwasdeterminedusingPDBeFOLD[21].

2.4. Ligand

The ligand (BDQ) chosen in this studywas obtained from PDB structure of 4V1F[14].

2.5. Molecular docking

TheGOLDprotocolisbasedontheprincipleofgeneticalgorithm,withtherigidrecep-torandtheflexibleligandduringtherefinementprocess,detailsofwhichhavebeendescribedelsewhere[22].DockingwasperformedbetweenBDQandsevenMTproteinsofAtpE(D28G,D28A,D28P,D28V,E61D,A63P,andI66M)incomparisontoWTwiththehelpofsoftwareGOLD(Version-4.0.1).Theinputatomfilesforboththeproteinsandtheligandwerecreated.Theligandandthemodelswereaddedwithhydrogenatomsbeforedocking.Thecavityatomfilecontainingtheatomnumberofbindingresidues(Gly58,Ala62,Phe65,Ile66andAla69)waspreparedforBDQ.ThebindingresidueswereselectedoncomparisonbetweenbindingregionsofBDQwithcrystalstructuresof4VIF.Dockingswereperformedunder‘Standarddefaultsettings’mode.-numberofislandswas5,populationsizeof100,numberofopera-tionswas100,000,anichesizeof2,andaselectionpressureof1.1.Tendockingposeswereobtainedforeachligand.PoseswithhighestGOLDscorewereusedforfurtheranalysis.Ide-

DiagnosisandManagementofTuberculosis

5

ally,thescoreshouldcorresponddirectlytothebindingaffinityoftheligandfortheprotein,sothatthebestscoringligandposearethebestbinders.

2.6. BIOVIA software

Biovia-2015wasusedwasusedforvisualizationpurposeofmodelledproteins,dock-ingdataandtodeterminetheinteractionsbetweentheligandandproteins[23].

3. Results and Discussion

BDQorTMC207isanewanti-TBdrugbelongingtotheclassofdiarylquinoline,whichselectivelyinhibitsthemycobacterialenergymetabolismi.e.ATPsynthesis.BDQisfoundtobeeffectiveagainstallstatesofM. tuberculosislikeactive,dormant,replicating,non-replicat-ing,intracellularandextracellular.AlthoughthecontributionofBDQagainstthetreatmentofdrugresistantTBseemssignificanttotheTBcontrolprogram,however,somein vitro studies [12,13]haveshowntheemergenceofresistantstrainstoBDQ.Inareport[12],in vitroresis-tantmutantsofBDQfromM. tuberculosisanddiverseatypicalmycobacteriawereisolated.Six distinct mutations,Asp28→Gly,Asp28→Ala, Leu59→Val, Glu61→Asp,Ala63→Pro,andIle66→MethavebeenidentifiedinthesubunitcformingaCringintheATPsynthase,inordertomaptheaminoacidresiduesinvolvedinthebindingofBDQ[12,13].

Thecatalyticcoreofthemembrane-embeddedrotorringofthesodiumion–translocatingATPsynthasecontainsα3β3γsubunitsarrangedinahexagonofalternatingαandβsubunitswithhelicesofγinthecenter.ATPsynthesisandhydrolysisreactionsoccuratthreecatalyticsites[9].Inthepresentstudy,ATPsynthasewasmodeledbecauseitscrystalstructureofM. tuberculosisisnotsofardeducedduetothecomplexityinvolvedincrystallizingtheprotein.Thetemplatechosenwasthecrystalstructureof4V1FfromM. Phleiat1.7Angstrom(Å)resolutionthatshowed90%identitywiththetargetM. tuberculosisprotein(Figure 1).More-over,4V1FwasincomplexedwithBDQ,whichenabledthedockingprocesseasier[14].

Inthisstudy,sevenMTmodelsofAtpEwerebuiltbasedontheWTsequenceofAtpEprotein(Rv1305)throughsubstitutionatposition28withfourmutants:Gly,Ala,Pro,ValinplaceofAsp,theremainingthreemutantsatcodon61,63and66withAsp,Pro,MetinplaceofGlu,AlaandIle,respectively.(Figure 2).Theywerevalidatedbyramachandranplot(RM)whichshowed100%andabove97%ofresiduesinthefavouredregionsforWTandsevenmutants,respectively(Figure 3 and Table 1).InadditiontoevaluationbyRMplot,theywerealsovalidatedbystructuralsuperimposition(Figure 4)whichshowedarootmeansquarede-viation(RMSD)of1.2Åbetween4V1FandWTsuggestingareliablemodel.

DiagnosisandManagementofTuberculosis

6

Figure 2:Three-dimensionalmodelsofAtpEshowingWTandmutatedresiduesatrespectivecodonpositionsinM. tuberculosis.

Figure 3:RamachandranPlotforWTandMTmodelsofAtpEfromM. tuberculosis

Figure 1:pBLASTresultsshowingmaximumidentity(90%)betweenthetemplate4V1FofATPsynthasefromM. PhleiandthetargetWTproteinsequence-Rv1305ofATPsynthaseofM. tuberculosis.

DiagnosisandManagementofTuberculosis

7

Figure 4:Validationofmodelingthroughsuperimpositionoftemplate4V1F(green)withWTproteinofAtpEfromM. tuberculosis (darkblue). ThemutantmodelssuchasD28G,D28A,D28PD28V,E61D,A63P,andI66MwerecreatedanddockedwithBDQalongwithWT,thedockedBDQandAtpEscomplexareillus-tratedinFigure 5.

Figure 5:DockingofBDQ(red)withWTandmutantsofAtpE.

ThedockingofBDQwithAtpEsresultedintenposes.Ofthetenposesproduced,thebestligandposewasselectedbasedontopGOLDscore.Amongthesevenmutants,thehighscoreof-76.04kcal/molwasobtainedfortheMT-D28P,followedbyD28A,I66M,A63P,E61D andD28V compared to theWT and theD28G-MT showed relatively low score of-22.86kcal/molcomparedtotheWT.ThebindingenergybetweenWTandBDQwasfoundtobe-32.1kcal/mol(Figure 6).Thus,thedockingresultssuggeststhatincomparisontotheWT,bindingaffinityofD28PwithBDQwasshowntobemore,followedbyothers.Incontrast,

Evaluation of resi-dues

WT D28A D28G D28P D28V E61D A63P I66M

Numberofresiduesinfavouredregion(~98.0%expected)

79(100) 77(97.5) 78(98.7) 77(97.5) 77(97.5) 78(98.7) 78(98.7) 78(98.7)

Numberofresiduesinallowedregion(~2.0%expected)

0 1 0 1 1(1.3) 0 0 0

Numberofresiduesinoutlierregion

0 1 1 1 1 1(1.3) 1(1.3) 1(1.3)

Table 1:RMplotanalysisofWTandMTmodelsofAtpEfromM. tuberculosis

DiagnosisandManagementofTuberculosis

8

Figure 6:DockingscoreofBDQwithWTandmutantsofAtpE

TheinteractionprofileofWTandMT-AtpEswithBDQatitsbindingsiteareillustratedin Figure 7.Ingeneral,theinhibitorandenzymemakeapatternofcomplementaryhydrogen(H)bondsbetween their respectivebackbone atoms. In caseofWT-AtpEcomplexedwithBDQ,singlecarbonHbondwasseenfollowedbytwoPi-pistackedinteractionswithPhe65,andotherresiduesinvanderWaalscontactdistanceasshowninFigure 7.IncaseofMT-D28G,threecarbonHbondswereobservedbetweenBDQandresiduesGlu61,Phe62andPhe65,respectively,followedbyweakPi-alkylinteractions.IncontrasttoWTinMT-D28A,severalinteractionsprincipallyofalkyl,Pi-alkyl,Pi-PistackedwerefoundandPi-sigmabe-tweenthedrugandPhe66wasalsofound.InMT-D28Pcomplexedwiththedrugmolecule,twocarbonHbondswereformedwithresiduesPhe54andGly58,followedbyalkyl,Pi-alkylPi-pistackedinteractions.Surprisingly,inMT-D28VcomplexedwithBDQnoHbondswerefound.Notably,bromineatomoftheBDQwasinvolvedinalkylinteractionwiththeresidueLeu68.Inaddition,alkyl,Pi-alkylandtwoamidePi-stackedtypesofinteractionswereob-served.Interestingly,bromineatomoftheBDQwasinvolvedinalkylinteractionwiththeresi-dueLeu68(Figure 7).IncaseofMT-E61D,acarbonHbondwithAsp142wasformedandPi-alkyl,Pi-pistackedinteractionswerefoundasshowninFigure 7.InMT-A63PcomplexedwithBDQ,anamide-PistackedandPialkylinteractionswerefoundwithGlu61andAla62residues.InMT-I66MwithBDQ,interestingly,sulphurbasedinteractionswithMet66itselfwasfound.Then,api-SigmawithLeu59andPi-PiT-shapedwithPhe146werealsofound(Figure 7).

The reason for thehigh score in allmutants (D28A,D28PD28V,E61D,A63P, andI66MexceptD28G)withBDQcouldbeattributedtothepresenceoffavourableinteractionsthatlacksinWT.ThiscouldinturnbeduetothesubstitutionofAla/Pro/ValintheMTproteinsinplaceofAspinWTatposition28.Inthesemutants(D28A,D28PD28V),Alacontainsoneextramethylgroup,ProcontainsaheterocyclicgroupandValcontainstwomethylgroupsinsteadofAspwhichisanacidicaminoacid,mighthaveinducedmorestructuralchangesintheprotein’ssidechain,whichwasobviousinchangesinthepatternofinteractions(manyalkyl, Pi-alkyl, Pi-Pi stacked andPi-sigma inD28A; two carbonHbonds, alkyl, Pi-alkyl,Pi-pi stacked interactions inD28Pandalkyl,Pi-alkyl and twoamidePi-stacked inD28V)

D28G,displayedlowbindingaffinitycomparedtotheWTprotein.

DiagnosisandManagementofTuberculosis

9

andconsequentlyreflectedinhighscore.IncaseofE61D,A63P,andI66M,Aspcontainsanacidicgroup,ProcontainsaheterocyclicgroupandMetcontains functionalsidechainofmethylgroupandsulphuratominplaceofGlu61,Ala63andIle66,mighthaveinducedmorestructuralchangesintheprotein’ssidechain,whichwasobviousinchangesinthepatternofinteraction.

Figure 7:DifferencesinnetworkofinteractionsbetweenWTandmutantsofAtpEfromM. tuberculosis with BDQ

AcarbonHbond,Pi-alkyl,Pi-pistackedinteractionsinE61Dandamide-PistackedandPialkylinteractionsinA63P;interactionssuchassulphurbasedwithMet66itselfwasfound.Api-SigmawithLeu59andPi-PiT-shapedwithPhe146werefoundinMT-I66M.ThesetypesofinteractionsinallmutantsconsequentlyreflectedinhighscorethantheWT.IncontrastincaseMT-D28G,asGlydoesnotcontainsafunctionalgrouporsidechain,resultedinlowerscore(onlyPi-alkyl)thantheWT(twoPi-pistackedinteractions).Thus,basedonthesefind-ingsitcanbeassumedthatthemutants-D28A,D28P,D28V,I66M,A63P,andE61Dcouldlead tohigh level resistancecompared toMT-D28G thatmay lead to low level resistance.Therefore,inthispilotstudy,aprimaryeffortwastakentounderstandtheeffectofbindingaffinityofWTandMTproteinsofAtpEwithBDQ,whichshowedmoreaffinitytowardsthemutantscomparedtotheWT.Therefore,itcanbesuggestedthatthemutantsdisplayedmoreaffinitywithBDQ,becauseofthesubstitutionthatinducesstructuralchangesandBDQbindsvery tightly leading to thesloworno releaseof thedrug tomediate its inhibitoryactivity,therebyleadstoBDQresistance.

DiagnosisandManagementofTuberculosis

10

4. Conclusions

Thefindingsinthisconcisereporthaveprovidedsomeusefulinsightstowardsunder-standingthebasisofin vitro-BDQresistanceinM. tuberculosis.Although,theeffectofdock-ingcouldbebetterexplainedafterperformingmoleculardynamics forunderstanding theirfunctionprecisely,yet, the informationprovidedoverherecanbeuseful tounderstand theimpactofsuchsubstitutionandconsequentchangesinbindingability.However,furtherstruc-turalstudiesareneededtogetadeeperunderstandingofthemechanismofATPresistancetoBDQthatwillaidindevelopmentofinhibitorsthatareselectiveagainstATPsynthasewhichcancircumvent theproblemofBDQresistance. In addition, tounderstandmoreabout thebindingaspectsofBDQwithothernon-targetbased resistance i.e., efflux-based resistancetoBDQwhichwasidentifiedinpairedisolatesfrompatientstreatedwithBDQ,aswellasinmice,showingcross-resistancetoclofazimine[24].Thus,structuralstudiesrelatedtoefflux-basedresistancemutantsleadingtocauseofBDQresistancearealsoneeded.

5. References

1.WorldHealthOrganization.Tuberculosis.Geneva,Switzerland:WHO;2017FactsheetNo.104.http://www.who.int/mediacentre/factsheets/fs104/en/.

2.AndriesK,VerhasseltP,GuillemontJ,GöhlmannHW,NeefsJM,WinklerH,etal.AdiaryquinolinedrugactiveonATPsynthaseofMycobacteriumtuberculosis.Science2005;307:223–237.

3.IbrahimM1,Truffot-PernotC,AndriesK,JarlierV,VezirisN.SterilizingactivityofR207910(TMC207)-containingregimensinthemurinemodeloftuberculosis.AmJRespirCritCareMed2009;180:553-7.

4.DiaconAH,PymA,GrobuschM,PatientiaR,RustomjeeR,Page-ShippL,etal.ThediarylquinolineTMC207formultidrug-resistanttuberculosis.NEnglJMed2009;360:2397–405.

5.PymAS,DiaconAH,TangSJ,ConradieF,DanilovitsM,ChuchottawornCetal.Bedaquilineinthetreatmentofmul-tidrug-andextensivelydrug-resistanttuberculosis.EurRespirJ2016;47:564-74.

6.vanHeeswijkRPG,DannemannB,HoetelmansRMW.Bedaquiline:areviewofhumanpharmacokineticsanddrug–druginteractions.JAntimicrobChemother2014;69:2310–8.

7.GuglielmettiL,JaspardM,LeDûD,LachâtreM,Marigot-OuttandyD,BernardC,etal.Long-termoutcomeandsafetyofprolongedbedaquilinetreatmentformultidrug-resistanttuberculosis.EurRespirJ2017;49:1601799.

8.KoulA,DendougaN,VergauwenK,MolenberghsB,VranckxL,Willebrordsetal.DiarylquinolinestargetsubunitcofmycobacterialATPsynthase.NatChemBiol2007;3:323–4.

9.LeslieAGW,WalkerJE.StructuralmodelofF1-ATPaseandtheimplicationsforrotarycatalysis.Biophys.2000,355.465–72.

10.MatteelliA,CarvalhoACC,DooleyKE,KritskiA.TMC207: thefirstcompoundofanewclassofpotentanti-tuberculosisdrugs.FutureMicrobiol2010;5:849–58.

11.HaagsmaAC,PodascaI,KoulA,AndriesK,GuillemontJ,LillH,etal.ProbingtheinteractionofthediarylquinolineTMC207withitstargetmycobacterialATPsynthase.PLoSOne2011;6:e23575.

DiagnosisandManagementofTuberculosis

11

12.SegalaE,SougakoffW,Nevejans-ChauffourA,JarlierV,PetrellaS.NewmutationsinthemycobacterialATPsyn-thase:newinsightsintothebindingofthediarylquinolineTMC207totheATPsynthaseC-ringstructure.AntimicrobAgentsChemother2012;56:2326–34.

13.HuitricE,VerhasseltP,KoulA,AndriesK,HoffnerS,AnderssonDI.Ratesandmechanismsofresistancedevelop-mentinMycobacteriumtuberculosistoanoveldiarylquinolineATPsynthaseinhibitor.AntimicrobAgentsChemother2010;54:1022–8.

14.PreissL,LangerJD,YildizÖ,Eckhardt-StrelauL,GuillemontJE,KoulA,etal.StructureoftheMycobacterialATPSynthaseFoRotorRinginComplexwiththeAnti-TBDrugBedaquiline.SciAdv2015;1:106.

15.DeJongeMR,KoymansLHM,GuillemontJEG,KoulA,AndriesK.AcomputationalmodeloftheinhibitionofMycobacteriumtuberculosisATPasebyanewdrugcandidateR207910.Proteins.2007;67:971e980.

16.Sali,A.MODELLER:Implementing3DProteinModeling.mc2.MolecularSimulationsInc.,1995.vol2:5.

17.AltschulSF,MaddenTL,SchäfferAA.GappedBLASTandPSI-BLAST:anewgenerationofproteindatabasesearchprograms.Nuc.acidsRes.1997;25:3389-402.

18.HuitricE,VerhasseltP,AndriesK,HoffnerSE.Invitroantimycobacterialspectrumofa

diarylquinolineATPsynthaseinhibitor.Antimicrob.AgentsChemother2007;51:4202–4.

19.PearlmanDA,CaseDA,CaldwellJW,RossWS,CheathamTE,DeBoltS,etal.AMBER,apackageofcomputerprogramsforapplyingmolecularmechanics,normalmodeanalysis,moleculardynamicsandfreeenergycalculationstosimulatethestructuralandenergeticpropertiesofmolecules,CompPhysCommun1995;91:1–41.

20.LovellSC,DavisIW,ArendallWB,deBakkerPI,WordJM,PrisantMG,etal.StructurevalidationbyCαgeometry:ϕ,ψandCβdeviation.ProStrucFuncBioinfor2003;50:437-50.

21.KrissinelE,HenrickK.Proteinstructurecomparisonin3Dbasedonsecondarystructurematching(PDBeFold)fol-lowedbyCαalignment,scoredbyanewstructuralsimilarityfunction.InProceedingsofthe5thInternationalConfer-enceonMolecularStructuralBiology,Vienna.2003;Vol.88.

22.JonesG,WillettP,GlenRC.Molecularrecognitionofreceptorsitesusingageneticalgorithmwithadescriptionofdesolvation,JMolBiol1995;245:43–53.

23.DassaultSystemesBIOVIA,DiscoveryStudioModelingEnvironment,Release4.5,SanDiego:DassaultSystèmes,2015.

24.AndriesK,VillellasC,CoeckN,ThysK,GeversT,VranckxL,etal.AcquiredresistanceofMycobacteriumtuber-culosistobedaquiline.PLOSone.2014;9:e102135.