University of Groningen Penetratin and derivatives acting as antifungal agents Masman, Marcelo F.

18 Pages·2008·1.44 MB·English

University of Groningen

Penetratin and derivatives acting as antifungal agents
Masman, Marcelo F.; Rodriguez, Ana M.; Raimondi, Marcela; Zacchino, Susana A.; Luiten,
Paul G. M.; Somlai, Csaba; Kortvelyesi, Tamas; Penke, Botond; Enriz, Ricardo D.
Published in:
European Journal of Medicinal Chemistry
DOI:
10.1016/j.ejmech.2008.02.019
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from
it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date:
2009
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Masman, M. F., Rodriguez, A. M., Raimondi, M., Zacchino, S. A., Luiten, P. G. M., Somlai, C., Kortvelyesi,
T., Penke, B., & Enriz, R. D. (2009). Penetratin and derivatives acting as antifungal agents. European
Journal of Medicinal Chemistry, 44(1), 212-228. https://doi.org/10.1016/j.ejmech.2008.02.019
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the
author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license.
More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne-
amendment.
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately
and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the
number of authors shown on this cover page is limited to 10 maximum.
Download date: 17-03-2023
Available online at www.sciencedirect.com
EuropeanJournalofMedicinalChemistry44(2009)212e228
http://www.elsevier.com/locate/ejmech
Original article
Penetratin and derivatives acting as antifungal agents
Marcelo F. Masman a,b,c, Ana M. Rodr´ıguez a,b, Marcela Raimondi d, Susana A. Zacchino e,
Paul G.M. Luiten c, Csaba Somlaif, Tamas Kortvelyesi g,
Botond Penke f, Ricardo D. Enriz a,b,*
aFacultaddeQu´ımica,Bioqu´ımicayFarmacia,UniversidadNacionaldeSanLuis,Chacabuco915,5700SanLuis,Argentina
bIMBIO-SL,CONICET,UNSL,Chacabuco915,5700SanLuis,Argentina
cDepartmentofMolecularNeurobiology,CentreforBehaviourandNeurosciences,UniversityofGroningen,Kerklaan30,9751NNHaren,TheNetherlands
dMicrobiolog´ıa,FacultaddeMedicina,UniversidadNacionaldeRosario,SantaFe3100,2000Rosario,Argentina
eFarmacognosia,FacultaddeCienciasBioqu´ımicasyFarmace´uticas,UniversidadNacionaldeRosario,Suipacha531,2000Rosario,Argentina
fDepartmentofMedicalChemistry,UniversityofSzeged,Do´mter8,H-6720Szeged,Hungary
gDepartmentofPhysicalChemistry,UniversityofSzeged,RerrichSq.1,H-6720Szeged,Hungary
Received26December2007;receivedinrevisedform8February2008;accepted11February2008
Availableonline29February2008
Abstract
Thesynthesis,invitroevaluation,andconformationalstudyofRQIKIWFQNRRMKWKKeNH (penetratin)andrelatedderivativesacting
2
asantifungalagentsarereported.Penetratinandsomeofitsderivativesdisplayedantifungalactivityagainstthehumanopportunisticpathogenic
standardizedATCCstrainsCandidaalbicansandCryptococcusneoformansaswellasclinicalisolatesofC.neoformans.Amongthecompounds
tested,penetratinalongwiththenonapeptideRKWRRKWKKeNH andthetetrapeptideRQKKeNH exhibitedsignificantantifungalactivities
2 2
against the Cryptococcus species. A comprehensive conformational analysis on the peptides reported here using three different approaches,
molecularmechanics,simulatedannealingandmoleculardynamicssimulations,wascarriedout.Theexperimentalandtheoreticalresultsallow
us to identify a topographical template which may provide a guide for the design of new compounds with antifungal characteristics against
C. neoformans.
(cid:2)2008Elsevier MassonSAS. Allrightsreserved.
Keywords:Penetratin;Cationicpeptides;Cryptococcosis;Candidainfections;Moleculardynamics;Conformationalanalysis
1. Introduction patients [1,2]. Invasive fungal infections as well as dermato-
mycoses produced by fungal organisms with even low viru-
Fungal infections are a persistent major health problem, lence can be life threatening [3] for individuals with
which especially affect and threaten immunocompromised increased vulnerability such as neonates, cancer patients re-
ceivingchemotherapy,organtransplantpatients,andburnspa-
tients, apart from those with acquired immunodeficiency
Abbreviations: AIDS, acquired immunodeficiency syndrome; CPP, cell- syndrome (AIDS). Other risk factors include corticosteroid
penetrating peptide; SA, simulated annealing; MD, molecular dynamics;
and antibiotic treatments, diabetes, lesions of epidermis and
MIC,minimuminhibitoryconcentration;MFC,minimumfungicideconcen-
dermis, malnutrition, neutropenia and surgery [2]. Many fun-
tration; EDMC, electrostatically driven Monte Carlo; RMSD, route mean
squaredeviation;R,radiusofgyration;RMSF,rootmeansquarefluctuation; gal infections are caused by opportunistic pathogens that
g
SASA,solventaccessiblesurfacearea;TFEd2,trifluoroethanol-d2;MEPs,mo- may be endogenous or acquired from the environment (Can-
lecularelectrostaticpotentials. dida, Cryptococcus, Aspergillus infections). Patients with sig-
* Correspondingauthor.FacultaddeQu´ımica,Bioqu´ımicayFarmacia,Uni-
nificant immunosuppression frequently develop Candida
versidad Nacional de San Luis, Chacabuco 915, 5700 San Luis, Argentina.
esophagitis.Cryptococcosis,causedbytheencapsulatedyeast
Tel.:þ542652423789;fax:þ542652431301.
E-mailaddress:[email protected](R.D.Enriz). Cryptococcus neoformans, has been the cause of fungal
0223-5234/$-seefrontmatter(cid:2)2008ElsevierMassonSAS.Allrightsreserved.
doi:10.1016/j.ejmech.2008.02.019
M.F.Masmanetal./EuropeanJournalofMedicinalChemistry44(2009)212e228 213
mortality among HIV-infected patients. This organism has amino acid homeodomain of the Antennapedia protein of
a predilection for the central nervous system and leads to se- Drosophila was able to translocate over cell membranes. In
vere, life-threatening meningitis. In addition, an increasing order to understand the driving force for the internalization,
number of normal individuals, including children in third- the homeodomain was modified by site-directed mutagenesis
worldnations[4]thatsufferdeficientsanitationandeducation, leading to the discovery that its third helix was necessary
are prone to fungal infections, especially those involving the and also sufficient for membrane translocation, which re-
skin and mucosal surfaces. sulted in the development of a 16 amino acid-long CPP
Although it appears that many drugs are available for the called penetratin (1) [19]. Thus, peptide 1, a synthetic 16
treatment of systemic and superficial mycoses, there are in amino acid peptide from the third helix of Antennapedia ho-
fact only a limited number of effective antifungal drugs [1]. meodomain [19,20], is a cationic amphipathic peptide which
Many of the antifungal compounds currently available have might penetrate cell membranes via a proposed ‘‘inverted
undesirableeffectsorareverytoxic(amphotericinB);arefun- micelle’’ pathway. However, the mechanism of membrane
gistaticandnotfungicidal(azoles),orleadtothedevelopment translocation is not well known. The question is whether
of resistance, as with 5-fluorocytosine [5]. Amphotericin B, the peptide is internalized via endocytosis which is energy-
developedinthe1950s,stillremainsasawidelyusedantifun- dependent or via direct transport, while the latter mechanism
gal drug, most recently gaining renewed applications through is scarcely known at present, it is believed that the process is
lipidbasedformulations.AccordingtoPolak[6]idealdrugsto non-receptor mediated [20,21]. In addition, we previously
cure fungal infections have not been discovered yet. In the provide evidence on the energy-dependent and lipid raft-me-
meantime, resistance to currently available antifungal agents diated endocytic uptake of penetratin [22]. Peptide 1 has
continues to grow [7]. Although combination therapy has been proposed as a universal intracellular delivery vehicle
emerged as a good alternative to bypass these disadvantages [23]. Since 1 possesses 16 amino acids and a charge of
[6,8], there is an urgent need for a next generation of safer þ8, it mightbeincludedinthegeneralclassificationof‘‘cat-
and more potent antifungal agents [1,8]. These explorations ionic antimicrobial peptide’’. To the best of our knowledge
have resulted in the identification of novel molecules, which this is the first study reporting the antifungal activity of 1
couldprovepromisingforfurtherfuturedevelopment.Among and structurally related derivatives.
them, some natural peptides were recently described as The aim of the present investigation is exploring the
antifungal compounds, inhibiting a broad spectrum of fungi antifungalpotentialof1anditsderivativesagainstCandidaal-
[9e11]. It has also been reported that a group of cationic an- bicans and C. neoformans. To better characterize the struc-
timicrobial peptides are major players in the innate immune tureeantifungal activity relationship of peptide 1 and related
response [12,13]. These peptides are very ancient elements compounds under study the present analysis explored influ-
of the immune response of all species of animal and plant encesofamino acidsubstitutionsanddeletionsonitsantifun-
life,andtheinductionpathwaysforthesecompoundsinverte- gal activity. In addition, an extensive conformational analysis
brates, insects and plants [12e14] are highly conserved. of1anditsderivativeswascarriedoutusingthreedifferentap-
Furthermore, it is becoming increasingly clear that cationic proaches:molecularmechanics,simulatedannealing(SA)and
antimicrobial peptides play manypotential roles in inflamma- molecular dynamics (MD) simulations. The ability of each
tory responses, which represent an orchestration of the mech- method to obtain the different conformations is tested and
anisms of innate immunity. compared.Conformationalandelectronicstudieswerecarried
Small cationic peptides [15,16] are abundant in nature and outinordertoidentifyatopographicaland/orasub-structural
havebeendescribedas‘‘nature’santibiotics’’or‘‘cationican- template,whichmaybethestartingstructureforthedesignof
timicrobial peptides’’. These peptides are 12e50 amino acids new antifungal compounds.
longwithanetpositivechargeofþ2orþ9,whichisduetoan
excess of basic arginine and lysine residues, and approxi- 2. Results and discussion
mately 50% hydrophobic amino acids [15]. These molecules
arealsofoldedinthreedimensionssothattheyhavebothahy- 2.1. Antifungal activity
drophobic face comprising non-polar amino acid side chains,
and a hydrophilic face of polar and positively charged resi- Toevaluatetheantifungalpotential,concentrations ofpep-
dues:thesemoleculesareamphipathic.Despitethesetwosim- tides up to 200mM were incorporated in the growth media
ilarities these compounds vary considerably in length, amino accordingtopreviously reportedprocedures[11,24,25].Com-
acid sequence and secondary structure. The different spatial pounds producing no inhibition of fungal growth at 200mM
orderings include small b-sheets stabilized by disulphide wereconsideredinactive.Table1givestheantifungalactivity
bridges, amphipathic a-helices and, less commonly, extended obtainedforpeptide1againstC.albicans andC.neoformans.
and loop structures. Peptide 1 displayed a significant antifungal activity against
Cell-penetrating peptides (CPPs) are defined as peptides both fungi being C. neoformans the more sensitive species.
with a maximum of 30 amino acids, which are able to enter It is interesting to note that 1 displayed a significant degree
cells in a seemingly energy-independent manner, thus being ofinhibitionagainstC.neoformansevenatlowconcentrations
able to translocate across membranes in a non-endocytotic (90% of inhibition was observed at 12.5mM and 100% at
fashion [17]. In 1991, Joliot et al. [18] reported that the 60 25mM). The inhibitory effect observed against C. albicans
214 M.F.Masmanetal./EuropeanJournalofMedicinalChemistry44(2009)212e228
M 3 3 was slightly lower than that obtained for C. neoformans at
m 2. 2.
2.5 0(cid:2) 2(cid:2) 0000 12.5mM but similar at 25e200mM.
1 90000003 11 In order to gain insight into the spectrum of activity, pep-
tide 1 was tested against a panel of clinical isolates of
2978801489 4 C. neoformans. The minimum inhibitory concentration (MIC)
M 0.(cid:2)2.(cid:2)1.(cid:2)3.(cid:2)2.(cid:2)2.(cid:2) 7.(cid:2) values of 1 were determined against this new panel by using
m25 10028169610062 100100 threeendpoints:MIC100,MIC80andMIC50(theminimumcon-
d) centrationofcompoundsthatinhibit100,80and50%ofgrowth,
ol
b respectively).Theapplicationofalessstringentendpointsuch
n 2253
wni M 0(cid:2)3.4(cid:2)1.3(cid:2)2.3(cid:2)4.3(cid:2)6.7(cid:2) 3.6(cid:2) asMIC80andMIC50hasbeenshowntoconsistentlyrepresent
ho m 064319 2 00 theinvitroactivityofcompoundsandmanytimesprovideabet-
s 0 031111 9 00
e 5 1 0 11 ter correlation with other measurements of antifungal activity
ar
0 such as the minimum fungicide concentration (MFC) [26]. In
6
erthan mans mM 0(cid:2)5.7(cid:2)21.22(cid:2)3.92(cid:2)4.79(cid:2)2.09(cid:2) 4.0(cid:2) aadgdaiintisotnthtiospMaInCeldweatesramcicnoamtipolniss,hethdebeyvsauluba-ctiuolnturoifngMaFsCamopfl1e
nshigh neofor 100 10092681133140100 100100 oagfamrepdlaiatesfr.oSmo,MpeIpCtidtueb1eswsahsotwesitnegdnagoaginrostw1t0h,colinntiocadlriusgo-lafrteees
o s ofC.neoformans,allprovidedbytheMalbra´nInstitute(Buenos
ofinhibiti yptococcu m0M 00.1(cid:2)62.7(cid:2)91.68(cid:2)31.51(cid:2)00.36(cid:2)02.94(cid:2) 01.2(cid:2) 00 Asimireilsa)r.Ttoh,eoserrleoswueltrstahraens,htohwosneinobTtaaibnleed2aagnadinthstetahcetisvtiatyndwarads
% Cr 20 10997106010 1010 strain(ATCC32264).
4( Peptide 2 possesses the same amino acids of peptide 1 but
6
22 M 1 0 located in a completely different sequence. In fact, the se-
ATCC3 m12.5 40.(cid:2)0151.(cid:2)00000 100100 qshuoewnceedonfothaisntpifeupntigdaelwacatsivriatnydoamgaliynsgtenbeortahtefdu.nTghiistepsteepdtidaet
s 12.5mMbutinhibited96and60%ofthegrowthofC.neofor-
n
a
m 59 mans and of C. albicans at 200mM, respectively. It is clear
r 6 77
neofo m5M 11.(cid:2) 01.(cid:2)33.(cid:2) 0000 twhhaetrtehaesa1ntisfhuonwgaeldacativsiigtynifiofca1ntanadnt2ifuinsgmalarakcetdivlyitydifafgeareinnstt,
s 2 90420000 11
u
cc bothC.albicansandC.neoformans;peptide2waspractically
o
oc 12 inactive.Onthebasisofthese resultsitcan beconcluded that
ypt M 1.2 0.77.4 the sequence as well as the different spatial orderings of the
dCr m0 5(cid:2) 8(cid:2)3(cid:2) 0000 cationic, polar and hydrophobic residues are important deter-
n 5 90530000 11
a minants for the antifungal activity. In contrast, the positive
1
3
2 charge (þ8) of 1 appears to be a necessary requirement but
0 705
1 6090540
C M 0.3.0.7.0.0.1. not by itself sufficient to produce the antifungal response.
ATC m00 00(cid:2)52(cid:2)76(cid:2)44(cid:2)13(cid:2)12(cid:2)2(cid:2) 0000 To study the structureeantifungal activity relationship on
Candidaalbicans Candidaalbicans m200M1 10010.2(cid:2)604.2(cid:2)1000.08(cid:2)611.16(cid:2)291.05(cid:2)434.96(cid:2)92.3(cid:2)61.20(cid:2) 10011001 tTMmhauiebnmlsiemef2uunmpgeiicpnidhtiaibdliectoosrn,ycectnhotnreacteionentfraf(eMticoFntCss)(oMoffIpCe1ns0et0rt,ruaMctitInuC(r81a0)laagnadcihnMastnICcgl5ien0)sicaanlidnisomlatihtneies-
ofC.neoformans
st
n
agai H2H2 Vspoeuccihmeern MIC100 MIC80 MIC50 MFC AMmICph.B Itz.MIC100
s NN 100
e ee
Table1Antifungalactivity(%inhibition)ofpeptid PeptideSequence 1RQIKIWFQNRRMKWKK2WKQKNIKWRFRQKMIRe3RKWRRKWKKNH2e4RKFRRKFKKNH2e5RKRRKWKKNH2e6RKRRKKKNH2e7KWKKNH2e8RQKKNH2 aAmph.BbKet. aAmphotericinB.bKetoconazole. omBIIIIIIIIIIMMMMMMMMMMMre¼I5Cn99090900900:a18748073365%0m32233211120,0700176794pIr4273953057hMMe044611610odIt¼uCecr8Mit0111cioi3666666222aann1222222555nlb.......dBo2555555raf;MnItthIzCe.5I3666336333¼g0n1222112111:rsio..........tc2555522222twirotauntchtcoeencnao3333313113tznr1111161661(oatB.......rtl2222222ioeuo.len,nrooesfs111pa3666666222ecc1222222555Aot.......i2555555imvreepsloy,0000000000.u..........nW1021212205dA3653535560ittrhhgiaentntcvianouaus<<<<<<<)ce;h0000000000de..........2250000000r1A55011111110s55555550mp,epc8hi0-.
M.F.Masmanetal./EuropeanJournalofMedicinalChemistry44(2009)212e228 215
sequenceofpeptide1wereconsidered.Ourprincipalgoalwas peptide 1 is of paramount importance. Linear peptides are
tosynthesizeshorterderivativesof1tryingtomaintainthean- highlyflexibleandthereforetodeterminethebiologicallyrel-
tifungal activity as much as possible. In the first step we syn- evant conformations is a matter of high complexity, which
thesized compound 3, a nonapeptide possessing the last four requires an exhaustive conformational analysis of these struc-
amino acids of 1. In this peptide we maintain the same num- tures. Consequently, we carried out calculations using three
bers of cationic amino acids (Arg and Lys) deleting Gln-2, different approaches: electrostatically driven Monte Carlo
Ile-3, Ile-5, Phe-7, Gln-8, Asn-9 and Asn-12. Peptide 3 dis- (EDMC) calculations implemented in the ECCEPAK [29]
played a lower antifungal activity with respect to 1. The anti- package, SA calculations using the Tinker Molecular Model-
fungal activity against C. neoformans and C. albicans was ling package [30] and MD simulations from the GROMACS
moderately effective but still significant. We decided to per- program [31,32].
form changes on peptide 3 and then we synthesized peptides
4e6.Inpeptide4wereplacedthetwoTrpresiduesofpeptide 2.2.1. EDMC results
3(Trp-3andTrp-7)byPhe.Thisstructuralchangeyieldedare- EDMC results are summarized in Table 3 and Fig. 1 and
duction of antifungal activity (compare the % of inhibition of more details are given in Tables SIeSVIII in Supplementary
3 and 4 in Table 1), which is not an unexpected result; a role material. Calculations yielded a large set of conformational
for Trpastranslocation determinantofpeptideshasbeen pro- familiesforeachpeptidestudied.Thetotalnumberofconfor-
posed [27] andmutationof both tryptophans in peptide 1 was mationsgeneratedforeachpeptidevariedbetween47391and
found to abolish internalization [19]. In addition, it has previ- 129922,andthenumberofthoseacceptedwas5000forallthe
ouslybeenreportedthattryptophansarepoorlyreplaceableby cases. In the clustering procedure, an RMSD (root mean
phenylalanine in 1 and derivativeswhen tested for their pene- square deviation) of 0.75A˚ and a cutoff of 30kcalmol(cid:3)1
trating properties [28]. Our results lend support to previously were used. The number of families after clustering varied be-
reported findings, but in addition demonstrate the antifungal tween 137 and 1001. The total number of families accepted
activity of these peptides. Octapeptide 5 was obtained by de- with a relative population higher than 0.20% varied between
leting Trp-3 from peptide 3; in turn heptapeptide 6 was ob- 11and86.Theirpopulationssumuptoca.80%ofallconfor-
tained by deleting Trp-3 and Trp-7 from peptide 3. Whereas mations in each case (see Table 3).
octapeptide 5 displayed only a marginal antifungal activity, All low-energy conformers of the peptides studied here
peptide 6 was practically ineffective in comparison to their were then compared to each other. The comparison involved
congeners. the spatial arrangements, relative energy and populations.
In order to further understand the above experimental re- A total of 639 different families were found for peptide 1.
sults,weperformedaconformationalstudyofthepeptidesre- However, 82.92% of total population of this peptide corre-
ported here using different approaches. sponded to only 11 families (Table 3). It is interesting to
note that the energetically preferred family comprises
2.2. Conformational study of peptide 1 and derivatives 70.48% of the entire population. Thus, this family which
adopts an a-helix structure is the most representative form
Alargenumberofstudieshavebeen performed inorderto of this molecule. This conformation is characterized by stabi-
shed light on the structural aspects and mechanism of action lizinghydrogenbondsbetweenthecarbonylicoxygen(residue
for translocation of 1. However, compared to these mechanis- i) and the NH group (residue iþ4). The first and the last res-
ticproperties,theconformationalintricaciesofthiscompound idues do not present a stable structure. A spatial view of this
havereceivedrelativelylittleattention.Itis,however,obvious conformationisshowninFig. 1a.The second mostpopulated
that a better understanding of the conformational behavior of family (7.46%) corresponds to a structure possessing the first
Table3
Selectedconformationalsearchandclusteringresultsforpeptides1e8optimizedattheEDMC/SRFOPT/ECCEP/3leveloftheory
Peptide Generateda Acceptedb #Fc #F d %Pe
0.20%
Electrostatical Random Thermal Total Electrostatical Random Thermal Total
1 8973 119535 575 129083 1431 3245 324 5000 639 11 82.92
2 9200 120229 493 129922 1349 3373 278 5000 703 19 81.92
3 8050 107710 304 116064 1372 3412 216 5000 270 6 88.94
4 7490 102380 245 110115 1121 3697 182 5000 288 11 88.72
5 7905 106150 294 114349 1176 3606 218 5000 242 11 89.76
6 7191 98483 213 105887 1192 3635 173 5000 137 6 91.02
7 3007 44352 32 47391 508 4466 26 5000 505 74 82.82
8 3939 54133 54 58126 579 4379 42 5000 481 45 83.34
a Numberofconformationsgeneratedelectrostatically,randomlyandthermallyduringtheconformationalsearch.
b Numberofconformationsacceptedfromthosegeneratedelectrostatically,randomlyandthermallyduringtheconformationalsearch.
c #Frepresentsthetotalnumberofconformationalfamiliesasresultoftheclusteringrun.
d #F representsthenumberofconformationalfamilieswithpopulationsabove0.20%.
0.20%
e %Prepresentsthesumofthepercentrelativepopulationof#F .
0.20%
216 M.F.Masmanetal./EuropeanJournalofMedicinalChemistry44(2009)212e228
To better characterize the peptide spatial orientations, we
plotted Edmundson wheel representations of peptides 1e6
(Fig. 2). The representation obtained for peptide 1 displays
two clearly differentiated faces: the ‘‘charged face’’ (denoted
in dashed blue line in Fig. 2) and a more extended ‘‘non-
chargedface’’(denotedinfullgreenline).Thefirstfaceiden-
tifiesresiduesR11,K4,K5andR1asthoseaccountingforthe
mutualcoulombicbinding.Thefirstthreeresiduesarelocated
on the same side of the helical peptide and hence we desig-
nated it as the ‘‘charged face’’. These positively charged resi-
duesareabletoproducesaltbridgeswiththehydrophilicpart
of the lipids. The non-charged face is more extensive and is
formed by six hydrophobic (M12, I5, W6, I3, W14 and F7)
and two polar residues (N9 and Q2). However, it should be
noted that 1 displays a homogeneously distributed remainder
of the positively charged residues. Thus, residues K16, K13
andK10 are strategically intercalated along the ‘‘non-charged
face’’. This is a striking difference with respect to peptide 2,
whichdisplaystwo‘‘chargedfaces’’whereallthecationicres-
iduesareconcentrated.Inpeptide2therearetwonon-charged
faces; however, it should be noted that even adding the two
non-chargedfacesofpeptide2,thisportionismarkedlylower
with respect to the only non-charged face obtained for 1. The
Edmundsonwheelrepresentationsobtainedforpeptides3and
4areverysimilardisplayingaveryextensivechargedfaceand
Fig. 1. Spatial view of the preferred forms obtained for peptide 1. (a) The a markedly reduced non-charged face. Peptide 5 in turn gives
global minimum (a-helix structure) and (b) the second more populated only a minimal non-charged zone corresponding to the W6
conformation.
residue and the rest is ‘‘charged face’’. Obviously, peptide 6
displays a completely charged face because only cationic res-
iduesformit.PreviouslyLensinketal.[33]reportedthataho-
fourresidueswithoutastablestructure;residues5e7inaturn mogeneous distribution of positively charged residues along
structure and residues 10e15 with a typical a-helix structure. the axis of the helical peptide, and especially K4, R5, and
The residues 8 and 9 present a bend structure connecting the K11 contribute to the association of peptide 1 with lipids.
turn moiety with the a-helix portion. The last residue does Our EDMC results are in a complete agreement with those
not display a stable form (Fig. 1b). However, this family has MD simulations. In addition a very good correlation between
an energy gap of 22.61kcalmol(cid:3)1 with respect to the global the antifungal activities and the potential penetrating proper-
minimum. ties of these peptides are particularly striking.
For peptide 2 a total of 703 different families were ob-
tained, from which 19 families comprise 81.92% of the total 2.2.2. SA results
population. The most populated family (61.06%) presents an The initial structure of peptides 1e6 was extended. The
a-helixconformationfromresidues3to15.Thisconformation secondary structures of the lowest energy conformers calcu-
hasanenergygapof1.27kcalmol(cid:3)1withrespecttotheglobal lated by DSSP program [34] are summarized in Table 4.
minimum,which has6.04% ofpopulation. The lowestenergy The best structure of peptide 1 contains bends (I3eW6), 3 -
10
conformation possesses the following structure:from residues helix (F7eN9) and b-turn (R10eR11, K13eK15) and bends
2 to 5 in a turn structure; residue 6 in a bend form and from (F7eN9, R11eK13) in AMBER99 and OPLS-AA force field
residues 7 to 15 in a typical a-helix structure. Residues 1 calculations,respectively.Hierarchicalclusteringshowsbend-
and16donotshowanystablestructure.Ingeneraltheconfor- ingand helical backbone structures for peptide 1 for the most
mationalbehaviorof1and2iscomparable.However,inpep- representative clusters. OPLS-AA calculations, as in peptides
tide1themostpopulatedfamily(atypicala-helixstructure)is 2e6, predict H-bonds in other positions than AMBER99.
alsotheenergeticallypreferredone.Incontrast,forpeptide2, Four H-bonds were formed between O(i) and HeN(j), two
the fully a-helix structure is not the most preferred form. H-bonds between O(i) and HeN(iþ3), one H-bond between
Compounds 3e6 display a closely related conformational O(i) and HeN(iþ2), and O(i) and HeN(iþ4) (AMBER99
behavior preferring a helical structure for the most populated results). The results of OPLS-AA calculations predict eight
families. Peptides 3e6 are somewhat more rigid with respect H-bonds formed between O(i) and HeN(j), in peptide
to1.This fact might beappreciatedcomparingthe total num- 1 four H-bonds between O(i) and HeN(iþ2) and two H-
ber of conformational families obtained for each compound bonds between O(i) and HeN(iþ4), one between O(i) and
(Table 3). HeN(iþ3). In peptide 2, bend (Q3eK4, I6, Q12eK13) and
M.F.Masmanetal./EuropeanJournalofMedicinalChemistry44(2009)212e228 217
CF
CF
(1) NCF (2) NCF
NCF
CF
NCF NCF
(3) (4)
CF CF
CF CF
(5) (6)
NCF
Fig.2.Edmundsonwheelrepresentationsofpeptides1e6.Thenumberinthecenterofthewheelcorrespondstothepeptidenumber.The‘‘charged’’(CF)and
‘‘non-charged’’(NCF)facesareshowninbluedashedlinesandfullgreenlines,respectively.Positivelychargedaminoacidsaredenotedwithbluedots,thepolar
oneswithorange,andthehydrophobiconeswithyellow.(Forinterpretationofthereferencestocolourinthisfigurelegend,thereaderisreferredtotheweb
versionofthisarticle.)
b-turn (W8eF10, M14eI15) alternate. Almost all of the bend in Supplementary material). The root mean square fluctuation
structures remained (Q3eN5, F10, K13) and a-helix formed (RMSF) of the backbone atoms (Fig. S5 in Supplementary
(I6eR9) which includes H-bonds. Peptides 3e6 contain a-he- material) and the hydrophilic and hydrophobic solvent acces-
lix, 3 -helix and b-turn in the central residues in AMBER99 siblesurfaceareas(SASAs)werealsocalculated.Thesecond-
10
results. OPLS-AApredictsbendstructuresalmost atthesame ary structures of peptides were analyzed by sampling
residues. trajectories every 10ps with the DSSP program [34].
In all peptides simulated here, the initial 3 -helix was de-
10
2.2.3. MD simulations stroyed in the first 50e100ps. The RMSD and the RMSF of
Inthetrajectoryanalysisofpeptides1e6,thetotalandpo- the backbone during the simulation characterize this change
tential energies, radius of gyration (R ) and the RMSD of the in their secondary structure. The relative small change in the
g
backbone (NeC eC ) atoms related to the structure at RMSDofthepeptidesinthetrajectoryisevidenceforthesta-
a (carbonyl)
theendofequilibration(100ps)werecalculated(Figs.S1eS4 bilization of the backbone structure. RMSD for simulations
218 M.F.Masmanetal./EuropeanJournalofMedicinalChemistry44(2009)212e228
Table4 initial conformation of 3 -helix was destroyed and a-helix,
10
Secondarystructuresofthebestconformationofpeptides1e8obtainedfrom b-turn/bend and a stable random meander structures at the
simulated annealing calculations by using AMBER99 and OPLS-AA force
N-andC-terminalregionsfluctuatedduringthewholesimula-
fields(FFs)
tion.Theresidues5e10haveshownthehighestpreferenceof
Peptidea FF Secondarystructureb
3 -helix conformation. Here, in peptide 2, the b-turn and
10
Residuenumber bend conformations were mainly formed at residues 2e4
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 and 11e15. Also in this simulation the initial and final resi-
1 AMBER99 _ _ S S S S G G G T T _ T T T _ duesappeartohaverandomcoilstructurebecauseoftheflex-
OPLS-AA _ _ _ _ _ _ S S S _ S S S _ _ _ ibility of these residues. A mixture of coil, bend and b-turn
2 AMBER99 _ _ S S _ S _ T T T _ S S T T _ conformationswasformedafter86nsofsimulation.Thismix-
OPLS-AA _ _ S S S H H H H S _ _ S _ _ _
turewasobserveduntiltheendofthesimulation.Thisisadif-
3 AMBER99 _ _ G G G T T T _
ferent result with respect to the simulation performed on
OPLS-AA _ _ _ S S S _ _ _
4 AMBER99 _ _ T T T T T _ _ peptide 1.
OPLS-AA _ _ _ S S S S _ _ In simulations 3e6, all derivatives adopt a helix-like con-
5 AMBER99 _ _ _ H H H H _ formation. However, while peptide 4 displays both a-helical
OPLS-AA _ _ _ S S _ _ _
and 3 -helical features, the structure of peptide 3 is predom-
6 AMBER99 _ _ G G G _ _ 10
inantly a 3 -helix.
OPLS-AA _ _ S S S S _ 10
7 AMBER99 _ T T _ It is interesting to note that the vectors of the dipole mo-
OPLS-AA _ _ _ _ ment of the solvent molecules had no definite direction in
8 AMBER99 _ T T _ theperiodicbox.Thisindicatesthatthesolventhadnoelectro-
OPLS-AA _ _ _ _
static directional effect on the peptide structure. The largest
a PeptidecodesusedinTable1. deviation of the dipole moment at the wall of the periodic
b ThesecondarystructurecodeobtainedfromDSSPprogram.H:4-helix(a-
box was (cid:2)0.05D.
helix);S:bend;G:3-helix(3 -helix);T:H-bondedturns;_:loopsorirregular
10 InsummaryourMDsimulationsindicatethatpeptides1e6
elements.
adopt a helix-like conformation. However, whereas peptide 1
displays a marked preference for an a-helix structure, peptide
1e6isshowninFig.S4inSupplementarymaterial.RMSDin- 2 shows a mixture of beta turn, bend and 3 -helix, being the
10
creased to 0.2e0.8nm in all cases and remained almost con- preferred form the 3 -helix features.
10
stant in simulations 3e6. The fluctuation in RMSD is
attributed to slight changes in structure. In simulations 1 and
2 the conformations fluctuated between helical and turn/bend 2.3. Comparison of theoretical results obtained from
structures (Figs. 3 and 4). different approaches
TheN-andC-terminalresiduesinallsimulationsappearto
have a large flexibility as indicated by the change of RMSF The energetically preferred cluster obtained with EDMC
values during the simulations. The change at central residues calculations for peptide 1 (70.48% of the total population)
is moderate in simulations 1 and 2 (about 0.25nm). The sim- containsana-helixstructure,thesecondmostpopulatedclus-
ulations 3e5 have shown a 0.15e0.20nm change at residues ter(7.46%ofthetotalpopulation)displaysa-helixandb-turn
2e4 and 7. Also, the simulations 3 and 5 in the region which structures. On the other hand, the energetically most stable
seems to be the most sensitive in the conformational change form obtained in SA with minimization showed bend, 3 -
10
have a w0.35nm change at residues 5 and 4, respectively. helix and b-turn structures using AMBER99 force field, and
In all simulations, R s remained almost constant, except in only bend and coil features using OPLS-AA force fields (see
g
1 and 2, where the values fluctuated, but the secondary struc- Table4).Theenergeticallypreferredformofpeptide2showed
tures might be considered stable (Fig. S3). bendandb-turnstructuresforAMBER99forcefield,andbend
The conformational changes in simulation 1 are shown in anda-helixstructuresforOPLS-AAforcefield.Peptides3e6
Fig. 3. The initial conformation returned and remained stable showed structures with helical or consecutive turn secondary
in simulation 1, suggesting that the starting helical structure structures(AMBER99force field)andbendwithcoilfeatures
was destroyed to form a mixture of a-helix, b-turn and bend (OPLS-AA force field). These peptides have shown a higher
in the structure at residues 2e15. Such a conformational be- flexibilitythanpeptides1e2duetotheirsmallersize.Insum-
havior was observed until the end of the simulation. The res- mary,AMBER99forcefieldhasshownaslightpreferencefor
idues 4e6 have shown the highest preference for a-helix helicalstructures.Thus,thisforcefieldhasabettercorrelation
conformation. The initial and final residues appear to have with ECCEP/3 force field than OPLS-AA force field. The
a random coil structure because of the flexibility of these res- OPLS-AA results differ significantly showing some differ-
idues. For the same peptide in water Czajlik et al. [35] found ences in 4 and j angle values and H-bond positions. Despite
a significant amount of helix-like conformation, even in thisfact,allforcefieldsusedherepredictahelix-likestructure
a membrane-mimetic solvent system (TFE /water¼9:1), by forpeptides1e6.TheN-andC-terminalresidueshaveshown
d2
1HNMR.Thus,ourMDsimulationisinverygoodagreement a high flexibility, since no regular stable structure could be
with the experimental results. In simulation 2 (Fig. 4), the observed.
M.F.Masmanetal./EuropeanJournalofMedicinalChemistry44(2009)212e228 219
Fig.3.Changeinthesecondarystructureduringmoleculardynamicssimulationforpeptide1.
These results support the use of the MD simulations for to generate stereoelectrostatic forces. Thus, the electronic
thesepeptides.Itisclear,however,thatinordertoobtainarel- study of peptides was performed using MEPs [36]. MEPs
atively complete picture about the conformational intricacies have been shown to provide reliable information, both on
of peptide 1 and derivatives at least 100ns of simulation ap- the interaction sites of molecules with point charges and on
pearstobenecessary.Suchsimulationscanprovideusefulin- the comparative reactivities of those sites [36,37]. More posi-
formationabouttheconformationalpreferencesandmolecular tive potentials reflect nucleus predominance, while less posi-
flexibility of 1 and derivatives, which might be useful to get tive values represent rearrangements of electronic charges
a more profound understanding of the biological response of and lone pairs of electrons. The fundamental application of
these peptides. this study is the analysis of non-covalent interactions [37],
Comparing the results obtained for the conformational mainly by investigating the electronic distribution in the mol-
analysis using the different approaches, we can conclude ecule. Thus, this methodology was used to evaluate the elec-
that, in general, these methods predict a helix-like structure tronic distribution around molecular surface for peptides
for peptide 1 and derivatives. These results are also in agree- reported here. The MEPs of peptide 1 are shown in Fig. 5
ment with the experimental results obtained from NMR [35]. and the MEPs of peptides 3 and 6 are plotted in Fig. 6. We
evaluated the MEPs of all peptides tested but we show here
2.4. Molecular electrostatic potentials (MEPs) only the MEPs obtained for the three peptides, which dis-
played a significant antifungal activity.
Knowledgeofthestereoelectronicattributesandproperties To better appreciate the electronic behavior of 1 and con-
ofpeptide1andderivativeswillcontributesignificantlytothe sidering that two different faces were signalled in Fig. 2, we
elucidation of the molecular mechanism involved in the anti- present the MEPs of 1, showing the two faces of this peptide
fungal activity. Molecular electrostatic fields and molecular (Fig. 5). Fig. 5a gives the ‘‘charged face’’ (CF) characterized
electrostatic potentials (MEPs), which are their visualisation, by the presence of four cationic residues (R1, K4, R11 and
offer an informative description of the capacity of peptides K15). Although it is possible to visualise residue K16 near
220 M.F.Masmanetal./EuropeanJournalofMedicinalChemistry44(2009)212e228
Fig.4.Changeinthesecondarystructureduringmoleculardynamicssimulationforpeptide2.
to this face, in fact this residue is somewhat shifted in the di- residues I3, W6, F7 and W14. It appears that a kind of p-
rection of the non-charged face. It has been previously re- stacking cluster through W6/R10/W14 occurs in this portion
ported that peptideelipid association occurs through of1.Lensinketal.[33]reportedthattheseresiduescouldpro-
formation of salt bridges between the positively charged resi- tectthepeptidefromthewaterphase.Aclearhydrophobicin-
duesK4,R11andK15andthelipidphosphategroups[33].In teractionbetweenI3andW6mightbealsoappreciatedinthis
addition, tryptophan fluorescence studies previously showed figure. Fig. 5c displays a more polar face in comparison to
the importance of peptide with positively charged residues Fig. 5b, since it possesses the three polar residues of 1 (Q2,
for the initial binding to negatively charged vesicles, since Q8 and N9). Interestingly, I5 is located in an intercalated po-
double R/K/A mutations involving the residues K4/R10/ sitionwithrespecttopolarresiduesandthereforethereareno
R11/K13/K15 significantly decreased the binding affinity interactions between them. These results suggest that these
[38]. The MEPs of 1 suggest that the above-mentioned resi- polar residues could be highly solvated. Mutation of either
dues (R1, K4, R11 and K15) could be responsible for the ini- tryptophandecreasesinternalization,whereasdoublesubstitu-
tial binding. The previously reported [33] p-stacking tion completely inhibits peptide internalization [19,22,39].
interactionbetweenF7andR11residuesmightbealsoappre- Theseresultsindicate thatpeptide1is notsufficientlyhydro-
ciatedonthisface.Althoughthemainpositivepotentials(V(r) phobic to insert deeply into phospholipid model membranes
ranging from 0.60 to 0.43elau(cid:3)3) are concentrated on this [21,40]. Therefore charge neutralization is required for
face,itshouldbenotedthatthereisarelativelyhomogeneous a deeper insertion of the peptide into the hydrophobic core
distribution of positively charged residues along the entire of the membrane. The extended non-charged face alternating
structure. Thus, residues R10, K16 and K13 are strategically cationic residues among the hydrophobic and polar ones ob-
located in an alternated fashion within the non-charged face. served in the MEPs of 1 appears to be operational in this
Since the non-charged moiety is too large, two different sense.
views of the MEPs were plotted in order to better visualise Fig.6agivestheMEPsobtainedforpeptide3.Thissurface
this face (Fig. 5b and c). Fig. 5b displays four hydrophobic showsamoreextendedpositivelychargedfacewithrespectto