The effects of ring-size analogs of the antimicrobial peptide gramicidin S on phospholipid bilayer model membranes and on the growth of Acholeplasma laidlawii B Monika Kiricsi 1 , Elmar J. Prenner 1,2 , Masood Jelokhani-Niaraki 1,2, *, Ruthven N. A. H. Lewis 1 , Robert S. Hodges 1,2,† and Ronald N. McElhaney 1,2 1 Department of Biochemistry, University of Alberta, Edmonton, Alberta, Canada; 2 Protein Engineering Network of Centers of Excellence, University of Alberta, Edmonton, Alberta, Canada We have examined the effects of three ring-size analogs of the cyclic b-sheet antimicrobial peptide gramicidin S (GS) on the thermotropic phase behavior and permeability of phos- pholipid model membranes and on the growth of the cell wall-less Gram-positive bacteria Acholeplasma laidlawii B. These three analogs have ring sizes of 10 (GS10), 12 (GS12) or 14 (GS14) amino acids, respectively. Our high-sensitivity differential scanning calorimetric studies indicate that all three of these GS analogs perturb the gel/liquid-crystalline phase transition of zwitterionic phosphatidylcholine (PtdCho) vesicles to a greater extent than of zwitterionic phosphatidylethanolamine (PtdEtn) or of anionic phos- phatidylglycerol (PtdGro) vesicles, in contrast to GS itself, which interacts more strongly with PtdGro than with Ptd- Cho and PtdEtn bilayers. However, the relative potency of the perturbation of phospholipid phase behavior varies markedly between the three peptides, generally decreasing in the order GS14 > GS10 > GS12. Similarly, these three GS ring-size analogs also differ considerably in their ability to cause fluorescence dye leakage from phospholipid vesi- cles, with the potency of permeabilization also generally decreasing in the order GS14 > GS10 > GS12. Finally, these GS ring-size analogs also differentially inhibit the growth of A. laidlawii with growth inhibition also decreasing in the order GS14 > GS10 > GS12. These results indicate that the relative potencies of GS and its ring-size analogs in perturbing the organization and increasing the permeability of phospholipid bilayer model membranes, and of inhibiting the growth of A. laidlawii B cells, are at least qualitatively correlated, and provide further support for the hypothesis that the primary target of these antimicrobial peptides is the lipid bilayer of the bacterial membrane. The very high anti- microbial activity of GS14 against the cell wall-less bacteria A. laidlawii as compared to various conventional bacteria confirms our earlier suggestion that the avid binding of this peptide to the bacterial cell wall is primarily responsible for its reduced antimicrobial activity against such organisms. The relative magnitude of the effects of GS itself, and of the three ring-size GS analogs, on phospholipid bilayer organi- zation and cell growth correlate relatively well with the effective hydrophobicities and amphiphilicities of these peptides but less well with their relative charge density, intrinsic hydrophobicities or conformational flexibilities. Nevertheless, all of these parameters, as well as others, may influence the antimicrobial potency and hemolytic activity of GS analogs. Keywords: antimicrobial peptides; gramicidin S; phospholi- pid bilayers; membranes. Gramicidin S (GS) is a cyclic decapeptide of primary structure [cyclo-(Val-Orn-Leu- D -Phe-Pro) 2 ] first isolated from Bacillus brevis [1] and is one of a series of antimicrobial peptides produced by this microorganism [2,3]. GS exhibits potent antibiotic activity against a broad spectrum of both Gram-negative and Gram-positive bacteria, as well as against several pathogenic fungi [4– 7]. Unfortunately, GS is rather nonspecific in its actions Correspondence to R. N. McElhaney, Department of Biochemistry, University of Alberta, Edmonton, Alberta, Canada T6G 2H7. Fax: +1 780 4920095, E-mail: rmcelhan@gpu.srv.ualberta.ca Abbreviations: GS, gramicidin S; Myr 2 Gro-PCho, dimyristoylglycerophosphocholine; Myr 2 Gro-PEtn, dimyristoylglycerophosphoethanolamine; Myr 2 Gro-PGro, dimyristoylglycerophosphoglycerol; DSC, differential scanning calorimetry; L a , lamellar liquid-crystalline phase; L b or L b¢ , lamellar gel phase with untilted or tilted hydrocarbon chains, respectively; L C or L C¢ , lamellar crystalline phase with untilted or tilted hydrocarbon chains, respectively; P b¢ , lamellar rippled gel phase with tilted hydrocarbon chains; PtdCho, phosphatidylcholine; PtdEtn, phosphatidylethanol- amine; PtdGro, phosphatidylglycerol; PamOleGro-PCho, 1-palmitoyl-2-oleoyl-glycerophosphocholine; PamOleGro-PEtn, 1-palmitoyl- 2-oleoyl-glycerophospholamine; PamOleGro-PGro, 1-palmitoyl-2-oleoyl-glycerophosphoglycerol. *Present address: Department of Chemistry, Wilfred Laurier University, Waterloo, Ontario, Canada N2L 3C5. Present address: Department of Biochemistry and Molecular Genetics, University of Colorado, Health Sciences Center, 4200 East Ninth Avenue, Denver, CO 80262, USA. (Received 9 August 2002, revised 9 October 2002, accepted 15 October 2002) Eur. J. Biochem. 269, 5911–5920 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03315.x and exhibits appreciable hemolytic as well as antimicrobial activity, thus restricting the potential use of GS as an antibiotic to topical applications at present. However, recent work has shown that structural analogs of GS can be designed with markedly reduced hemolytic activity and enhanced antimicrobial activity (see below), suggest- ing the possibility that appropriate GS derivatives may be used as potent oral or injectable broad-spectrum antibiotics [4–7]. GS has been extensively studied by a wide range of physical techniques [2,3] and its 3D structure is well determined. In this minimum energy conformation, the two tripeptide sequences Val-Orn-Leu form an antiparallel b-sheet terminated on each side by a type II¢ b-turn formed by the two D -Phe-Pro sequences. Four intramolecular hydrogen bonds, involving the amide protons and carbonyl groups of the two Leu and two Val residues, stabilize this rather rigid structure. The GS molecule is amphiphilic, with the two somewhat polar and positively charged Orn sidechains and the two D -Phe rings projecting from one side of this molecule, and the four hydrophobic Leu and Val sidechains projecting from the other side. Moreover, the amphiphilic nature of GS is required for its antimicrobial activity [2,3]. A number of studies have shown that this conformation of the GS molecule is maintained in water, in protic and aprotic organic solvents of widely varying polarity, and in detergent micelles and phospholipid bilay- ers, even at high temperatures and in the presence of agents which often alter protein conformation. There is good evidence from studies of the interaction of GS and its analogs with bacterial cells that the destruction of the integrity of the lipid bilayer of the inner membrane is the primary mode of antimicrobial action of this peptide [8]. In support of this hypothesis, GS has been found to interact strongly with phospholipid model membranes, perturbing their organization and increasing their per- meability. In recent years we have extended these studies of GS/phospholipid bilayer interactions considerably. Specifically, we have investigated the strength and nature of the interactions of GS with phospholipid bilayers by utilizing differential scanning calorimetry (DSC) to mon- itor the effect on this peptide on phospholipid thermo- tropic phase behavior [9]. We demonstrated by 31 P-NMR spectroscopy [10] and X-ray diffraction [11] that GS can induce inverted nonlamellar cubic phases in various phospholipid vesides by producing negative monolayer curvature stress and that GS causes thinning of phosphol- ipid bilayers. We showed by densitometry and sound velocimetry that GS binding to PtdCho bilayers decreases the temperature and cooperativity of their gel/liquid- crystalline phase transition and increases the volume compressibility and decreases the density of the host bilayer [12], indicating that GS increases the motional freedom of the lipid hydrocarbon chains. We also showed that cholesterol decreases the effect of GS on phospholipid bilayers, at least primarily by reducing the penetration of the peptide into the phospholipid model membrane [13]. We demonstrated by Fourier transform infrared spectros- copy that GS is located at the polar/apolar interfacial region of phospholipid bilayers near the glycerol backbone region below the polar headgroups and above the fatty acyl chains, and that GS penetrates more deeply into anionic and more fluid phospholipid bilayers [14]. Finally, using solid-state 19 F-NMR spectroscopy and a 19 F-labeled GS analog, we showed that GS is aligned with its cyclic b-sheet ring lying flat in the plane of the bilayer, consistent with its amphiphilic character, although the peptide molecules rotate rapidly and wobble in liquid-crystalline PtdCho bilayers [15]. We have recently shown that there are several structural variations of the GS molecule which can lead to a dissociation of antimicrobial and hemolytic activities [4–7], including variations in ring size [5]. The secondary structures of these ring-size analogs exhibit a definite periodicity in b-sheet structure, with rings containing six, 10 and 14 residues having the conventional antiparallel b-sheet structure of GS, and those containing eight or 12 residues having largely distorted structures [5,16]. Although GS analogs containing fewer than 10 residues exhibit no significant antimicrobial or hemolytic activities, the 12-residue peptide (GS12) retains appreciable activity against Gram-negative bacteria and fungi, exhibits consid- erably reduced activity against Gram-positive bacteria, but most importantly also displays a significantly reduced hemolytic activity, resulting in a significant improvement in microbial specificity (therapeutic index) for Gram- negative bacteria. In contrast, the 14-residue analog (GS14) shows markedly reduced antimicrobial activity against both Gram-positive and Gram-negative bacteria and increased hemolytic activity as compared to GS itself and thus a much lower therapeutic index [5]. These results are important because they establish that it is possible to dissociate the antimicrobial and hemolytic activities of GS by ring-size alterations. In this paper, we present our initial results dealing with the interactions of the three biologically active ring- size analogs of GS (GS10, GS12 and GS14) with phospholipid bilayer model membranes. We first investi- gated the effects of these GS ring-size analogs on the thermotropic phase behavior of LMVs composed of dimyristoylglycerophosphocholine (Myr 2 Gro-PCho), dimy- ristoylglycerophosphoethanolamine (Myr 2 Gro-PEtn) and dimyristoylglycerophosphoglycerol (Myr 2 Gro-PGro) by DSC, in order to determine their effects on phospholipid bilayer organization in the gel and liquid-crystalline states. We then studied the effect of these analogs on the permeability of LUVs composed of 1-palmitoyl-2-oleoyl- glycerophosphocholine (PamOleGro-PCho), 1-palmitoyl- 2-oleoylglycerophosphoethanolamine (PamOleGro-PEtn) and 1-palmitoyl-2-oleoylglycerophosphoglycerol (PamOle- Gro-PGro), in order to assess their relative abilities to disrupt lipid membranes. We chose to study the zwitter- ionic, bilayer-preferring phospholipid PtdCho as it is abundant in the outer monolayer of the lipid bilayers of mammalian plasma membranes [17], while the zwitter- ionic, nonbilayer phase-preferring PtdEtn and the anionic, bilayer phase-preferring PtdGro are thought to be common components of the outer monolayer of the lipid bilayer of bacterial membranes [18,19]. Finally, we investigated the relative abilities of GS10, GS12 and GS14 to inhibit the growth of A. laidlawii B, a cell wall- less Gram-positive bacteria. The goal of this work is to understand the relationship between the structure of these GS ring-size analogs, their interactions with phospholipid bilayer model membranes, and their anti- microbial and hemolytic activities. 5912 M. Kiricsi et al. (Eur. J. Biochem. 269) Ó FEBS 2002 MATERIALS AND METHODS The three ring-size analogs of GS studied here were synthesized and purified as described previously [4–6]. The phospholipids utilized in this study were purchased from Avanti Polar Lipids (Alabaster, AL, USA) and the calcein from Molecular Probes (Eugene, OR, USA) and all were used without further purification. All of the other chemicals utilized here were highest purity reagent grade purchased from BDH Inc. (Toronto, ON, Canada) and were used as received. The lipid/peptide mixtures for the DSC studies were prepared by mixing the appropriate amounts of phospho- lipid and peptide dissolved in methanol and ethanol, respectively, removing the solvent under a stream of N 2 , and exposing the resultant lipid/peptide films to high vacuum overnight to remove any traces of solvent. Fully hydrated peptide-containing MLVs were then prepared by vortexing in excess aqueous buffer (10 m M Tris/HCl, 100 m M NaCl, 2 m M EDTA,pH7.4)atatemperature above the main phase transition temperature of the phospholipid component. This procedure, and the rationale for utilizing it, have been described previously [9]. The calorimetry was performed on a Nano-DSC Calori- meter (Calorimetry Sciences Corp., Spanish Fork, UT, USA) utilizing a scan rate of 10 °CÆh )1 . Sample runs were repeated at least three times to insure reproducibility. Data acquisition and analysis was carried out using MICROCAL DA -2 (MicroCal LLC, Northampton, MA, USA) and ORIGIN software (OriginLab Corporation, Northampton, MA, USA). Samples containing the GS ring-size analogs alone, dissolved in buffer at a peptide concentration corresponding to that present in the phospholipid-peptide mixtures, exhibit no detectable thermal events over the temperature range 0–90 °C. This indicates that these peptides do not undergo any cooperative thermal denatur- ation over this temperature range and thus that the endothermic events observed by DSC arise exclusively from phase transitions of the phospholipid. The calcein leakage experiments were performed essen- tially as previously described [20]. Briefly, the phospholipid vesicles were prepared by drying chloroform solutions of PamOleGro-PCho or PamOleGro-PEtn/PamOleGro- PGro (7 : 3 molar ratio) under N 2 in a round-bottomed flask and removing traces of the solvent by overnight vacuum. The dry lipid film was then hydrated by the same buffer used for the DSC experiments, but in this case also containing a high concentration (70 m M )ofcalcein,by vortexing at room temperature. The resulting MLVs were then freeze-thawed several times and extruded through a 100-mm filter using a LipoFast apparatus (Avestin Inc., Ottawa, ON, Canada). The resulting LUVs were then passed through a Sephadex G-50 column to remove calcein not trapped inside the phospholipid vesicles. The peptide- induced leakage of the self-quenched calcein from the LUVs was then monitored by measuring the fluorescence of calcein released into the aqueous buffer as a function of time at 25 °C. The fluorescence intensity was measured with a Perkin-Elmer LS50B spectrophotometer (Beaconsfield, UK) utilizing slit widths for both excitation and emission of 3 nm and quartz cells of 1-cm path length; the excitation and emission were recorded at wavelengths of 496 and 516 nm, respectively. The A. laidlawii B cells were grown and cell growth was monitored turbidometriedly, all as previously described [21]. The effect of the ring-size analogs studied here on cell growth was monitored by adding various concentrations of these peptides to the culture medium just prior to the addition of a 10% by volume inoculation with cells in the mid log phase of growth. RESULTS Structure and biological activities of GS ring-size analogs The amino acid sequences and the 3D structures of the three ring-size analogs of GS studied here are presented in Fig. 1. These three peptides are all based on the structure of GS itself except that the D -Phe residue in each of the two type II¢ b-turns has been replaced by a D -Tyr residue and the Orn residues have been replaced by Lys residues. The former replacement was made to increase the water solubility of these compounds and the latter to decrease the cost of chemical synthesis [4–7]. Note that these conservative amino acid substitutions do not by themselves significantly alter the conformation or biological activity of these peptides, as shown by the fact that the structure in aqueous solution, and antimicrobial and hemolytic activities, of GS and GS10 are similar, although GS is slightly more active against both Gram-positive and Gram-negative bacteria than is GS10 [5]. Perturbation of phospholipid thermotropic phase behavior by GS ring-size analogs We studied the effects of concentrations of these GS ring- size analogs ranging from 1 to 4 mole percent on the thermotropic phase behavior of aqueous dispersions of two zwitterionic phospholipids and one anionic phospholipid by DSC. In each case the result of progressively increasing the Fig. 1. The amino acid sequence, structure and conformation of the ring- size analogs of GS studied here (GS10, GS12 and GS14) in aqueous solution. Ó FEBS 2002 Gramicidin S analog–membrane interactions (Eur. J. Biochem. 269) 5913 peptide concentration was simply to progressively increase the magnitude of the characteristic effects of each particular analog on the thermotropic phase behavior of each phospholipid system examined. We have thus chosen to present DSC thermograms for each GS ring-size analog and each phospholipid system at only the highest peptide concentration tested, as the characteristic differences in their effects are most clear under these circumstances. We also point out that a peptide concentration of 4 mole percent (phospholipid/peptide molar ratio of 25 : 1) is well within the physiologically relevant concentration range for GS itself [2,3,8]. The initial DSC heating thermograms, illustrating the thermotropic phase behavior of large MLVs composed of Myr 2 Gro-PCho alone and of Myr 2 Gro-PCho/peptide mixtures, are presented in Fig. 2. The MLVs composed of Myr 2 Gro-PCho alone exhibit two transitions on heating, a less enthalpic, less cooperative pretransition centered at 14 °C and a more enthalpic, more cooperative main phase transition centered at 24 °C. The pretransition corresponds to the conversion of a planar lamellar gel phase with tilted hydrocarbon chains (the L b¢ phase) to the rippled gel phase with tilted hydrocarbon chains (the P b¢ phase) and the main phase transition to the conversion of the P b¢ phase to the lamellar liquid-crystalline (L a) phase. A subtransition cen- tered at 16 °C is not observed here because this sample was not extensively annealed at low temperatures. For a more complete description of the thermotropic phase behavior of Myr 2 Gro-PCho and other members of the homologous series of linear disaturated PCs, the reader is referred to Lewis et al. [22]. The effect of the incorporation of 4 mole percent (peptide/phospholipid molar ratio 1 : 25) of the three GS ring-size analogs studied here on the thermotropic phase behavior of the host Myr 2 Gro-PCho bilayer varies greatly, as illustrated in Fig. 2. For GS12, a single DSC endotherm is observed whose temperature and enthalpy are essentially unchanged from that of Myr 2 Gro-PCho alone and whose cooperativity is only moderately reduced. In contrast, the incorporation of both GS10 and GS14 produce much broader, lower enthalpy DSC endotherms, particularly in the case of the latter peptide. In fact in both instances, these two peptides produce two-component DSC traces consist- ing of a relatively more cooperative, higher enthalpy component centered at a lower temperature than the main phase transition temperature of Myr 2 Gro-PCho alone, and a much less cooperative, more enthalpic component centered at a higher temperature (see Fig. 3). Moreover, the magnitude in the downward shift in the temperature of the sharp component, and of the upward shift in the temperature of the broad component, is greater for GS14 than for GS10. Also, the relative enthalpy of the higher temperature DSC component is considerably greater and the cooperativity of this component is considerably less in the GS14/Myr 2 Gro-PChothanintheGS10/Myr 2 Gro- PCho MLVs. According to our prior studies of GS/ Myr 2 Gro-PCho mixtures, we interpret the sharp and broad components of the two-component DSC endotherms as the chain-melting phase transition of peptide-poor and peptide- enriched phospholipid domains, respectively [9]. Note also that the incorporation of GS10 and GS14 abolish the pretransition of Myr 2 Gro-PCho whereas the incorporation of GS12 does not. These results suggest that GS12 perturbs the thermotropic phase behavior of Myr 2 Gro-PCho bilay- ers to a much lesser extent than does GS10 and GS14, and that GS14 is more potent in this regard than is GS10. Fig. 2. Initial high-sensitivity DSC heating scans illustrating the effect of the addition of 4 mole percent GS10, GS 12 or GS14 on the thermotropic phase behavior of Myr 2 Gro-PCho MLVs. Fig. 3. A DSC heating thermogram of Myr 2 Gro-PCho MLVs con- taining 4 mole percent GS14 (– – —) and its deconvolution into sharp and broad components (- - - -). 5914 M. Kiricsi et al. (Eur. J. Biochem. 269) Ó FEBS 2002 Interestingly, GS14 and GS10 alter the phase behavior of Myr 2 Gro-PChoMLVstoagreaterextentthandoesGS itself [9]. DSC heating scans of MLVs composed of Myr 2 Gro- PEtn alone, or of Myr 2 Gro-PEtn containing 4 mole percent of one of the three ring-size analogs of GS, are presented in Fig. 4. Aqueous dispersions of Myr 2 Gro-PEtn alone, which have not been extensively incubated at low temperatures prior to calorimetric analysis, exhibit a single fairly cooper- ative, relatively energetic L b /L a phase transition centered near 50 °C (see [23] for a more complete description of the thermotropic phase behavior of Myr 2 Gro-PEtn and other members of the homologous series of linear saturated PtdEtns). If peptide-containing Myr 2 Gro-PEtn MLVs are exposed to temperatures above but near to the L b /L a phase transition of Myr 2 Gro-PEtn, the presence of these peptides has only a very small effect on the main phase transition, causing a slight reduction in the phase transition tempera- ture and a modest decrease in the cooperativity of the phase transition with no detectable change in the overall transition enthalpy (see Fig. 4A). However, if these peptide-containing Myr 2 Gro-PEtn vesicles are exposed to temperatures well above the L b /L a phase transition temperature, then subse- quent DSC heating scans reveal an additional decrease in the temperature and cooperativity, but still little change in the enthalpy, of the chain-melting phase transition (see Fig. 4B). Interestingly, however, GS14 and GS10 again exhibit larger effects on the thermotropic phase behavior of Myr 2 Gro-PEtn than does GS12. However, in all cases the presence of these peptides produces only a slight destabi- lization of the L b phase relative to the L a phase of Myr 2 Gro- PEtn bilayers, as also observed previously with GS itself [9]. Also, repeated recycling through the phase transition temperature actually increases the magnitude of the effect of these peptides on Myr 2 Gro-PEtn phase behavior, in contrast to the situation with Myr 2 Gro-PCho MLVs. This effect, which was also observed to a lesser extent with GS itself [9], suggests that repeated exposure to high tempera- tures facilitates peptide incorporation into Myr 2 Gro-PEtn bilayers. The initial DSC heating scans of MLVs of Myr 2 Gro- PGro alone, or Myr 2 Gro-PGro MLVs containing 4 mole percent of one of the three GS ring-size analogs studied here, are shown in Fig. 5. Aqueous dispersions of Myr 2 Gro- PGro alone, which have not been extensively annealed at low temperatures, exhibit two endothermic events upon heating, a less energetic pretransition near 14 °C and a more energetic main transition near 24 °C. Again, a subtransition (L C¢ /L a phase transition) centered near 25 or 40 °C is not observed under these conditions. The pretransition arises form a conversion of the (L b¢ )tothe(P b¢ ) phase and the main transition from the conversion of the P b¢ to the L a phase. For a more detailed discussion of the thermotropic phase behavior of Myr 2 Gro-PGro and other members of the homologous series of linear saturated PGs, see Zhang et al. [24]. The addition of 4 mol percent of one of the three GS ring-size analogs of GS studied has a relatively modest effect on the thermotropic phase behavior of Myr 2 Gro-PGro MLVs. In all cases the presence of peptide decreases the cooperativity of the main transition. Also, each peptide induces the presence of a second, less enthalpic, broad component of the DSC endotherm which occurs at a higher temperature than does the more enthalpic sharp compo- nent. As well, in the case of GS14 only, additional endothermic events are noted at temperatures near 31 and 39 °C. As before, the magnitude of the effect of these peptides on the cooperativity of the main phase transition decreases in the order GS14 > GS10 > GS12. Moreover, GS14 decreases the enthalpy of the main phase transition of Myr 2 Gro-PGro substantially whereas GS12 and GS10 actually appear to slightly increase the total enthalpy of the two-component main phase transition. We note that GS itself, however, has a greater effect on the thermotropic phase of Myr 2 Gro-PGro MLVs than do any of the three Fig. 4. Initial high-sensitivity DSC heating scans illustrating the effect of the addition of 4 mol percent GS10, GS12 or GS14 on the thermotropic phase behavior of Myr 2 Gro-PEtn MLVs. (A) Myr 2 Gro-PEtn MLVs not exposed to high temperatures (i.e. temperatures above 65–70 °C). (B) Myr 2 Gro-PEtn MLVs exposed to high temperatures (i.e. temper- atures of 75 °Corhigher). Fig. 5. Initial high-sensitivity DSC heating scans illustrating the effect of the addition of 4 mole percent GS10, GS12 or GS14 on the thermotropic phase behavior of Myr 2 Gro-PGro MLVs. Ó FEBS 2002 Gramicidin S analog–membrane interactions (Eur. J. Biochem. 269) 5915 ring-size analogs studied here [9]. Also, as noted previously for GS, recycling through the phase transition temperature has little effect on phospholipid phase behavior in contrast to the situation with Myr 2 Gro-PCho and Myr 2 Gro-PEtn MLVs, suggesting that these peptides readily incorporate into Myr 2 Gro-PGro bilayers in the liquid-crystalline state and remain incorporated in the gel state. Permeabilization of phospholipid bilayers by GS ring-size analogs In order to determine the relative abilities of these three ring- size analogs of GS to permeabilize phospholipid bilayers, we determined the amount of entrapped calcein dye released by the addition of 4 mole percent peptide to LUVs composed of either PamOleGro-PCho or a mixture of PamOleGro- PEtn/PamOleGro-PGro (7 : 3 molar ratio). PamOleGro- PCho was selected to mimic the phospholipid composition of the outer monolayer of eukaryotic plasma membranes [17] and the PamOleGro-PEtn/PamOleGro-PGro mixture tomimicthephospholipid composition of the Escherichia coli inner membrane [18]. Although we intended to study vesicles composed of PamOleGro-PEtn or PamOleGro-PGro alone, the former did not form well defined LUVs under our experimental conditions [25] and PamOleGro-PGro formed only small unilamellar vesicles [26]; as antimicrobial peptide binding can be influenced by the degree of curvature strain in phospholipid [27], such a size difference would have made comparisons between the three individual phosphol- ipids difficult. However, the PamOleGro-PEtn/PamOle- Gro-PGro mixture formed well-behaved LUVs. As illustrated in Fig. 6, the addition of each of the three ring-size analogs of GS cause considerable fluorescence dye leakage when added to PamOleGro-PCho LUVs at a final peptide/phospholipid molar ratio of 1 : 25, with the extent of dye leakage decreasing in the order GS14 > GS10 > GS12 [28]. Moreover, at lower peptide concentra- tions, GS14 is perhaps 10-fold more potent at releasing calcein than is GS10 or GS12 [28]. Interestingly, the addition of the same amount of these three peptides to PamOleGro-PEtn/PamOleGro-PGro LUVs is generally less effective at releasing entrapped calcein, particularly in the case of GS10, and the relative effectiveness of the three peptides now decreases in the order GS14 > GS12 > GS10. Moreover, in this vesicle system GS14 does not exhibit a relatively much greater potency at lower concen- trations than do the other peptides. GS14 was thus the most effective peptide in both phospholipid vesicle systems with GS10 exhibiting the greatest phospholipid compositional selectivity, much more strongly affecting the PamOleGro- PCho system in comparison to the PamOleGro-PEtn/ PamOleGro-PGro mixed system. Inhibition of growth of A. laidlawii B by GS ring-size analogs In order to extend the above studies to a living microbial system, we investigated the effect of these three ring-size analogs of GS on the growth of A. laidlawii B, a cell wall-less Gram positive bacteria (Mollicute). The absence of a lipopolysaccharide-containing cell wall or outer membrane or a lipopeptidoglycan outer layer is a major advantage of utilizing this organism for such studies, as the antimicrobial peptides added should have free physical access to the surface of the limiting membrane and extracellular structures should not compete with the membrane lipid bilayer for peptide binding. Moreover, the membrane lipid composition [29], and the organization and dynamics of the membrane lipid bilayer [30] of this organism, have been extensively studied by ourselves and others, potentially facilitating a molecular interpretation of any results obtained. We present in Fig. 7 growth curves for A. laidlawii in the presence or absence of various concentrations of the three GS ring-size analogs studied here. It is clear from these curves that a considerable difference in the growth inhib- itory potency of these three peptides exists. For example GS10 is a fairly potent antimicrobial agent, inhibiting A. laidlawii B growth slightly at the lowest concentration tested (0.25 l M ), strongly at the next highest peptide Fig. 6. A bar graph illustrating the percentage of entrapped calcein dye leakage at equilibrium from LUVs composed of either PamOleGro- PCho (white bar) or PamOleGro-PEtn/PamOleGro-PGro(7:3molar ratio) (hatched bar) upon the addition of 4 mol percent GS10, GS12 or GS14. Fig. 7. Growth curves at 37 °CofA. laidlawii B in the absence or presence of various concentrations of GS10, GS12 or GS14. The sym- bols utilized are: (·), absence of peptide, and (h), (s), (n), (,), and (e), peptide concentrations of 0.25, 0.50, 1.0, 2.0 and 4.0 l M , respectively, in the growth medium. 5916 M. Kiricsi et al. (Eur. J. Biochem. 269) Ó FEBS 2002 concentrated (0.50 l M ), and completely suppressingly growth at concentrations of 1.0 l M and higher. In con- trast, GS12 is a much weaker growth inhibitory agent, with significant inhibition of growth being observed only at peptide concentrations of 1.0–2.0 l M with complete growth inhibition occurring only at the highest peptide concentration tested (4.0 l M ). On the other hand, GS14 is a very potent inhibitor of the growth of this organism, with significant growth suppression being observed at the lowest peptide concentration tested (0.25 l M ) and the complete inhibition of growth at all higher peptide concentrations. Thus the potency of growth inhibition of these peptides decreases in the order GS14 > GS10 > GS12. We note also that when sufficient peptide was added to these A. laidlawii B cultures to completely inhibit cell growth, the initial turbidity of the 10% (v/v) innoculum of mid-log phase cells added to fresh culture media was reduced to blank values. This result indicates that these GS ring-size analogs of GS exert a cidal rather than a static effect on A. laidlawii cells, presumably by causing cell lysis to occur, ashasalsobeenreportedtobethecaseforGSitselfin mycoplasma and bacterial systems [8]. Interestingly, we find that GS itself is slightly less effective at inhibiting the growth of A. laidlawii than is GS10 (data not presented), in contrast to results with most species of conventional bacteria, where GS is slightly more effective than GS10 [5]. DISCUSSION We showed previously that the effect of GS on the thermotropic phase behavior of phospholipid bilayers depends markedly on both the structure and charge of the lipid polar headgroup [9,10]. Specifically, the presence of GS has only a very small effect on the thermotropic phase behavior of Myr 2 Gro-PEtn bilayers, even at very high peptide concentrations and after multiple cycling through the gel/liquid-crystalline phase transition. Only upon expo- sure of the Myr 2 Gro-PEtn bilayers to high temperature is a small decrease in the temperature, enthalpy and coopera- tivity of the main phase transition and the induction of a minor lower temperature shoulder on this endotherm observed. The addition of similar amounts of GS to Myr 2 Gro-PCho bilayers results in a somewhat greater but still rather small decreases in the temperature, enthalpy and cooperativity of the main phase transition and induces a new broad component of the DSC endotherm centered at a slightly higher temperature. In contrast, the addition of GS to Myr 2 Gro-PGro bilayer produces a considerably larger decrease in the temperature, enthalpy and cooperativity of the main phase transition and induces the presence of a second transition at a considerably higher temperature, but whose temperature decreases more rapidly than that of the main phase transition with increasing peptide concentra- tion. We thus concluded that GS interacts more strongly with anionic phospholipids such as PtdGro than with zwitterionic phospholipids, and more strongly with more fluid zwitterionic phospholipids like PtdCho than with less fluid zwitterionic phospholipids like PtdEtn. However, the three ring-size analogs of GS studied here thus exhibit a somewhat different phospholipid polar head group specif- icity, as discussed below. The overall phospholipid specificity of the three GS ring- size analogs studied here is broadly similar to that of GS itself in that the degree of perturbation of phospholipid thermotropic phase behavior increases in the order Myr 2 Gro-PEtn < Myr 2 Gro-PCho < Myr 2 Gro-PGro. Moreover, the magnitude of the decrease in temperature, enthalpy and cooperativity of the main phase transition of the three phospholipids studied here generally decreases in the order GS14 > GS10 > GS12, as does the temperature and the relative magnitude of the new endotherm or endotherms induced by the addition of the peptide. However, the magnitude of the effects of GS [9] and its three ring-size analogs on the thermotropic phase behavior of the phospholipids studied here depends on the specific phospholipid vesicle system being studied. Specifically, the order of decreasing perturbation of phospholipid phase behavior by GS itself, and by the three ring-sized analogs studied here, is GS14 > GS10 > GS > GS12 in Myr 2 Gro-PCho MLVs, GS14 @ GS10 > GS @ GS12 in Myr 2 Gro-PEtn MLVs, and GS > GS14 >GS10 > GS12 in MLVs of Myr 2 Gro-PGro. Thus, although GS12 has the weakest effect in all three vesicles systems, the relative order of effectiveness varies with polar headgroup structure for the other three peptides, with GS14 and GS10 exhibiting a greater effect than GS itself in the two zwitterionic phospholipid bilayers studied here but a smaller effect in the anionic phospholipid bilayer system. The potency of the three ring-size analogs of GS in inducing the leakage of calcein dye entrapped in PamOle- Gro-PChoLUVsalsodecreasesintheorderGS14> GS10 > GS12, which is the same decreasing order as exhibited by these three peptides in perturbing the thermo- tropic phase behavior of Myr 2 Gro-PCho MLVs. Interest- ingly, the three GS ring-size analogs are generally less potent at releasing entrapped calcein from PamOleGro-PEtn/ PamOleGro-PGrothanfromPamOleGro-PCho LUVs, particularly in the case of GS10, so that the order of decreasing potency of vesicle permeation in PamOleGro- PEtn/PamOleGro-PGro LUVs is GS14 > GS12 > GS10. Thus, although GS14 is the most potent ring-size analog in both perturbing the thermotropic phase behavior and permeabilizing PtdCho vesicles and GS10 is the second most potent analog, the behavior of GS12 is somewhat anomalous in that its ability to induce dye leakage in PamOleGro-PEtn/PamOleGro-PGro vesides, but not in PamOleGro-PCho vesicles, is much greater than predicted by its smaller effect on the thermotropic phase behavior of all three phospholipids examined. Interestingly, GS itself is less potent than GS14 but more potent than GS10 and GS12 in permeabilizing PamOleGro-PCho LUVs [31]. However, in PamOleGro-PEtn/PamOleGro-PGro LUVs, GS is less potent than both GS14 and GS12, but remains more potent than GS10 [31]. The effectiveness of the three ring-size analogs of GS studied here on the growth inhibition and killing of A. laidlawii B cells also decreases in the order GS14 > GS10 > GS12, again paralleling the decreasing relative potency of these peptides in perturbing the thermotropic phase behavior of the three phospholipid MLVs studied and of the permeabilization of PamOleGro-PCho LUVs. As well, the relative antimicrobial potency of this series of peptides also mirrors the order of the decreasing extent of dye leakage from PamOleGro-PEtn/PamOleGro-PGro LUVs, except that the order of GS10 and GS12 are reversed in this system. Nevertheless, there is generally a Ó FEBS 2002 Gramicidin S analog–membrane interactions (Eur. J. Biochem. 269) 5917 good correlation overall between the relative perturbation of host bilayer organization as measured by DSC, the permeabilization of phospholipid vesicles as measured by calcein leakage, and the inhibition of the growth of A. laidlawii B. We note also that GS is less potent at inhibiting the growth of this organism than is GS14 but more potent than GS10 and especially GS12 (data not presented). Overall, then, these results indicate that studies of the interactions of other analogs of GS with phospholipid vesicles may be useful for predicting the antimicrobial potency of these analogs and possibly also for understand- ing the molecular basis for their differential antimicrobial potencies again different classes and species of bacteria, many of which may differ considerably in the lipid compositions of their membranes [18,19]. It is instructive to compare the relative antimicrobial potencies of these three GS ring-size analogs against various Gram-positive and Gram-negative bacteria and against A. laidlawii, as the former types of bacteria possess either a lipopeptidoglycan outer barrier or a lipopolysaccharide- containing cell wall or outer membrane, respectively, which is lacking in the latter organism. Against conventional bacteria, the order of decreasing antimicrobial potency is GS10 > GS12 > GS14 [5], whereas against A. laidlawii B the order is GS14 > GS10 > GS12. This result would appear to confirm our previous suggestion that the low effective antimicrobial activity of GS14, particularly against Gram-negative bacteria, is due to its strong binding to the lipopolysaccharide component of the bacterial cell wall [5], which effectively competes for the binding of available peptide with the lipids of the inner membrane [5]. Thus, when an outer cell wall is absent, as in A. laidlawii B,such competition is not observed and GS14 is then able to exert its intrinsically high antimicrobial activity. However, the aggre- gation of GS14 in solution may also reduce its ability to penetrate the cell wall of Gram-negative bacteria, which could also reduce its effective antimicrobial activity. What- ever the reason for its low activity against conventional bacteria, the high antimicrobial potency of GS14 against A. laidlawii suggests that this peptide might potentially be clinically useful in treating the many serious diseases of man and animals caused by various Mollicutes [32,33]. It is also interesting to note that with these particular ring-size analogs of GS, their relative potencies at perturbing the organization and increasing the permeability of phospholipid bilayer model membranes, and of causing the lysis of A. laidlawii B and human erythrocytes, are at least qualitatively correlated. As discussed earlier, GS10 and GS14 both exist in aqueous solution as antiparallel b-sheet structures separated by type II¢ turns, as does GS itself, although GS14 has a somewhat less rigid structure as compared to GS10, presumably due to its expanded ring; in contrast, GS12 exists in a distorted b-sheet and b-turn structure and is conformationally much more flexible [5,16,28]. Although the intrinsic hydrophobicities of these three peptides are predicted to decrease in the order GS12 > GS14 > GS10, based on the ratios of the number of charged polar Lys residues to hydrophobic Val and Leu residues (4 : 4, 4 : 6 and 2 : 4, respectively), the actual measured solubilities in water decrease in the order GS12 > GS10 > GS14. The lower solubility of GS14 as compared to GS10 in water appears to be related to its slightly greater amphiphilicity and to its significantly greater exposed hydrophobic surface area, which results in GS14 forming aggregates in aqueous solution above a concentration of about 50–60 m M , whereas GS10 and GS12 remain monomeric even at much higher concentrations [22]. Note also that GS14 and GS10 are considerably more amphiphilic than is GS12, as in the former two peptides the more polar charged Lys and the less polar Val and Leu residues project on opposite sides of the ring, whereas this is not the case for GS12. We can ask whether or not the observed order of biophysical or biological potencies of GS itself and of the three ring-size analogs of GS studied here, namely GS @ GS14 > GS10 > GS12, correlates well with any of the physical properties of these peptides which we have previ- ously measured. In terms of the relative conformational flexibility of this series of peptides, there is not a particularly good correlation with the observed results, as conforma- tional rigidity decreases in the order GS > GS10 > GS14 > GS12. Similarly, an even poorer correlation is observed between the intrinsic hydrophobicities of these peptides, which decrease in the order GS > GS10 > GS14 > GS12, and their biophysical and biological activ- ities or with the ratio of positively charged Orn or Leu residues to the total number of residues (GS12 > GS14 > GS10 ¼ GS). However, a reasonably good correlation is observed between the effective hydrophobicities of these peptides, as assessed by their decreasing solubilities in water, and their decreasing amphiphilicities, as measured by their retention times on reversed-phase high-performance liquid chromatographic columns, which are both related to the accessible nonpolar surface areas of these peptides [5]. Both effective hydrophobicity and amphiphilicity decrease in the order GS > GS14 > GS10 > GS12, which correlates well with their decreasing ability to perturb the organization of lipid bilayer model and A. laidlawii membranes. These results are in only partial agreement with previous studies of GS analogs with 10-membered rings, which indicated that a high b-sheet content, as well as a high effective hydrophob- icity and amphiphilicity, are correlated with a high antibac- terial activity [2,3,8]. However, the present results may not be surprising, as the requirement for an ordered b-sheet structure was rationalized previously by assuming that in disordered GS analogs, the absence of the interstrand hydrogen bonds present in the b-sheet ring results in solvation of the amide NH and CO groups, in turn causing a decrease of partitioning into the lipid bilayer and reducing their effectiveness [5]. However, because the three ring-size analogs of GS studied here have fairly markedly different intrinsic hydrophobicities due to their variable ratios of Lys to Val and Leu residues, these amino acid compositional differences may dominate the phospholipid/water partition- ing process, thus overshadowing any smaller changes arising from conformational effects. However, the generally positive correlation between the effective hydrophobicity and amphiphilicity of the GS ring-size analogs and the magnitude of their perturbation of phospholipid bilayer membranes and the growth of A. laidlawii, generally agrees well with the results of previous studies of the antimicrobial activity of these [5] and other [4–7] ring-size analogs of GS. This finding is important in that it adds further support to the hypothesis that GS and its analogs kill bacteria primarily through their disruption of the lipid bilayer of the cell membrane. In the absence of complications arising from the differ- ential interactions of GS and its ring-size analogs with the 5918 M. Kiricsi et al. (Eur. J. Biochem. 269) Ó FEBS 2002 outer membrane or cell wall of conventional bacteria, we can identify at least three factors which can determine the degree to which a particular antimicrobial peptide will perturb the organization and integrity of phospholipid bilayer membranes and inhibit the growth of A. laidlawii B. These are the phospholipid bilayer/water partition coeffi- cient, the localization and orientation of the peptide within the phospholipid bilayer, and the degree to which the peptide disturbs phospholipid packing once inserted into the bilayer. The fact that the relative order of both decreasing effectiveness in perturbing the thermotropic phase behavior and compromising the integrity of phosp- holipid model membranes, as well as inhibiting the growth of A. laidlawii B (GS @ GS14 > GS10 > GS12) correlates well with the increasing water solubility of these peptides, can be explained in part by the fact that the phospholipid bilayer/water coefficient should also decrease in the above order, so that the effective concentration of peptide in the target membrane also progressively decreases. Similarly, the good correlation observed between the biophysical and biological effects of GS and its ring-size analogs with their degree of amphiphilicity may be related to the fact that the degree of amphiphilic character may determine the good- ness of the characteristic interfacial location of these and some other antimicrobial peptides, which is thought to be at the polar/apolar region of the phospholipid bilayer near the glycerol backbone, where the polar, positively charged Orn or Lys residues can interact with the negatively charged phosphate polar headgroups of the lipid bilayer and the nonpolar Val and Leu sidechains with the upper portions of phospholipid hydrocarbon chains [9,14,15]. Although we have no independent information at present about the intrinsic perturbing effects of these peptides on phosphol- ipid bilayers (i.e. the degree perturbation per peptide molecule actually present in the bilayer), we might expect that this would be related to the asymmetry of shape and possibility also to the size of the peptide molecule. A crude estimate of these two parameters, based simply on the structure and conformation of these peptides in water as illustrated in Figs 1 and 2, might suggest that the intrinsic perturbation of these peptides would decrease in the order GS14 > GS > GS10 > GS12. Experiments are currently underway to actually determine the phospholipid bilayer/ water partition coefficient, the localization and orientation in the phospholipid bilayer, and the effects of the presence of these peptides on phospholipid organization and pack- ing. The results of these experiments should allow us to quantitate the above parameters and gain additional insight into the molecular basis of the structure/activity correlations reported here. ACKNOWLEDGEMENTS This work was supported by operating grants from the Protein Engineering Network of Centers of Excellence and the Canadian Institutes of Health Research, and by major equipment grants from the Alberta Heritage Foundation for Medical Research. MK was sup- ported in part by a Hungarian Eo ¨ tvo ¨ s Fellowship. REFERENCES 1. Gause, G.G. & Brazhnikova, M.G. (1944) Gramicidin S and its use in the treatment of infected wounds. Nature 154,703. 2. Izumiya, N., Kato, T., Aoyaga, H., Waki, M. & Kondo, M. (1979) Synthetic Aspects of Biologically Active Cyclic Peptides: Gramici- din S and Tyrocidines. Halsted Press, New York. 3. Waki, M. & Izumiya, N. (1990) Recent advances in the bio- technology of b-lactams and microbial bioactive peptides. In Biochemistry of Peptide Antibiotics (Kleinhaug, H. & van Doren, H., eds), pp. 205–244. Walter de Gruyter Company, Berlin. 4. Kondejewski, L.H., Farmer, S.W., Wishart, D., Kay, C.M., Hancock, R.E.W. & Hodges, R.S. (1996) Gramicidin S is active against both gram-positive and gram-negative bacteria. Int. J. Pept. Protein Res. 47, 460–466. 5. Kondejewski, L.H., Farmer, S.W., Wishart, D., Kay, C.M., Hancock, R.E.W. & Hodges, R.S. (1996) Modulation of structure and antibacterial and hemolytic activity by ring size in cyclic gramicidin S analogs. J. Biol. Chem. 271, 25261–25268. 6. Kondejewski,L.H.,Jelokhani-Niaraki,M.,Farmer,S.W.,Lix,B., Kay, C.M., Sykes, B.D., Hancock, R.E.W. & Hodges, R.S. (1999) Dissociation of antimicrobial and hemolytic activitities in cyclic peptide diasteeomers by systematic alteration in amphipathicity. J. Biol. Chem. 274, 13181–13192. 7. Jelokhani-Niaraki, M., Kondejewski, L.H., Farmer, S.W., Han- cock, R.E.W., Kay, C.M. & Hodges, R.S. (2000) Diaster- eoisomeric analogues of gramicidin S: structure, biological activity and interaction with lipid bilayers. Biochem. J. 349, 747–755. 8. Prenner, E.J., Lewis, R.N.A.H. & McElhaney, R.N. (1999) The interaction of the antibacterial peptide gramicidin S with lipid bilayer model and biological membranes. Biochim. Biophys. Acta 1462, 201–221. 9. Prenner, E.J., Lewis, R.N.A.H., Kondejewski, L.H., Hodges, R.S. & McElhaney, R.N. (1999) Differential scanning calorimetric study of the effect of the antimicrobial peptide gramicidin S on the thermotropic phase behavior of phosphatidylcholine, phosphati- dylethanolamine and phosphatidylglycerol lipid bilayer mem- branes. Biochim. Biophys. Acta 1417, 211–223. 10. Prenner, E.J., Lewis, R.N.A.H., Newman, K.C., Gruner, S.M., Kondejewski, L.H., Hodges, R.S. & McElhaney, R.N. (1997) Nonlamellar phases induced by the interaction of gramicidin S with lipid bilayers. A possible relationship to membrane disrupting activity. Biochemistry 36, 7906–7916. 11. Staudegger, E., Prenner, E.J., Kriechbaum, M., Degovics, G., Lewis,R.N.A.H.,McElhaney,R.N.&Lohner,K.(2000)X-ray studies of the interaction of Gramicidin S with microbial lipid extracts: Evidence for cubic phase formation. Biochim. Biophys. Acta 1468, 213–230. 12. Krivanek, R., Rybar, P., Prenner, E.J., McElhaney, R.N. & Hianik, T. (2001) Interaction of the antimicrobial peptide grami- cidin S with dimyristoyl-phosphatidylcholine bilayer membranes: a densitometry and sound velocimetry study. Biochim. Biophys. Acta 1510, 452–463. 13. Prenner, E.J., Lewis, R.N.A.H., Jelokhani-Niaraki, M., Hodges, R.S. & McElhaney, R.N. (2001) Cholesterol attenuates the inter- action of the antimicrobial peptide gramicidin S with phospholipid bilayer membranes. Biochim. Biophys. Acta 1510, 83–92. 14. Lewis, R.N.A.H., Prenner, E.J., Kondejewski, L., Flach, C.R., Mendelsohn, R., Hodges, R.S. & McElhaney, R.N. (1999) Fourier transform infrared spectroscopic studies of the interaction of the antimicrobial peptide gramicidin S with lipid micelles and with lipid monolayer and bilayer membranes. Biochemistry 38, 15193–15203. 15. Salgado, J., Grage, S.L., Kondejewski, L.H., Hodges, R.S., McElhaney, R.N. & Ulrich, J. (2001) Membrane-bound structure and alignment of the antimicrobial b-sheet peptide gramicidin S derived from angular and distance constraints by solid-state 19F- NMR. J. Biomol. NMR 21, 191–208. 16. Gibbs, A.G., Kondejewski, L.H., Gronwald, W., Nip, A.M., Hodges, R.S. & Wishart, D.S. (1998) Unusual b-sheet periodicity in small cyclic peptides. Nat. Struct. Biol. 5, 284–288. Ó FEBS 2002 Gramicidin S analog–membrane interactions (Eur. J. Biochem. 269) 5919 17. Verkleij, A.J., Zwaal, R.F.A., Roelofsen, B., Comfurius, P., Kastelijn, D. & van Deenen, L.L.M. (1973) Asymmetric dis- tribution of phospholipids in the human red cell membrane. A combined study using phospholipases and freeze-etch electron microscopy. Biochim. Biophys. Acta 323, 178–193. 18. Wilkinson, S.G. (1988) Gram-negative bacteria. In Microbial Lipids (Ratledge, C. & Wilkinson, S.G., eds), Vol. 1, pp. 299–488. Academic Press, London. 19. O’Leary, W.M. & Wilkinson, S.G. (1988) Gram-positive bacteria. In Microbial Lipids (Ratledge, C. & Wilkinson, S.G., eds), Vol. 1, pp. 117–201. Academic Press, London. 20. New, R.R.C. (1990) Characterization of liposomes. In Liposomes: a Practical Approach (New, R.R.C., ed.), pp. 105–161. IRL Press, Oxford. 21. Silvius, J.R. & McElhaney, R.N. (1978) Lipid compositional manipulation in Acholeplasma laidlawii B. Effect of exogenous fatty acids on fatty acid composition and cell growth when endogenous fatty acid production is inhibited. Can. J. Biochem. 56, 462–469. 22. Lewis, R.N.A.H., Mak, N. & McElhaney, R.N. (1987) A differ- ential scanning calorimetric study of the thermotropic phase behavior of model membranes composed of phosphatidylcholines containing linear saturated acyl chains. Biochemistry 26, 6118– 6126. 23. Lewis, R.N.A.H. & McElhaney, R.N. (1993) Calorimetric and spectroscopic studies of the polymorphic phase behavior of the homologous series of N-saturated 1,2-diacyl phosphatidyletha- nolamines. Biophys. J. 64, 1081–1096. 24. Zhang, Y P., Lewis, R.N.A.H. & McElhaney, R.N. (1997) Calorimetric and spectroscopic studies of the thermotropic phase behavior of the n-saturated 1,2-diacylphosphatidylglycerols. Bio- phys. J. 72, 779–793. 25. Allen, T.M., Hong, K. & Papahadjopoulos, D. (1990) Membrane contact, fusion and hexagonal (H II ) transitions in phosphatidyl- ethanolamine liposomes. Biochemistry 29, 2976–2985. 26. MacDonald, R.C., MacDonald, R.I., Menco, B.Ph, Takeshita, K., Subbarao, N.K. & Hu, L R. (1991) Small-volume extrusion apparatus for preparation of large unilamellar vesicles. Biochim. Biophys. Acta 1061, 297–303. 27. Beschiaschvili, G. & Seelig, J. (1992) Peptide binding to lipid bilayers. Nonclassical hydrophobic effect and membrane-induced pK shifts. Biochemistry 31, 10044–10053. 28. Jelokhani-Niaraki, M., Prenner, E.J., Kondejewski, L.H., Kay, C.M., McElhaney, R.N. & Hodges, R.S. (2001) Conformation and other biophysical properties of cyclic antimicrobial peptides in aqueous solution. J. Peptide Res. 58, 293–306. 29. Smith, P.F. (1992) Membrane lipid and lipopolysaccharide structures. In Mycoplasmas: Molecular Biology and Pathogenesi (Maniloff, J., McElhaney, R.N., Finch, L.R. & Baseman, J.B., eds), pp. 79–91. American Society for Microbiology, Washington, D.C. 30. McElhaney, R.N. (1992) Membrane structure. In Mycoplasmas: Molecular Biology and Pathogenesis (Maniloff, J., McElhaney, R.N., Finch, L.R. & Baseman, J.B., eds), pp. 113–155. American Society for Microbiology, Washington, D.C. 31. Jelokhani-Niaraki, M., Prenner, E.J., Kay, C.M., McElhaney, R.N. & Hodges, R.S. (2002) Conformation and interaction of the cyclic cationic antimicrobial peptides in lipid bilayers. J. Peptide Res. 60, 23–36. 32. Simecka, J.W., Davis, J.K., Davidson, M.K., Ross, S.E., Stadt- lander, C.T.K H. & Cassell, G.H. (1992) Mycoplasma diseases of animals. In Mycoplasmas: Molecular Biology and Pathogenesis (Maniloff, J., McElhaney, R.N., Finch, L.R. & Baseman, J.B., eds), pp. 391–415. American Society for Microbiology, Wash- ington, D.C. 33. Krause, D.C. & Taylor-Robinson, D. (1992) Mycoplasmas which infect humans. In Mycoplasmas: Molecular Biology and Patho- genesis (Maniloff, J., McElhaney, R.N., Finch, L.R. & Baseman, J.B., eds), pp. 417–444. American Society for Microbiology, Washington, D.C. 5920 M. Kiricsi et al. (Eur. J. Biochem. 269) Ó FEBS 2002 . The effects of ring-size analogs of the antimicrobial peptide gramicidin S on phospholipid bilayer model membranes and on the growth of Acholeplasma laidlawii B Monika. with phospholipid bilayer model membranes. We first investi- gated the effects of these GS ring-size analogs on the thermotropic phase behavior of LMVs composed of dimyristoylglycerophosphocholine