Synthetic Multivalent Antifungal Peptides Effective against Fungi

18 Pages·2014·3.88 MB·English

Synthetic Multivalent Antifungal Peptides Effective
against Fungi
RajamaniLakshminarayanan1,2*.,ShoupingLiu1,2.,JianguoLi1,3,MurugananthamNandhakumar1,Thet
TunAung1,EuniceGoh1,JamieYaTingChang1,PadhmanabanSaraswathi1,CharlesTang5¤,SitiRadiah
BinteSafie1,LimYihLin1,HowardRiezman4,ZhouLei1,2,ChandraS.Verma3,6,7,RogerW.Beuerman1,2,6*
1SingaporeEyeResearchInstitute,Singapore,Singapore,2DepartmentofNeuroscienceandBehaviouralDisorders,Duke-NUSGraduateMedicalSchool,Singapore,
Singapore, 3Bioinformatics Institute (A*STAR), Singapore, Singapore, 4Department of Biochemistry and NCCR Chemical Biology, University of Geneva, Geneva,
Switzerland,5DepartmentofPathology,SingaporeGeneralHospital,Singapore,Singapore,6SchoolofBiologicalSciences,NanyangTechnologicalUniversity,Singapore,
Singapore,7DepartmentofBiologicalSciences,NationalUniversityofSingapore,Singapore,Singapore
Abstract
Takingadvantageoftheclustereffectobservedinmultivalentpeptides,thisworkdescribesantifungalactivityandpossible
mechanismofactionoftetravalentpeptide(B4010)whichcarries4copiesofthesequenceRGRKVVRRthroughabranched
lysine core. B4010 displayed better antifungal properties than natamycin and amphotericin B. The peptide retained
significantactivityinthepresenceofmonovalent/divalentcations,trypsinandserumandtearfluid.Moreover,B4010isnon-
haemolytic and non-toxic to mice by intraperitoneal (200mg/kg) or intravenous (100mg/kg) routes. S. cerevisiae mutant
strains with altered membrane sterol structures and composition showed hyper senstivity to B4010. The peptide had no
affinityforcellwallpolysaccharidesandcausedrapiddissipationofmembranepotentialandreleaseofvitalionsandATP
when treated with C. albicans. We demonstrate that additives which alter the membrane potential or membrane rigidity
protectC.albicansfromB4010-inducedlethality.Calceinreleaseassayandmoleculardynamicssimulationsshowedthatthe
peptide preferentially binds to mixed bilayer containing ergosterol over phophotidylcholine-cholesterol bilayers. The
studiesfurthersuggestedthatthefirstarginineisimportantformediatingpeptide-bilayerinteractions.Replacingthefirst
arginineledtoa2–4 folddecreaseinantifungalactivitiesandreducedmembranedisruptionproperties.Thecombinedin
silico andinvitro approachshould facilitaterationaldesign ofnewtetravalent antifungalpeptides.
Citation:LakshminarayananR,LiuS,LiJ,NandhakumarM,AungTT,etal.(2014)SyntheticMultivalentAntifungalPeptidesEffectiveagainstFungi.PLoSONE9(2):
e87730.doi:10.1371/journal.pone.0087730
Editor:RizwanH.Khan,AligarhMuslimUniversity,India
ReceivedOctober16,2013;AcceptedDecember28,2013;PublishedFebruary3,2014
Copyright:(cid:2)2014Lakshminarayananetal.Thisisanopen-accessarticledistributedunderthetermsoftheCreativeCommonsAttributionLicense,which
permitsunrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalauthorandsourcearecredited.
Funding:ThisresearchissupportedbytheSingaporeMinistryofHealth’sNationalMedicalResearchCouncilunderitsTranslationalResearchInnovationsin
OcularSurgery(TRIOS)Programme:DefensinsprojectNMRC/TCR/002-SERI/2008(RWB)andNMRC/NIG/1020/2010(RL).Thefundershadnoroleinstudydesign,
datacollectionandanalysis,decisiontopublish,orpreparationofthemanuscript.
CompetingInterests:Theauthorshavedeclaredthatnocompetinginterestsexist.
*E-mail:[email protected](RL);[email protected](RWB)
¤Currentaddress:ALSFoodandPharmaceuticals,Singapore,Singapore.
.Theseauthorscontributedequallytothiswork.
Introduction Since current antifungals have limited number of targets and
increasing reports of resistance to the available drugs, peptide-
Resistance tocurrent antifungaldrugs hasbecome commonin
based antifungals, which disable the membrane physiology, offer
recent years, mainly due to increased number of immuno-
an important opportunity for rational drug design [8–10].
compromised patients, and has underscored the need for new
However, their antimicrobial properties are antagonized by
classesofantifungals[1,2].Pathogenicfungalinfectionsarethe7th
physiological concentration of salts and polyanionic polymers
mostcommoncauseofinfection-relateddeathsintheUSA[3].In
(e.g., DNA, glycosaminoglycans and mucins) as well as being
fact, recent surveys suggest that fungal diseases are posing a
rapidly degraded by proteolytic enzymes in complex biological
growing threat to the global biota [4]. The number of available
milieu, thus limiting their therapeutic potential [11,12]. Several
antifungals is limited and developing new antifungal drugs is
different strategies have been attempted to circumvent the
challengingsincecommondrugtargetsinfungiarehomologuesof
drawbacks of antimicrobial peptides [13–17]. The development
similar molecular types in human which might also be inhibited.
of multivalent peptides by assembling multiple copies of mono-
Amongthefivedifferentclassesofantifungalspolyenes,azolesand
meric peptides around a core molecule has attracted significant
allylamines target ergosterol or ergosterol biosynthetic pathways
interest as a new molecular phenotype for potential therapeutic
whereas 5-fluorocytosine (5-FC) targets DNA synthesis and
use[18–20].However,theantimicrobialpropertiesofmultivalent
caspofungins inhibit b-glucan synthase [5]. The extent of drug
peptides have been limited to Gram-negative and/or Gram-
resistance also varies, as resistance to 5-FC and azoles are more
positivebacteriaandtheirutilityaspotentantifungalshasnotbeen
common when compared to polyenes, although there are now
explored in greatdetail [21–24].
manyreports of resistancetopolyenesas well [6,7].
PLOSONE | www.plosone.org 1 February2014 | Volume 9 | Issue 2 | e87730
MultivalentAntifungalPeptides
We have previously reported the properties of a 10-residue out as described above. C. albicans ATCC10231 cells and salts in
peptide (RGRKVVRRKK) which showed potent activity against brothweretakenasapositivecontrol.Saltsalone(NaCl,KCl,and
P. aeruginosa, but poor activity against fungi [25]. The anti- CaCl and MgCl ) in broth or with 22mM peptide served as a
2 2
pseudomonas activity was attributed to the concentration-depen- negativecontrol.Theeffectof5%humanserumontheactivities
dent formation of a non-covalent dimer in buffer as well as in of B4010 against two C. albicans clinical isolates was determined.
liposome model of a prokaryotic membrane [26]. The present The human male serum was centrifuged at 13,000 rpm for
work extends the analysis to higher order covalently linked 10mins in order to remove the lipids and the supernatant was
peptides.Theantifungalpropertiesofapotenttetravalentpeptide collected [27]. The MIC values against clinical isolates of C.
(B4010) have been examined in physiological concentrations of albicans 2672R and C. albicans 1976R were determined both in
salts and in complex biological fluids. We have also assessed the standard medium (SD broth) as well as in standard medium
biocompatibility of B4010 including its effect on corneal reepi- containing 5%serumsupernatant asbefore.
thelialization in rabbit and systemic toxicity in mice. The
antifungal action of this molecule on fungal membrane was DeterminationofMICagainstWildType(WT)andMutant
ascertained using environmental-sensitive fluorescent probes.
S. cerevisiae Strains
Finally, we demonstrate that coupling MD simulations with
A few identical colonies from WT and mutant strains were
in vitroexperimentsshedlightontheimportantresiduesthatare
immediately inoculated and grown overnight in SD broth
essential for mediating membrane-lipid interactions and rational-
containing 0.1 mg/mL ampicillin at 30uC [28,29]. Cells were
ization of theantifungalactivity.
harvestedbycentrifugation,washedwithsterilewaterandfrozen.
CellswereplatedagainonSDagarplatesandincubatedat37uC
Materials and Methods
for 48h. Two or three identical colonies of the mutant strains
Ethics Statement werecollectedandinoculatedinSDbroth.Peptidesweredissolved
in sterile water and mixed with WT and mutant strains and
All animals used in this study were treated in agreement with
incubated at 37uC for 48 or 72h in a test tube. The lowest
the tenets of the Association for Research in Vision and
concentrationofthepeptidethatinhibitedthegrowthoftheyeast
Ophthalmology (ARVO) Statement for the Use of Animals in
strain was determined byvisual inspection.
OphthalmicandVisionResearch,andtheprotocolwasapproved
by the Singhealth Institutional Animal Care and Use Committee
(IACUC; AALACaccredited, #2012/SHS/775). Time-kill Kinetics Assay
The time-kill kinetics of B4010 was determined against two
Chemicals and Peptides strains of C. albicans (ATCC 10231 and DF2672R). The cultures
weregrownovernightinSDbrothandthecellconcentrationwas
Sabouraud’s Dextrose Agar was purchased from Acumedia
adjustedto105–106CFU/mLwithphosphatebuffer.Thepeptide
(Michigan, MI, USA). Peptides were purchased from EZBiolabs
Inc., (Carmel, IN, USA). Lipids such as L-a-phophatidylcholine or antifungals were added tothe individual cultures. Appropriate
(PC),L-a-phostidylethanolamine(PE),L-a-phosphatidylserine(PS) final concentrations of B4010 (1.4–11mM for ATCC and 0.37–
and L-a-phosphatidylinositol (PI) were bought from Avanti Polar 3.7 mM for 2672R) or amphotericin B (0.55–11mM for 2672R
LipidsInc.(AL,USA).Ergosterol,sodiumazide(NaN3),Carbonyl strains) and natamycin (4.7–75 mM for ATCC and 30mM for
cyanide m-chloro phenylhydrazone (CCCP), 4-aminopyridine (4- 2672Rstrains)wereadjustedfortheinocula.Thetestsolutionwas
AP), 5-nitro-2-(3-phenylpropylamino) benzoic acid (NPPB), Gad- incubatedat37uCwithconstantshaking.100mLofthesuspension
olinium III Chloride, tetra ethyl ammonium chloride (TEA) and waswithdrawnatpredeterminedtimepoints,seriallydiluted(102
39,39-dipropylthiadicarbocyanine (diS-C3-5) dye were purchased or 103 fold) and poured into the SDA plate. The plate was
fromSigma-AldrichCorp.(MO,USA).ATPbioluminescencekit incubated for 48h at 37uC for colony counting. The data were
wasobtainedfromMolecularProbesInc(OR,USA).Amphoter- expressed in terms of % fungal viability relative to the positive
icinBandNatamycinwereobtainedinpowderformfromSigma- control.
Aldrich (S)Pte Ltd (Singapore).
Antifungal Activity of B4010 in Presence of Trypsin and
Determination of Minimum Inhibitory Concentration 50% Tear Fluid (TF)
(MIC) For all the experiments, unless otherwise stated, C. albicans
TheyeaststrainswerecultivatedandsuspendedinSabouraud’s ATCC 10231 strain was used. B4010 (1 mg/mL) was incubated
Dextrose (SD) broth diluted to one sixth at a starting OD = with trypsin (enzyme:peptide ratio 1:100) at 37uC. 20mL aliquot
600
,0.08 in a flat-bottomed microtitre plate. A serial dilution of ofthismixturewaswithdrawnatvarioustimeintervals(0.5,1,2,4
peptideinthesamebrothwasmixedwiththeinoculumtogivea and6h)andmixedwith1 mLoftrypsininhibitor.Theantifungal
finalpeptideconcentrationof0.4–22 mM.Theantifungalactivity activity of the peptide was determined by adding this mixture to
wasassessedbymonitoringtheOD600incyclesof30minutesand 180mL of inoculum and monitoring the growth at OD600 for
an orbitalshakingat 100rpmusinganInfinite M200microplate 24h. Similar experiments performed in the presence of trypsin/
reader (Tecan Group Ltd., Switzerland) for 24/48h at 37uC. trypsin inhibitor (without B4010) were used as a positive control.
Cultureswithoutpeptideswereusedaspositivecontrolsandbroth For assessing the antifungal activity in TF, the peptide was
alone or with 22mM peptide served as negative controls. The dissolvedinfreshlycollectedrabbitTFandincubatedat37uCfor
minimum concentration required for complete inhibition was 6 h.Afterincubation,anequalvolumeofanovernightcultureof
assessed by both visible observations as well as by measuring the C.albicans(,106CFU/mL)wasmixedwithpeptidesintearfluid
OD and taken as the MIC. Each experiment was repeated in and incubated at 37uC for 24h. The final concentrations of the
600
triplicates. For studying the effects of metal ions, SD broth was peptide were at 4.4, 8.8 and 22 mM. A 100mL aliquot of serial
adjusted with appropriate concentrations of salts to yield a final dilutions (102 or 103 times) of this mixture were inoculated on a
concentration of 100 and 135mM for NaCl and KCl and 0.5– SDagarplateandthenincubatedfor48hat37uC.Culturealone
2 mM for CaCl and MgCl . MIC determinations were carried andculturewith50%tearfluidservedasthepositivecontrol.The
2 2
PLOSONE | www.plosone.org 2 February2014 | Volume 9 | Issue 2 | e87730
MultivalentAntifungalPeptides
datawasexpressedin%killingwithrespecttoculturewithouttear CandidacidalActivityofB4010inPresenceofMetalIons,
fluid. About 6–10% killing was observed in the presence of 50% Energy Poisons and Ion Channel Inhibitors
tear fluid.
Tostudytheeffectsofmetalions,thebrothsolutioncontaining
C. albicans (OD =0.4–0.6) was adjusted to appropriate final
600
Haemolytic Assay concentrationsofmetalions,incubatedwith5.5mMofB4010for
Haemolytic activity of peptides and antifungal drugs were 6 handthecellviabilitywasdeterminedasbeforeafter48h.For
determined against rabbit red blood cells [30]. Briefly, serial theeffectsofenergypoisonsandionchannelinhibitors,yeastcells
dilution of peptides/antifungals in PBS was mixed with rRBC (105–106CFU/mL) were incubated for 2 h with or without
(final concentration 4% v/v), incubated at 37uC for 1 h and additivesat37uC.Afinalconcentrationof5.5mMB4010peptide
centrifuged at 3000rpm for 10 minutes. The release of wasaddedtothecellsuspensionandincubatedfurtherforanother
hemoglobin in the supernatant was monitored by measuring the 1.5 h at 37uC. Fungal viability was obtained by counting the
hemoglobin absorbance at 576nm. The readings from cell numberofcoloniesformedineachplateafter48hincubationat
suspension in PBS (without any additives) or 1% Triton-X100 37uC. SDA plates with additives (and no peptide) served as the
were usedas 0%or 100%haemolysis. positive control. The final concentrations of the additives were:
CCCP (5 mM), NaN (5 mM), 4-AP (1 mM), NPPB (0.5mM),
3
Cytotoxicity on Human Conjunctival Epithelial Cells Gadolinium II chloride (1 mM) and tetra ethyl ammonium
Immortalized normal human conjunctival epithelial cells chloride (15 mM). The reported values were averages of two
(IOBANHC) were a gift from Yolanda Diebold at the University independent duplicate experiments. Control experiments with all
of Valladolid, Spain [31]. IOBA-NHC cells were cultured under the additives were also performed to assess the toxicity of the
standardconditions(humidifiedatmosphereof5%CO at37uC) additives to C.albicans.
2
in DMEM/F12 supplemented with 1 mg/ml bovine pancreas
insulin, 2 ng/ml mouse epidermal growth factor, 0.1mg/ml SDS-PAGE Pull down Assay
cholera toxin, 5 mg/ml hydrocortisone, 10% fetal bovine serum B4010 (5.5–22mM) was incubated with chitin (from shrimp
(FBS), 50UI/ml penicillin, and 50UI/ml streptomycin. Cells shells)orb-D-glucanfor2 hat37uC.Themixturewascentrifuged
from passages 50–80 were used in all experiments. Every day, at10,000 gandthesupernatantwasanalyzedbySDS-PAGE(4–
normal culture development was observed by phase-contrast 20%). A control experiment without chitin/b-D-glucan was also
microscopy. Cells were removed by gentle trypsin incubation at performed to quantify peptide binding to the carbohydrate
confluence and counted. They were seeded into 96 well culture polymers.
plates (Corning, Schiphol-Rijk, The Netherlands) for microtitra-
tion analysis of flow cytotoxicity assay (,10,000 cells/well). DiS-C -5 Membrane Depolarization and SYTOX Green
3
Cultures were kept at 37uC for 24h. Subconfluent cells (culture (SG) uptake Assays
surfacecoveringnearly70%)werethenexposedtovariouspeptide
ChangesinmembranepotentialofC.albicansuponadditionof
concentrations (0.22–225 mM). Cytotoxicity was measured peri- B4010 were monitoredby the releaseof potential sensitiveprobe
odically using MultiTox-Fluor Multiplex assay kit (Promega, WI,
diS-C -5 dye. Briefly, 1 ml of overnight grown mid-log phase
3
USA) by measuring the AFC fluorescence (l =485 and
ex Candida cells in 5 mM HEPES buffer (pH7.0) was mixed with
l =520nm). We report the cell viability after 24h incubation
em 10mM diS-C -5 and incubated at 37uC for 1–2h in a thermo
3
ofthecellswiththepeptidessincetherewasnodetectabletoxicity
shaker.800mLofthedye-loadedcellsuspensionwastransferredto
even after8 h of incubation. a stirred quartz cuvette and the change in fluorescence intensity
was monitored at an emission wavelength (l ) of 670nm
em
Rabbit Model of Corneal wound Healing (excitation wavelength, 622nm) using a Quanta Master spectro-
FourNewZealandwhiterabbitsweredividedintotwogroups, fluorimeter (Photon Technology International, NJ, USA). The
two each for control (saline) and two each for the study group excitation and emission bandwidths were set at 1 and 2nm,
treated with B4010. Rabbits were tranquilized by intra muscular respectively. Once a constant fluorescence level was achieved,
injectionof1 mLofketamine(100 mg/mL)and0.5mLofxylazil 10mL of concentrated peptide solution in HEPES buffer was
(20 mg/mL).Corneaswereanesthetizedbytopicaladministration addedsothatthefinalconcentrationofpeptidewas0.22-22mM.
of1%xylocaine.A5-mmtrephinewasusedtooutlinethewound Thechangeinfluorescenceintensitywasmonitoredcontinuously
marginandmechanicalremovalofepithelialcellswascarriedout for 1 or 2 h. To study the effects of energy poisons and ion-
bysterileminiblade(BD-Beaver)leavingthebasallaminaintactas channelinhibitors,theadditiveswereaddedpriortotheaddition
previously described [32]. The animals were treated by topical of 5.5mM B4010. The change in fluorescence intensity was
administrationof22mMofB4010at3times/day.Corneawound monitored asbefore.
wasvisualizedbystainingwithfluoresceinsodium,whichisusedin For SG uptake assay, overnight cultures of Candida were
theophthalmologyclinicasanon-toxicdyetodisclosewoundsof harvested by centrifugation, washed three times with HEPES
the cornea, and using a cobalt-blue filter with slit lamp buffer and resuspended in the same buffer. The cells were
biomicroscopy. Measurements of the residual wound area were incubated with 1 mM SG in the dark prior to the addition of
performedduringthereepithelializationprocessbyImage-J1.440. peptide. The increase in emission intensity at 520nm was
monitored after addition of various concentrations of peptide
In vivo Toxicity (l =485nm). 1% Triton-X100 was added at the end of the
ex
Acute toxicity of B4010 was assessed with C57BL6 (6–8 weeks experiment todeterminethe%uptake.
old) wildtype mice.Twohealthywild typemice werechosenfor
eachroutesofadministration.B4010wasdeliveredthroughintra- Measurement of Extracellular Cations
peritoneal (200mg/kg) or (100 mg/kg) intra-venous routes. Two Overnight cultured late logarithmic phase C. albicans was
animalswereusedforeachadministrationandmonitoredthrough harvestedbycentrifugation,washed5timeswith10mMHEPES
24/48 htodetermine mortality, orsigns of toxicity. (pH 7.0), resuspended in the same buffer and adjusted to an
PLOSONE | www.plosone.org 3 February2014 | Volume 9 | Issue 2 | e87730
MultivalentAntifungalPeptides
OD =0.4. To 5 mL of this suspension, B4010 (final concen- Percentage calcein released at each time point was calculated
600
tration 5.5mM) was added and incubated at 37uC for 2 h. The using theequation,
mixturewascentrifugedat3,000 gandthepresenceofK+,Ca2+
andMg2+inthesupernatantwasestimatedbyPerkinElmerDual- A {A
view Optima 5300DV inductively coupled plasma – optical %CalceinRelease~ 0 min |100
A {A
emission spectrometry (Massachusetts, USA) available at max min
CMMACfacilities(DepartmentofChemistry,NationalUniversity
of Singapore). Where A is the observed fluorescence intensity upon addition
0
of peptides, A is the average intensity at the baseline (before
min
ATP Bioluminescence Assay peptide addition), A is the average intensity at the saturation
max
The extracellular ATP levels upon challenging C. albicans with phase after adding 1% Triton X-100. Similar protocol was
B4010 were determined as reported before [33]. Cells followed forthepreparation of PC/Cholesterol SUVs.
(OD <0.6) were incubated with or without various additives
600
for 2 hours at 37uC with orbital shaking. B4010 (5.5mM) was Isothermal Titration Calorimetry (ITC)
addedandincubatedforanother1.5hat37uCwithshaking.Each ITC experiments were performed using a Microcal VP-ITC
tubewasthencentrifugedat5,000 gfor5mins.Then225mLof Calorimeter (Microcal, Northampton, MA, USA). SUVs, B4010,
boilingTEbuffer(50 mMTris,2mMEDTA,pH 7.8)wasadded and polyene antifungals were dissolved in 10mM HEPES
tothe25mlofthesupernatantandmixedwell.Thismixturewas buffered saline at pH 7.2. SUV (PC containing 15% ergosterol)
boiledagainforanother2minutesandstoredat4uCuntilfurther waspreparedasbeforeinthesamebuffer.Thepeptide(20 mM)or
examination. 100mL of a luciferin-luciferase ATP assay mixture antifungal solution (natamycin at 10mM and amphotericin B at
was added to 100mL of the supernatant and luminescence was 5 mM) was placed in the calorimetric cell (V=1.4mL) and the
monitored using the Infinite M200 microplate reader (Tecan lipidvesicles(300 mM)weretakeninthesyringe.Aliquotsof7 mL
Group Ltd., Switzerland). For the time-course experiment, cells of liposome (300 mM for B4010/natamycin and 75mM for
(OD600=0.4) were treated with 5.5mM peptide for various time amphotericin B) were injected into the cell with constant stirring
intervals. For dose-dependent studies, the concentration of the (310 rpm)at25uC.Atotalof40injectionswereperformedandthe
peptide was varied from 0.4–44mM. The extracellular ATP heat of reaction produced by each injection was determined by
concentrationwasdeterminedfromcalibrationcurveobtainedby integrationofheatflowtraces.Inaseparateexperiment,theheat
ATP assay kit (Molecular Probes, OR, USA) as per the of dilution was determined by titrating SUVs into the buffer
manufacturer’s instruction. solution (without peptide). The heat of dilution was subtracted
from the heat changes determined in the peptide/antifungal-
Circular Dichroism (CD) Spectropolarimetry liposomeexperiments.
FarUV-CDspectraofthepeptides(0.6mg/mL)wererecorded
on a JASCO J810 spectropolarimeter (JASCO, Tokyo, Japan) in Molecular Dynamics Simulation Studies
10mMPBS(pH 7.0)usinga0.1 cmpathlengthquartzcuvetteat Moleculardynamicssimulationswereusedtostudythemodeof
25uC.Spectrawererecordedfrom260nmto190nmin0.1nm interaction of B4010 with a fungal model membrane. Fungal
steps at a scan rate of 50nm/min. The final spectrum is the membranes have complicated lipid compositions such as PS/PI,
average of 4 scans. The CD data is expressed as mean residual PC and PE and ergosterol. We constructed a model fungal
ellipticity ([h]mrw,deg cm2dmol21). membraneconsistingof288lipidswithfourtypicallipids:POPC,
POPE, POPS, and erogosterol, at a ratio of 4:2:1:2. The
Calcein Leakage Assay CHARMMforcefieldwasusedtodescribeB4010conformation,
Smallunilamellarvesicles(SUVs)werepreparedusingPC/PE/ apreviousstudyhasdemonstratedthatitcanconsistentlypredict
PI or PS lipids (5:2.5:2.5) containing 15wt% ergosterol or PC/ conformations of branched antimicrobial peptide B2088 in both
Cholesterol(10:1)[17].ForthepreparationofPC/PE/PIorPS/ aqueous and lipid environments [34]. For the lipids, the recently
Erg SUV, thelipids were dissolved inchloroform/methanol (2:1, released CHARMM36 force field was used because it predicts
v/v)inaglasstube.Ergosterolwasdissolvedinsamesolventand reasonable membrane properties in a tensionless ensemble [35].
addedtolipidmixturetomakefinalergosterolcontentof15wt%. Initially,thepeptidewasplacedclosetothefungalmembranewith
TheLipid-ergosterolmixturewasdriedusingnitrogengastoform theinitialconformationtakenfromtheendofa100nssimulation
lipid layer. The film was hydrated in a buffer containing 20mM inpurewater.Thepeptide-membranecomplexwassolvatedwith
PBS (pH 7), vortexed and sonicated. Each cycle of sonication TIP3P water and counter ions were added to neutralize the
(5 sec, 40uC) was followed by freezing in liquid nitrogen and system. Before the production stage of the MD simulation, the
thawing on a water bath kept at 37uC. The procedure was systemwassubjectedto500stepsofenergyminimizationusingthe
repeated 5–6 times until an optically clear dispersion appeared. steep descent algorithm, followed by 100ps of NVT simulation,
The SUV was divided into two parts. To the first part 50mM which was followed by an MD simulation of 400ns using the
calcein wasadded andincubated for1 h.Theexcess calceinwas GROMACS4.5package[36].DuringtheMDsimulations,both
removed by injecting 100mL of the calcien-loaded SUV’s into a LJ and real-space electrostatic interactions were treated using a
UltrahydrogelTM 250 (7.8mm 6300mm) column equilibrated cutoff potential at a distance of 1.2nm, while the long-range
with20mMPBS(pH 7.0)andelutedisochraticallyataflowrate electrostaticinteractionsinreciprocalspacewerecalculatedusing
of 1 mL/min for 1.5 column volumes. The calcein-loaded the particle-mesh Ewald algorithm [37]. The covalent bonds
liposomes were mixed with calcein-free liposome (1:1) to adjust involving hydrogen atoms were constrained using the LINCS
finalliposomeconcentration.Peptideconcentrationwasaddedto algorithm which allowed us to use a time step of 2fs in all the
obtain 1:30 or 1:15 peptide:liposome ratio. Changes in the simulations[38].ThesimulationswererunintheNPTensemble
fluorescence signalafteraddition ofpeptideorTritonX-100was with the temperature coupled to an external heat bath at 310K
measured on a PTI spectrofluorometer, using an excitation through the Nose-Hoover method and pressure maintained at
wavelength of 480nm and emission wavelength of 512nm. The 1 atm using the Parrinello-Rahman method with semi-isotropic
PLOSONE | www.plosone.org 4 February2014 | Volume 9 | Issue 2 | e87730
MultivalentAntifungalPeptides
coupling [39]. The electrostatic potential maps for peptide lysine did not alter the secondary structure. Similarly no
adsorbed or peptide-free bilayer were calculated as previously discernable differences were observed for B4010 and Sc_B4010
reported [40]. tetrabranched peptides, suggesting that scrambling the sequence
did not influence the secondary structure (Figure S1). Concen-
Scanning Electron Microscopy tration-dependent time-kill experiments were conducted after
Candidaalibicans(ATCC10231)wasincubatedwith20mg/mLof exposing C. albicans to B4010 and polyene antifungals
B4010 (MIC) at 370C and 200ml of a fungal suspension was (Figure 2A and 2B). For the ATCC strains, B4010 induced
removed after 30 minutes. The suspension was centrifuged at ,91% killing in 1 h at 16 MIC whereas ,97% killing was
2000rpm for 2 minutes; the pellet was washed twice with observed as the concentration was increased to 26 MIC.
phosphatebuffer,prefixedin0.5mlofamixedaldehydefixative IncreasingtheconcentrationofB4010to46and86MICvalues
(2% glutaraldehyde and 2% paraformaldehyde in 0.1 M sodium led to 4 log reduction (99.99%) in viable Candida cells within 20
cacodylatebuffer)for24hours.Thenthepelletwaswashedonce minutesand10minutes,respectively.Incontrast,natamycin,even
againinsodiumcacodylatebuffer(ElectronMicroscopySciences, at166MIC,required24htoachievesimilarendpoints.Wehave
Washington, USA) and the fungal cells suspension was subse- also compared the kill-kinetics of B4010, natamycin and ampho-
quently mounted on poly-L-Lysine coated cover-slips and post- tericinBagainsttheclinicalisolatesC.albicansDF2672Rstrains.At
fixed in 1% osmium tetroxide (Electron Microscopy Sciences). 46 MIC, B4010 elicited complete killing in 1 h whereas for
Following dehydration in a graded series of ethanol, the samples amphotericin B maximal effect was achieved in less than 8 h at
were critical-point-dried and sputter coated with 10nm of gold. 166 MIC. However, no fungicidal action was observed for
AllsampleswereviewedandphotographedonaFESEM(Supra natamycin at 26MICevenafter 24h exposure.
55VP - Zeiss) with an accelerating voltage of 3 kv at Carl Zeiss Next,weexaminedtheeffectsofphysiologicalconcentrationsof
facility, National University of Singapore. Candida incubated in metalions(NaCl,CaCl2andMgCl2)ontheantifungalactivityof
PBSwasconsideredaspositivecontrolandprocessedinthesame B4010.Thepresenceof135mMNaClamioleratedtheantifungal
way. activity, decreasing the MIC by 2-fold whereas physiological
concentrationsofdivalentcationsresultedinonlya2-foldincrease
in MIC values (Figure 3A). However, addition of 100mM KCl
Results
resulted inan increase inMICby a factor .16compared tothe
Antifungal Properties of Linear and Multivalent Peptides medium not supplemented with KCl. We have also determined
The peptide RGRKVVRRKK displayed weak antifungal the cation sensitivity on the candidacidal activity of B4010.
activity (Table 1) against C. albicans strains. However, upon AdditionofB4010(at5.5mM)resultedinacompletelossofviable
linking two copies of the sequence through a branched lysine cells in the presence of monovalent and divalent metal ions
(B2088, Figure 1), substantial decrease in MIC values were (Figure3B).However,inthepresenceofahighconcentrationof
observed as compared to the monomer or linear retrodimer KCl,slight decreaseinthecandidacidal activitywas observed.
(RGRKVVRRKKRRVVKRGR) peptides (Table 1). Further
assembling the two copies of branched dimer through branched Antifungal Activity of B4010 in Trypsin, Tear Fluid and
lysinesproducedatetrabranchedpeptide(B4010)whichdecreased Serum
the MIC by an additional 4–10 fold compared to the co-valent Antifungal properties of the peptide were examined in three
dimer (Table 1). The MIC values of B4010 (0.37mM) for two complexbiologicalenvironments.Thepeptidewasincubatedwith
clinicalisolatesofC.albicanswerealsolowerwhencomparedtothe trypsin (trypsin:peptide ratio=1:100) at 37uC. An aliquot of this
MICvaluesforamphotericinB(1.4mM)andnatamycin(15 mM), mixturewaswithdrawnat varioustimeintervals, addedtotheC.
thelatteristheonlyUSFDAapprovedantifungalforophthalmic albicansinoculumandthegrowthwasmonitored(Figure3C).No
applications[41].Interestingly,scramblingthesequenceofB4010 loss of antifungal activity was found up to 6 h incubation with
(Sc_B4010) led to 2–4-fold increase in MIC values, yet retained trypsin, suggesting improved proteolytic resistance of tetra-
significant potency. branched peptides. Mass spectrometry analysis indicated signifi-
CD spectra displayed strong negative minima around 200nm cant presence (,20%) of intact B4010 after 6 h incubation with
for both linear as well as branched dimers, suggesting branched trypsin (FigureS2).
Table1. MICofsynthetic linearand branched peptidesagainst various yeastsand fungi.
StrainsofYeasts/Fungi MICinmMof
Monomer LinearRetrodimer B2088 B4010 Sc_B4010 Natamycin
C.albicansATCC10231 78 22 5.5 1.4 5.6 15
C.albicansATCC24433 n.da 5.5 2.7 1.4 5.6 15
C.albicansATCC2091 n.d 11 1.4 0.7 2.7 7.5
C.albicansDF2672R 78 2.7 0.8 0.34 0.8 15
C.albicansDF1976R n.d. 2.7 0.8 0.34 0.8 15
F.solaniATCC36031 .78 n.d. 10.9 1.4 n.d. 4.7
F.solaniDM3782 n.d. n.d. n.d. 1.4 n.d. 2.4
F.solaniDF1500 n.d. n.d. n.d. 1.4 n.d. 2.4
aNotdetermined.
doi:10.1371/journal.pone.0087730.t001
PLOSONE | www.plosone.org 5 February2014 | Volume 9 | Issue 2 | e87730
MultivalentAntifungalPeptides
Figure1.Schematicrepresentationofthepeptidesusedinthisstudy.Thebranchedlysineiscoloredinblue.Forthescrambledpeptide
(Sc_B4010)eachcopycontainsthesequenceVRGRVRKR.
doi:10.1371/journal.pone.0087730.g001
The antifungal efficacy of B4010 was also monitored in the experiments were performed using S. cerevisiae strains that carry
presenceofcomplexbiologicalfluidssuchastearfluidandhuman specific mutations in the ergosterol (ergD) pathways. Table 2
serum. In the absence of tear fluid, about 9862% loss of viable summarizes the MIC values of B4010 against wild type and
cells were observed at a peptide concentration of 5.5mM. various ergD mutants which carry altered sterol structure and
However, the candidacidal activity was moderately suppressed compositions. Interestingly, all the ergD mutants displayed hyper
by the presence of 50% tear fluid at lower concentration susceptibility to B4010 i.e. 2–4 fold decrease in MIC values
(Figure3D).At22mM,however,nolossofactivitywasobserved compared tothewildtype strains.
suggesting a higher concentration of peptide was needed to
achieve more than 3 log reduction of viable C. albicans in the Effect of B4010 on Haemolysis, Cytotoxicity, Corneal
presenceoftearfluid.Theeffectofserumonantifungalefficacyof Reepithelialization Rate and in vivo Toxicity
B4010wasassessedbymeasuringtheshiftintheMICofB4010in The ability of B4010 in disrupting the membrane integrity of
thepresenceof5%humanserumagainsttwoclinicalisolatesofC. mammalian cells was evaluated by haemolytic assay using rabbit
albicans. For both the strains the MIC values were shifted from redbloodcells.Thepeptidehadnosignificanthaemolyticactivity
0.37 mM without serum to 5.5 mM in the presence of serum, (,1% haemolysis) to rabbit erythrocytes even at 440mM
suggesting 16-foldincrease intheMICs in5%sera. (Figure 4A). On the other hand amphotericin B and natamycin
displayedsignificanthaemolyticactivityatlowconcentrationand
Antifungal Activity of B4010 against ergD Strains the HC values were 59.467.8mM and 139.660.18mM,
50
Theemergenceofresistantpathogensposesagreaterthreatto respectively. Cytotoxicity was assessed after 24h exposure of
the management of fungal infections [5–7]. To determine if the humanconjunctivalepithelialcelllinestothepeptides/drugsand
sterolalterationshaveanaffectontheantifungalactivityofB4010, the EC (effective concentration of peptide that decreased the
50
Figure2.TimekillcurvesforB4010againstC.albicans.(A)ATCC10231and(B)clinicalisolateDF2672R.Theyeastscellswereincubatedwith
variousconcentrationsofpeptideorantifungalsinSDbroth.Attheindicatedtimes,survivorsweredilutedandplatedtoallowcolonycounts.
doi:10.1371/journal.pone.0087730.g002
PLOSONE | www.plosone.org 6 February2014 | Volume 9 | Issue 2 | e87730
MultivalentAntifungalPeptides
Figure3.AntifungalpropertiesofB4010inthepresenceofmetalionsandcomplexbiologicalfluids.(A)MICvaluesdeterminedinthe
presenceofmonovalentanddivalentmetalions.ThenumbersabovethebarsindicatethedeterminedMICvalues.(B)Candidacidalpropertiesof
B4010inthepresenceofmetalions.TheconcentrationofB4010was5.5mM.(C)Antifungalactivityinthepresenceoftrypsin.(D)Candidacidal
activityofB4010inthepresenceof50%rabbittearfluid.
doi:10.1371/journal.pone.0087730.g003
viable cells by 50%) was determined. Exposure of B4010 caused better than amphotericin B (134616mM, Figure S3). We have
50%lossofcellviabilityat220mM(Figure4B).TheEC value furtherexaminedwhethertopicalapplicationofB4010influences
50
of B4010 was comparable to natamycin (211.7620.5 mM) and the corneal wound healing invivo. The epithelial wound healing
Table2. MICofB4010 against S.cerevisiae mutants carrying alteredsterol structure and composition.
YeastStrains Relevantgenotype SterolContent(%)a MIC,mM
RH448 Wildtype Ergosterol(77%) 5.5
RH5812 erg2D Ergosta-8,24(28)-dienol(24.5%),Ergosta-8-enol(23.8%) 1.4
RH4213 erg3D Ergosta-7,22-dienol(45.9%) 1.4
RH5930 erg3Derg6D Zymosterol(40%),Cholesta-7,24-dienol(39.5%) 1.4
RH5873 erg4Derg5D Ergosta-5,7,24-trienol(72.2%) 1.4
RH3616 erg2Derg6D Zymosterol(85.6%) 2.8
RH5684 erg6D Zymosterol(39.4%),Cholesta-5,7,24-trienol(32.3%) 2.8
aTakenfrom[29].
doi:10.1371/journal.pone.0087730.t002
PLOSONE | www.plosone.org 7 February2014 | Volume 9 | Issue 2 | e87730
MultivalentAntifungalPeptides
rate was not altered by topical application of B4010 (22 mM) 3 which have intact cell membrane. However, in membrane-
times a day compared to the control groups (Figure 4C and compromised cells, the dye fluoresces strongly upon binding to
FigureS4).Thisisanimportantresultascornealepithelialwound intracellular nucleic acids [42]. Addition of B4010 to C. albicans
healingre-establishesacriticalcomponentofinnateimmunityfor incubatedwith1mMSGresultedinrapiduptakeofthedyewitha
theeye.Preliminaryinvivotoxicitytestsonmiceindicatednosign concomitantincreaseinfluorescenceintensityinaconcentration-
of mortality, morbidity or toxicity when injected by IP (200 mg/ dependentmanner(Figure5B).AtK6MIC(13%)and16MIC
kg) orIV(100 mg/kg)routes. values (24%), a weak permeation was observed. About 70% cells
were membrane compromised after the addition of B4010 at 26
Membrane Disrupting Activity of B4010 MIC in 10mins. However, a higher level of dye uptake (.90%)
To probe the mechanism of action, we first examined the was observed within 25mins when the concentration of B4010
affinity of B4010 for insoluble polysaccharides such as chitin and was increased to46MIC.
b-D-glucan, the major components of fungal cell walls. The Since yeasts maintain a negative resting membrane potential
peptide showed no affinity for cell wall polysaccharides as no co- insidethecell,wedeterminedifthecandidacidalactionofB4010
precipitation was observed with either chitin or b-D-glucan was due to dissipation of electrochemical gradients across the
(Figure 5A). Therefore, two complementary fluorometry assays membrane,usingapotentialsensitiveprobe,diS-C3-5.Figure5C
wereperformedtodetermineifB4010hadamembranetargeting shows the concentration-dependent dissipation of membrane
action, a characteristic feature of cationic antimicrobial peptides. potential upon addition of B4010 using C. albicans loaded with
Tocheckifthepeptidepermeabilizedthecytoplasmicmembrane, the potential sensitive probe. The time course of depolarization
we assayed uptake of SYTOX Green dye. This is a membrane occurred instantaneously upon addition of B4010, indicating
impermeabledyeanddoesnotfluorescewhenincubatedwithcells release of probe from the cells. However, addition of peptide 26
Figure 4. Toxicity and corneal reepithelialization of B4010. (A) Haemolytic activity of B4010 and other antifungals against 4% rabbit
erythrocytes. (B) CytotoxicityofB4010and twoalaninesubstitutedB4010to HCEcells.(C) EffectofB4010 oncornealreepithelializationof New
Zealandwhiterabbitsasmeasuredbyfluoresceinstaining.Thepeptideconcentrationwas22mM.
doi:10.1371/journal.pone.0087730.g004
PLOSONE | www.plosone.org 8 February2014 | Volume 9 | Issue 2 | e87730
MultivalentAntifungalPeptides
Figure5.EffectofB4010oncytoplasmicmembranepotential,membranepermeabilizationandmorphologyofC.albicans.(A)SDS-
PAGEshowinglackofaffinityofB4010forinsolublechitin.(B)SYTOXGreenuptakeofC.albicansinducedbyvaryingconcentrationofB4010.(C)
B4010-mediatedmembranedepolarizationmonitoredbydiSC 5assay.(D)B4010-inducedextracellularATPreleaseofC.albicans.Theinsetshowsthe
3
kineticsofATPrelease.B4010concentrationwas5.5mM.(E)SEMofuntreatedC.albicans.Scalebaris2mm(insetscalebar=200nm).(F)SEMimage
ofC.albicanstreatedwith5.5mMB4010.Scalebaris1mm(insetscalebar=100nm).
doi:10.1371/journal.pone.0087730.g005
PLOSONE | www.plosone.org 9 February2014 | Volume 9 | Issue 2 | e87730
MultivalentAntifungalPeptides
abovetheMIChadnosignificantinfluenceonthefinalmembrane B4010 in the presence of NaN is reversed by benzyl alcohol.
3
potential dissipation. To verify if the dissipation of membrane AdditionofB4010toC.albicanspretreatedwithNaN and60mM
3
potentialislinkedtocandidacidalactivity,wehavealsoperformed benzylalcohol resulted insignificant lossofviability (8567%).
the fungal viability assay under identical conditions without the TheeffectofCCCP/NaN onATPreleasebyB4010wasalso
3
probe. A complete loss of viability was observed when the investigated. Consistent with the cell viability assays, ATP
concentration of peptide exceeded 46 MIC (data not shown). bioluminescence assay indicated significant increase in the
Together with the SG uptake assay, these results confirmed that extracellular ATP levels in B4010-treated cells (Figure 6C).
B4010causedrapidmembranepotentialdissipationandpermea- However, cells pretreated with CCCP/NaN followed by the
3
bilizesthecytoplasmicmembraneofC.albicansinaconcentration- addition of B4010 resulted inprofound reduction inATP release
dependent manner. compared toB4010treatedcells (Figure 6C).
As the perturbation of membrane would affect its barrier
function and accompanied by release of essential intracellular EffectsofIon-channelInhibitorsonAntifungalActivityof
components, we have assessed the extracellular release of metal
B4010
ions and ATP upon challenging C. albicans with B4010. The
OurobservationsthatB4010causedrapidreleaseofK+aswell
background K+ and Ca2+ concentration in the supernatant were
asATPandthattheadditionofexternalK+decreasedtheextent
43.462 mMand2.560.5mM,respectively.AfterincubationofC.
of cell death with .16-fold increase in MIC at elevated ion
albicanswithB4010(5.5mM)for2 h,morethantwofoldelevation
concentrations prompted the question whether ion-channel
of potassium (10464.2mM) and calcium (5.861.3mM) concen-
inhibitors could protect C. albicans from B4010. We have tested
trationwasobserved,whereasnosignificantchangesinthelevels
of Na+ and Mg2+ ions were observed. The release of ATP was effectsofnon-specificorganiccationicion-channelinhibitors(TEA
and 4-AP) as well as yeast stretch-activated ion channel blocker
dependentontheconcentrationofpeptidewithamaximumATP
Gd3+ on candidacidal activity of B4010. The peptide caused
releasewasachievedabove46MICofB4010(Figure5D).At46
significant loss of viability in cells that were pretreated with TEA
MIC,ATPbioluminescenceassayfurtherindicatedarapidrelease
and 4-AP (Figure 6B). Consistent with the loss of viability, a
of ATP (,30 minutes) from C. albicans upon addition of B4010
similaramountofATPwasreleasedfromthecellspretreatedwith
(Figure 5Dinset).
TEAand4-APafterB4010additionaswasobservedwithoutthese
Morphological changes in C. albicans treated with B4010 were
inhibitors (Figure 6C). On the other hand, cells incubated with
investigated by scanning electron microscopy. Untreated cells
Gd3+providedpartialprotection(5466%lossofviability)against
displayed round, dome-shaped morphologies with a smooth
B4010-induced killing. Addition of B4010 to C. albicans cells
surface (Figure 5E). However, the cell surface of C. albicans
pretreated with the anion channel inhibitor, NPPB, conferred
incubated with B4010 for 30mins revealed severe damage with
significant protection (1264% loss of viability) of C. albicans from
bud scars appearing on the surface (Figure 5F). In certain cells
B4010 (Figure 6B). ATP release assays further confirmed the
release of irregular material was also visible (Figure 5F inset),
absence or reduced ATP release from additives that conferred
suggesting membrane-lytic actionof B4010.
complete orpartial protection(Figure 6C).
To determine if the loss of B4010 activity in the presence of
EffectsofProtonUncouplersandMetabolicInhibitorson
NPPBwasassociatedwithalterationsinthemembranepotential,
the Antifungal Activity of B4010
weperformeddiS-C -5assayinthepresenceofadditivesfollowed
To study the effect of metabolic activity of C. albicans on the 3
by the addition of B4010. TEA or 4-AP did not alter the
susceptibility to B4010, the diS-C -5 loaded cells were incubated
3 membrane potential whereas subsequent addition of B4010
with 5 mM CCCP (an uncoupler of proton gradients) or 5mM
resulted in complete loss of membrane potential and the
NaN (which blocks both classical and alternative pathways of
3 magnitude of intensity changes was similar to the one observed
mitochondrial inhibition) and monitored the changes in fluores-
in cells without prior addition to 4-AP or TEA (Figure 6D).
cence intensity upon addition of B4010. Addition of CCCP When Gd3+ was added to dye-loaded C. albicans, a significant
resulted in a strong reduction in fluorescence intensity indicating
depolarization was observed. Further addition of B4010 caused
collapseofthemembranepotential.SubsequentadditionofB4010
weak dissipation of membrane potential, augmenting the fungal
had weak changes in the transmembrane potential (Figure 6A).
viabilityandATPreleaseassays(datanotshown).Asimilareffect
As shown before by Veerman etal., addition of NaN to dye-
3 wasobservedwhen100mMKClwasused(FigureS5).C.albicans
loadedcellscausedweakdepolarization[43].Subsequentaddition
incubated with NPPB resulted in collapse of membrane potential
of B4010 resulted in an increase in fluorescence intensity
without affecting the fungal viability. Further addition B4010 to
(Figure 6A). However, ,25 fold reduction in fluorescence
NPPB-treated cells prevented dissipation of membrane potential
intensity was observed in the azide-treated cells compared to
(Figure6D)andabrogatedthecandidacidalpropertiesofB4010.
control cells. Removal of NaN by centrifugation reversed the
3 The dye release assays in the presence of various additives
intensity change, indicating that the energy poison inhibits the
suggested that a resting negative membrane potential and
candidacidal activity reversibly. These results suggested that both
metabolicactivity are crucialforcandidacidal activity of B4010.
proton uncoupler and energy poison significantly decreased the
B4010-induced depolarization of thecytoplasmic membrane.
B4010 Induced Calcein Leakage from Phospholipid
The dependence of candidacidal activity on the energy
metabolism was further assessed by fungal viability and extra Vesicles
cellular ATP release assays. B4010 (5.5mM) added to C. albicans The ability of B4010 to induce membrane perturbation in
resulted in 9862% loss of viable cells (Figure 6B). However, phospholipidvesiclesthatdidnotcontainproteinswasinvestigated
addition of B4010 to cells preincubated with 5 mM CCCP or by calcein release from SUVs containing various lipid composi-
5 mM NaN resulted in only 4162.5% or 662% loss of viable tions. Two lipid mixtures, PC containing cholesterol and
3
cells, respectively (Figure 6B). When tested under identical PC:PE:PI/PScontainingvariousamountsofergosterolwereused
conditions without B4010, both the additives did not impair the tostudytheabilityofB4010toinducemembranelysis.About60–
viabilityofC.albicans.Interestingly,thelossofantifungalactivityof 70% calcein-release was observed from PC:PE:PI SUVs contain-
PLOSONE | www.plosone.org 10 February2014 | Volume 9 | Issue 2 | e87730