Sendai virus N-terminal fusion peptide consists of two similar repeats, both of which contribute to membrane fusion Sergio G. Peisajovich 1 , Raquel F. Epand 2 , Richard M. Epand 2 and Yechiel Shai 1 1 Department of Biological Chemistry, Weizmann Institute of Science, Rehovot, Israel; 2 Department of Biochemistry, McMaster University Health Sciences Centre, Hamilton, Ontario, Canada The N-terminal fusion peptide of Sendai virus F 1 envelope glycoprotein is a stretch of 14 amino acids, most of which are hydrophobic. Following this region, we detected a segment of 11 residues that are strikingly similar to the N-terminal fusion peptide. We found that, when anchored to the mem- brane by palmitoylation of its N-terminus, this segment (WT-palm-19–33) induces membrane fusion of large unila- mellar liposomes to almost the same extent as a segment that includes the N-terminal fusion peptide. The activity of WT- palm-19–33 was dependent on its specific sequence, as a palmitoylated peptide with the same amino-acid composi- tion but a scrambled sequence was inactive. Interestingly, two mutations (G7A and G12A) known to increase F 1 - induced cell-cell fusion, also increased the homology between the N-terminal fusion peptide and WT-palm-19–33. The role of the amino-acid sequence on the fusogenicity, secondary structure, and mechanism of membrane fusion was analyzed by comparing a peptide comprising both homologous seg- ments (WT 1–33), a G12A mutant (G12A 1–33), a G7A– G12A double mutant (G7A–G12A 1–33), and a peptide with a scrambled sequence (SC 1–33). Based on these experiments, we postulate that replacement of Gly 7 and Gly12 by Ala increases the ahelical content of the N-terminal region, with a concomitant increase in its fusogenic activity. Furthermore, the dissimilar abilities of the different peptides to induce membrane negative curvature as well as to promote isotropic 31 P NMR signals, suggest that these mutations might also alter the extent of membrane penetration of the 33-residue peptide. Interestingly, our results serve to explain the effect of the G7A and G12A mutations on the fusogenic activity of the parent F 1 protein in vivo. Keywords: viral entry; peptide–lipid interactions; spectro- scopic studies. A key step in the infection by enveloped viruses is the fusion between the viral and the cellular plasma or endosomal membranes. Most of the specialized viral envelope proteins directly involved in the fusion process, contain a discrete region of apolar amino acids, termed the Ôfusion peptideÕ, which is believed to play an important role in the merging of the membranes [1]. Although much is known about the 3D structure of fragments of fusion proteins in the absence of membranes [2–8], the intimate interplay between fusion peptides and the membrane is still unknown. Fusion peptides’ insertion into the cell membrane [9,10], viral membrane [11,12], or both [13,14] is believed to facilitate local dehydration [15] and to promote increased negative curvature strain in the bilayer (reviewed in [16]), factors that can help to overcome the energetic barriers associated with the fusion process. In addition, fusion peptides can serve as membrane anchors that facilitate partition of other regions of the viral envelope proteins to the membrane, which can subsequently participate in membrane merging [17]. Viruses from the Paramyxoviridae family are important respiratory tract pathogens of humans [18]. A salient feature of Paramyxoviridae infection is the fusion between infected and noninfected cells [19], a process mediated by the paramyxovirus envelope glycoprotein F. The F protein is synthesized as an inactive precursor, which is cleaved by a host protease, producing two fusion-active subunits, F 1 and F2 [20]. F 1 remains attached to the membrane by a transbilayer segment, whereas F2 and F 1 are disulfide bonded. Although it was initially thought that viral fusion glycoproteins contained a single fusogenic region respon- sible for the actual merging of the membranes, over the last years a more complex view has emerged. Both the region consecutive to the N-terminal fusion peptide and the one immediately before the transmembrane domain of HIV- 1 gp41 were shown to facilitate membrane fusion [17,21]. Furthermore, the F 1 subunit of Sendai and Measles virus (two distantly related members of the Paramyxovirus family) were shown to contain, in addition to the N-terminal fusion domain, an internal fusogenic segment, located downstream of the N-terminal heptdad repeat [22,23]. The structural organization of this internal fusogenic region, postulated based on studies using protein segments [22,24], was recently confirmed by the X-ray determined structure of the prefusion conformation of Newcastle disease virus F protein [25]. Both in the cases of Paramyxovirus and Correspondence to Y. Shai, Department of Biological Chemistry, Weizmann Institute of Science, Rehovot 76100, Israel. Fax: + 972 8 344112, Tel.: + 972 8 342711, E-mail: Yechiel.Shai@weizmann.ac.il Abbreviations: ATR-FTIR, attenuated total reflection Fourier transformed infrared spectroscopy; BOC, butyloxycarbonyl; Cho, cholesterol; DiPoPtdEth, dipalmitoleoylphosphatidylethanol- amine; DOPtdCho, dioleoylphosphatidylcholine; DOPtdEth, diol- eoylphosphatidylethanolamine; LUV, large unilamellar vesicles; NBD-PtdEth, N-(7-nitro-2,1,3-benzoxadiazol-4-yl) phosphatidyleth- anolamine; Rho-PtdEth, N-(lissamine rhodamine B sulfonyl) phos- phatidylethanolamine; T H , bilayer to hexagonal phase transition temperature; DTGS, deuterated triglyceride sulfate. (Received 1 April 2002, revised 14 July 2002, accepted 24 July 2002) Eur. J. Biochem. 269, 4342–4350 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03132.x Retrovirus internal fusogenic regions, the mechanism by which these segments destabilize membranes remains unknown. In an attempt to further our understanding of the process of viral infection and to determine whether the presence of a fusogenic region consecutive to the N-terminal fusion peptide is a characteristic common to unrelated viral families, here we analyzed the role in membrane merging of the N-terminal domain of Sendai virus F 1 protein. The first 14 amino acids of this region are termed Ôthe N-terminal fusion peptideÕ. Following this apolar segment, we detected a stretch of about 11 residues strikingly similar to the N-terminal fusion peptide (Fig. 1). We found that, when anchored to the membrane by palmitoylation of its N-terminus, this segment (WT-palm-19–33) induces membrane fusion of large unilamellar liposomes to almost the same extent as a longer fragment that also includes the N-terminal fusion peptide (WT 1–33); whereas a palmitoylated peptide with the same amino-acid composition of WT- palm-19–33, but a scrambled sequence (SC-palm-19–33), was inactive. In addition, we analyzed the role of the amino-acid sequence on the fusogenicity, secondary structure, and mechanism of membrane fusion exerted by the F 1 N-terminal region. EXPERIMENTAL PROCEDURES Materials BOC-amino acids were purchased from Novabiochem AG (La ¨ ufelfingen, Switzerland), and BOC-amino acid phenyl- acetamidomethyl (PAM)-resin was obtained from Applied Biosystems (Foster City, CA, USA). Reagents for peptide synthesis were obtained from Sigma. Dioleoylphosphatidyl- choline (DOPtdCho), dioleoylphosphatidylethanolamine (DOPtdEth), and dipalmitoleoylphosphatidylethanolamine (DiPoPtdEth) were purchased from Avanti Polar Lipids (Alabaster, AL, USA); cholesterol (Cho) was purchased from Lipid Products (South Nutfield, UK. NBD-PtdEth and Rho-PtdEth were purchased from Molecular Probes (Eugene, OR). All other reagents were of analytical grade. Buffers were prepared using double glass-distilled water. NaCl/KCl/P i is composed of NaCl (8 gÆL )1 ), KCl (0.2 gÆL )1 ), KH 2 PO 4 (0.2 gÆL )1 ), and Na 2 HPO 4 (1.09 gÆL )1 ), pH 7.3. Peptide synthesis The peptides (derived from Sendai virus F 1 protein Swiss- prot entry P04856) were synthesized by a standard solid phase method using a Boc-strategy on PAM-resin as described [26]. The peptides were cleaved from the resin by HF treatment and purified by RP-HPLC. Purity ( 99%) was confirmed by analytical HPLC. The peptide compositions were determined by amino-acid analysis and mass spectrometry. Preparation of lipid vesicles Large unilamellar vesicles (LUV) were prepared from DOPtdCho, DOPtdEth, and Cho (1 : 1 : 1) and when necessary with different amounts of Rho-PtdEth and NBD- PtdEth, as follows: dry mixed lipid films were suspended in NaCl/KCl/P i buffer by vortexing to produce large multi- lamellar vesicles. The lipid suspension was freeze-thawed six times and then extruded 20 times through polycarbonate membranes with 0.1 lm-diameter pores (Nuclepore Corp., Pleasanton, CA, USA). Peptide-Induced lipid mixing Lipid mixing of large unilamellar vesicles was measured using a fluorescence probe dilution assay [27]. Lipid vesicles containing 0.6 mol% each of NBD-PtdEth (energy donor) and Rho-PtdEth (energy acceptor) were prepared in NaCl/ KCl/P i as described above. A 1 : 4 mixture of labeled and unlabeled vesicles (110 l M total phospholipid concentra- tion) was suspended in 400 lL of NaCl/KCl/P i ,andasmall volume of peptide in dimethylsulfoxide was added. The increase in NBD fluorescence at 530 nm was monitored Fig. 1. Sendai virus F 1 N-terminal fusion domain is composed of two repeats. Panel A, the segment ranging from residue 7 to residue 17 is homologous to the segment 21–31. The mutations G7A and G12A, known to increase cell-cell fusion between cells expressing the F 1 protein and normal cells [30], increase also the homology between the two segments: G7 matches A21 and G12 matches A26. Panel B, sequence alignment of the segment 7–31 of Sendai virus with homol- ogous regions in other paramyxoviruses (HPIV 1, Human Parainflu- enza virus 1; Measles virus; Rinderpest virus; CDV, Canine Distemper virus;Mumpsvirus;SV5,SimianParainfluenzavirus5;NDV,New- castle Disease virus). The consensus sequence consists of those amino acids that are present in at least 50% of the aligned sequences. Panel C, alignment between the segments 7–17 and 21–31 of the consensus sequence. Ó FEBS 2002 Sendai virus-mediated membrane fusion (Eur. J. Biochem. 269) 4343 with the excitation set at 467 nm. The fluorescence intensity before the addition of the peptide was referred to as zero percent lipid mixing, and the fluorescence intensity upon the addition of Triton X-100 (0.05% v/v) was referred to as 100% lipid mixing. Electron microscopy The effects of the peptides on liposomal suspensions were examined by negative-staining electron microscopy. A drop containing DOPtdCho/DOPtdEth/Cho (1 : 1 : 1) LUV alone or a mixture of LUV and peptide was deposited onto a carbon-coated grid and negatively stained with 2% uranyl acetate. The grids were examined using a JEOL JEM 100B electron microscope (Japan Electron Optics Laboratory Co., Tokyo, Japan). Differential scanning calorimetry A Nanocal instrument from Calorimetry Sciences Corpora- tion (Spanish Fork, UT, USA) was used for all scans. Films composed of DiPoPtdEth and increasing mole fractions of peptide were prepared by dissolving the lipid in chloroform: methanol (2 : 1) and adding appropriate amounts of a dilute methanolic solution of peptide. The lipid DiPoPtdEth was used for determining the effects of the peptides on curvature because this lipid has a sharp bilayer to hexagonal phase transition at a moderate temperature of 43 °Csothat small shifts in the temperature of the transition of this lipid can easily be measured. This is not the case for the DOPtdCho/DOPtdEth/Cholesterol mixture used for other purposes in this manuscript. The films were dried in a test tube under a stream of nitrogen and then kept for 2–3 h in a vacuum dessicator. They were hydrated with Pipes buffer pH 7.40 (20 m M Pipes, 0.15 M NaCl, 1 m M EDTA and 20 mgÆL )1 NaN 3 ) to give a final lipid concentration of 7mgÆmL )1 , vortexed extensively and loaded into the calorimeter sample cell. The same buffer was placed in the reference cell. Heating scan rates of 0.75 °Cmin )1 were used. The bilayer to hexagonal phase transition was fitted using parameters to describe an equilibrium with a single van’t Hoff enthalpy and the transition temperature reported as that for the fitted curve. Data was analyzed with the program ORIGIN 5.0. 31 P NMR Spectroscopy The 31 P NMR spectra were measured using suspensions of about 10 mg of a lipid mixture containing equimolar amounts of DOPtdCho, DOPtdEth and cholesterol, with or without the addition of peptide at a lipid to peptide molar ratio of 200 : 1. The lipids and peptide were mixed in organic solvent and dried, as described for the DSC. The lipid film was hydrated with 200 lLof20m M Pipes, 1m M EDTA, 150 m M NaCl with 20 mgÆL )1 NaN 3 , pH 7.40. Spectra were obtained using a Bruker AV-500 spectrometer operating at 202.456 MHz in a 5-mm broadband inverse probe with triple axis gradient capa- bility. The spectra were acquired over a 48.544-kHz sweep width in 32K data points (0.338 s acquisition time). A 90° pulse width of 9.9 ls(90° flip angle) and a relaxation delay of 3.0 s were used. Composite pulse decoupling was used to remove any proton coupling. Generally, 700 free induction decays were processed using an exponential line broadening of 100 Hz and were zero-filled to 64K prior to Fourier transformation. Probe temperature was main- tained at 25 °C by a Bruker B-VT 3000 variable tem- perature unit. Temperatures were monitored with a calibrated thermocouple probe placed in the cavity of the NMR magnet. ATR-FTIR Measurements Spectra were obtained with a Bruker equinox 55 FTIR spectrometer equipped with a deuterated triglyceride sulfate (DTGS) detector and coupled with an ATR device. For each spectrum, 150 scans were collected, with resolution of 4cm )1 . Samples were prepared as previously described [28]. Briefly, DOPtdCho/DOPtdEth/Cho (1 : 1 : 1) (0.78 mg) alone or with peptide (23 lg) were deposited on a ZnSe horizontal ATR prism (80 · 7 mm). Prior to sample preparation the trifluoroacetate (CF 3 COO – ) counterions, which strongly associate with the peptide, were replaced with chloride ions through several washings of the peptides in 0.1 M HCl and lyophilizations. This allowed the elimin- ation of the strong C¼O stretching absorption band near 1673 cm )1 [29]. Peptides were dissolved in methanol, and lipids in a 1 : 2 methanol/CHCl 3 mixture. Lipid-peptide mixtures or lipids with the corresponding volume of methanol were spread with a Teflon bar on the ZnSe prism. Drying under vacuum for 30 min eliminated the solvents. Polarized spectra were recorded and the respective spectra corresponding to pure phospholipids in each polarization were subtracted from the sample spectra to yield the difference spectra. The background for each spectrum was a clean ZnSe prism. Hydration of the sample was achieved by introduction of excess of deuterium oxide ( 2 H 2 O) into a chamber placed on top the ZnSe prism in the ATR casting and incubation for 30 min prior to acquisition of spectra. Any contribution of 2 H 2 O vapor to the absorbance spectra near the amide I peak region was eliminated by subtraction of the spectra of pure lipids equilibrated with 2 H 2 O under the same conditions. ATR-FTIR Data analysis Prior to curve fitting, a straight base line passing through the ordinates at 1700 cm )1 and 1600 cm )1 was subtracted. To resolve overlapping bands, the spectra were processed using PEAKFIT TM (Jandel Scientific, San Rafael, CA, USA) software. Second-derivative spectra were calculated to identify the positions of the component bands in the spectra. These wavenumbers were used as initial parameters for curve fitting with Gaussian component peaks. Positions, bandwidths, and amplitudes of the peaks were varied until good agreement between the calculated sum of all compo- nents and the experimental spectra were achieved (r 2 > 0.995), under the following constraints: (a) the resulting bands shifted by no more than 2 cm )1 from the initial parameters, and (b) all the peaks had reasonable half- widths (< 20–25 cm )1 ). The relative contents of different secondary structure elements were estimated by dividing the areas of individual peaks, assigned to particular secondary structure, by the whole area of the resulting amide I band. The experiments were repeated twice and were found to be in good agreement. 4344 S. G. Peisajovich et al. (Eur. J. Biochem. 269) Ó FEBS 2002 RESULTS Sendai virus F 1 N-terminal fusion domain is composed of two repeats The N-terminal fusion peptide of Sendai virus F 1 envelope glycoprotein is formed by the 14 most N-terminal amino acids. Following this apolar region, we detected a segment of 11 residues strikingly similar to the N-terminal fusion peptide. As shown in Fig. 1A, the segment ranging from residue 7 to residue 17 is similar to the segment 21–31. Interestingly, the mutations G7A and G12A, known to increase cell–cell fusion between cells expressing the F 1 protein and normal cells [30], increase also the identity between the two segments: G7 matches A21 and G12 matches A26. Furthermore, as shown in Fig. 1B,C, homo- logous regions exist in other paramyxoviruses. This intrigu- ing finding prompted us to investigate the role played by the region consecutive to the Sendai virus N-terminal fusion peptide in membrane fusion. To this end, we synthesized a peptide corresponding to amino acids 19–33 from Sendai F 1 protein (WT-palm-19–33, see Table 1) and compared its ability to induce lipid mixing of large unilamellar liposomes with that of a longer segment (WT 1–33) that includes the N-terminal fusion peptide. In order to facilitate partition of the short WT-palm-19–33 to the membrane, its N-terminus was palmitoylated. To ensure that palmitoylation did not cause lipid mixing per se, we used as a control a palmitoy- lated peptide with the same amino-acid composition of WT-palm-19–33, but with a scrambled sequence (SC-palm- 19–33). WT-palm-19–33 induces lipid mixing of large unilamellar vesicles The ability of the peptides to induce lipid mixing of DOPtdCho/DOPtdEth/Cho (1 : 1 : 1) large unilamellar vesicles (LUV) was determined by the probe-dilution assay [27]. As depicted in Fig. 2, WT-palm-19–33 induces lipid mixing of LUV in a dose-dependent manner, although it is not as potent as the longer WT 1–33. On the contrary, SC- palm-19–33 is poorly active and palmitic acid alone did not induce any significant lipid mixing (not shown), reflecting that the WT-palm-19–33¢s potency is not solely a conse- quence of palmitoylation. It has been previously reported that the mutations G7A and G12A increase the cell–cell fusion activity of the full-length F protein [30]. Accordingly, the G12A mutation enhanced the lipid mixing ability of a peptide corresponding to the N-terminal segment of Sendai F 1 protein toward negatively charged LUV composed of PS [31]. As we wanted to determine the role of these mutations on the mechanism of membrane fusion exerted by Sendai F 1 protein, we also tested here the ability to induce lipid mixing of zwitterionic DOPtdCho/DOPtdEth/Cho (1 : 1 : 1) LUV of the mutant G12A 1–33, the double mutant G7A G12A 1–33, and a peptide with the same amino-acid composition of WT 1–33, but a scrambled sequence (SC 1–33). We found that successive replacement of Gly 7 and Gly 12 by Ala results in higher fusogenic activity (Fig. 2), whereas scram- bling of the wild-type sequence renders the SC 1–33 peptide inactive, indicating that the fusogenicity of the peptides Table 1. Sequences of the peptides and lipopeptides. WT 1–33 FFGAVIGTIALGVATSAQITAGIALAEAR EAKR G12A 1–33 FFGAVIGTIALAVATSAQITAGIALAEAR EAKR G7A–G12A 1–33 FFGAVIATIALAVATSAQITAGIALAEAR EAKR SC 1–33 VILEQRAFAVGGAILTSKFAIGGRTAAIA TAEA WT-palm-19–33 palmitoyl – ITAGIALAEAREAKR SC-palm-19–33 palmitoyl -AERATAELGIKAIAR Fig. 2. Peptide-promoted membrane fusion of DOPtdCho/DOPtdEth/ Cho (1 : 1 : 1) LUV as determined by lipid mixing. Panel A, dose dependence of lipid mixing. Peptide aliquots were added to mixtures of LUV (22 l M ), containing 0.6% NBD-PtdEth and Rho-PtdEth, and unlabeled LUV (88 l M )inNaCl/KCl/P i . The increase in the fluores- cence was measured 15 min after the addition of the peptide. The fluorescence intensity upon the addition of reduced Triton-X-100 (0.25% v/v) was referred to as 100%. Symbols: WT 1–33, empty cir- cles; G12A 1–33, empty triangles; G7A–G12A 1–33, filled squares; SC 1–33, empty squares; WT-palm-19–33, filled circles; SC-palm-19–33, filled triangles. Panel B, kinetics of lipid mixing for a peptide to lipid ratio of 0.06. Ó FEBS 2002 Sendai virus-mediated membrane fusion (Eur. J. Biochem. 269) 4345 depends on their specific sequences. Note that when the experiments were repeated using different liposome prepa- rations, differences of 15–20% were observed. However, with any given single liposome preparation the relative activities of the different peptides remained unchanged and the error was never higher than 5–10%. Lipid mixing is a result of membrane fusion In order to confirm that the observed intervesicular lipid mixing was the result of membrane fusion, suspensions of LUV were directly visualized under an electron microscope, before and after the treatment with the peptides. Briefly, DOPtdCho/DOPtdEth/Cho (1 : 1 : 1) LUV of 100 nm- diameter (200 l M ) were incubated for 15 min alone, or with each of the peptides (peptide lipid )1 molar ratio of 0.05) in NaCl/KCl/P i , before examination by electron microscopy. Figure 3 shows representative micrographs of the LUV without any peptide (Panel A), with WT 1–33 (Panel B), with G12A 1–33 (Panel C), with G7A G12A 1–33 (Panel D), with WT-palm-19–33 (Panel E), and with SC 1–33 (Panel F). The activity of SC-palm-19–33 was similar to that of SC 1–33 and therefore it is not shown. It is evident from the micrographs that the lipid mixing observed with G7A G12A 1–33, G12A 1–33, WT 1–33, and WT-palm-19–33 appear concurrently with an increase in the size of the vesicles, confirming that the ability of the peptides to induce lipid mixing is the result of membrane fusion. In order to shed light into their mechanism of action, we analyzed the ability of the peptides to lower the T H of DiPoPtdEth, to give rise to isotropic 31 P NMR signals, and determined the secondary structure of the membrane-bound full-length peptides by ATR-FTIR spectroscopy. Peptide effects on DiPoPtdEth transition temperature T H is a measure of the relative stability of the L a and H II lipid phases. A reduction in T H with the addition of a peptide can be interpreted as a tendency of the peptide to promote negative curvature of the membrane. As indicated by shifts in T H , at high peptide concentrations, WT 1–33 is the most potent in lowering the transition temperature, followed by G12A 1–33, SC 1–33, and by G7A–G12A 1–33. On the other hand, WT-palm-19–33 slightly increases T H (Fig. 4). 31 P NMR spectroscopy The shape of the 31 P NMR powder pattern of lipid mixtures serves as a good criterion for their morphology. The mixture of DOPtdCho:DOPtdEth:cholesterol (1 : 1 : 1) exhibits a spectrum typical of a membrane bilayer (see Fig. 5). Upon addition of only 0.05 mol% peptide, WT 1–33 and to a lower extent G12A 1–33 cause the formation of a structure that gives rise to an isotropic component at the chemical shift of phosphoric acid (Fig. 5). This is typical for highly curved membrane structures and the appearance of such peaks has been associated with higher rates of membrane fusion [32,33]. Structures such as hemifusion intermediates and fusion pores have highly curved surfaces that would allow for the motional averaging of the chemical shift anisotropy of the phospholipid. Interestingly, G7A G12A 1–33 and Palm-WT 19–33, although active in lipid mixing, did not give rise to significant isotropic components. A similar lack of isotropic component was observed for the scrambled peptide, SC 1–33, as well as the short and lipid- mixing inactive peptides. Fig. 3. Electron micrographs of negatively stained vesicles. Panel A, DOPtdCho/DOPtd- Eth/Cho(1:1:1)LUValone;PanelB,WT 1–33; Panel C, G12A 1–33; Panel D, G7A– G12A 1–33; Panel E, WT-palm-19–33; Panel F, SC 1–33. The vesicles were incubated with the peptides ([peptide]/[lipid] 0.05) for 15 min prior to visualization. The bar represents 200 nm. The effect of SC-palm-19–33 was similar to that of SC 1–33, therefore it is not shown. 4346 S. G. Peisajovich et al. (Eur. J. Biochem. 269) Ó FEBS 2002 Replacement of Gly7 and Gly12 by Ala results in higher a-helical content in the membrane As structure has been shown to be important for the activity of fusion peptides [31,34,35] and in order to investigate whether the different behavior of the peptides in inducing lipid mixing of LUV, lowering the T H of DiPoPtdEth, and promoting isotropic 31 P NMR signals could be related to differences in their structures, we determined the secondary structure of the membrane-bound peptides by ATR-FTIR spectroscopy. The spectra of the different peptides and the respective second derivatives calculated to identify the positions of the components bands are shown in Fig. 6. The percentages corresponding to the different structures are listed in Table 2. As expected, the combined G7A and G12A mutations significantly increase the a-helical content of the peptides, likely lowering the conformational flexibility associated with the higher amount of random structures observed in WT 1–33. On the other hand, scrambling of the wild-type sequence resulted in a very different spectrum with a predominance of aggregated b strands. DISCUSSION The region consecutive to the N-terminal fusion peptide participates in the fusion process In this study we report that the N-terminal domain of Sendai F 1 protein can be considered as two consecutive repeats (Fig. 1). Remarkably, the mutations G7A and G12A, known to enhance F 1 -induced cell-cell fusion [30], augment also the identity between the two segments. Interestingly, we found that, when partitioned into the membrane by palmitoylation of its N-terminus, the most C-terminal repeat induces fusion of large unilamellar liposomes (Figs 2 and 3). It should be noted that membrane anchoring via a C-terminal cysteine coupled to a modified phospholipid has been shown before to lead to fusogenic properties in a short synthetic model peptide [36]. However, the fusogenic activity of WT-palm-19–33 is not solely a consequence of its palmitoylation, as a palmitoylated peptide with the same amino-acid compo- sition of WT-palm-19–33, but a scrambled sequence, as well as palmitic acid alone, are not active. These results suggest that the Sendai virus F 1 N-terminal domain is composed of two repeats, both of which participate in the actual merging of the viral and cellular membranes. We do not believe that palmitoylation fulfills all of the functions of the initial 18 amino acids, as addition of these residues results in a longer peptide with a substan- tially increased fusogenic activity. The finding that homologous regions exist in other paramyxoviruses (Fig. 1, panels B and C) suggests that the structural Fig. 5. 31 P NMR spectra of an equimolar mixture of DOPtdCho:DOPtdEth:cholesterol with and without the addition of 0.05 mol per- centage peptide at 25 °C. The samples were hydrated with 20 m M Pipes, 1 m M EDTA, 150 m M NaCl with 0.002% NaN 3 , pH 7.40. Fig. 4. Shift of the bilayer to the hexagonal phase transition temperature of DiPoPtdEth as a function of the mole fraction of peptide. Symbols: WT 1–33, filled circles; G12A 1–33, filled squares; SC 1–33, filled diamonds; G7A–G12A 1–33, empty triangles; WT-palm-19–33, empty circles. Ó FEBS 2002 Sendai virus-mediated membrane fusion (Eur. J. Biochem. 269) 4347 and functional organization reported here for Sendai N-terminal fusion peptide, may be common to other members of the Paramyxoviridae family. Furthermore, recently it was shown that the polar region consecutive to the HIV-1 gp41 N-terminal fusion peptide also enhances its fusogenic activity, presumably by promoting self- association of the fusion peptide [17]. The similarity between what was found in HIV-1, a retrovirus, and what we report here in Sendai, a paramyxovirus, suggests that the N-terminal fusogenic domains from these distantly related viruses share a common mechanism. Secondary structure modulates the fusogenic activity of the peptides Changes in the secondary structure of fusogenic peptides have been shown to alter their activity [34]. Interestingly, the mutations Gly7 to Ala and Gly12 to Ala were shown to increase F 1 -induced cell-cell fusion [30]. Here, we analyzed how these mutations, which increase the identity between the two repeats, affect the structure and activity of the N-terminal region of Sendai F 1 protein. We found that replacement of Gly7 and Gly12 by Ala, which increased the a helical content of the peptide when bound to DOPtdEth/ DOPtdCho/Cho (1 : 1 : 1) membranes, enhanced the fuso- genicity of the peptide. On the contrary, scrambling of its amino-acid sequence resulted in an inactive peptide with a significantly reduced amount of a helix. The presence of Gly at position 7 and 12 in the wild-type Sendai F 1 protein imparts a greater flexibility to this region. Replacement of the two Gly by Ala, a residue with a higher helical propensity, may result in a longer or more stable helix. The G7A and G12A mutations are associated in vivo with a severe cytopathic effect, thus glycine may have been selected to balance high fusion activity with successful viral repli- cation [30]. It should be mentioned that Rapaport and Shai [31] did not observed a significant difference in the a helical content of WT 1–33 and its G12A mutant, as determined by circular dichroism in 70% TFE and methanol. However, these are only Ômembrane mimeticÕ environments, whereas here we measured the peptides’ secondary structure in the presence of phospholipid membranes. Unlike organic solvents, aqueous dispersions of phospholipids allow seg- ments of peptides to partition simultaneously into both aqueous and nonpolar solvent environments. As observed with HIV-1 gp41 fusion peptide, we cannot rule out that other secondary structures play also some role during the fusion process [17,35,37,38]. The mechanism of membrane fusion According to the stalk model [36] in its modified form [39,40], both a membrane fusion pore and the inverted Fig. 6. FTIR spectra deconvolution of the fully deuterated amide I band (1600–1700 cm 1 ) and their respective second derivatives. PanelA,WT 1–33; panel B, G12A 1–33; panel C, G7A–G12A 1–33; panel D, SC 1–33. The component peaks are the result of a curve fitting using a Gauss line shape. The amide I frequencies characteristic of the various secondary-structure elements were taken from [41]. The sums of the fitted components superimpose on the experimental amide I region spectra. The solid lines represent the experimental FTIR spectra after Savitzky-Golay smoothing; the broken lines represent the fitted com- ponents of the spectra. A 100 : 1 lipid:peptide molar ratio was used. Table 2. Secondary structure of the membrane-bound peptides according to FTIR spectroscopy. A 100 : 1 lipid:peptide molar ratio was used. The amide I frequencies characteristic of the various secondary-structure elements were taken from Jackson and Mantsch [41], mean values ± standard deviation are given. Sample Secondary Structure (%) a helix Random coil Aggregated strands Other structures WT 1–33 50 ± 8 32 ± 4 17 ± 4 1 ± 1 G12A 1–33 65 ± 1 13 ± 2 18 ± 1 4 ± 2 G7A–G12A 1–33 72 ± 1 13 ± 1 10 ± 1 5 ± 1 SC 1–33 31 ± 3 16 ± 2 43 ± 4 10 ± 4 4348 S. G. Peisajovich et al. (Eur. J. Biochem. 269) Ó FEBS 2002 hexagonal phase arise through a common intermediate. The first step is hemifusion between the outer leaflets of two opposing membranes that results in a stalk with high negative curvature. Subsequently, joining of the opposing monolayers leads to formation of the more stable trans monolayer contact (TMC) intermediate. The pathway after formation of TMCs diverges, leading either to membrane fusion or to the formation of an inverted hexagonal phase. Rupture of a single TMC produces a fusion pore, whereas transition to inverted hexagonal phase requires aggregation of numerous TMCs [40]. Several fusion peptides have been shown to lower the transition temperature from lamellar to inverted hexagonal phases (reviewed in [16]), indicating that they promote negative curvature in the membrane, thus favoring formation of the highly curved stalk intermediate. This property correlated well with the infectivity of influenza virus containing single amino-acid mutations in the fusion peptide segment of hemagglutinin [16]. Here we observed an inverse correlation between the lipid mixing ability of WT 1–33, G12A 1–33, and G7A G12A 1–33 and their ability to lower T H or to give rise to an isotropic peak in the 31 PNMR spectra. It should be noted that in the NMR and DSC experiments the peptide was added to both sides of the bilayer starting from a solution in organic solvent. This is different from the procedure for the lipid mixing assay in which the peptide is added to one side of the bilayer in buffer. This difference in the methodology that had to be used could contribute to a different behaviour of the peptides in the different system. An additional factor, however, that could contribute to the higher fusogenic activity of the single mutant G12A and the double mutant G7A G12A, despite their smaller effect on T H and on the 31 P NMR isotropic peak, as compared to the wild-type peptide, may be related to their more shallow penetration into the membrane, as shown for different constructs of the HIV-1 fusion peptide [17]. This possibility is supported by the effect of WT-palm-19–33, which due to the polar nature of its amino-acid composition, is likely to be located on the surface and, indeed, causes a slight increase in T H .As noticed before for other viral fusion peptides [1], when the amino-acid sequence of Sendai F 1 N-terminal region is represented as an a helix, it forms a sided helix, with most of the Gly and Ala residues lying on the same face. We can speculate that the presence of Ala at position 7 and/or 12 reduces the flexibility of G12A 1–33 and G7A G12A 1–33 by extending an a helix that runs closer to the membrane surface, thus diminishing the insertion into the membrane. Then, G12A 1–33 and more markedly G7A G12A 1–33 protrude from the membrane more than WT 1–33; thus, at high mole fraction, G12A 1–33 and even more G7A G12A 1–33 may sterically prevent aggregation of TMCs and the concomitant transition to the inverted hexagonal phase. Fusion between two opposing membranes requires the formation of only one fusion pore and therefore it is not affected by a protruding peptide. Alternatively, this intrigu- ing observation might be related to their different potency in facilitating the rupture of the dimple in the center of the TMC. The peptide that better promotes the rupture of the TMC will favor formation of fusion pore-like structures more easily, therefore lowering the chances of TMC aggregation and subsequent hexagonal phase formation. The current study generalizes the finding that, consecutive to their N-terminal fusion peptides, the envelope glycopro- teins from Paramyxo- and Lentiviruses have a relatively polar helical segment that facilitates membrane fusion but do not insert deeply into the membrane, suggesting that unrelated viral families share common mechanisms of cell entry. How such surface seeking helices promote fusion remains to be determined but could include lowering the degree of hydration of the membrane surface. ACKNOWLEDGEMENTS Sergio G. Peisajovich is supported by fellowships from The Mifal Ha’paiys Foundation of Israel and the Feinberg Graduate School of the Weizmann Institute of Science. This work was supported in part by the Canadian Institutes of Health Research (grant MT-7654). REFERENCES 1. White, J., Kielian, M. & Helenius,A. (1983) Membrane fusion pro- teins of enveloped animal viruses. Q. Rev. Biophys. 16, 151–195. 2. Bullough, P.A., Hughson, F.M., Skehel, J.J. & Wiley, D.C. (1994) Structure of influenza haemagglutinin at the pH of membrane fusion. Nature 371, 37–43. 3. Chan, D.C., Fass, D., Berger, J.M. & Kim, P.S. (1997) Core structure of gp41 from the HIV envelope glycoprotein. Cell 89, 263–273. 4. Weissenhorn, W., Dessen, A., Harrison, S.C., Skehel, J.J. & Wiley, D.C. (1997) Atomic structure of the ectodomain from HIV-1 gp41. Nature 387, 426–430. 5. Malashkevich, V.N., Chan, D.C., Chutkowski, C.T. & Kim, P.S. (1998) Crystal structure of the simian immunodeficiency virus (SIV) gp41 core: conserved helical interactions underlie the broad inhibitory activity of gp41 peptides. Proc. Natl Acad. Sci. USA 95, 9134–9139. 6. Yang, Z.N., Mueser, T.C., Kaufman, J., Stahl, S.J., Wingfield, P.T. & Hyde, C.C. (1999) The crystal structure of the SIV gp41 ectodomain at 1.47 A resolution. J. Struct. Biol. 126, 131–144. 7. Kobe, B., Center, R.J., Kemp, B.E. & Poumbourios, P. (1999) Crystal structure of human T cell leukemia virus type 1 gp21 ectodomain crystallized as a maltose-binding protein chimera reveals structural evolution of retroviral transmembrane proteins. Proc. Natl Acad. Sci. USA 96, 4319–4324. 8. Baker, K.A., Dutch, R.E., Lamb, R.A. & Jardetzky, T.S. (1999) Structural basis for paramyxovirus-mediated membrane fusion. Mol. Cell. 3, 309–319. 9. Harter, C., James, P., Bachi, T., Semenza, G. & Brunner, J. (1989) Hydrophobic binding of the ectodomain of influenza hemagglu- tinin to membranes occurs through the Ôfusion peptideÕ. J. Biol. Chem. 264, 6459–6464. 10. Stegmann, T., Doms, R.W. & Helenius, A. (1989) Protein-medi- ated membrane fusion. Annu. Rev. Biophys. Biophys. Chem. 18, 187–211. 11. Wharton, S.A., Martin, S.R., Ruigrok, R.W., Skehel, J.J. & Wiley, D.C. (1988) Membrane fusion by peptide analogues of influenza virus haemagglutinin. J. Gen. Virol. 69, 1847–1857. 12. Ruigrok, R.W., Aitken, A., Calder, L.J., Martin, S.R., Skehel, J.J., Wharton, S.A., Weis, W. & Wiley, D.C. (1988) Studies on the structure of the influenza virus haemagglutinin at the pH of membrane fusion. J. Gen. Virol. 69, 2785–2795. 13. Guy, H.R., Durell, S.R., Schoch, C. & Blumenthal, R. (1992) Analyzing the fusion process of influenza hemagglutinin by mutagenesis and molecular modeling. Biophys. J. 62, 95–97. 14. Hughson, F.M. (1995) Structural characterization of viral fusion proteins. Curr. Biol. 5, 265–274. 15. Zimmerberg, J., Curran, M. & Cohen, F.S. (1991) A lipid/protein complex hypothesis for exocytotic fusion pore formation. Ann. NY Acad. Sci. 635, 307–317. Ó FEBS 2002 Sendai virus-mediated membrane fusion (Eur. J. Biochem. 269) 4349 16. Epand, R.M. (1998) Lipid polymorphism and protein–lipid interactions. Biochim. Biophys. Acta. 1376, 353–368. 17. Peisajovich, S.G., Epand, R.F., Pritsker, M., Shai, Y. & Epand, R.M. (2000) The polar region consecutive to HIV-1 fusion peptide participates in membrane fusion. Biochemistry. 39, 1826–1833. 18. Collins, P.L., Chanock, R.M. & McIntosh, K. (1996) Parainflu- enza Viruses. In Virology (Fields, B.N., Knipe, D.M. & Howley, P.M., eds), pp. 1205–1241. Lippincott-Raven, Philadelphia, PA, USA. 19. Lamb, R.A. & Kolakofsky, D. (1996) Paramyxoviridae: The VirusesandTheirReplication.InVirology (Fields, B.N., Knipe, D.M. & Howley, P.M., eds), pp. 1177–1204. Lippincott-Raven, Philadelphia, PA, USA. 20. Homma, M. & Ohuchi, M. (1973) Trypsin action on the growth of sendai virus in tissue culture cells. J. Virol. 12, 1457–1465. 21. Suarez, T., Gallaher, W.R., Agirre, A., Goni, F.M. & Nieva, J.L. (2000) Membrane interface-interacting sequences within the ectodomain of the human immunodeficiency virus type 1 envelope glycoprotein: putative role during viral fusion. J. Virol. 74, 8038– 8047. 22. Peisajovich, S.G., Samuel, O. & Shai, Y. (2000) The F 1 protein of paramyxoviruses has two fusion peptides: implications for the mechanism of membrane fusion. J. Mol. Biol. 296, 1353–1365. 23. Samuel, O. & Shai, Y. (2001) Participation of two fusion peptides in measles virus-induced membrane fusion: emerging similarity with other Paramyxoviruses. Biochemistry 40, 1340–1349. 24. Ghosh,J.K.,Peisajovich,S.G.&Shai,Y.(2000)Sendaivirus internal fusion peptide: structural and functional characterization and a plausible mode of viral entry inhibition. Biochemistry 39, 11581–11592. 25. Chen, L., Gorman, J.J., McKimm-Breschkin, J., Lawrence, L.J., Tulloch, P.A., Smith, B.J., Colman, P.M. & Lawrence, M.C. (2001) The structure of the fusion glycoprotein of Newcastle dis- ease virus suggests a novel paradigm for the molecular mechanism of membrane fusion. Structure 9, 255–266. 26. Merrifield, R.B., Vizioli, L.D. & Boman, H.G. (1982) Synthesis of the antibacterial peptide cecropin A(1–33). Biochemistry 21, 5020– 5031. 27. Struck, D.K., Hoekstra, D. & Pagano, R.E. (1981) Use of resonance energy transfer to monitor membrane fusion. Bio- chemistry 20, 4093–4099. 28. Gazit, E., Miller, I.R., Biggin, P.C., Sansom, M.S. & Shai, Y. (1996) Structure and orientation of the mammalian antibacterial peptide cecropin P1 within phospholipid membranes. J. Mol. Biol. 258, 860–870. 29. Surewicz, W.K., Mantsch, H.H. & Chapman, D. (1993) Determi- nation of protein secondary structure by Fourier transform infrared spectroscopy: a critical assessment. Biochemistry 32,389– 394. 30. Horvath, C.M. & Lamb, R.A. (1992) Studies on the fusion peptide of a paramyxovirus fusion glycoprotein: roles of conserved residues in cell fusion. J. Virol. 66, 2443–2455. 31. Rapaport, D. & Shai, Y. (1994) Interaction of fluorescently labeled analogues of the amino-terminal fusion peptide of Sendai virus with phospholipid membranes. J. Biol. Chem. 269, 15124– 15131. 32. Siegel, D.P., Banschbach, J., Alford, D., Ellens, H., Lis, L.J., Quinn, P.J., Yeagle, P.L. & Bentz, J. (1989) Physiological levels of diacylglycerols in phospholipid membranes induce membrane fusion and stabilize inverted phases. Biochemistry 28, 3703–3709. 33. Ellens, H., Siegel, D.P., Alford, D., Yeagle, P.L., Boni, L., Lis, L.J., Quinn, P.J. & Bentz, J. (1989) Membrane fusion and inverted phases. Biochemistry 28, 3692–3703. 34. Lee, S., Aoki, R., Oishi, O., Aoyagi, H. & Yamasaki, N. (1992) Effect of amphipathic peptides with different alpha-helical con- tents on liposome-fusion. Biochim. Biophys. Acta. 1103, 157–162. 35. Nieva,J.L.,Nir,S.,Muga,A.,Goni,F.M.&Wilschut,J.(1994) Interaction of the HIV-1 fusion peptide with phospholipid vesicles: different structural requirements for fusion and leakage. Bio- chemistry 33, 3201–3209. 36. Markin, V.S., Kozlov, M.M. & Borovjagin, V.L. (1984) On the theory of membrane fusion. The stalk mechanism. Gen. Physiol. Biophys. 5, 361–377. 37. Pritsker, M., Jones, P., Blumenthal, R. & Shai, Y. (1998) A syn- thetic all D -amino acid peptide corresponding to the N-terminal sequence of HIV-1 gp41 recognizes the wild-type fusion peptide in the membrane and inhibits HIV-1 envelope glycoprotein-mediated cell fusion. Proc. Natl Acad. Sci. USA 95, 7287–7292. 38. Pritsker, M., Rucker, J., Hoffman, T.L., Doms, R.W. & Shai, Y. (1999) Effect of nonpolar substitutions of the conserved Phe 11 in the fusion peptide of HIV-1 gp41 on its function, structure, and organization in membranes. Biochemistry 38, 11359–11371. 39. Siegel, D.P. & Epand, R.M. (1997) The mechanism of lamellar-to- inverted hexagonal phase transitions in phosphatidylethanol- amine: implications for membrane fusion mechanisms. Biophys. J. 73, 3089–3111. 40. Siegel, D.P. (1999) The modified stalk mechanism of lamellar/ inverted phase transitions and its implications for membrane fusion. Biophys. J. 76, 291–313. 41. Jackson, M. & Mantsch, H.H. (1995) The use and misuse of FTIR spectroscopy in the determination of protein structure. Crit. Rev. Biochem. Mol. Biol. 30, 95–120. 4350 S. G. Peisajovich et al. (Eur. J. Biochem. 269) Ó FEBS 2002 . Sendai virus N-terminal fusion peptide consists of two similar repeats, both of which contribute to membrane fusion Sergio G. Peisajovich 1 ,. 269) Ó FEBS 2002 RESULTS Sendai virus F 1 N-terminal fusion domain is composed of two repeats The N-terminal fusion peptide of Sendai virus F 1 envelope glycoprotein