F901318 represents a novel class of antifungal drug that inhibits dihydroorotate dehydrogenase

15 Pages·2016·2.06 MB·English

F901318 represents a novel class of antifungal drug
that inhibits dihydroorotate dehydrogenase
JasonD.Olivera,1,GrahamE. M.Sibleya,NicolaBeckmanna,Katharine S.Dobba,MartinJ.Slaterb,LauraMcEnteec,
SaskiaduPréa,JoanneLivermorec,MichaelJ.Bromleya,2,NathanP.Wiederholdd,WilliamW.Hopec,
AnthonyJ.Kennedya,DerekLawa,andMikeBircha
aF2GLtd.,ManchesterM300LX,UnitedKingdom;bCresset,LitlingtonSG80SS,UnitedKingdom;cAntimicrobialPharmacodynamicsandTherapeutics,
DepartmentofMolecularandClinicalPharmacology,UniversityofLiverpool,LiverpoolL693GE,UnitedKingdom;anddFungusTestingLaboratory,
DepartmentsofPathologyandMedicine/InfectiousDiseases,UniversityofTexasHealthScienceCenteratSanAntonio,SanAntonio,TX78229
EditedbyCarlF.Nathan,WeillMedicalCollegeofCornellUniversity,NewYork,NY,andapprovedSeptember27,2016(receivedforreviewMay24,2016)
There is an important medical need for new antifungal agents Issuesexistwithcurrenttherapies,includingoverttoxicity,drug–
withnovelmechanismsofactiontotreattheincreasingnumberof drug interactions, variable pharmacokinetics, and increasing levels
patients with life-threatening systemic fungal disease and to of drug resistance (5, 6). In particular, the development of resis-
overcomethegrowingproblemofresistancetocurrenttherapies. tance to the azole class of antifungals is concerning, because they
F901318,theleadingrepresentativeofanovelclassofdrug,the are currently the sole orally available antifungal agent for the
orotomides, is an antifungal drug in clinical development that treatmentofaspergillosis (7).Azole-resistantclinicalisolatesof
demonstratesexcellentpotencyagainstabroadrangeofdimorphic Aspergillusfumigatushavebeenobservedandisolatedfrompatients
andfilamentousfungi.InvitrosusceptibilitytestingofF901318against worldwide, including Europe, the United States, Asia, Africa,
morethan100strainsfromthefourmainpathogenicAspergillus Australia, and the Middle East (8, 9). Apparently exacerbated by
spp.revealedminimalinhibitoryconcentrationsof≤0.06μg/mL— the environmental use of azole fungicides in agriculture (10),
greaterpotencythantheleadingantifungalclasses.Aninvestiga- reportedratesofazoleresistancehaveapproached30%atcertain
tionintothemechanismofactionofF901318foundthatitactsvia sites in Europe, with rates outside Europe varying between 0.6%
inhibition of the pyrimidinebiosynthesis enzymedihydroorotate and11.2%(9).
dehydrogenase (DHODH) in a fungal-specific manner. Homology
modeling of Aspergillus fumigatus DHODH has identified a pre- Results
dicted binding mode of the inhibitor and important interacting DiscoveryofF901318.With the aim of identifying new antifungal
aminoacidresidues.Inamurinepulmonarymodelofaspergillosis, chemistries, a library of 340,292 small molecules was screened
F901318displaysinvivoefficacyagainstastrainofA.fumigatus in vitro against A. fumigatus, and multiple chemical series with
sensitivetotheazoleclassofantifungalsandastraindisplaying antifungal activity were identified. The initial hits in one such
anazole-resistantphenotype.F901318iscurrentlyinlatePhase1 series, originally termed the “F3 series,” were developed by a
clinical trials, offering hope that the antifungal armamentarium
medicinal chemistry program that was driven by classical struc-
can be expanded to include a class of agent with a mechanism ture–activity relationships based on in vitro activity. This series
ofactiondistinctfromcurrentlymarketedantifungals.
wascharacterizedbyexcellentinvitropotencyagainstAspergillus
| | | spp.,butwasdevoidofactivityagainstCandidaspp.Thisunusual
antifungaldrug Aspergillusfumigatus mechanismofaction
dihydroorotatedehydrogenase
Significance
Arecentestimateputstheannualdeathtollfromseriousfungal
New antifungal drugs that act via novel mechanisms are ur-
infectionsat1.5million(1).Asoneofthefourgreatestkillers,
gentlyneededtocombatthehighmortalityofinvasivefungal
Aspergillusspeciesareopportunistichumanpathogens,particularly
diseaseandtheemergenceofresistancetoexistingtherapies.
affecting immunocompromised persons, such as transplant recipi-
Wedescribethediscovery,structure,activity,andmechanism
entsandthosewithhematologicmalignancies.Invasiveaspergillosis
hasahighmortality(30–90%)andisestimatedtoaffectmorethan ofactionofF901318,anewantifungalagent.Amemberofa
novel class of antifungals, the orotomides, F901318 acts via
200,000peopleannually.OtherdiseasescausedbyAspergillusspe-
inhibitionofdihydroorotatedehydrogenase,anenzymeofde
cies,includingallergicbronchopulmonaryaspergillosis(2)and
novo pyrimidine biosynthesis. F901318 is currently in clinical
chronicpulmonaryaspergillosis(3),haveasignificantglobalimpact,
developmentforthetreatmentofinvasiveaspergillosis.
affectingmillionsofpatients.
Therehasbeenadearthofnewdrugclassesdevelopedtotreat
Authorcontributions:J.D.O.,G.E.M.S.,M.J.S.,M.J.B.,N.P.W.,W.W.H.,A.J.K.,D.L.,andM.B.
systemicfungalinfectionsarrivingintheclinic,withthemostrecent designedresearch;J.D.O.,G.E.M.S.,N.B.,K.S.D.,M.J.S.,L.M.,S.d.P.,J.L.,N.P.W.,andD.L.
being the echinocandins, introduced in 2001. Only three other performedresearch;M.J.B.contributednewreagents/analytictools;J.D.O.,G.E.M.S.,N.B.,
classesofantifungaldrugsarecurrentlyavailableforthetreatment M.J.S.,J.L.,N.P.W.,W.W.H.,A.J.K.,D.L.,andM.B.analyzeddata;andJ.D.O.,G.E.M.S.,M.J.S., GY
ofinvasivefungaldisease:polyenes(amphotericinB),azoles(e.g., N.P.W.,W.W.H.,A.J.K.,D.L.,andM.B.wrotethepaper. OLO
voriconazole, posaconazole, and the recently licensed isavucona- Conflictofintereststatement:J.D.O.,G.E.M.S.,N.B.,S.d.P.,A.J.K.,D.L.,andM.B.areem- OBI
ployeesofF2GLtd.K.S.D.andM.J.B.areex-employeesofF2GLtd.J.D.O.,G.E.M.S.,K.S.D., R
zole),andflucytosine(4).Theseagentsworkviaalimitedrangeof A.J.K.,D.L.,andM.B.areshareholdersofF2GLtd.W.W.H.hasreceivedresearchfundingfrom MIC
cellulartargets.Echinocandins,suchascaspofungin,inhibitβ-(1,3)- Pfizer,Gilead,Astellas,AiCuris,andF2G,andhasactedasaconsultantand/orgiventalksfor
glucansynthase,exploitingthemoststrikingdifferencebetweenthe Pfizer,Basilea,Astellas,F2G,NordicPharma,MaynePharma,andPulmocide.G.E.M.S.,D.L.,J.D.O.,
fungalcellanditshumancounterpart—thecellwall.Twoantifungal andM.B.arelistedasinventorsonpatentWO2009130481.J.D.O.,M.J.B.,G.E.M.S.,and
M.B.arelistedasinventorsonpatentWO200913379.
drug classes target the cell membrane: azoles inhibit ergosterol
ThisarticleisaPNASDirectSubmission.
biosynthesis,andpolyenesdisruptfungalmembranesviaergosterol
1Towhomcorrespondenceshouldbeaddressed.Email:[email protected].
binding.Flucytosineisapyrimidineanalog,convertedto5-fluoro-
2Presentaddress:ManchesterFungalInfectionGroup,InstituteofInflammationandRe-
uracil within fungal cells, that disrupts DNA and RNA synthesis;
pair,UniversityofManchester,ManchesterM139NT,UnitedKingdom.
however, owing to rapid development of resistance, it is used pri-
Thisarticlecontainssupportinginformationonlineatwww.pnas.org/lookup/suppl/doi:10.
marilyincombinationtherapy. 1073/pnas.1608304113/-/DCSupplemental.
www.pnas.org/cgi/doi/10.1073/pnas.1608304113 PNAS | November8,2016 | vol.113 | no.45 | 12809–12814
F resistantcloneswereobtained,pAMA1DNAwasisolated,andthe
N genomicDNAinsertwassequenced.Allresistantclonescontained
insertsthatmappedtothesameregionofchromosomeI(Fig.S1A).
Although sequence data from five genes was retrieved, only one
gene—ANIA_05909—was intact in all four genomic fragments.
N N
Thisgene, namedpyrEin Aspergillus spp., encodesthe pyrimidine
biosynthesis enzyme dihydroorotate dehydrogenase (DHODH;
N
EC1.3.5.2).
ToconfirmthatextracopiesofpyrEledtoF901318resistance,
therecoveredplasmidpAMA1_18.1wastreatedwithabacterial
transposon (Tn5) to disrupt either pyrE or a neighboring gene,
HN
ANIA_05910, and the resulting plasmids were transformed
into A. nidulans. Strains carrying the intact pAMA1_18.1 or the
O ANIA_05910 disruptant exhibited resistance to F901318; how-
O ever,ondisruptionofpyrE,thestrainreturnedtowild-type(WT)
levels of susceptibility to F901318 (Fig. S1B). This finding con-
firmedthatextracopiesofthegeneencodingDHODHwerere-
N
sponsible for the resistance to F901318, implicating DHODH as
thetargetofthedrug.
DHODHIstheTargetofF901318.DHODHisanoxidoreductasethat
catalyzes the fourth step of the pyrimidine biosynthesis pathway
Fig.1. StructureofF901318.
(Fig. S2), the conversion of dihydroorotate to orotate. Confirma-
tion that the drug disrupts pyrimidine biosynthesis was obtained
patternperhapsexplainswhysimilarchemicalshavenotbeenfound aftertheadditionofexogenouspyrimidines(uridineanduracil)to
previously.Typically,antifungalscreenshavedependedonfirstfinding themediaduringsusceptibilitytesting.Areversaloftheantifungal
activity against Candida. Modifications to improve physicochemical effect of F901318 on A. fumigatus was observed, but only at mil-
properties, antifungal potency, pharmacokinetics, ADMET (absorp- limolarconcentrationsofpyrimidines(5mMandabove;Fig.S3).
tion,distribution,metabolism,excretion,andtoxicity)properties, Interestingly,humanserumcontainslowlevelsofpyrimidines,es-
invivoefficacyininfectionmodels,andtoxicologyhaveledtothe timated to be ∼15 μM (13), which are insufficient to reverse
developmentofF901318(Fig.1).Antifungalsusceptibilitytestingof theeffectofF901318onA.fumigatusinvivo.Indeed,mutantsof
F901318usingstandardizedtechniquesdemonstrateditspotentac- A. fumigatus (14), Candida albicans (15), Histoplasma capsulatum
tivities against clinicalisolatesofaspergilli,with minimuminhib- (16),andCryptococcusneoformans(17),whicharedisruptedin
itoryconcentrations(MICs)of<0.1μg/mLagainstmultiplestrainsof pyrimidinebiosynthesis,haveexhibitedattenuatedvirulencein
A.fumigatus,Aspergillusterreus,Aspergillusniger,andAspergillusfla- animal models of infection, indicating that targeting pyrimidine
vus,includingisolatesresistanttootherantifungalagents(Table1). synthesisisavalidantifungalstrategy.
Biochemical evidence confirming that the target of F901318
Mechanism of Action Screen. Initially, owing to the method of dis- was DHODH was obtained from in vitro enzyme assays that
covery,themechanismofactionofthisserieswasunknown.A were set up with recombinant A. fumigatus DHODH using 2,6-
combination of microbiological, genetic, and biochemical ap- dichloroindophenol as a redox indicator. F901318 inhibited
proacheswereusedtoidentifythetargetofthisdrugseries.A A.fumigatusDHODHinadose-dependentmanner,withanIC of
50
genetic screen, similar to a multicopy suppressor screen, was per- 44±10nM(n=11;Fig.2).DHODHisalsopresentinmammals,
formed to identify genes that, when present in multiple copies, althoughthereisalowoverallidentitytoAspergillusDHODH(∼30%;
providedresistancetoF901318.Thisapproachwaspreviouslyval- Fig. S4). A known inhibitor of human DHODH, teriflunomide
idated for the antifungal drugs itraconazole and terbinafine, by (18),which isusedto treat multiple sclerosisinman, did not in-
demonstratingthatthepresenceofadditionalcopiesofcytochrome hibitA.fumigatusDHODHinvitro.SpeciesselectivityofF901318
P450 C-14 lanosterol α-demethylase and squalene epoxidase, re- was confirmed in an assay in which little inhibition of human
spectively,ledtoresistancetothoseagents(11,12).Inthepresent DHODH was observed, whereas teriflunomide inhibited human
study,Aspergillusnidulanssporesthathadbeentransformedwithan DHODH,asexpected.Infact,theIC valueforF901318against
50
A. nidulans genomic library carried by the autonomously replicat- human DHODH was not reached at 100 μM, the highest con-
ingplasmidpAMA1wereexposedtoF901318.Fourindependent centrationusedintheseexperiments,indicatingthatF901318was
Table1. AntifungalpotencyofF901318andotherantifungaldrugsagainstthemajorAspergillusspecies
Drug Parameter A.fumigatus(n=55) A.terreus(n=21) A.flavus(n=19) A.niger(n=19)
F901318 MICmean,μg/mL 0.029 0.014 0.021 0.031
MICrange,μg/mL 0.008–0.06 0.004–0.03 0.015–0.06 0.015–0.06
AmphotericinB MICmean,μg/mL 1.55 2.07 1.39 0.46
MICrange,μg/mL 0.5–2 1–4 1–2 0.25–1
Caspofungin MEC*mean,μg/mL 0.096 0.112 0.06 0.062
MEC*range,μg/mL 0.06–0.12 0.06–0.12 0.06 0.06–0.12
Voriconazole MICmean,μg/mL 0.69 0.59 0.96 0.77
MICrange,μg/mL 0.25–16 0.25–1 0.5–1 0.5–16
TheMICs(inμg/mL)ofF901318,amphotericinB,andvoriconazoleweredeterminedfortheAspergillusspp.indicated(n,numberof
differentstrainstested).*Forcaspofungin,theMECisshown,becausegrowthisnotcompletelyinhibitedwiththisdrug.Dataare
displayedasthegeometricmeanoftheMICsandtherangeofMICfromlowesttohighestforthestrainsofaparticularspecies.
12810 | www.pnas.org/cgi/doi/10.1073/pnas.1608304113 Oliveretal.
A A.fumigatus B Human
100 100
F901318 teriflunomide
90 90
80 80
70 70
n
bitio 60 60
hi 50 50
n
i
% 40 40
F901318
30 30
20 20
teriflunomide
10 10
0 0
0.01 1 100 0.01 1 100
[drug](µM) [drug](µM)
Fig.2. F901318inhibitsA.fumigatusDHODHinvitro.RecombinantA.fumigatusDHODH(A)andhumanDHODH(B)wereincubatedinthepresenceand
absenceofvaryingconcentrationsofF901318andteriflunomide.Theactivityoftheenzymeswasmeasuredforeachdrugconcentration,andthepercent
inhibitioncomparedwithnodrugcontrolswascalculated.
>2,200-foldmorepotentagainsttheA.fumigatusenzyme.Thus, preventing reoxidation of the dihydroflavin mononucleotide
fungal DHODH was confirmed as the target of F901318, and (FMNH)cofactoressentialforthereactiontoproceed(Fig.S2).
2
despite the presence of a mammalian version of the enzyme,
no target-based toxicity was predicted. On elucidation of the Spectrum,Resistance,andinVivoEfficacyofF901318.Although we
mechanism of action, the F3 series was renamed the orotomides, observednoactivityagainstCandidaspp.andthezygomycetesin
combining the mechanism (dihydroorotate) with the chemistry (α- antifungal susceptibility testing, F901318 displayed excellent
ketoamide). potency against a broad range of pathogenic filamentous and
Further enzyme kinetic experiments revealed that F901318 is a dimorphicfungi,includingPenicilliumspp.,Coccidiodesimmitis,
reversible inhibitor of A. fumigatus DHODH (Fig. S5A) and is a H. capsulatum, Blastomyces dermatitidis, Fusarium spp., and the
competitiveinhibitorwithrespecttotheubiquinone(coenzymeQ) difficult-to-treatScedosporiumspp.Thisspectrumappearstobe
cofactor, which functions as an electron acceptor in the reaction sequence-driven, given that the sensitive organisms were found
(Fig. S5B). This latter property is perhaps not unexpected, given to be grouped together in a phylogenetic analysis of DHODH
thatstructuralstudieshaveshownthatknowninhibitorsofhuman (Fig.S7).Thisfindingisconsistent withobservationsthat other
DHODH(teriflunomideandbrequinar)(19)andthePlasmodium DHODH inhibitors exhibit specificity of action owing to interspe-
falciparumenzyme(DSM265)(20)bindinaregionoftheprotein cies variations in the architecture of the hydrophobic channel in
predictedtobeachannelthroughwhichtheubiquinoneentersthe whichtheinhibitorsarepredictedtobind(19–21).Thus,theunique
moleculefromtheinnermitochondrialmembrane. spectrum of F901318 is a reflection of the structure of the target,
DHODH,andthebindingmodeoftheorotomides.
StructuralInsightsofF901318BindingtoA.fumigatusDHODH.Inthe We investigated the resistance to F901318 in A. fumigatus by
absence of a crystal structure, we investigated the binding of repeatedexposuretoaconcentrationgradientofthedrugonan
F901318toA.fumigatusDHODHbycreatingahomologymodel agar plate and selection from the margins of growth. This was
ofA.fumigatusDHODH(Fig.S6)usingthestructuralinformation carriedoutfor50passages,withnochangeintheMICobserved
providedbyotherclass2DHODHenzymes,includingthestructure for the first 40 passages and only a modest increase thereafter
ofhumanDHODH(19).F901318andothermembersoftheseries (Fig.S8).Incontrast,voriconazoleexhibitedanincreaseintheMIC
were used to identify a likely binding mode. Key residues for between10and15passages.Basedontheseresults,itappearsthat
bindingwereidentified(Fig.3A).Validationoftheimportanceof F901318resistanceisnoteasilyinducedinA.fumigatus.
two of these residues was obtained by mutagenesis of C. albicans Pharmacokinetic studies in mice have identified good distri-
Y
DHODH. The WT C. albicans DHODH was not inhibited by butionofF901318totissuesincludingthekidney,liver,andlung, OG
L
F901318, but mutation of two residues, Phe and Val , to the withdetectioninthebrain,albeitatlowerlevels,suggestingthat O
162 171 BI
residuespredictedtooccupythesamepositionsintheA.fumigatus the drugreaches keysites of infection. The efficacy of F901318 O
R
enzyme,Val200andMet209,respectively,createdamutantC.albi- was demonstrated in a neutropenic murine model of invasive MIC
cansDHODHthatwasinhibitedbyF901318(Fig.3B).TheIC of pulmonaryaspergillosis.Afterinfectionwithawell-characterized
50
the mutant C. albicans_V _M was still ∼40-fold higher than A.fumigatusstrain(NIH4215),survivalwassignificantlyimproved
162 171
theIC ofF901318againsttheA.fumigatusenzyme,indicatingthe by F901318 treatment (Fig. 4A). Treatment with the triazole
50
existence of other important differences in DHODH between the drug posaconazole also increased survival after infection with
twospecies;however,bothoftheseresiduesareclearlyimportant this strain.
for inhibition. These data, along with the observed competition Mutations in the gene encoding the target molecule of the
betweenF901318andcoenzymeQintheinvitroassay,suggestthat azoleclassofantifungaldrugs,Cyp51A,thatcauseresistancehave
the orotomides bind in the “quinone channel,” where ubiquinone beenidentified.Severalazole-resistantstrainsofA.fumigatuscarry
enters the enzyme from the inner mitochondrial membrane, a combination of a tandem repeat in the promoter and a point
Oliveretal. PNAS | November8,2016 | vol.113 | no.45 | 12811
A
Met209
Arg202
Val200
Tyr213
Tyr510
His116
B F901318
IC (µM)
50
PLPQEGNPRPRVFRLPSQKAMINRYGLNSLG
A.fumigatus 189 0.044
PEPQPGNPQPRFFRLPKDDAVINRYGFNSSG
C. albicans 151 >90
C. albicans PEPQPGNPQPRVFRLPKDDAMINRYGFNSSG
151 1.85
V M
162- 171
Fig.3. (A)BindingofF901318toA.fumigatusDHODH.AhomologymodelofA.fumigatusDHODHwascreated,andthebindingmodeofF901318(cyan)
wasestimated.Theproductorotate(orange)andthecofactorflavinmononucleotide(FMN;magenta)arealsoshown.Residuespredictedtobeclosetothe
moleculearehighlighted.(B)F901318inhibitsamutantversion,butnottheWTversion,ofC.albicansDHODH.RecombinantC.albicansDHODHresidues
Phe162andVal171weremutatedtoValandMet,respectively(theirpredictedequivalentsinA.fumigatusDHODH).TheIC50ofF901318inhibitionoftheWT
andmutantDHODHproteinsisdisplayedintheright-handcolumn(A.fumigatusIC50,n=11,SD=0.01μM;WTC.albicansIC50,n=7,allsevenreplicateshad
anIC50>90μM;C.albicans_V162_M171IC50,n=7,SD=0.91μM).
mutation in the coding sequence. One such strain, A. fumigatus early stage projects into clinical candidates has proven elusive.
F16216,carryingtheTR34/L98HmutationofCyp51A,hasproven This difficulty has mirrored the issues with target-based screen-
toberesistanttomultipleazoledrugs,includingitraconazole,vor- ingencounteredintheantibacterialarena(24).Infact,areview
iconazole, and posaconazole (22). In vitro, A. fumigatus F16216 of new mechanism, first-in-class medicines approved by the US
displayed no resistance to F901318, with an MIC of 0.03 μg/mL, FoodandDrugAdministrationbetween1999and2008revealed
whichiscomparabletothedatapresentedinTable1.Invivo,inthe that target-based screens were responsible for the discovery of
pulmonary aspergillosis model, A. fumigatus F16216 caused an in- only 3 of 10 drugs for infectious disease, with the majority dis-
fection that cannot be treated with posaconazole (Fig. 4B); how- covered by phenotypic screening (i.e., “whole-cell screens” for
ever,F901318therapyledtoasignificantincreaseinsurvivalinthis antibiotics/antifungals)(25).Theorotomideswerediscoveredvia
severemodel,demonstratingthatthedifferentmechanismofaction awhole-cellscreeningapproach,providinghitsthatwereknown
of the orotomides enables F901318 to overcome azole resistance tohaveantifungalactivityfromthestart,butwithnoknowledge
causedbyCyp51Amutations. of mechanism of action. This classical approach was coupled
PreclinicalsafetypharmacologyandtoxicologystudiesofF901318 withageneticscreentoidentifythetargetofthedrug,DHODH.
havesupportedtheprogressionandevaluationofthisantifungalin A recent reviewof antifungal drugdiscovery suggested that sim-
Phase1oralandi.v.singleandrepeateddosetrials. ilar approaches, taking advantage of genetic tools such as hap-
loinsufficiency strain collections and new technologies such as
Discussion next-generation sequencing, may accelerate the translation of
As highlighted by Denning and Bromley (23), the antifungal antifungal chemistries toward the clinic (26).
pipeline has failed to produce new antifungal drugs with mech- Pyrimidinesareessentialtothecell,notjustforthesynthesisof
anisms of action different from those of existing classes in the DNAandRNA,butalsofortheformationofprecursorsforlipid
years since caspofungin was licensed in 2001. Many potential and carbohydrate metabolism. For example, synthesis of the cell
antifungal targets have been investigated, but translating these wall requires UDP-activated sugars at multiple stages, including
12812 | www.pnas.org/cgi/doi/10.1073/pnas.1608304113 Oliveretal.
A NIH4215 IdentifyingDHODHasthetargetoftheorotomideshashelped
explainthespectrumofantifungalactivitythatweobserved.F901318
100
hasactivityagainstnumerouspathogenicfilamentousanddimorphic
90 F901318 fungi,includingAspergillusspp.,H.capsulatum,Blastomycesderma-
80 titides,andC.immitis,alongwiththedifficult-to-treatScedosporium
al 70 prolificans.TheseF901318-susceptibleorganismsgrouptogetheron
v posaconazole
urvi 60 tChaenpdhidyalosgpepn.e,tCic.trneeeofoofrmDaHnOs,DanHd(tFhieg.hSu7m),awnhearnedasPDlaHsmOoDdiHumfroenm-
s 50 control
nt zymes are more distantly related while still classified as class 2
e 40 DHODHenzymes.DHODHfromthezygomycota,suchasRhizo-
c
er 30 pusandMucor,alignmorecloselywithclass1ADHODHs,cytosolic
P
20 enzymesthatoccurinGram-positivebacteriaandthetrypanosma-
tidsthatusealternativecofactors,suchasfumarate.
10
DHODHhasbeensuggestedasatargetfortherapyinmultiple
0
diverse disease areas, including oncology, rheumatoid arthritis,
0 1 2 3 4 5 6 7 8 9 10
multiplesclerosis,andinfectiousdiseasescausedbysuchagentsas
Days post infection
Plasmodium,bacteria,andviruses(21,29).Therearecurrentlytwo
B F16216 marketedagentsthathaveactivityagainsthumanDHODH:leflu-
nomide for rheumatoid arthritis and teriflunomide for multiple
100 sclerosis. DSM265 is an antimalarial drug targeting plasmodial
90 DHODHthatiscurrentlyinPhase2clinicaltrials(20);however,to
F901318
80 our knowledge, no other human antifungal therapies have pro-
val 70 gressed with DHODHas a target.
vi Although at first consideration, the breadth of therapeutic
ur 60 control areasforwhichDHODHhasbeenproposedasadrugtargetis
s 50
nt surprising,ineachcaselimitingthepoolofpyrimidinesprevents
ce 40 proliferationofapopulationofcells.Insomecases,thehostcells
er 30 are targeted, such as lymphocytes in autoimmune diseases and
P
20 posaconazole proliferating cancerous cells in oncology. Alternatively, the
10 DHODH of invading pathogens is targeted to selectively limit
the pyrimidine pools of the infective agent. Between these two ef-
0
fects,antiviralactionhasbeenreportedforhumanDHODHinhib-
0 1 2 3 4 5 6 7 8 9 10
itors,giventhatvirusesrequirehostpyrimidinesforreplication(29).
Days post infection
Inconclusion,tocombattheincreasingproblemofresistance
to existing antifungal therapies, the discovery of new cellular
Fig.4. InvivoefficacyofF901318inamurinemodelofA.fumigatusin-
fection. Groups of 10 immunosuppressed mice were infected intranasally targetsforantifungals,alongwithviablechemistryagainstthese
withA.fumigatusNIH4215(A)ortheazole-resistantA.fumigatusF16216 new targets (23), is vitally important. F901318 is an antifungal
(B) conidia on day 0.TreatmentwithF901318 (15 mg/kg i.v., three times drug,currentlyinbothintravenousandoralPhase1clinicaltrials
daily;bluecircles),posaconazole(7.5mg/kgorally,oncedaily;redtriangles), (ClinicalTrials.govidentifiers:NCT02142153,NCT02342574,
orcontrol(greensquares)wasinitiatedat6hpostinfection.Kaplan–Meier NCT02394483, and NCT02737371), that acts via inhibition of the
curvesofsurvivingmiceineachgroupwereplotted.Afterinfectionwith pyrimidine biosynthesis enzyme DHODH, validating an alternative
NIH 4215, F901318 treatment significantly improved survival compared
targetforantifungaldrugdiscovery.
with controls (P < 0.001, Mantel–Haenszel test). After infection with
F16216,F901318treatmentsignificantlyimprovedsurvivalcomparedwith
MaterialsandMethods
controls (P < 0.001) and compared with posaconazole treatment (P <
0.005). Primers.ThesequencesoftheprimersusedinthisworkarelistedinTableS1.
PrimersweresuppliedbyEurofinsMWG.
UDP-glucose for β-(1,3)-glucan synthesis. Pyrimidines are synthe- SynthesisofF901318.The2-(1,5-dimethyl-3-phenyl-1H-pyrro-2-yl)-N-(4-[4-(5-
sized in the de novo pyrimidine biosynthesis pathway (Fig. S2), in fluoro-pyrimidin-2-yl-piperazin-1yl]-phenyl)-2-oxo-acetamide (F901318) was
preparedasdescribedinSIMaterialsandMethods.
whichDHODHisakeyenzyme,buttheyalsocanbescavengedby
fungi from the environment via the salvage pathway. The pyrimi-
InVitroAntifungalSusceptibilityTesting.MICsofantifungaldrugswerede-
dine salvage pathway appears to be inefficient for A. fumigatus,
termined according to Clinical and Laboratory Standards Institute (CLSI)
however (Fig. S3). In animal models of infection, pyrimidine bio- protocolM38-A2inRPMI1640mediumbufferedtopH7.0withMOPSbuffer
synthesismutantsfromseveralpathogenicfungiarehighlyattenu- at35°C.Forcaspofungin,theminimumeffectiveconcentration(MEC)was
atedforvirulence,includingstudiesonA.fumigatus(14),C.albicans defined as the lowest drug concentration causing abnormal growth (i.e., Y
G
(15),H.capsulatum(16),andC.neoformans(17).InSaccharomyces short-branchinghyphae). LO
O
cerevisiae,aura3deletionstrainlackingtheorotidine-5′-phosphate BI
(OMP) decarboxylase enzyme of pyrimidine biosynthesis was un- MpReGch3a-AniMsmA1o-fNAocttIiovencStocrrewena.sAonbtAa.inneidduflraonmsgtheenoFmunicglaiblrGaerynectaicrrSietdocoknCtehne- MICRO
abletosurviveinvivo,exhibitingadecreaseinthecompetitiveindex
ter.ProtoplastsfrompyrG-strainsofA.nidulans(A767)weretransformed
versusaWTorreconstitutedstrainobservedafterjust4h(27).In withthegenomiclibrary byPEG-mediated transformation.Transformants
C.albicans,URA3hasbeencommonlyusedasaselectablemarker, wereexposedtolethalconcentrationsofF901318onVogel’sminimalagar
but concerns have been raised that in some virulence studies, the (30).PlasmidDNAwasextractedfromresistantcoloniesandsequenced.
ectopic expression of URA3, leading to reduced OMP decarbox-
DHODH Assays. These assays were carried out using recombinant DHODH
ylaseactivity,hadagreatereffectonvirulencecomparedwiththe
prepared from A. fumigatus cDNA, C. albicans gDNA, or, for the human
disruptionofthetargetgeneofinterest(28).Thus,theevidencein
protein, IMAGE clone 6064723 (Geneservice Ltd.) cloned into the vector
theliteraturesupportsthetargetingofpyrimidinebiosynthesisasa pET44(Novagen)minustheN-terminal88(A.fumigatus),56(C.albicans),or
validantifungalstrategy. 28(human)aminoacids.ForC.albicans,CTG-encodedserinesweremutated
Oliveretal. PNAS | November8,2016 | vol.113 | no.45 | 12813
toTCG.FurthermutationsalteredPhe162andVal171tobecomeValandMet a concentration gradient. After 4 d of incubation at 35 °C, a zone of in-
to create the mutant protein C. albicans_V162_M171. Primers are listed in hibition was observed, and conidia were collected from the margins of
TableS1,andfurthermethodologicaldetailsareprovidedinSIMaterialsand growthandthenusedtocreatethenextplate.Aftereveryfifthpassage,the
Methods.Theassaywasconductedasdescribedelsewhere(31). MICwasdeterminedasdescribedabove.
HomologyModelingofA.fumigatusDHODH.HumanDHODH[ProteinData InVivoEfficacyTesting.AllexperimentswereconductedunderU.K.Home
Bank(PDB)IDcode1D3G]servedastheproteintemplateforconstructionof Officeprojectlicense40/3630andapprovedbytheUniversityofLiverpool’s
theA.fumigatusDHODHmodel.OtherDHODHstructuresfromhuman(PDB
Animal Welfare Committee. Groups of 10 CD-1 mice were immunosup-
ID codes 1PRH, 2PRL, 2PRM, 3G0X, 3KVM, 2WV8, and 2FQI), rat (PDB ID
pressedwith200mg/kgcyclophosphamidei.p.at4dbeforeinfectionand
codes1UUMand1UUO),Trypanosomacruzi(PDBIDcode2E68),P.falciparum
withcyclophosphamideand250mg/kgcortisoneacetates.c.at1dbefore
(PDBIDcodes3I68,1ITV,and3O8A)Leishmaniamajor(PDBIDcode3MJY),
infection.A.fumigatusF16216carriesanL98Hmutationofcyp51Aanda
and Escherichia coli (PDB ID code 1F76) also informed the process. Coarse
refinementofthestructurewithDiscoveryStudio4.1(Accelerys)wasfollowed 34-bptandemrepeatinthecyp51Apromoter,leadingtoresistancetoazole
by fine refinement with XEDraw (Cresset). More details of the homol- drugs(22).ConidiafromthisstrainandfromtheWTA.fumigatusNIH4215
ogy modeling process and ligand binding are provided in SI Materials wereadministeredintranasallyonday0.TreatmentwithF901318(15mg/kg
andMethods. i.v.threetimesdaily)orposaconazole(7.5mg/kg/dayorally)wasinitiatedat
6hpostinfection.
ResistanceTesting.A.fumigatus210conidiawereinoculatedontoSabour-
audagar(Oxoid)ina9-cmPetridish.An8-mm-diametercircleofagarwas ACKNOWLEDGMENTS. We thank John Rex, Richard White, and Markus
removedfromthecenteroftheplatetocreateawell.100μLof500μg/mL Heepfortheircriticalreadingofthemanuscript.S.d.P.issupportedbythe
drugwasloadedintothewellandallowedtodiffuseintotheagar,creating EuropeanUnionGrantPITN-GA-2013-607963.
1. BrownGD, et al. (2012) Hiddenkillers: Human fungalinfections. Sci Transl Med 18.Bar-OrA,PachnerA,Menguy-VacheronF,KaplanJ,WiendlH(2014)Teriflunomide
4(165):165rv13. anditsmechanismofactioninmultiplesclerosis.Drugs74(6):659–674.
2. DenningDW,PleuvryA,ColeDC(2013)Globalburdenofallergicbronchopulmonary 19.LiuS,NeidhardtEA,GrossmanTH,OcainT,ClardyJ(2000)Structuresofhumandi-
aspergillosiswithasthmaanditscomplication,chronicpulmonaryaspergillosis,in hydroorotatedehydrogenaseincomplexwithantiproliferativeagents.Structure8(1):
adults.MedMycol51(4):361–370. 25–33.
3. DenningDW,PleuvryA,ColeDC(2011)Globalburdenofchronicpulmonaryasper- 20.PhillipsMA,etal.(2015)Along-durationdihydroorotatedehydrogenaseinhibitor
gillosis as a sequel to pulmonary tuberculosis. Bull World Health Organ 89(12): (DSM265)forpreventionandtreatmentofmalaria.SciTranslMed7(296):296ra111.
864–872. 21.Vyas VK, Ghate M (2011) Recent developments in the medicinal chemistry and
4. Pound MW, Townsend ML, Dimondi V, Wilson D, Drew RH (2011) Overview of therapeuticpotentialofdihydroorotatedehydrogenase(DHODH)inhibitors.Mini
treatmentoptionsforinvasivefungalinfections.MedMycol49(6):561–580. RevMedChem11(12):1039–1055.
5. DenningDW,HopeWW(2010)Therapyforfungaldiseases:Opportunitiesandpri- 22.HowardSJ,PasqualottoAC,DenningDW(2010)Azoleresistanceinallergicbron-
orities.TrendsMicrobiol18(5):195–204. chopulmonaryaspergillosisandAspergillusbronchitis.ClinMicrobiol Infect16(6):
6. NettJE,AndesDR(2016)Antifungalagents:Spectrumofactivity,pharmacology,and 683–688.
clinicalindications.InfectDisClinNorthAm30(1):51–83. 23.DenningDW,BromleyMJ(2015)Infectiousdisease:Howtobolstertheantifungal
7. VerweijPE,ChowdharyA,MelchersWJ,MeisJF(2016)AzoleresistanceinAspergillus pipeline.Science347(6229):1414–1416.
fumigatus:Canweretaintheclinicaluseofmold-activeantifungalazoles?ClinInfect 24.PayneDJ,GwynnMN,HolmesDJ,PomplianoDL(2007)Drugsforbadbugs:Con-
Dis62(3):362–368. frontingthechallengesofantibacterialdiscovery.NatRevDrugDiscov6(1):29–40.
8. Rivero-MenendezO,Alastruey-IzquierdoA,MelladoE,Cuenca-EstrellaM(2016)Tri- 25.SwinneyDC,AnthonyJ(2011)Howwerenewmedicinesdiscovered?NatRevDrug
azoleresistanceinAspergillusspp.:Aworldwideproblem?JFungi2(3):21. Discov10(7):507–519.
9. GonçalvesSS,SouzaAC,ChowdharyA,MeisJF,ColomboAL(2016)Epidemiologyand
26.RoemerT,KrysanDJ(2014)Antifungaldrugdevelopment:Challenges,unmetclinical
molecularmechanismsofantifungalresistanceinCandidaandAspergillus.Mycoses
needs,andnewapproaches.ColdSpringHarbPerspectMed4(5):a019703.
59(4):198–219.
27.GoldsteinAL,McCuskerJH(2001)DevelopmentofSaccharomycescerevisiaeasa
10.ChowdharyA,KathuriaS,XuJ,MeisJF(2013)Emergenceofazole-resistantAsper-
modelpathogen:Asystemforthegeneticidentificationofgeneproductsrequired
gillusfumigatusstrainsduetoagriculturalazoleusecreatesanincreasingthreatto forsurvivalinthemammalianhostenvironment.Genetics159(2):499–513.
humanhealth.PLoSPathog9(10):e1003633.
28.BrandA,MacCallumDM,BrownAJ,GowNA,OddsFC(2004)Ectopicexpressionof
11.OsherovN,KontoyiannisDP,RomansA,MayGS(2001)Resistancetoitraconazolein
URA3caninfluencethevirulencephenotypesandproteomeofCandidaalbicansbut
AspergillusnidulansandAspergillusfumigatusisconferredbyextracopiesofthe
A. nidulans P-450 14α-demethylase gene, pdmA. J Antimicrob Chemother 48(1): canbeovercomebytargetedreintegrationofURA3attheRPS10locus.EukaryotCell
75–81. 3(4):900–909.
12.LiuW,MayGS,LionakisMS,LewisRE,KontoyiannisDP(2004)Extracopiesofthe 29.Munier-LehmannH,VidalainPO,TangyF,JaninYL(2013)Ondihydroorotatedehy-
Aspergillus fumigatus squalene epoxidase gene confer resistance to terbinafine: drogenasesandtheirinhibitorsanduses.JMedChem56(8):3148–3167.
Genetic approach to studying gene dose-dependent resistance to antifungals in 30.VogelHJ(1956)AconvenientgrowthmediumforNeurospora(mediumN).Microbiol
A.fumigatus.AntimicrobAgentsChemother48(7):2490–2496. GenetBull13:42–43.
13.ParryTE,BlackmoreJA(1974)Serum“uracilplusuridine”levelsinnormalsubjects 31.ZameitatE,Gojkovi(cid:1)cZ,KnechtW,PiskurJ,LöfflerM(2006)Biochemicalcharacter-
andtheirpossiblesignificance.JClinPathol27(10):789–793. ization of recombinant dihydroorotate dehydrogenase from the opportunistic
14.D’EnfertC,etal.(1996)Attenuatedvirulenceofuridine-uracilauxotrophsofAsper- pathogenicyeastCandidaalbicans.FEBSJ273(14):3183–3191.
gillusfumigatus.InfectImmun64(10):4401–4405. 32.SibleyGEM,etal.(2009)Pyrroleantifungalagents.PatentCooperationTreatyAppl
15.NobleSM,JohnsonAD(2005)Strainsandstrategiesforlarge-scalegenedeletion WO2009130481.
studiesofthediploidhumanfungalpathogenCandidaalbicans.EukaryotCell4(2): 33.CopelandRA,etal.(1995)Recombinanthumandihydroorotatedehydrogenase:Ex-
298–309. pression, purification, and characterizationof a catalytically functionaltruncated
16.RetallackDM,HeineckeEL,GibbonsR,DeepeGS,Jr,WoodsJP(1999)TheURA5gene enzyme.ArchBiochemBiophys323(1):79–86.
isnecessaryforHistoplasmacapsulatumgrowthduringinfectionofmouseandhu- 34.DereeperA,etal.(2008)Phylogeny.fr:Robustphylogeneticanalysisforthenon-
mancells.InfectImmun67(2):624–629. specialist.NucleicAcidsRes36(WebServerissue):W465–469.
17.deGontijoFA,etal.(2014)Theroleofthedenovopyrimidinebiosyntheticpathway 35.RawlsJ,KnechtW,DiekertK,LillR,LöfflerM(2000)Requirementsforthemito-
inCryptococcusneoformanshigh-temperaturegrowthandvirulence.FungalGenet chondrialimportandlocalizationofdihydroorotatedehydrogenase.EurJBiochem
Biol70:12–23. 267(7):2079–2087.
12814 | www.pnas.org/cgi/doi/10.1073/pnas.1608304113 Oliveretal.
Supporting Information
Oliver et al. 10.1073/pnas.1608304113
SIMaterialsandMethods thenfiltered.Thefiltratewasconcentratedinvacuotoyieldthe
Fungal Strains. For susceptibility testing, a range of clinical isolates crudeproduct,whichwaswashedsuccessivelywithdiethylether
weretestedforeachspeciesplusstandardstrains:A.fumigatusAF210, (300 mL) and hexane (200 mL) to afford 60 g (73%) of 2-(1,5-
A. niger AN1, A. terreus AT4, A. flavus AFl1, C. albicans CA6862, dimethyl-3-phenyl-1H-pyrro-2-yl)-N-{4-[4-(5-fluoro-pyrimidin-2-
E.coliNCTC10418,andStaphylococcusaureusOxfordstrainNCTC yl-piperazin-1yl]-phenyl}-2-oxo-acetamide F901318 as a yellow-
6571.A.nidulansA4(WT)andA.nidulansA767(pyrG89nicA) colored solid (TLC system: EtOAc:pet ether, 1:1; Rf value,
were used for the mechanism of action screen. A. fumigatus 0.65).1HNMR(400MHz,CDCl3):δ8.21(s,2H),8.19(broad
NIH4215andA.fumigatusF16216wereusedforefficacytesting. s,1H),7.32–7.22(m,5H),7.10(d,J=8.7Hz,2H),6.83(d,J=
8.7 Hz, 2 H), 6.13 (s, 1 H), 3.92–3.88 (m, 4 H), 3.81 (s, 3 H),
Synthesis of F901318. The synthesis of 1,2-dimethyl-4-phenyl- 3.19–3.15(m,4H),2.33(s,3H).m.p.194–197°C.(m/z,APCI,
1H-pyrrole was done as described previously (32). positive scan):499.41 (100%;M+H).
(1,5-Dimethyl-3-phenyl-1H-pyrrol-2-yl)-oxo-acetyl chloride. Oxalyl chlo-
ride(78.42mL,0.91mol)wasslowlyaddedtoacooledsolution Mechanism of Action Screen. A. nidulans A767 protoplasts were
of 1,2-dimethyl-4-phenyl-1H-pyrrole (105.5 g, 0.61 mol) in dry prepared from fungal biomass by digestion with 5% (wt/vol)
dichloromethane(3×200mL)at0°C.Thereactionmixturewas glucanex (Novozymes) in 0.6 M KCl and 50 mM CaCl2 for 60–
then warmed to room temperature and stirred for 1 h. The 120minat30°Cwithshakingat100rpm.Theprotoplastswere
solventwasevaporatedtodrynessinvacuotoafford124g(78% washed in 0.6MKCland50mMCaCl2, isolated bycentrifuga-
yield) of (1,5-dimethyl-3-phenyl-1H-pyrrol-2-yl)-oxo-acetyl chlo- tion at 820 × g for 9 min at 4 °C, and then washed and re-
ride as a brown oily liquid, which was used directly in the next suspendedinSTCbuffer(1.2Msorbitol,50mMCaCl2,and10mM
step(TLCsystem: EtOAc:petether, 3:7; Rfvalue,0.65). Tris·HCl pH 7.5). Then 2–5 μg of an A. nidulans genomic library
5-Fluoro-2-[4-(4-nitro-phenyl)-piperazin-1-yl]-pyrimidine.2-Chloro-5- carried on the pRG3-AMA1-NotI vector (Fungal Genetic Stock
fluoropyrimidine(78.76mL,0.64mol)wasaddedtosuspensionof Center) was transformed into protoplasts in the presence of
1-(4-nitro-phenyl)-piperazine (110 g, 0.53 mol) and potassium 40% (wt/vol) PEG 6000 in STC buffer for 5 min, followed by
carbonate (367.2 g, 2.65 mol) in acetonitrile (3 L). The reaction vortexing and inoculation onto Vogel’s minimal medium, 1.2 M
mixturewasrefluxedfor48h.Oncompletionofreflux,themixture sorbitol,and5μg/mLnicotinicacid.Theplateswereincubatedat
wascooledto0°Candthenfiltered,andtheresiduewastakenin 37 °C for 5 d, after which the spores were collected in PBS-0.1%
water (3 L) and stirred for 30 min. The suspension was filtered, Tween80andfilteredthroughglasswool.Atotalof105sporeswere
the solid cake was washed with water (1 L) and pet ether (1 L) inoculatedontoVogel’sminimalmedium(30)plus0.25μg/mL
andthendriedundervacuumtoafford148g(92%)of5-fluoro-2- F901318, and the plates were incubated at 37 °C for 7 d. DNA
[4-(4-nitro-phenyl)-piperazin-1-yl]-pyrimidineasayellow-colored was isolated from resistant colonies using the FastDNA Spin
solid.(TLC system: EtOAc:pet ether, 3:7; Rf value,0.70).1H Kit(MPBiomedicals),andplasmidDNAwastransformedinto
NMR(300MHz,CDCl3):δ8.24(s,2H),8.17–8.13(d,J=9.6 E. coli (MegaX DH10B T1R Electrocomp cells; Invitrogen).
Hz, 2 H), 6.87–6.84 (d, J = 9.6 Hz, 2 H), 3.97–3.94 (m, 4 H), DNA was mutagenized using the EZ-Tn5 bacterial transposon
3.55–3.52(m,4H).m.p.205–206°C.(m/z,ES,positivescan): system (Epicentre Biotechnologies). A DNA fragment comprising
304.21 (100%; M+H). A.fumigatuspyrG(GenBankaccessionno.Y11303.1),followedby
4-[4-(5-Fluoro-pyrimidin-2-yl)-piperazin-1-yl]-phenylamine.Asolutionof the EM7 promoter upstream of the She ble ORF (codes for re-
5-fluoro-2-[4-(4-nitro-phenyl)-piperazin-1-yl]-pyrimidine(80g, sistance to zeocin: 1,924–2,426 bp of pEM7/zeo; Invitrogen) with
0.26 mol) in methanol (1 L) was hydrogenated over Raney Tn5 mosaic ends at both ends, was included in the transposase
Nickel(30g)under70-psipressureatroomtemperaturefor6h. reaction. After transformation into E. coli, insertion of the Tn5
The reaction mixture was filtered over celite, and the filtrate transposon was detected by growth in the presence of 50 μg/mL
wasconcentratedinvacuotoyieldacrudeproduct,whichwas zeocin,andPCRestablishedtheinsertionsite.
washedwithhexane(500mL)toafford52g(72%)of4-[4-(5-
fluoro-pyrimidin-2-yl)-piperazin-1-yl]-phenylamineasabrown- Cloning,Expression,andPurificationofDHODH.Toovercomeinitial
coloredsolid(TLCsystem:methanol:chloroform,1:9;Rfvalue, problems with solubility of A. fumigatus DHODH expressed in
0.50). 1H NMR (300 MHz, CDCl3): δ8.21 (s, 2 H), 6.86–6.84 E. coli, the N-terminal 88 amino acids consisting of the mito-
(d,J= 8.7 Hz,2H),6.68–6.65(d,J =9.0 Hz,2H),3.92–3.89 chondrial targeting sequence and transmembrane domain were
(m, 4 H), 3.45 (broad s, 2 H, NH2), 3.09–3.06 (m, 4 H). m.p. omitted, and the truncated DHODH (amino acids 89–531) was
138–139 °C. (m/z, ES, positive scan): 288.14 (100%; M+H). clonedintopET44(Novagen)that codesforanN-terminalNusA
2-(1,5-Dimethyl-3-phenyl-1H-pyrro-2-yl)-N-{4-[4-(5-fluoro-pyrimidin-2-yl)- fusion protein to aid solubility. A. fumigatus cDNA was prepared
piperazin-1-yl]-phenyl}-2-oxo-acetamide F901318. A solution of 1,5- from A. fumigatus RNA using AMV reverse transcriptase (Prom-
dimethyl-3-phenyl-1H-pyrrol-2-yl)-oxo-acetyl chloride (64.65 g, ega).PCRofN-truncatedA.fumigatusDHODHcodingcDNAwas
0.24 mol) in dichloromethane (500 mL) was added to a stirred carriedoutusingprimersAF_DHODH_F1andAF_DHODH_R1
solution of 4-[4-(5-fluoro-pyrimidin-2-yl)-piperazin-1-yl]-phenyl (primersequencesgiveninTableS1)andKODDNApolymerase
amine (45 g, 0.16 mol) and triethylamine (46 mL, 0.32 mol) in (Novagen). This product was cloned into pET44 by ligation-
dichloromethane (150 mL) at 0 °C. The reaction mixture was independentcloning(Novagen).N-truncated(aminoacids29–395)
warmed to room temperature and then stirred for 30 min. The humanDHODHcDNAwasclonedfromIMAGEclone6064723
reaction mixture was quenched with water and extracted with (MGC70636; Geneservice) into pET44 after PCR with primers
dichloromethane (6 × 500 mL). The combined organic layers HS_DHODH_F1andHS_DHODH_R1.
were washed successively with saturated sodium bicarbonate C.albicansDHODHcontainstwoCTG-encodedserineresidues
solution(1.5L),water(1L),andbrine(1.5L),andfinallydried at codons 11 and 78 that were mutated to TCG codons by incor-
over anhydrous sodium sulfate. The organic layer was stirred poratingthenewcodonsintwooverlappingfragments,whichwere
withneutralalumina(500g)atroomtemperaturefor30minand subsequentlyjoinedbyfusionPCR.Thefragmentswerecreatedby
Oliveretal.www.pnas.org/cgi/content/short/1608304113 1of9
PCRfromgDNAwithprimerpairsCA_DHODH_mutF1withCA_ charged residues to solvent. The final protein was coarse mini-
DHODH_mutR1 and CA_DHODH_mutF2 with CA_DHODH_ mized(0.1 rmsd) using the XEDforcefield.
mutR2. The fusion PCR used primers CA_DHODH_fusF1 with In parallel to the initial protein analysis, the fungal inhibitor
CA_DHODH_fusR1, and the product was cloned into pGEM- data were compared with known DHODH activities for which
Teasy(Promega)andsequenced.N-truncated(aminoacids57–444) X-ray datawere available. Full multiple ligand alignments were
C. albicans DHODH cDNA was cloned into pET44 for protein generated,and3DQSARwasperformedforthefungalinhibitors
expression.PreviousstudiesonrecombinantDHODHfromavari- (Forge;Cresset; www.cresset-group.com/products/forge/).
ety of organisms, including human (33) and C. albicans (31), have All of the available information for the ligand centric and
foundthatremovaloftheNterminus,includingthemitochondrial proteincomplexdatawerecombinedtoprovideabestestimateof
targeting sequence and transmembrane regions, improves expres- correct placement of the F901318 inhibitor structure in the ho-
sion,simplifiespurification,andhaslittleeffectonkinetics. mology model. The resultant complex was then reminimized in
C.albicansDHODHwasfurthermutatedattworesidues.Phe placeinitiallyusingshellminimizationinXEDrawandthenusing
162
andVal werealteredtobecomeValandMet,respectively,using full minimization to0.01 rmsd(Fig.S6A).
171
the Phusion Site-Directed Mutagenesis Kit (Thermo Fisher Scien- The homology model was assessed using a Ramachandran
tific). PCR was carried out with phosphorylated primers CA_ plot (Fig. S6C). This plot exhibited good geometries for the
DHODH_VM_F and CA_DHODH_FV_R and template pET44_ core, with theonlydiscrepanciesbeing distant fromtheligand
CA_DHODH. The product was religated into a plasmid using T4 andconfinedtomorespeculativeregionsofthemodel,i.e.,where
nohomologyexisted.
DNAligasetocreatepET44_CA_DHODH_V _M .
162 171
Bybringingboththeproteinsequenceandstructuraldatatogether
Protein expression plasmids were transformed into E. coli
with ligand SAR and field data, we were able to conclude with
BL21 (DE3) cells, and expression was stimulated with 0.5 mM
isopropylβ-D-1-thiogalactopyranoside,followedbyincubationin reasonableconfidencethemostlikelynatureoftheinteractionsof
LB broth in the presence of 100 μg/mL ampicillin and 100 μM this DHODH inhibitor chemotype. As with known inhibitors, a
flavinmononucleotideat18–20°Cfor16–18h.Bacterialpellets bidirectional H-bonding arrangement was observed for these
fungalligands,inkeepingwiththelikelyinteractionswithinthe
werelysedwithBugBuster(Novagen),andHis-taggedDHODH
putative ubiquinone-binding site. These interactions, from ob-
protein was isolated by immobilized metal affinity chromatog-
servationacrossmultipleexamples,involveequivalentsofHis ,
raphy using His-Bind resin (Novagen) according to the manu- 116
facturer’sinstructions. Arg202,Tyr213,and Tyr505 (A. fumigatus residue numbering) as
the corresponding H-bonding partners (Fig. 3A). From the se-
quence,structureanalysis,andligandfieldanalysis,furtherlipophilic
DHODHAssays.Theseassayswerecarriedoutbroadlyasdescribed
elsewhere(31).Approximately2μgofrecombinantA.fumigatus contacts—Val200, Met209, and Tyr510—were defined. This ligand-
DHODH or 0.5 μg of recombinant human DHODH was in- binding model is supported by the mutagenesis of the C. albicans
recombinant protein (Fig. 3B). Phe andVal of C.albicans
cubatedinthepresenceof1mML-dihydrooroticacid,0.05mM 162 171
alignwithVal andMet ofA.fumigatusDHODH,respectively.
coenzyme Q2, and 0.1 mM 2,6-dichlorindophenol in 50 mM 200 209
Tris·HCl,150mMKCl,10%(wt/vol)glycerol,and0.1%(wt/vol) MutatingtheseresiduestotheirA.fumigatusequivalentsintroduced
inhibitionofDHODHactivitybyF901318inthemutant,whereno
Triton X-100, pH 8 at 25 °C. DHODH activity was determined
inhibitionhadoccurredwiththeWTenzyme.
using the redox indicator 2,6-dichloroindophenol by measuring
thedecreaseinabsorbanceat600nm.Initialreactionvelocities
Creation of DHODH Phylogenetic Trees. DHODH sequences were
were determined, and the percent inhibition in the presence of
obtainedfromsearchingsequencedatabasesheldbytheNational
thedrugwascalculatedcomparedwithcontrolsintheabsenceof
Center for Biotechnology Information, EnsemblFungi, and the
drug. The inhibition curve data were fitted to a four-parameter
Joint Genome Initiative (JGI). The accession codes for the
sigmoidalplot(r2>0.998forIC plots)usingXLFit(IDBS),and
50 proteins are as follows, with alternative fungal names in paren-
the IC50 value was calculated as the concentration of drug theses: Absidia idahoensis (Lichtheimia ramosa), CDS13591.1;
causing50%inhibition. Acremonium_alcalophilum,jgijAcral2j2024193;Alternariaalternata,
jgijAltal1j982905;Alternariabrassicicola,jgijAltbr1j1659;A.flavus,
HomologyModelingofA.fumigatusDHODH.A range of DHODH
XP_002379410.1; A. fumigatus, XP_755434.1; Aspergillus lentulus,
structures, including those from human, rat, P. falciparum, and
GAQ06647.1; A. niger, XP_001399431.2; A. terreus, protein se-
E. coli, were compared with the fungal sequences for Candida
quence derived from gene ATEG_01477 (EnsemblFungi); Bipo-
and Aspergillus strains. Sequences were aligned manually with
laris sorokiniana, XP_007699125.1; B. dermatitidis (Ajellomyces
referencetotheproteinsuperimpositions(PyMOL).RCSBPDB
dermatitidis), EEQ85598.1; C. albicans, XP_723522.1; Candida
IDcodesforthestructuresusedwereasfollows:1D3G,1PRH,
glabrata, XP_449903.1; Candida krusei (Pichia kudriavzevii),
2PRL, 2PRM, 3G0X, 3KVM, 2WV8, 2FQI (human); 1UUM, KGK37337.1; Candida parapsilosis, CCE41941.1; Candida tropicalis,
1UUO (rat); 2E68 (Trypanosoma cruzi); 3I68, 1ITV, 3O8A XP_002549089.1;Chaetomiumglobosum,XP_001223311.1;C.immitis,
(P. falciparum); 3MJY (Leishmania major); and 1F76 (E. coli) XP_001240778.2; C. neoformans, XP_569657.1; E. coli, WP_
(www.rcsb.org). 001330055.1; Exophiala dermatitidis, XP_009159400.1; Fusarium
Human DHODH (PDB ID code 1D3G) (19) served as the oxysporum, EGU79150.1; Fusariumsolani(Nectriahematococca),
protein template for construction of the A. fumigatus DHODH XP_003046434.1;H.capsulatum(Ajellomycescapsulatus),EGC45332.1;
model. Residue replacement, bump checking, and coarse re- Homosapiens,NP_001352.2;Magnaporthe_oryzae,XP_003719157.1;
finement were done in Discovery studio 4.1 (Accelrys). Fine Malassezia globosa, XP_001729892.1; Microsporum canis (Arthro-
residue refinement and new secondary structure or loop inser- derma otae), XP_002849178.1; Microsporum gypseum, XP_
tionswerebuiltinXEDraw(Cresset’sinternalresearchsoftware, 003175518.1; Mucor circinelloides, EPB81175.1; Neurospora crassa,
results of which can be visualized in Torch; www.cresset-group. XP_961804.1;Paecilomycesvariotii(Byssochlamysspectabilis),
com/products/torch/). When no information was known for a GAD93826.1; Paracoccidioides brasiliensis, XP_010757420.1; Peni-
givenloop,theloopwasconstructedmanuallybasedonthelocal cillium chrysogenum, jgijPench1j33932; Penecillium marneffei (Ta-
environmentandthenatureoftheprimarysequence(i.e.,using laromyces marneffei), KFX48065.1; Phycomyces blakesleeanus,
the predicted beta sheet or helical secondary structural prefer- jgijPhybl2j177367; P. falciparum, XP_966023.1; Pneumocystis
ence)toyieldappropriatepackingthatminimizedtheexposure jirovecii,CCJ31177.1;Rhizopusdelemar,EIE79352.1;Scedosporium
of lipophilic residues and maximized the exposure of polar and apiospermum,JOWA01000121.1;Sporothrixschenckii,ERT03286.1;
Oliveretal.www.pnas.org/cgi/content/short/1608304113 2of9
Trichoderma reesei, XP_006962857.1; Trichophyton mentagrophytes Class 2 DHODH sequences are present in all mammals, many
(Arthroderma benhamiae), XP_003012466.1; Trichophyton rubrum, fungi,andsomebacteria,includingE.coli(29).Theyaremembrane-
XP_003234264.1;andUstilagomaydis,XP_011389794.1. bound(integralmitochondrialmembraneproteinsinmammalsand
SequencesinFASTAformatweresubmittedforphylogenetic somefungi;membrane-associatedinsomebacteria)monomersthat
analysisatwww.phylogeny.frusingthe“Advanced”analysis(34). have a catalytic serine and use ubiquinone as a terminal electron
Sequences were aligned using MUSCLE (“Full mode”) and acceptor.Class1Aand1BDHODHsaresolubleandmultimeric,
curated with GBlocks (default settings). Phylogeny was estab- have a catalytic cysteine, and usefumarate or NAD as aterminal
lished by the maximum likelihood method in PhyML using the electron acceptor. Class 2 DHODHs (Fig. S7A) were analyzed
approximate likelihood ratio test set to SH-like with the WAG separately from the zygomycete DHODHs (Fig. S7B), which ap-
substitution method. Thetree was visualized inTreeDyn. peartobefromclass1A.
A
B [F901318](µg/ml)
0 0.125 0.25 0.5 1 2
(i)A.nidulansA4
(ii)A.nidulans #18.1
(iii)A.nidulans
#18.1_Tn5_pyrE
(iv)A.nidulans
#18.1_Tn5_AN5910
Fig.S1. (A)MapofAMAinsertsrecoveredfromthemechanismofactionscreen.AnA.nidulansgenomicscreenwasperformedforgenesthatwhenpresent
inahighcopynumberprovidedresistancetoF901318(MaterialsandMethods).pAMA1clones18.1–18.4wererecoveredfromfourindependentcoloniesthat
wereresistanttoF901318.DNAwasextractedfromtheresistantA.nidulansandtransformedintoE.coli.PlasmidDNAwasisolatedandsequenced,andthe
genomicinserts(redbars)foundweremappedtotheA.nidulansgenomeusingEnsemblFungi.(B)A.nidulanscarryingextracopiesofpyrEisresistantto
F901318.ThepAMA1_18.1plasmidwastransformedintoA.nidulansA767before(#18.1;ii)andaftermutagenesiswiththeEZ-Tn5bacterialtransposon
system.ThetransposondisruptedthepyrEgene(#18.1_Tn5_pyrE;iii)ortheneighboringANIA_05910gene(#18.1_Tn5_AN5910;iv).Conidiafromtheresulting
transformantsandfromtheWTA.nidulansA4(i)wereinoculatedontomediainthepresenceandabsenceofvaryingconcentrationsofF901318andin-
cubatedat37°Cfor4d.
Oliveretal.www.pnas.org/cgi/content/short/1608304113 3of9
Fig.S2. DHODHisanenzymeofpyrimidinebiosynthesis.(A)PathwayofdenovopyrimidinebiosynthesisinA.fumigatus.Enzymenames,classification
(accordingtotheNomenclatureCommitteeoftheInternationalUnionofBiochemistryandMolecularBiology;www.chem.qmul.ac.uk/iubmb/enzyme/),and
predictedA.fumigatusgenes(fromKEGG;www.genome.jp/kegg/andEnsemblFungi;www.fungi.ensembl.org/index.html/)arelistednexttoeachstepofthe
pathway.(B)ReactioncatalyzedbyDHODH.Inmammalsandmanyfungi,DHODHislocatedatthemitochondria,whereatransmembranedomainanchorsthe
enzymetotheoutsideoftheinnermitochondrialmembrane.Intheintermembranespace,DHODHcatalyzestheoxidationofdihydroorotatetoorotate.FMN
andubiquinonefromthemitochondrialmembrane(coenzymeQ)actaselectronacceptorsinthereaction.
Oliveretal.www.pnas.org/cgi/content/short/1608304113 4of9