Copyright 0 Munksgaard, 1998
J . Peptide R e f . 5 1 , 1 9 9 8 , 142-148 Printed in the United Stares of America - ail rights reserved
JOURNAL OF PEPTIDE RESEARCH ISSN 1397-002X
A peptide of nine amino acid residues from a-sarcin cytotoxin is a membrane-perturbing structure JOSE M. MANCHENO,’ ALVARO MARTfNEZ DEL POZO,’ JUAN P.ALBAR? MERCEDES ONADERRA’ and JOSE G. GAVILANES’ 1
Departamento de Bioquimica y Biologia Molecular, Facultad de Quirnica, Universidad Cornplutense, 28040 Madrid; Centro Nacional de Biotecnologia, Departamento de Inrnunologia y Oncologi‘a, CSIC, Canto Blanco, 28049 Madrid, Spain
2
Received 8 July, revised 26 August, accepted for publication 28 September 1997
A water-soluble synthetic peptide with only nine amino acid residues, comprising the 131-139 sequence region of the cytotoxic protein a-sarcin (secreted by the mold Aspergillus giganteus), interacts with large unilamellar vesicles composed of acid phospholipids. It promotes lipid mixing between bilayers and leakage of vesicle aqueous contents, and it also abolishes the phospholipid phase transition. Other larger peptides containing such an amino acid sequence also produce these effects. These peptides acquire a-helical conformation in the presence of trifluoroethanol, but display P-strand conformation in the presence of sodium dodecyl sulfate. The interaction of these peptides with the lipid vesicles also results in p-structure. The obtained data are discussed in terms of the involvement of the 131-139 stretch of a-sarcin in its interaction with lipid membranes. 0 Munksgaard 1998. Key words: circular dichroism; lipid vesicle; peptide conformation; trifluoroethanol
a-Sarcin, a single polypeptide chain protein (150 amino acid residues) secreted by the mold Aspergillus giganteus, is the most representative member of a family of highly similar ribotoxins produced by different Aspergillus strains (1). a-Sarcin is cytotoxic for several human tumour cell lines (2). No protein membrane receptors have been detected; thus, the entering of asarcin into the target cells may be related to its ability to strongly interact with model phospholipid membranes. In fact, the protein promotes fusion of membranes, leakage of intravesicular aqueous contents and alterations of the phospholipid phase transition (3-7), and it translocates across phospholipid bilayers (8). The involved hydrophobic interactions are difficult to explain because the protein is highly polar (9). A 24-mer peptide, comprising the 116-139 sequence region of asarcin, promotes effects on phospholipid vesicles simi-
lar to the entire cytotoxin, which may suggest that this is a membrane-interacting portion of the protein (10). This peptide was selected because two consecutive hydrophobic P-strands were predicted in that region (1l), which was corroborated by ‘H and ”N nuclear magnetic resonance analysis of the protein (12, 13). We have studied in detail such a sequence region by considering several peptides. A nine amino acid residue portion seems to be essential for the vesicle-interacting properties of the polypeptide segment of a-sarcin being considered. EXPERIMENTAL PROCEDURES
Peptides (Table 1) were synthesized as amide in the Cterminus on an automated multiple peptide synthesizer (AMS 422, Abimed) by using the solid-phase procedure and standard Fmoc-chemistry in a base of 25 pmol. The synthesis was carried out on a N-a-FmocAbbreviations: ANTS, 8-aminonaphthalene-1,3,6-trisulfonic acid; DMP resin [4-(2’,4’-dimethoxyphenyl-Fmoc-aminoCD, circular dichroism; DMPC, dimyristoyl-Ptd-choline; DMPG, methyl)-phenoxy resin] (Novabiochem, La Jolla, CA), dimyristoyl-Ptd-glycerol; DPH, 1,6-diphenyl-1,3,5-hexatriene; with Fmoc-protected amino acids activated in situ with DPX, N,N’-p-xylene-bis-pyridiniumbromide; DTNB, 5 , 5 ’ PyBOP (benzotriazole- 1-yl-oxy-tris-pyrrolidinophosdithiobis-2-nitrobenzoic acid; Fmoc, 9-fluorenylmethoxycarbonyl; NBD-PE, N-(7-nitro-2-1,3-benzoxadiazol-4-yl)-dimyristoyl-Ptd-phonium hexafluorophosphate) in the presence of ethanolamine; PG, egg Ptd-glycerol; PS, bovine brain Ptd-serine; N-methylmorpholine and 20% piperidine/dimethylRh-PE, N-(lissamine rhodamine B sulfony1)-diacyl-Ptd-ethanola- formamide for deprotection. The protecting side-chain mine; SDS, sodium dodecyl sulfate; TFE, trifluoroethanol. groups were as follows: Ser, Thr and Tyr (tBu), Arg 142
Lipid interaction of a-sarcin peptides (pmc), Cys and His (Trt), Lys (Boc). Peptides were tion of the fluorescence intensity at 530 nm for excitacleaved from the resin with 2.5% ethanedithiol (King tion at 450 nm (4 nm slit width for both excitation and et al., 1990) as scavenger, precipitated and washed with emission beams) was monitored continuously on a cold methyl-tert-butyl ether, water-extracted and lyo- SLM Aminco 8OOO spectrofluorimeter.The percentages philized. They were subjected to further purification by of energy transfer (%ET) were calculated according to reverse-phase HPLC analysis on a Ultrasphere-ODS the expression (%ET) = [ l -(F/F,,)] x 100, where F is CI8column (10 x 150 mm) with a linear gradient wa- the final fluorescence intensity at 530 nm after additer/35% acetonitrile in 0.1 % trifluoroacetic acid. The tion of peptide and Fo is the fluorescence intensity at ) amino acid composition of the purified peptides exactly 530 nm of a vesicle population of DMPG (75 p ~ conmatches what was expected. Peptide concentration was taining 0.1% NBD-PE. To eliminate the potential condetermined from amino acid analyses. The peptides tribution of the sample turbidity to the fluorescence were hydrolyzed at 108°C in evacuated and sealed measurements, 10-mm Glan-Thompsonpolarizers (90"/ tubes for 24 h with 5.7 N HC1, containing 0.1% (w/v) 0") were used. Leakage of vesicle aqueous contents was measured phenol and nor-leucine as internal standard. The analyses were performed on a Beckman System 6300 amino by using the ANTSDPX (ANTS, g-aminonaphthaleacid analyzer. All cysteine-containingpeptides were ti- ne-l,3,6-trisulfonic acid; DPX, N,N'-p-xylene-bistrated with 5,5 '-dithiobis-2-nitrobenzoicacid (DTNB) pyridinium bromide; Molecular Probes Europe, Leiden, to determine the concentration of the thiol groups. Only The Netherlands) assay (17). Bovine brain phosphafree -SH groups were found except for the C8-peptide tidylserine (PS) vesicles containing 12.5 mM ANTS, 45 (see below), which contains cysteine at the NH,-termi- mM DPX, 20 mM NaC1, 10 mM Tris, pH 7.4, were prenal position. This peptide showed a time-dependent pared as described (18). In a typical assay a small voloxidation process; therefore, it was maintained under ume of peptide in the Tris buffer was added to the argon atmosphere. In all cases, only freshly prepared vesicles (75 PM lipid concentration) at zero time. The variation of fluorescence intensity measured through a peptide solutions were used. Circular dichroism (CD) spectra were obtained on 3-68 cut-off filter (>530 nm) upon excitation at 386 nm a Jasco 5-715 spectropolarimeter, The instrument was was registered on a SLM Aminco 8000 spectrofluoricalibrated using (+)-10-camphorsulfonic acid, AE = meter. The 0% and 100% leakage were taken as the +2.37 M-' cm-' at 290.5 nm and AE = 4 . 9 5 M-' cm-' at fluorescence intensity of a PS vesicle suspension be192.5 nm (14). Measurements were performed at 25°C fore and after addition of 1% Triton X-100 final conwith thermostated cylindrical cells of 0.10 or 0.05 cm centration, respectively. A single peptide stock solution pathlength. Data were collected every 0.2 nm at 50 nm/ was used; the stirring conditions were not altered durmin; CD results were expressed in units of degree x ing the study, because the mode of mixing vesicle and peptide solutions in leakage assays can affect the kicm2x dmol-' of amino acid residue. The different vesicles (except those for the leakage netics of the process (19). Measurements of the fluorescence depolarization of assay) were formed by hydrating a dry lipid film with Mops buffer (50 mM Mops, pH 7.0, containing 0.1 M 1,6-diphenyl-l,3,5-hexatriene(DPH) (Aldrich, MilwauNaCl and 1 mM EDTA) for 60 min at 37°C. The lipid kee, WI) were made on a SLM Aminco 8000 spectrosuspension was subjected to five cycles of extrusion fluorimeter equipped with 10-mm Glan-Thompson through two stacked 0.1 pm (pore diameter) polycarbo- polarizers. Labelling of the vesicles with DPH was carnate membranes (Nuclepore Costar, Cambridge, MA) ried out as described previously (20). Peptide-vesicle in an Extruder (Lipex Biomembranes Inc., Vancouver, mixtures were incubated for 1 h above the transition B.C. Canada). An additional series of five cycles of ex- temperature of the phospholipid, and later cooled down. trusion through two stacked 0.05 pm (pore diameter) The fluorescence emission was measured at 425 nm and membranes was used for the preparation of the lipid excitation at 365 nm, after equilibration of the samples vesicles for the CD studies. Phospholipid concentration at each required temperature. was determined as described (15). The lipid-mixing was assayed according to Struck RESULTS et al. (16). A vesicle population of dimyristoylphosphatidylglycerol (DMPG) containing 1% N-(7-nitro- The peptide comprising the 116-1 39 sequence region 2- 1,3-benzoxadiazol-4-yl)-dimyristoylphosphatidyl- of the cytotoxic protein a-sarcin largely perturbs phosethanolamine (NBD-PE, donor) and 0.6% N-(lissamine pholipid vesicles (10). Analysis of the hydrophobicity rhodamine B sulphony1)-diacylphosphatidylethanol- of this sequence according to the Eisenberg normalized amine (Rh-PE, acceptor) (Avanti Polar Lipids, Alabas- consensus scale (21) reveals three regions, two of them ter, AL, USA) was mixed with unlabelled vesicles at a with hydrophobicity values above the average of such 1:9 molar ratio (75 p~ final lipid concentration) in the a scale (Table 1). Three peptides accounting for the Mops buffer. At zero time, a small volume of peptide three mentioned regions and covering the complete sein the same buffer was added. The time-course varia- quence have been synthesized, N-, M- and C-peptide 143
J. M. Mancheiio et al. (from NH2-terminus,medium and COOH-terminus, respectively; thus, the 116-139 peptide is named NMCpeptide according to this nomenclature) (Table 1). We have compared the effects promoted by these peptides on lipid vesicles with those produced by the NMC-peptide. The N- and M-peptides do not affect phospholipid vesicles in terms of thermotropic behaviour, lipid-mixing between membranes and leakage of aqueous contents. However, the C-peptide completely abolishes the thermal transition of acidic phospholipid vesicles. Figure 1A shows the anisotropy variation of DPH-labelled DMPG vesicles on addition of the peptide. It produces a progressive decrease of the amplitude of the phospholipid phase transition. This peptide also induces extensive lipid-mixing between these membranes. A summary of the results obtained is given in Figure 2A. The variation of the steady-state energy transfer (%ET) is consistent with a dilution of the two fluorescent probes (donor and acceptor, NBD and Rh, respectively). The C-peptide also promotes leakage of the aqueous contents (Fig. 2B), as deduced from the time-dependent increase of the ANTS fluorescence. This is consistent with the dequenching promoted by dilution of the probe, ANTS, and the quencher, DPX, both entrapped in the same vesicle population. Considering the vesicle-interacting properties of this sequence, we have also synthesized many derived structures, C9- to C6-peptides(Table l), whose sequences are based on that of the C-peptide (Clo-peptide,ac-
t
0.4
r
I
0.4
r
03
0.3
02
0.2
0.1
0.1 1
1
20
30
40
‘c
I
I
I
20
30
40
I OC
I
I
I
20
30
40
FIGURE 1 Effect of peptides on the thermotropic behaviour of lipid vesicles. Fluorescence anisotropy (r) thermal variation of DPH-labelled DMPG vesicles in the presence of (A) C-peptide; (B) C9-peptide; and (C)C8-peptide, at the different peptide-to-lipid molar ratios indicated for each plot.
cording to this nomenclature) but containing from one to four residues less at the NH2-terminus,respectively. The C,-peptide behaves essentially as the C-peptide in
peptidellipid 0.10 0.20 0.30 0.40 0.50 0.60 0.70
%
TABLE 1 Amino acid sequences, structural propensities‘ and hydrophobicityb of the synthetic peptides studied Peptide
Amino acid sequence‘
N- NPGPAR (116-121) M- VIYTYPNK (122-129) C- VFCGIIAHTK (130-139) Cs- FCGIIAHTK (131-139) Cg- CGIIAHTK (132-139) C7- GIIAHTK (133-139) Cs-IIAHTK(134139) MCVIYTYPNKVFCGIIAHTK (122-139) NMCNPGPARVIYTYPhKVFCGIIAHTK (116-139)
b
a
0.80 0.85 1.00 1.00 0.98 1.02 1.09
0.75 -0.328 1.20 0.096 1.19 0.447 1.13 0.377 1.10 0.275 1.08 0.273 1.14 0.238
0.93
1.19
0.291
0.90
1.08
0.136
aStructural propensities for u-helix and for P-strand Q p > , calculated by averaging the Pa and P B values (39) across the complete peptide sequence. bHydrophobicity(normalized consensus scale) (2 1) averaged across the complete sequence. ‘Numbers in parentheses correspond to the extreme sequence positions in the cytotoxin a-sarcin. 144
0.10 0.20 0.30 0.40 0.50 0.60 1.00
peptidellipid FIGURE 2 Effect of peptides on lipid-mixing and leakage of lipid vesicles. (A) Fluorescence energy transfer (%ET) variation and (B) leakage of aqueous contents (%L) (see “Experimental Procedures”) versus peptide-to-lipid molar ratio: (D)C-peptide; (0)C9-peptide.
Lipid interaction of a-sarcin peptides terms of interaction with lipid vesicles (Fig. 1B; Fig. 2, A and B). The C,-peptide also abolishes the thermal transition of DMPG vesicles (Fig. 1C) although at higher peptidebipid molar ratio (about 6-fold) than that required for both Clo-and C9-peptides.Moreover, C8peptide does not promote any effect in terms of lipidmixing and leakage of aqueous contents. C7- and C6-peptidesdo not induce any change on the phospholipid vesicles. The conformation of all peptides so far considered has been analyzed by circular dichroism in the far-UV region. Spectral changes are not observed up to 250 p~ peptide concentration. Therefore, potential peptide-peptide interactions in solution should be discarded at the peptide concentrations herein studied (around 100 p~). In addition, no spectral changes are observed at different NaCl concentrations, from 0 to 1.O M, thus indicating that the peptide conformations are not sensitive to the ionic strength under the standard conditions used. Concentrated solutions of the C-sequence-containing peptides aged for more than 1 week display spectral changes consistent with an increase of the P-strand content (data not shown), which suggests a slow aggregation process. Therefore, the studies herein reported have been performed with freshly prepared peptide solutions. A quantitative estimation of the secondary structure of peptides from the CD spectra may not be accurate enough, because most of the calculation procedures are based on protein CD data. In addition, when peptides are considered, the obtained CD spectra might be the average of those corresponding to a heterogeneous population of different conformers at equilibrium in solution. Nevertheless, we have deconvoluted (22) the spectra to evaluate the conformational changes produced under the different conditions tested. In aqueous solution (50 mM Mops buffer, pH 7.0, containing 0.1 M NaCl and 1 mM EDTA), all of them display large negative ellipticity bands around 200 nm indicating that the peptides are mostly randomly organized, despite the sequence-based secondary structure propensities (Table 1). The CD spectrum of the M-peptide exhibits a positive ellipticity band centered at about 230 nm. This would be related to the presence of two tyrosine residues, because the La transition of this chromophore can produce positive bands in the far-UV region, with no specific requirements on stacking interactions among the aromatic side-chains (23). The magnitude of this band is +3,300 deg x cm2 x dmol-' on an amino acid residue basis. Thus, the ellipticity per tyrosine residue of the M-peptide is +13,200 deg x cm2 x dmol-', a value similar to those described for tyrosine-containing model structures (23). The conformation of the peptides has also been studied in the presence of phospholipid vesicles. Those composed of the zwitterionic phospholipid dimyristoylphosphatidylcholine (DMPC) do not produce any spectral variation on these peptides. However, nega-
tively charged phospholipids (phosphatidylglycerol or phosphatidylserine),either synthetic or natural, promote p-strand formation in the C-peptide (Fig. 3), whereas the N- and the M-peptides do not exhibit any change on their far-UV CD spectra. As the lipid-to-peptide molar ratio is increased, the curves become characteristic of p-structure. The existence of an isodichroic point at 212 nm suggests a random coil to P-strand twostate transition. A plot of the ellipticity at 220 nm versus the lipid-to-peptide molar ratio (Fig. 4A) indicates that 6:1 is a saturating ratio for the induced conformational change. Either egg phosphatidylglycerol (PG) or bovine brain phosphatidylserine (PS) vesicles promote effects qualitatively similar to DMPG vesicles in the C-peptide, although the resulting p-contents are slightly lower (data not shown). Thus, the effect of the lipid vesicles seems to be characteristic of acidic phospholipids and independent of other features of the lipid molecule as polar head or fatty acid composition. 210 220 230 240
210 220 230 240 I
I
'
"
5
0
40 .5
30
.10
30
.15
7
X
-
20
a
Y
10
0
-10
210
220
230
240
nm
FIGURE 3 Circular dichroism spectra of the C-peptide. CD spectra in the presence of DMPG vesicles at different lipid-to-peptide molar ratios (from 0 to 6). Inset: (A) Circular dichroism spectra of the C-peptide in the presence of different %TFE proportions (by volume), from 0 to 70% at 10% steps; (B) circular dichroism spectra of the C-peptide in the presence of different SDS concentrations (from 0 to 2 m~ at 0.5 mM steps; the spectrum at 5 mM SDS is also shown). [O],.r,,., mean residue weight ellipticity values in units of degree x cm2 x dmol-'. 145
J. M. Mancheiio et al.
DMPG/peptide 0
1
2
3
4
5
6
-2
To
-4
r
X
-
0 - 6 h( h(
2-8 -10 0
-2
random coil to a-helix transition. We have also studied the peptide conformation in the presence of SDS, a surfactant that can provide a hydrophobic environment also resembling membranes. A plot of the ellipticity values at 220 nm versus detergent concentration reveals a midpoint for the conformational transition at about 1.3 mM SDS (Fig. 4C), that is the reported value for the critical micellar concentration of the detergent at the salt concentration used (25). From 2 to 5 mM SDS (20:50 detergent-to-peptide molar ratio) there are no significant spectral changes, and the peptide contains S O % pstrand. The other derived C-peptides display a similar behaviour under both model environments, TFE and SDS, although showing their own characteristics: the ahelix induced by TFE depends on the peptide length and the proportion of p-structure induced by SDS depends on the detergent concentration (Fig. 4).
4
DISCUSSION
-6
5 - 8
3 Q,
Y
-10 -12
-14 -16
-1 8
0
20
40
% TFE
60
0
5
10 15 20 25
mM SDS
FIGURE 4 Ellipticity variation of C-peptides. [O]zzO, mean residue weight ellipticity values in units of degree x cm2 x dmol-’, at 220 nm versus (A) DMPG-to-peptide molar ratio; (B) Q (v/v) TFE; and (C) mM SDS. (0)C-peptide; (0) C9-peptide; ( A ) C8-peptide; (0) CIpeptide; (W) C6-peptide.
To explain this conformational change, we have studied the effect of “membrane mimicking” environments. Trifluoroethanol (TFE) is widely used for this purpose because of its ability to promote and strengthen intramolecular hydrogen bonding, as it would occur within the membranes (24). Thus, we have studied the effect of different TFE concentrations on the conformation of these peptides. An a-helical conformation is promoted in the C-peptide as deduced from these spectra. At 70% TFE, the a-helical content increases up to 59% (0% p-structure content) (Fig. 4). The C-peptide adopts a helical conformation although the average p-structure propensity is higher than the a-helix propensity (Table I). Thus, TFE induces a structure different from that expected from the intrinsic propensity of the amino acid sequence and induced by the phospholipid vesicles. The presence of an isodichroic point at 202 nm is also remarkable (Fig. 3, inset A), which suggests a two-state 146
The obtained results demonstrate that peptides containing the FCGIIAHTK amino acid sequence perturb acid phospholipid vesicles. The C9-peptide,the smallest vesicle-perturbing structure herein found, contains three ionizable groups (a-amine, imidazol and eamine of the single lysine residue, because the COOH-terminus is amidated). Thus, at pH 7.0 this peptide would contain one net positive charge from the E-amine group and two fractions of positive charge from the other groups. The charge values of these two groups depend on their particular pK values, although presumably the aamine group would be highly protonated. Therefore, this C-peptide would contain positive charges located at the two ends of the molecule. This feature may be an essential requirement for the interaction because of the acidic character of the interacting vesicles. In this regard, the effect of the peptide on the thermotropic behaviour of the vesicles suggests an increased acyl chain packing hindering the cooperative melting what may be interpreted in terms of charge neutralization of the polar head of the phospholipid. However, the positive charges cannot be the only structural requirement for lipid interaction because other peptides studied have a similar charge distribution, except the histidine residue that is only present in the C-region, and they are not interacting structures. An additional factor explaining the vesicle-perturbing ability of the C-region may be related to its p-structure propensity; although, the M-peptide, displaying a higher p-structure propensity, does not interact with the vesicles. However, the potential p-strand of the C-peptide is composed of more hydrophobic residues than that of the M-peptide (Table 1). This would be also valid for the C9-peptide,whereas the other C-sequence-based peptides, displaying decreased or absent abilities to perturb the vesicles, show lower hydrophobicity values (Table 1). This difference may result in a feasible distinct interaction with a bi-
Lipid interaction of a-sarcin peptides layer. A model has been proposed for the interaction of peptides such as somatostatin, melittin or Alzheimer P-amyloid peptides with lipid membranes (28,29). This model assumes two steps: an electrostatic attraction of the positively charged peptide to the membrane, resulting in an increased peptide concentration at the surface layer, and a hydrophobic binding step from the surface layer to the plane of binding. Then, it can be proposed that the vesicle-perturbing ability of the C9-peptideis caused by: (i) its positive charges at both the NH,- and COOH-terminal ends (interaction with acid vesicles); (ii) its propensity to form P-strand under conditions that enhance intermolecular hydrogen bonds (as would occur upon interaction with bilayers); (iii) the hydrophobicity of the resulting P-strand. The higher positive charge and hydrophobicity of the C9-peptide in comparison with the M-peptide would thus explain its interaction with the vesicles. The ability of the NMC-peptide to perturb phospholipid vesicles (10) would be related to the presence of the C-portion as elementary interacting unit. However, the NMC-peptide is more effective in perturbing the vesicles than the C-peptides, in terms of peptide concentration required for maximum leakage of intravesicular contents (10). This may be explained by the presence of secondary interacting units that would contribute to the perturbing action when incorporated to the C-portion. The NMC-peptide has a higher positive charge than the C-peptides what would favour the electrostatic attraction step. In addition, it might adopt a P-hairpin conformation because of the combination of the M- and C-portions. Thus, the NMC-peptide perturbs the vesicles at lower concentrations than the Cand C9-peptides,for which the P-sheet would result from peptide-peptide interactions at the bilayer surface, thus requiring higher peptide concentration in the bulk solution to promote leakage of the lipid vesicle. In this regard, we have also synthesized the MC-peptide and performed the same studies described above. The obtained data corroborate such an assumption. In fact, the peptide adopts a p-structure in the presence of DMPG vesicles (it behaves as the NMC-peptide), and it also produces lipid-mixing and leakage of intravesicular aqueous contents, being more effective than the C-peptides in terms of peptide-to-lipid molar ratio required for maximum effect (about 2- and 10fold, respectively; data not shown). The C- and the C,-peptides adopt a P-structure in the presence of SDS. It has been proposed that SDS first binds to the cationic groups of peptides, and additional detergent molecules then cluster around the peptide resulting in ordered structure (30). In this regard, signal peptide sequences are reported to form a-helix and/ or P-strand depending on whether they are inserted and/ or bound to the surface of a membrane (31, 32). This may suggest that the C-region-containing peptides would interact with the surface of the bilayer, result-
ing in a high peptide concentration at the surface layer of the vesicle and P-structure formation. Studies with a peptide from the presequence for subunit IV of cytochrome oxidase suggested a peptide concentration at the vesicle surface as high as 4 mM for a peptide concentration in a 25 p~ solution (33). This high concentration of peptide at the bilayer surface would facilitate its perturbation and the subsequent events, lipid-mixing and leakage of contents. CONCLUSIONS The results of this study demonstrate that: (i) the 131139 portion of the NMC-peptide is the smallest region displaying conformational changes in the presence of lipid vesicles (it is critical for the formation of p-structure in the presence of lipids); (ii) the formation of Pstructure is a requirement to perturb vesicle permeability (the C,-peptide abolishes the thermal transition of the phospholipidsbut it does not promote leakage of the vesicular contents or adopt p-structure conformation); (iii) the M-portion does not show any conformational change induced by the phospholipid vesicles except when incorporated to the C-portion (MC-peptide and NMCpeptide). Previous observations would agree with the participation of the C-region in the hydrophobic interaction of the cytotoxin with membranes. Thus, an overall destabilization of the protein structure resulted upon bilayer interaction (34). This has been observed for other proteins for which membrane binding is also governed by electrostatic interactions (35-37). In addition, a denatured form of a-sarcin, containing P-strands as the only regular secondary structure elements, promotes destabilization of the hydrophobic core of bilayers (38). Both observations make the participation of the C-region in the tertiary structure of the native protein compatible with its involvement in the vesicle-perturbing abilities of a-sarcin. ACKNOWLEDGMENTS Authors are indebted to Profs. F. Gavilanes and R. Rodriguez from this Department for critical reading of the manuscript and valuable suggestions. This work has been supported by grants from Direccion General Enseiianza Superior PB96-060 1 and Comunidad Aut6noma de Madrid AE00328-95 (Spain).
REFERENCES 1 . Kao, R. & Davies, J. (1995) Biochem. Cell Biol. 73,1151-1159 2. Tumay, J., Olmo, N., Jimknez, A., Lizarbe, M.A. & Gavilanes, J.G. (1993) Mol. CeN. Biochem. 122,3947 3. Gasset, M., Martinez del Pozo, A., Ofiaderra, M. & Gavilanes, J.G. (1989) Biochem. J . 258,569-575 4. Gasset, M., Ofiaderra, M., Thomas, P.G. & Gavilanes, J. G. (1990) Biochem. J . 265,815-822 5. Gasset, M., Ofiaderra, M., Martinez del Pozo, A,, Schiavo,
147
J. M. Manchefio et al. G.P., Laynez, J., Usobiaga, P. & Gavilanes, J.G. (1991) Biochim. Biophys. Acra 1068,9-16 6. Gasset, M., Oiiaderra, M., Goormaghtigh, E. & Gavilanes, J.G. (1991) Biochim. Biophys. Actu 1080,51-58 7. Mancheiio, J.M., Gasset, M., Lacadena, J., Rambn, F., Martinez del Pozo, A,, Oiiaderra, M. & Gavilanes, J.G. (1994) Biophys. J . 67, 1117-1125 8. Oiiaderra, M., Mancheiio, J.M., Gasset, M., Lacadena, J., Schiavo, G.P., Martinez del Pozo, A. & Gavilanes, J.G. (1993) Biochem. J . 295,221-225 9. Sacco, G., Drickamer, K. & Wool, I.G. (1983) J . B i d . Chem. 258,5811-5818 10. Mancheiio, J.M., Gasset, M., Albar, J.P., Lacadena, J., Martinez del Pozo, A., Oiiaderra, M. & Gavilanes, J.G. (1995) Biophys. J. 68,2387-2395 11. Mancheiio, J.M., Gasset, M., Lacadena, J., Martinez del Pozo, A., Oiiaderra, M. & Gavilanes, J.G. (1995) J . Theor. Biol. 172, 259-267 12. Campos-Olivas, R., Bruix, M., Santoro, J., Martinez del Pozo, A., Lacadena, J., Gavilanes, J.G. & Rico, M. (1996) Protein Sci. 5,969-972 13. Campos-Olivas, R., Bruix, M., Santoro, J., Martinez del Pozo, A., Lacadena, J., Gavilanes, J.G. & Rico, M. (1996) FEBS Lett. 397, 163-165 14. Chen, G.C. & Yang, J.T. (1977) Anal. Lett. 10, 1195-1207 15. Barlett, G.R. (1959) J . Biol. Chem. 234,466-468 16. Struck, D., Hoekstra, D. & Pagano, R.E. (1981) Biochemistry 20,4093-4099 17. Ellens, H., Bentz, J. & Szoka, F.C. (1985) Biochemistry 24, 3099-3 106 18. Mancheiio, J.M., Oiiaderra, M., Martinez del Pozo, A., DiazAchirica, P., Andreu, D., Rivas, L. & Gavilanes, J.G. (1996) Biochemistry 35,9892-9899 19. Polozov, I.V., Polozova, A.I., Anantharamaiah, G.M., Segrest, J.P. & Epland, R.M. (1994) Biochem. Mol. Biol. Int. 33, 1073-1079 20. Gavilanes, J.G., Lizarbe, M.A., Municio, A.M. & Oiiaderra, M. (1985) f n t . 1. Pept. Protein Res. 26, 187-194 21. Eisenberg, D., Wesson, M. & Wilcox, W. (1989) in Prediction of Protein Structure and the Principles of Protein C o n formation (Fasman, G.D., ed.), pp. 635-646, Plenum Press, New York 22. Woody, R.W. (1978) Biopolymers 17, 1451-1467 23. Perczel, A,, Hollbsi, M., Tusnhdy, G. & Fasman, G.D. (1991) Protein Eng. 4,669-679 24. Sonnichsen, F.D., Van Eyck, J.E., Hodges, R.S. & Sykes, B.D. (1992) Biochemistry 31,879CL8798
148
25. Helenius, A., McCaslin, D.R., Fries, E. & Tanford, C. (1979) Methods Enzyrnol. 56,734-749 26. Beschiaschvili, G. & Seelig, J. (1992) Biochemistry 31, 10044-10053 27. Seelig, J., Nebel, S., Ganz, P. & Bruns, C. (1993) Biochemistry 32,9714-9721 28. Terzi, E., Holzemann, G. & Seelig, J. (1994) Biochemistry 33, 1345-1350 29. Terzi, E., Holzemann, G . & Seelig, J. (1994) Biochemistry 33, 743&7441 30. Wu, C.S., Ikeda, K. & Yang, J.T. (1981) Biochemistry 20, 566-570 31. Briggs, M.S., Cornell, D.G., Dluhy, R.A. & Gierasch, L.M. ( 1986) Science 233,206208 32. Comell, D.G., Dluhy, R.A., Briggs, M.S., McKnight, J. & Gierasch, L.M. (1989) Biochemistry 28,2789-2797 33. Thorgeirson, T.E., Yu, Y.G. & Shin, Y-K. (1995) Biochemistry 34,5518-5522 34. Gasset, M., Mancheiio, J.M., Laynez, J., Lacadena, J., Femandez-Ballester, G., Martinez del Pozo, A,, Oiiaderra, M. & Gavilanes, J.G. (1995) Biochim. Biophys. Acra 1252, 126-134 35. Surewicz, W.K., Leddy, J.J. & Mantsch, H.H. (1991) Biochemistry 29, 8106-8111 36. Muga, A., Mantsch, H.H. & Surewicz, W.K. (1991) Biochemistry 30, 7219-7224 37. Zhang, F. & Rowe, E.S. (1994) Biochim. Biophys. Acta 1193, 2 19-225 38. Gasset, M., Manchefio, J.M., Lacadena, J., Martinez del Pozo, A,, Oiiaderra, M. & Gavilanes, J.G. (1995) Biochim. Biophys. Acta 1252,43-52 39. Chou, Y. & Fasman, G.D. (1978) Annu. Rev. Biochem. 47, 25 1-216
Address: Dr. Jose G . Gavilanes Departamento de Bioquimica y Biologia Molecular Facultad de Quimica Universidad Complutense 28040 Madrid Spain Phone: 34-1-3944158 Fax: 34-1-3944159 E-mail:
[email protected]
