Multiple lipid interactions of the Sendai virus fusogenic protein

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1987 hy The American Society for Biochemistry and Molecular Biology, Inc.

Val. 262, No. 24, Issue of August 25, pp. 11490-11496,19S7 Printed in U. S. A.

Multiple Lipid Interactions of the Sendai Virus Fusogenic Protein*

(Received for publication, October 14, 1986)

Nicola MoscufoS, Angelo GallinaS, Giampietro SchiavoB, Cesare Montecuccoj, and Maurizio TomasiS From the $Laboratorw di Biologia Cellulure, Istituto Superwre di Sanita, Viale Regina Elena 299, 00161 Roma and the SCentro Consiglw Nazionale delle Ricerche per le Bwmembrane and Istituto di Patologia Generale, Universita di Padova, Via Loredan 16, 35200 Padova, Italy

The membrane topology of the envelope of Sendai virus was investigated using various radioactive pho- toactivable hydrophobic reagents: 3-(trifluoromethy1)- 3-(m-[1261]iodophenyl)diazirine and the two phospholipid analogues, l-palmitoyl-2-(2-azido-4-nitro)benzoyl-sn -glycero-3- ph~spho[~H]choline and l-myristoyl-2,12-amino-(4-N-3 - nitro-l-azidopheny1)dodecanoyl-sn-glycero-3-phospho[14C]choline. The hemagglutinin-neuraminidase glycoprotein and the fusogenic (F) glycoprotein were labeled by all three probes, con- firming that these proteins are integral components of the viral envelope. The labeled F glycoprotein, com- posed of the two subunits F1 and Fa, was cleaved in situ with trypsin to yield two fragments, FS2 (32 kDa) and Fle (19 kDa). F2 was not labeled by any of the probes, suggesting an external location; whereas Fls was labeled by all probes and hence contains the portion of the F glycoprotein which traverses the viral envelope. Fragment F32 reacted both with 3-(trifluoromethyl)-3- (m-[ 12SI]iodophenyl)diazirine and with l-palmitoyl-2- (2 - azido - 4 - nitro)benzoyl- sn - glycero- 3 - pho~pho[~H] choline, but not with l-myristoyl-2,12-amino-(4-N-3- nitro- l-azidophenyl)dodecanoyl-sn-glycero-3-phos- ph~[’~C]choline. This result opens the possibility that the F glycoprotein is formed by a loop-like structure having multiple interactions with viral lipids.

~~~

Membrane fusion is an essential step in viral infection (1). Influenza and Semliki Forest viruses inject their genetic ma- terial into the cell cytoplasm after receptor-mediated endo- cytosis and membrane fusion induced by a low endosomal pH (1). Parainfluenza viruses, such as Sendai virus, fuse with the plasma membrane at neutral pH (1). The Sendai virus envelope contains lipids and proteins, and the lipid composition is rather similar to that of the plasma membrane from which the viral particle had budded (2). Two glycoproteins are inserted in the Sendai virus envelope. These are the hemagglutinin-neuraminidase (HN) glycoprotein,’ which is respon-

*This work was partly supported by Consiglio Nazionale delle Ricerche Grant “Ingegneria genetica e basi molecolari delle malattie ereditarie,” Contract 85.01425.51, and a grant from Regione Veneto. A preliminary account of this work has been presented at the IV Convegno dell’Associazione Italiana di Biologia Cellulare a del Dif- ferenziamento, Bardonecchia (TO), October 16-19, 1985 (Moscufo, N., Gallina, A., Montecucco, C., and Tomasi, M. (1986) Cell Biol. Int. Rep. 10, in press). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The abbreviations used are: HN, hemagglutinin-neuraminidase; F, fusogenic, HA, hemagglutinin; TID, 3-(trifluoromethyl)-3-(m-[’251] iodopheny1)diazirine; PCI, l-palmitoyl-2-(2-azido-4-nitro)benzoyl- sn-glycer0-3-phospho[~H]choline; PCII, l-myristoyl-2,12-amino-(4- N-3-nitro-l-azidophenyl)dodecanoyl-sn-glycero-3-phospho[“C]~ho- line; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis.

sible for the binding of the virus to sialic acid-containing receptors, and the fusogenic (F) glycoprotein, which plays an essential role in fusion (1). In newly formed viral particles, the F protein is present as a single glycopolypeptide chain, and the virus is not infectious and is not fusogenic (3, 4). Virus activation consists of the proteolytic cleavage of the F polypeptide chain to form two subunits, named F1 and Fz, linked by a single disulfide bridge (5,6). This cleavage occurs amino-terminal to a very hydrophobic sequence (7) known as the fusion peptide (7-9). Sendai virus mutated at the cleavage site is not fusogenic (8). Moreover, there is a similar sequence in the hemagglutinin protein of the influenza virus (10). For these and other reasons, it is believed that the fusion peptide plays a central role in the membrane fusion process.

It has been suggested that the fusion peptide is part of a water-exposed domain, but readily inserts into the target cell’s membrane to create a hydrophobic bridge that induces a lipid intermixing and, therefore, fusion (9).

To obtain more information on the localization of the F glycoprotein with respect to the viral membrane, we have used hydrophobic photolabeling reagents (11, 12). Trace amounts of these reagents, in radioactive form, are interspersed among the membrane lipids. On illumination, they produce highly reactive intermediates able to cross-link to and label membrane proteins, thus revealing which segments of the proteins are within or are close to the lipid bilayer.

The present results suggest that subunit Fz is external to the viral envelope, whereas subunit F1 spans the lipid bilayer. It was also found that a tryptic Fl fragment of 32,000 Da (F32) interacts with the polar head group region of the viral membrane and that the transmembrane section of the F protein is localized within a 19,000-Da (FIJ tryptic fragment. Possible biological consequences of this localization are discussed.

MATERIALS AND METHODS

Sendai virus, z strain, was propagated in 10-day-old chicken em- bryos. After 72 h, the virus was harvested and purified by various centrifugations as described previously (13).

[3H]- and [14C]nitroaryl azide phospholipids were prepared as previously reported (14) at the following specific activities: PCI, 2.8 Ci/mmob PCII, 174 Ci/mol. They were stored at -20 “C in absolute ethanol. Trypsin-treated with L-l-tosylamido-2-phenylethyl chloro- methyl ketone and soybean trypsin inhibitor were purchased from Worthington; Triton X-100 was from Pierce Chemical Co. 3-(Tri- fluoromethyl)-3-(m-[’251]iodophenyl)diazirine (TID; specific activity, 10 Ci/mmol) was obtained from Amersham Corp. All other reagents were purchased locally and were of the highest purity available.

Labeling Procedure-Experiments were carried out under nitrogen and red illumination. In a 0.5-ml volume, either TID (5.0 pCi) or a mixture of both PC1 (1.2 pCi) and PC11 (0.3 pCi) was added to a viral suspension (2 mg/ml protein) and incubated at 37 “C with gentle shaking for 90 min. The ethanol added with the reagents amounted to less than 1% of the final volume. The mixture was then irradiated with an ultraviolet lamp for 10 min at 4 “C as previously described (15).

11490

Lipid Interactions of Fusogenic Protein 11491

Trypsin Digestion-The labeled viral particles were treated with HN and F. To separate the proteins from the major part of the trypsin as described by Asano et al. (16). Briefly, viral particles detergent, lipids, and noncovalently bound probes, cold acetone dispersed in 20 mM Tris-HC1, 150 mM NaCl at pH 7.4 (TBS) were (-20 "C) was added to a final concentration of 90% (v/v). The acetone incubated with trypsin (1% (w/w) of total viral protein) for 2 h a t suspension was centrifuged at 800 X g for 15 min at 5 "C, and the 37 "C, and digestion was stopped by adding a 2-fold molar excess protein pellet was dried under a gentle nitrogen stream to eliminate (over trypsin) of soybean trypsin inhibitor. residual acetone. The recovered protein typically accounted for 18%

suspension (either treated or untreated with trypsin) were recovered Reconstitution of Sendai Virus Enuelop-The reconstitution pro- by ultracentrifugation (100,000 x g for 30 min at 5 "C) and resus- cedures were performed by the method described by Vainstein et al. pended with twice the viral protein weight of Triton X-100 (4% (w/ (17), which involves the following steps: i) incubation of viral particles v) in TBS). This detergent specifically dissolves the envelope, leaving with the mild detergent Triton X-100 (weight ratio, 1:2); ii) ultracen- intact the nucleocapsid (17). After 1 h at room temperature, the trifugation of the mixture a t 140,000 X g for 1 h at 5 "C to sediment detergent/virus mixture was ultracentrifuged at 140,000 X g for 1 h; the nucleocapsid and the polymeric matrix protein as a pellet; and the pellet contained nucleocapsid together with the polymeric matrix iii) removal of Triton X-100 by direct addition of Bio-Beads SM-2 protein. The clear supernatant contained the envelope glycoproteins (weight ratio of detergent to resin, 1:8). As a consequence of Triton

Purification of Sendai Virus Glycoproteins-The viral particles in of the total viral protein.

Not-reduced HN HN 41 2.

4-2

* "** *. .. I

-HN-

- F,-

- F=-

d

*

FIG. 1. TID labeling of Sendai virus glycoproteins. TID-labeled Sendai virus glycoproteins HN and F are shown before (a and b ) and after (c and d ) trypsin digestion and separated by bidimensional SDS-PAGE. The first dimension was done on 7% polyacrylamide gel without reducing agents. At the top of a, Coomassie Blue profile of the rod gel, run without reducing agents, is shown; the corresponding profile of c was identical to a and thus is not included. Then, the second dimension was run on 12.5% polyacrylamide slab gel with reducing agent; the slab gel also includes a sample which was directly run as reduced. a and c are the stained gels, and b and d are corresponding autoradiographies. The labeled glycoproteins (200 pg) were isolated from viral particles (1-mg protein content) by extraction with Triton X-100, precipitated by acetone, electrophoresed, and autoradiographed as described under "Materials and Methods."

11492 Lipid Interactions of Fusogenic Protein

I a

b

Y L c 0

d

F1 3 2 FlS F2

0 - +

b O=P-O-LN(C 'n , ),

l i ' a o ' i o

C f

0

I I

to 20 30

SLICE NUMBER SLlCE NUMBER FIG. 2. Labeling of Sendai virus F glycoprotein by photoreactive phopholipids. a, the rectangle

represents the vertical strip of the Coomassie Blue-stained gel which contains the F glycoprotein after bidimensional SDS-PAGE as described for Fig. 1 and under "Materials and Methods"; b, the corresponding distribution of PC1 radioactivity; c, that of PC11 determined after gel slicing and counting; d, the rectangle represents the vertical strip of stained gel containing the F glycoprotein fragments which were obtained by trypsin digestion and reduction; e, the corresponding distribution of the PC1 radioactivity; f, that of PCII. The structural formulas of PC1 and PC11 are shown to the right. Part of the slab gel sliced and counted is in the rectangles.

X-100 removal, the viral lipids reform vesicles wherein the glycoproteins become inserted. The recovery of hemagglutinating and fusion ability in reconstituted Sendai virus envelope suggests that the structure and orientation of glycoproteins are preserved.

Bidimenswnnl SDS-PAGE of Viral Glycoproteins-Samples (-200 pg of protein) dried as described above were resuspended in 50 pl of 100 mM Tris-HC1, pH 7.4, containing 10% SDS, loaded on 7% polyacrylamide rod gels (5 X 120 mm), and electrophoresed. At the end of the run, the gel was placed in electrophoresis buffer (25 mM Tris, 200 mM glycine, 0.1% SDS) containing 100 mM dithiothreitol for 30 min at 4 "C and held for a further 15 min at 40 "C. The rod

gels were then placed at the top of a 12.5% discontinuous slab gel for the second run with a running buffer containing 1% (v/v) j3-mercap- toethanol.

After the Coomassie Blue staining, each gel (140 X 160 mm) was divided in vertical strips (20 X 160 mm). The strips were cut in 3- mm-thick slices, and these were dissolved in 400 pl of Packard Soluene 350 and counted by liquid scintillation (15). A typical vertical strip, containing the subunits of F glycoprotein or its tryptic fragments, is represented in Fig. 2, a and d, respectively. Labeling with TID was determined in a y-counter or alternatively by autoradiog- raphy of the dried gels with Kodak X-Omat EO 222 films.


FIG. 3. Effects of trypsin and reconstitution on F glycoprotein labeling. Labeling patterns are shown of F

tagged with PC1 (a) and PC11 (b) and glycoprotein isolated from viral particles

digested with trypsin before photoacti- vation. Methods were as described for Fig. 2 and under “Materials and Meth- ods.” c and d show the distribution of PC1 and PC11 labeling, respectively, of F glycoprotein in reconsituted envelope.

TABLE I Hydrophobic labeling of HN glycoprotein of Sendai virus

The data represent the average of five experiments. The standard deviations are also reported.

Samples con- PC1 /protein PC11 /protein

taining HN Tetramer Dimer Tetramer Dimer

d m / @ “C dpm

SV-UV-TrSp 4.2 f 1.2 10.2 f 1.8 3.2 f 0.5 8.1 f 0.5 SV-Try-UVb 6.4 f 1.2 10.5 f 1.2 4.5 f 0.5 10.5 f 0.5 RSVE‘ 2.8 f 1.2 5.0 f 1.2 2.3 f 0.5 4.0 f 0.5 Sendai virus (SV) particles, tagged with PC1 and PCII, were UV-

irradiated and then trypsinized (Try). The radioactivity content of HN glycoprotein was analyzed on Triton X-100 extract as described under “Materials and Methods.”

* Same as in Footnote a, except that trypsinization was performed before irradiation.

RSVE, reconstituted Sendai virus envelope. Viral envelopes were reconstituted as described (17) and tagged with the phospholipid probes in the presence of Triton X-100. After illumination, the HN band radioactivity was determined as described under “Materials and Methods.”

Agglutinating and Hemolytic Sendai Virus Activities-Agglutinat- ing activity was estimated by determining the largest dilution of the virus still able to agglutinate human red blood cells. Hemolytic activity was evaluated by incubating viral particles with 0.6% (v/v) human red blood cells for 45 min at 37 “C. Samples were diluted 3- fold with cold TBS, and intact cells were removed by centrifugation. The released hemoglobin was estimated by evaluating the absorbance of supernatant at 540 nm (13).

RESULTS

F1 Trypsin Cleauuge-In order to study the membrane topology of the different parts of the F protein, several proteolytic agents and protocols were tested. As reported by Asano et ul. (16), the only reproducible fragmentation pattern was given by trypsin action on intact virus particles. Both the viral nucleocapsid and polymeric matrix proteins were removed by Triton X-100 treatment and ultracentrifugation (17). As shown in Fig. la, among the viral proteins, mainly the envelope glycoproteins were recovered. The F subunits were well resolved from the hemoagglutinin-neuraminidase glycoprotein, which before S-S reduction runs as dimers and

tetramers. Only F1 is cleaved by trypsin (Fig. IC) and gives rise to two fragments, F32 and F19, which migrate together until reduced (Fig. IC). Since F19 has only 1 cysteine residue (at position 308), the results suggest that F19 and FS2 are connected by a single S-S bridge. The maximal yield of FI9 and F32 was around 60% of F,; prolonged incubations, higher temperatures or trypsin:virus ratios, or different pH values lead to further decrease of F1 without concomitant increase of the F19 and F32 fragments (data not shown). The amino- terminal analysis of F19, performed by Asano et al. (16) and confirmed by us, together with the available primary structure of F protein (22) suggest that Arg-296 is the probable site of trypsin cleavage, as indicated in Fig. 4.

Inspection of Fig. 4 shows that our two fragments of F1 protein contain the two major hydrophobic stretches of this protein, residues 1-26 and 384-407, which are thought to be the fusion and anchor peptides, respectively (18).

Hydrophobic Photolubeling of Sendai Virus Envelope with TZD-To label and hence identify the membrane-spanning segments of the Sendai virus envelope, we used TID. This is a small lipophilic reagent which partitions within the membrane where it can be converted to a highly reactive carbene by UV irradiation (19). Membrane components accessible to the probe will be labeled and will become radioactive. Fig. 16 shows that HN glycoprotein and F1 are labeled, whereas F2 is not. This suggests that HN glycoprotein and Fl are integral proteins of the Sendai virus envelope and that F2 is either peripheral or inserted but shielded by other integral proteins. Fig. Id shows that both fragments of F1 are labeled; the largest share of radioactivity (70%) was bound by Fls. This result suggests that both fragments interact with lipids, but does not provide information on the depth of penetration of the proteins into the membrane because of the undefined location of TID in the lipid bilayer. Moreover, TID can partition into hydrophobic pockets present on the hydrophilic domains of membrane proteins. To overcome these limitations and to improve spatial resolution, we used phosphatidylcholine de- rivatives, each bearing a photoreactive group covalently bound to one fatty acid chain (11, 12).

Hydrophobic Photolubeling of Sendai Virus Envelope with Phospholipid Analogues-The structural formulas of the pho-


tosensitive phospholipid reagents used here are shown on the right of Fig. 2. They carry the nitroaryl azido photoreactive group at two different levels of one fatty acid chain in order to report on different membrane regions. PC1 will label only those protein segments intercalated with the phospholipid polar head groups, whereas PC11 labels protein regions present in the hydrophobic core of the lipid bilayer. Although these phospholipid analogues readily insert from unilamellar ves- cicles into the membranes of several cellular organelles and red blood cells (11), they do not readily insert into Sendai viruses. Hence, they were used here as an ethanolic solution (final concentration not exceeding 1%). In this way, 96% of PC1 and 47% of PC11 were found associated with the viruses after centrifugation. These different values correlate with the free monomer concentration of the two reagents.

Fig. 2 illustrates the patterns of radioactivity incorporated by the Sendai virus envelope glycoproteins and their fragments after photolabeling with PC1 and PCII. Fig. 2 (a-c) shows that both PC1 and PC11 labeled F1, but that neither of them labeled F2. This result provides strong evidence that F1 is an integral component of the viral membrane and suggests that F2 is external to the membrane, in agreement with the fact that F2 has no long hydrophobic stretches (see Fig. 4). It differs from the conclusion of Ozawa et al. (20) who suggested that F2 was an integral protein on the basis of its tight association with the viral particles even under strong de- naturation and reducing conditions. The distribution of radioactivity between the two fragments F,, and F32 was determined after selective trypsin cleavage of labeled F,. Fig. 2 (d- f ) shows that the F32 fragment was labeled only by the super-


ficial probe PCI, whereas FI9 is labeled by both the shallow and the deeper probes. In the latter fragment, the ratio between the amounts of PC1 to PC11 radioactivity was similar to the PC1:PCII labeling ratio of HN glycoprotein, which possesses only one hydrophobic stretch, presumably its transmembrane segment (28). These data confirm that F I ~ is a transmembrane polypeptide and indicate that fragment F32 may interact with the polar head groups of viral lipids. The latter result implies that the protein region protruding out of the viral envelope deflects back to the viral lipid surface.

To assess better that labeling by PC1 monitors the actual interaction between viral lipids and F32, we carried out further hydrophobic photolabeling experiments in which the F protein had been previously nicked with trypsin, which causes an alteration of lipid-protein interaction (21). The trypsin- inactivated viral particles were incubated with PC1 and PCII, photoactivated by UV irradiation, and analyzed as described above. Fig. 3 ( a and b ) shows that the labeling patterns were similar to those of the virus labeled in the native conditions. In order to test the procedure of insertion of photoreactive phospholipids, PC1 and PC11 were incorporated into viral envelopes as they reformed from detergent solution. The method consisted of adding the probes to a Triton X-100 extract of the viral envelope and removing the detergent by adsorption onto Bio-Beads SM-2 for 18 h. As the concentration of Triton X-100 decreased, the envelope of Sendai virus gradually reformed (17). The procedure involves the loss of about 30% of the viral envelope glycoproteins together with 40% of the total lipids. Fig. 3 (c and d ) shows that a very similar labeling pattern was obtained by inserting the probes during the process of envelope reconstitution.

The above experiments provide evidence of an interaction between F32 and viral lipids at a surface level; they cannot exclude the possibility of a deeper insertion of F32 in the lipids because a lack of PC11 labeling may result from exposure to the probe of poorly reactive amino acid residues uersus the nitroarylnitrene intermediate (1 1).

Table I shows the labeling of HN glycoprotein both in dimeric and tetrameric form. The amount of radioactivity covalently incorporated by PC1 and PC11 labeling does not vary either when trypsinization precedes the labeling step or when the labeling was performed on the reconstituted Sendai virus envelopes. The fact that the labeling of HN glycoprotein parellels that of F protein suggests that the interaction between the envelope glycoproteins and lipids does not change when fusion is abolished by trypsinization.

Structure Analysis of F Glycoprotein-To shed more light on the structure of F glycoprotein, we analyzed the amino acid sequence with respect to the following parameters: hy- drophobicity, proline and cysteine distribution, and secondary structure predicted by the algorithm of Garnier et al. (21) (Fig. 4). Only those segments which contain more than 7 residues arranged in the same secondary structure were considered. We utilized the sequence of F protein determined by Shioda et al. (22) on the z strain of Sendai virus. This has 10 cysteine residues, which would allow up to five disulfide bonds. Since we could not find any free SH group in the F glycoprotein,' there may, in fact, be five S-S bonds. This sequence is slightly different from that determined by Blumberg et al. (18) on the Harris strain of Sendai virus. Fig 4 indicates that the F3, fragment contains two regions having opposite fea- tures. The first one is segment 224-297, where there is a considerable cluster of cysteines and prolines (7 and 6, respectively); since these residues are indicative of bending, this region might be highly folded. The second one corresponds to -

* A. Gallina and M. Tomasi, unpublished results.

segment 20-88, which is predicted to be arranged in an a- helical structure. Such a long helix is unusual in proteins; however, the influenza hemagglutinin (HA) glycoprotein possesses a 53-residue-long a-helical structure in the HA2 subunit, right in that part of the molecule where sequence homology with F protein is large (18).

DISCUSSION

We report here some hydrophobic photolabeling studies on the localization of the F glycoprotein of Sendai virus with respect to the membrane envelope. We have been able to probe different regions of the lipid bilayer: a superficial part with PC1 and a deeper hydrophobic zone with PCII.

Of the two subunits composing the F glycoprotein, only F1 interacts at superficial and deep levels with the viral envelope lipids, whereas F P does not appear to be in contact with lipids. These experiments do not exclude the possibility that the lack of labeling is due to a too low reactivity of the Fz residues exposed to lipids. However, F2 is labeled under denaturing conditions in the presence of SDS (data not shown), support- ing the suggestion that Fz does not interact with lipids. Hence, either it is external to the membrane or it is in the membrane but shielded from contact with lipids by other proteins. For- merly, the F2 subunit has been considered to be embedded in the viral lipid membrane because of failure to separate it from the envelope (20).

The interaction of F1 with phospholipids appears to be localized at two different regions of the polypeptide chain. The first region is the hydrophobic segment which traverses the viral membrane and anchors the entire F glycoprotein; the second region is part of the protein that protrudes outside the envelope.

Trypsin cleaves F, of active viral particles in the two fragments FI9 and F32 (16). The distribution of labeling between F,, and F32 provides further evidence for the idea that FI9 traverses the viral membrane with the hydrophobic stretch starting at Val-385 and ending at Tyr-407 (Fig. 4). F32 appears to interact with lipids only at the superficial level since it is labeled by the superficial probe PCI, but not by the deeper one, PCII. This result suggests that F, extends out of the membrane and returns to the membrane without penetrating into the hydrophobic core. Despite the fact that protein lo- cations at the water-lipid interface of the membrane are rare, they have been suggested for a 9-residue segment of the light- harvesting protein B870 (23). Neither the extension of the loop so formed by F protein nor the residues involved in this superficial interaction with the membrane can at present be evaluated. Only some speculations about the place taken by Sendai virus fusion peptide may be inferred by comparing the x-ray-defined situation of HA glycoprotein with the F glycoprotein since these proteins possess a sequence homology (18). The Sendai fusion peptide is located at the end of the 62-residue-long segment which has been predicted to be arranged in an a-helix, whereas the HA fusion peptide is 67 residues away. In the HA structure, this a-helical segment is almost perpendicular to the plane of the membrane wjth one extremity close to the viral lipid and the other end 82 A away. If this long a-helical structure is actually present in F protein and if it has a topology similar to HA glycoprotein, the position of the fusion peptide, at one of the two extremities of the a-helix, is relevant for defining its role in !usion. In fact, if it is at the top of the a-helix (about 80-100 A from the viral lipid), this location may favor an interaction with the membrane of the target cell as previously hypothesized (9). On the contrary, if it is at the bottom of the a-helix, it is already likely to be in contact with viral lipids. Both these


configurations of fusion peptide do not require large conformational changes of F protein to enter in contact with the lipids. On the contrary, in the case of infJuenza hemagglutinin where the fusion peptide is located 30 A from the viral lipids and 100 A from the distal tip, a great conformational rear- rangement appears essential to relocate the fusion peptide either at the tip or at the viral lipid level. This hypothesis fits with the fact that HA glycoprotein promotes fusion at acidic pH values (24), where a structural change has been observed (25); whereas the F protein is effective at pH 5.5-9.0. The current hypothesis is that both these fusion peptides promote fusion by insertion in the host membrane (1, 9). It has been thought that a contemporary insertion in two adjacent lipid bilayers allows a rapid lipid exchange and then fusion. How- ever, an alternative mechanism cannot be excluded: the fusion peptide induces fusion interacting with its own viral lipids. This interaction might create a specific arrangement of viral membrane, locally destabilizing the bilayer and then facilitat- ing the fusion. For instance, a restricted area which is not arranged in bilayer structure has been proposed as a fusion promoter (29). The conformational change occurring in influenza hemagglutinin upon acidification serves the purpose of inserting the fusion peptide in the viral membrane, thus altering the viral envelope which become prearranged to fuse with the host membrane. In contrast, F protein does not need a pH-related conformational change to be active because the F fusion peptide is already interacting with viral lipids. The fact that Sendai virus is able to fuse with liposomes lacking specific receptors (27) seems in agreement with this idea.

The localization of the fusion peptide of F, with respect to the viral lipid phase assumes an important role in determining if the trigger of fusion involves the perturbation of viral lipid bilayer or of the host membrane. We are presently investigat- ing by other methodologies the topology of the fusion peptide of Sendai virus.

Acknowledgments-We are very grateful to Prof. D. M. Gill (Tufts University, Boston) for helpful criticism and particularly for revising the manuscript. We are greatly indebted to Prof. D. Bossa (University “La Sapienza,” Rome) for the analysis of the N terminus of Sendai virus F glycoprotein. We thank Prof. G. D’Agnolo for encouragement and Dr. J. Brunner for useful discussion. We also thank T. Squatriti for his skillful technical assistance and for the preparation of Sendai virus.

1.

2. 3. 4. 5.

6. 7.

8.

9.

10.

11.

12. 13. 14. 15.

16.

17.

18.

19. 20.

21.

22.

23.

24. 25.

26.

27.

28.

29.

REFERENCES White, J. M., Kielian, M., and Helenius, A. (1983) Q. Reu.

Rodriguez-Boulan, E. (1983) Mod. Cell Bwl. 1 , 119-170 Homma, M., and Ohuchi, M. (1973) J. Virol. 12, 1457-1465 Scheid, A., and Choppin, P. W. (1974) Virology 57,475-490 Ozawa, M., Asano, A., and Okada, Y. (1976) FEBS Lett. 70,145-

Scheid, A., and Choppin, P. W. (1977) Virology 80, 54-66 Gething, M.-J., White, J. M., and Waterfield, M. D. (1978) Proc.

Natl. Acad. Sci. U. S. A. 75 , 2737-2740 Scheid, A., Graves, M. C., Silver, S. M., and Choppin, P. W.

(1978) in Negative Strand Viruses and the Host Cell (Mahy, B. W. J., and Barry, R. D., eds) pp. 183-191, Academic Press, New York

Hsu, M-C., Scheid, A., and Choppin, P. W. (1981) J. Biol. Chem.

Wilson, I. A., Skehel, J. J., and Wiley, D. C. (1981) Nature 289 ,

Bisson, R., and Montecucco, C. (1985) in Progress in Protein- Lipid Interactions (Watts, A., and De Pont J. J., eds) pp. 259- 287, Elsevier Scientific Publishing Co., Amsterdam

Biophy~. 16 , 151-195

149

256,3557-3563

366-373

Montecucco, C. (1987) Methods Enzymol., in press Tomasi, M., and Loyter, A. (1981) FEBS Lett. 131,381-385 Bisson, R., and Montecucco, C. (1981) Biochem J. 193 , 757-763 Tomasi. M.. and Montecucco. C. (1981) J. Biol. Chem. 256. , . ,

1117i-lli81 Asano. K.. Murachi. T.. and Asano. A. (1983) J. Biochem. (Tokvo)

93,733-741 I , I~ I . I -

Vainstein. A.. Hershkovitz. M.. Israel. S.. Rabin. S.. and Lovter. A. (1984) Biochim. Biophys. Acta 773,’181-188 ’

J. Gem Virol. 6 6 , 317-331

“ ,

Blumberg, B. M., Giorgi, C., Rose, K., and Kolakofsky, D. (1985)

Brunner, J., and Semenza, G. (1981) Biochemistry 20,7174-7182 Ozawa, M., Asano, A., and Okada, Y. (1979) J. Biochem. (Tokyo)

Garnier, J., Osguthorpe, D. J., and Robson, B. (1978) J. Mol.

Shioda, T., Iwasaki, K., and Shibuta, H. (1986) Nucleic Acids

Meister, H., Bachofen, R., Semenza, G., and Brunner, J. (1985)

Maeda, T., and Onishi, S. (1980) FEBS Lett. 122, 283-287 Doms. R. W.. Helenius. A.. and White. J. 11985) J. Biol. Chem.

86,1361-1369

Biol. 120,97-120

Res. 14,1545-1563

J. Biol. Chem. 260, 16326-16331

260,-2973:2981 , . I

HODD. T. P.. and Woods. K. R. 11981) Proc. Natl. A c d . Sci. U. i.1. 78,3824-3828 ’

279-294

. ,

Citovsky, V., Zakai, N., and Loyter, A. (1986) Erp. Cell Res. 166,

Blumberg, B., Giorgi, C., Roux, L., Raju, R., Dowling, P., Chollet,

Cullis, P. R., Hope, M. J., and Tilcock, C. P. S. (1986) Chem.

”.

A., and Kolakofsky, D. (1985) Cell 41,269-278

Phys. Lipids 40 , 127-144

Multiple lipid interactions of the Sendai virus fusogenic protein

Documents

Transcript of Multiple lipid interactions of the Sendai virus fusogenic protein