The conserved glycine residues in the transmembrane domain of the Semliki Forest virus fusion protein are not required for assembly and fusion

The alphavirus Semliki Forest virus (SFV) infects cells via a low pH-triggered fusion reaction mediated by the viral E1 protein. Both the E1 fusion peptide and transmembrane (TM) domain are essential for membrane fusion, but the functional requirements for the TM domain are poorly understood. Here we explored the role of the five TM domain glycine residues, including the highly conserved glycine pair at E1 residues 415/416. SFV mutants with alanine substitutions for individual or all five glycine residues (5G/A) showed growth kinetics and fusion pH dependence similar to those of wild-type SFV. Mutants with increasing substitution of glycine residues showed an increasingly more stringent requirement for cholesterol during fusion. The 5G/A mutant showed decreased fusion kinetics and extent in fluorescent lipid mixing assays. TM domain glycine residues thus are not required for efficient SFV fusion or assembly but can cause subtle effects on the properties of membrane fusion.     

Fate of the 6K membrane protein of Semliki Forest virus during virus assembly

The Semliki Forest virus directs the synthesis of three virus-specific transmembrane proteins p62, 6K, and E1, which all are made in equimolar amounts from a polyprotein precursor molecule. The p62 and E1 spike proteins form heterodimeric complexes in the endoplasmic reticulum before being transported to the cell surface where virus budding occurs. In this study we show that the 6K protein becomes associated to the p62E1 complex in the endoplasmic reticulum and transported with the complex to the cell surface. During virus budding, E1 and p62 (which has matured into the E2 protein) are incorporated into new virions whereas the 6K is mostly excluded. Virus particles released from infected BHK cells contain only about 3% of 6K in their membrane as compared to the spike protein content. The relevance of these findings for the mechanism of SFV assembly is discussed.     

Differential roles of two conserved glycine residues in the fusion peptide of Semliki Forest virus

Semliki Forest Virus (SFV) is an enveloped alphavirus that infects cells by a low-pH-dependent membrane fusion reaction. SFV fusion is catalyzed by the spike protein E1 subunit, which contains a putative fusion peptide between residues 79 and 97. Prior mutagenesis studies demonstrated that an E1 G91D mutation blocks both virus-membrane fusion and the formation of a highly stable E1 trimer believed to be a critical fusion intermediate. We have here demonstrated that the G91D mutant was also inactive in hemifusion, suggesting that the E1 homotrimer is important in the initial stages of lipid mixing. Revertant analysis of a G91 deletion mutant indicated that G91 was crucial for the viability of SFV. In contrast, a G83D mutation produced infectious virus with both efficient fusion and homotrimer formation. Thus, the G83 position, although highly conserved among alphaviruses, was functional if replaced with a charged amino acid.     

Fusion of Semliki forest virus with red cell membranes

The interaction of Semliki Forest virus (SFV) and red cells was studied using biochemical and immunological methods. When [³H]uridine-labeled SFV was adsorbed at 0°, pH 5.8, to ribonuclease-loaded human red cell ghosts, the viral RNA became sensitive to ribonuclease within 2 to 5 min at 20–37°, but not at 0°. Maximal digestion of RNA was achieved within 30 min at 37° suggesting that the nucleocapsids had penetrated the red cell membranes. When the red cells that had been fused with [³⁵S]methionine-labeled SFV were treated with trypsin, about 70% of E2 protein and 50% of E1 protein was digested, demonstrating that the viral envelope proteins remained on the surface of the red cell. This was also shown by radioimmunoassay, by complement-dependent immune hemolysis assay, and by the ability of the red cells to bind Staphylococcus aureus bacteria to the cell surface after treatment with anti-envelope antiserum. When the red cells carrying SFV glycoproteins on their surfaces were transferred to pH 5.8, the cells were readily fused after the temperature had been raised to 42°. This shows that the fusion of virus envelope with red cell membranes precedes the fusion of red cells.     

Semliki Forest virus produced in the absence of the 6K protein has an altered spike structure as revealed by decreased membrane fusion capacity

We examined the kinetics of membrane fusion of wild type (wt) and Delta6K mutant Semliki Forest virus in a liposomal model system. The final extent of membrane fusion of the mutant (at pH 5.5) was approximately one third that of the wt virus, although the level of E1 (fusion protein) trimerization was, in fact, greater than that of the wt. Studies on the effect of exposure of the viruses to low pH revealed that the Delta6K mutant was inactivated much more rapidly than the wt virus. It is this instability of the mutant particles which probably accounts for the lower fusion levels. Moreover, fusion of the Delta6K mutant was significantly increased by the inclusion of lipid-conjugated heparin in the target liposomes. We conclude that the presence of the 6K protein either in the particle or during the assembly process is important for the correct assembly of the fully infectious SFV particle.     

Two mutations in the envelope glycoprotein E2 of Semliki Forest virus affecting the maturation and entry patterns of the virus alter pathogenicity for mice.

The prototype strain of Semliki Forest virus (SFV) of known sequence and virus produced by the cDNA clone derived from it were lethal following intranasal (i.n.) infection of 40-day-old and intraperitoneal (i.p.) infection of pregnant BALB/c mice; this lethality was related to neuronal necrosis in the central nervous system (CNS). We conclude that the virulence of the prototype strain, and virus from the cDNA clone derived from it, is similar to that of L10 (the original SFV isolate). The effects of two mutations in the p62 envelope protein region of the clone were determined. Substitution of Glu for Lys at position 162 (mut64) extended the mean time of death following i.n. inoculation of 40-day-old mice. Pregnant mice infected with this virus survived but lethal infection of some fetuses did occur. Substitution of Leu for Arg at position 66 (mL), the cleavage site of the E2 and E3 proteins, results in the production of particles containing uncleaved p62. These particles were less virulent than the prototype strain when inoculated i.n. and induced immunity to virulent SFV challenge. The virus also induced the formation of multifocal glial nodules in the CNS of surviving mice. The differences in pathogenicity between the two mutants and the virulent parental virus are probably related to differences in the efficiency of virus multiplication in infected mice. The mut64 mutation attenuated the virus and allowed survival of pregnant mice infected i.p. so that the effects of fetal infection could be detected. The mL mutation allowed survival of i.n.-infected mice so that the later effects of virus multiplication in the CNS could be assessed. In the former case, this is probably a result of reduced virus release, whereas in the latter case it is due to inefficient entry of host cells. The results are consistent with our previous suggestion that lethality for virulent SFV infection results from a lethal threshold of damage to neurons in the CNS and that attenuating mutations may reduce neuronal damage below this threshold level.     

Mutagenesis of the La Crosse Virus glycoprotein supports a role for Gc (1066-1087) as the fusion peptide

The La Crosse Virus (LACV) M segment encodes two glycoproteins (Gn and Gc), and plays a critical role in the neuropathogenesis of LACV infection as the primary determinant of neuroinvasion. A recent study from our group demonstrated that the region comprising the membrane proximal two-thirds of Gc, amino acids 860-1442, is critical in mediating LACV fusion and entry. Furthermore, computational analysis identified structural similarities between a portion of this region, amino acids 970-1350, and the E1 fusion protein of two alphaviruses: Sindbis virus and Semliki Forrest virus (SFV). Within the region 970-1350, a 22-amino-acid hydrophobic segment (1066-1087) is predicted to correlate structurally with the fusion peptides of class II fusion proteins. We performed site-directed mutagenesis of key amino acids in this 22-amino acid segment and determined the functional consequences of these mutations on fusion and entry. Several mutations within this hydrophobic domain affected glycoprotein expression to some extent, but all mutations either shifted the pH threshold of fusion below that of the wild-type protein, reduced fusion efficiency, or abrogated cell-to-cell fusion and pseudotype entry altogether. These results, coupled with the aforementioned computational modeling, suggest that the LACV Gc functions as a class II fusion protein and support a role for the region Gc 1066-1087 as a fusion peptide.     

Growth and stability of a cholesterol-independent Semliki Forest virus mutant in mosquitoes.

Semliki Forest virus (SFV) is an enveloped alphavirus that is transmitted in the wild by mosquito vectors. In tissue culture cells, SFV requires cholesterol in the cell membrane both for virus membrane fusion and for the efficient exit of progeny virus from the cell. A previously isolated SFV mutant, srf-3, is strikingly less cholesterol-dependent for virus fusion, exit, and growth due to a single amino acid change in the E1 spike protein subunit, proline 226 to serine. Here we show that when mosquitoes were infected by intrathoracic injection at a range of virus multiplicities, the growth of srf-3 was significantly more rapid than that of wild-type virus, particularly at low multiplicity infection. The differential cholesterol requirements for wild-type and srf-3 infection were maintained during virus passage through mosquitoes. The presence or absence of cholesterol in the srf-3 virus membrane did not affect its infection properties in mosquitoes. Thus the srf-3 mutation causes a growth advantage in the tissues of the mosquito host.     

Semliki Forest virus particles containing only the E1 envelope glycoprotein are infectious and can induce cell-cell fusion

Hydrophobic interaction chromatography (phenyl- and octyl-Sepharose) was performed with Semliki Forest virus to investigate the effect of low pH on its hydrophobicity. At neutral pH, the virus could be bound to the column and completely eluted by the detergent NP-40. Low pH treatment of virus prior to application to the column resulted in stronger binding as reflected by the increased amount of detergent necessary to totally elute the virus. If, however, the low pH treatment was done after binding of the virus to the column, only 15% of the input virus could be eluted by the detergent, indicating a drastic increase in hydrophobicity. Thus binding of the virus to a hydrophobic environment potentiates the effect of low pH on viral hydrophobicity. Trypsin digestion of column-bound virus after low pH treatment resulted in complete digestion of E2 and E3; however, E1 was totally resistant. From this result, we conclude that E1 alone is responsible for the hydrophobic interaction. We have made use of these observations to produce viral particles which were devoid of E2 and E3 by trypsin digestion in the presence of octyl glucoside. These E1 viral particles were infectious and could induce membrane fusion. We conclude that only E1 is necessary and sufficient to mediate membrane fusion. Acid pH induces a drastic increase in the hydrophobicity of E1 which probably facilitates its interaction with the lipid bilayers during the fusion event in endosomes.     

Attenuation of Recombinant Vesicular Stomatitis Viruses Encoding Mutant Glycoproteins Demonstrate a Critical Role for Maintaining a High pH Threshold for Membrane Fusion in Viral Fitness

A plasmid-based recovery system was used to generate four unique vesicular stomatitis virus (VSV) mutants that encode glycoproteins (G proteins) with single or double amino acid substitutions in two conserved acidic residues adjacent to the putative G protein fusion domain. Previously we demonstrated that three of the mutant G proteins (D137-L, E139-L, and DE-SS) have slightly reduced pH thresholds for membrane fusion activity. In this report we show that even though the viruses encoding D137-L, E139-L, and DE-SS were recovered with high efficiency, these mutants were attenuated for growth in cell culture. Plaque formation was significantly delayed with these mutants and the plaques were smaller and more diffuse than those produced by wild-type VSV. In addition, cells infected with these mutants produced approximately 5- to 10-fold less infectious virus than cells infected with a similarly recovered VSV encoding the wild-type G protein. Using R18-labeled virus we found that the mutant G proteins had approximately 50% of the fusion activity of wild-type G at pH 6.3 and only 75% activity at pH 5.8. We also show that the mutant viruses were more sensitive to chloroquine inhibition of infection than either wild-type VSV or the mutant E139-T, which has a fusion phenotype similar to wild-type G protein. Reduced fusion activity and attenuation of infectivity was not due to differences in the amount of G protein incorporated into virions, nor to differences in the amount of virus binding to cells at physiological pH. Although infectivity was assayed at neutral pH, we observed an increase in virus binding with both mutant and wild-type virions as the pH was lowered, and the increase in binding occurred near the pH threshold for membrane fusion activity. From these data we propose a model in which VSV entry involves an increase in virus binding to the inner leaflet of the endosomal membrane during endosome acidification. Concomitant with this higher affinity binding, G protein becomes primed to initiate fusion of the viral envelope with the endosomal membrane. Viruses with mutations that delay the onset of increased binding and fusion lag behind wild-type VSV in their ability to initiate a productive infection, potentially because the location within the cytoplasm where these viruses ultimately fuse is not optimal for either virus uncoating or replication of the viral genome.     

Investigation of the role of glycans for the biological activity of Semliki Forest virus grown in Aedes albopictus cells using inhibitors of asparagine-linked oligosaccharides trimming

Summary The effects of N-linked-oligosaccharide-processing inhibitors on the formation of Semliki Forest virus (SFV) in C 6/36 Aedes albopictus cells were investigated. The glycosidase inhibitors deoxynojirimycin, deoxymannojirimycin and swainsonine prevented the formation of Endo-H resistant structures, but had little effect on virus formation and on the biological activities of the virus. Tunicamycin greatly inhibited virus formation, but had little effect on cell-cell fusion from within and the cleavage of p 62.These results indicate that correct glycosylation is not a prerequisite for biological activities of SFV, whereas glycosylation per se is needed for virus production.     

Membrane fusion activity of vesicular stomatitis virus glycoprotein G is induced by low pH but not by heat or denaturant

The fusogenic envelope glycoprotein G of the rhabdovirus vesicular stomatitis virus (VSV) induces membrane fusion at acidic pH. At acidic pH the G protein undergoes a major structural reorganization leading to the fusogenic conformation. However, unlike other viral fusion proteins, the low-pH-induced conformational change of VSV G is completely reversible. As well, the presence of an alpha-helical coiled-coil motif required for fusion by a number of viral and cellular fusion proteins was not predicted in VSV G protein by using a number of algorithms. Results of pH dependence of the thermal stability of G protein as determined by intrinsic Trp fluorescence and circular dichroism (CD) spectroscopy show that the G protein is equally stable at neutral or acidic pH. Destabilization of G structure at neutral pH with either heat or urea did not induce membrane fusion or conformational change(s) leading to membrane fusion. Taken together, these data suggest that the mechanism of VSV G-induced fusion is distinct from the fusion mechanism of fusion proteins that involve a coiled-coil motif.     

The Murray Valley encephalitis virus prM protein confers acid resistance to virus particles and alters the expression of epitopes within the R2 domain of E glycoprotein.

To study the role of the precursor to the membrane protein (prM) in flavivirus maturation, we inhibited the proteolytic processing of the Murray Valley encephalitis (MVE) virus prM to membrane protein in infected cells by adding the acidotropic agent ammonium chloride late in the virus replication cycle. Viruses purified from supernatants of ammonium chloride-treated cells contained prM protein and were unable to fuse C6/36 mosquito cells from without. When ammonium chloride was removed from the cells, both the processing of prM and the fusion activity of the purified viruses were partially restored. By using monoclonal antibodies (MAbs) specific for the envelope (E) glycoprotein of MVE virus, we found that at least three epitopes were less accessible to their corresponding antibodies in the prM-containing MVE virus particles. Amino-terminal sequencing of proteolytic fragments of the E protein which were reactive with sequence-specific peptide antisera or MAb enabled us to estimate the site of the E protein interacting with the prM to be within amino acids 200 to 327. Since prM-containing viruses were up to 400-fold more resistant to a low pH environment, we conclude that the E-prM interaction might be necessary to protect the E protein from irreversible conformational changes caused by maturation into the acidic vesicles of the exocytic pathway.     

Effects of double-site mutations of vesicular stomatitis virus glycoprotein G on membrane fusion activity

Site-directed mutagenesis of specific amino acids within a conserved amino-terminal region (H2) and a conserved carboxyl-terminal region (H10/A4) of the fusion protein G of vesicular stomatitis virus have previously identified these two segments as an internal fusion peptide and a region influencing low-pH induced conformational change, respectively. Here, we combined a number of the substitution mutants in the H2 and H10/A4 regions to produce a series of double-site mutants and determined the effect of these mutations on membrane fusion activity at acid pH and on pH-dependent conformational change. The results show that most of the double-site mutants have decreased cell-cell fusion activity and that the effects appeared to be additive in terms of inhibition of fusion, except for one mutant, which appeared to be a revertant. The double-site mutants also had pH optima for fusion that were lower than those observed with wild-type G but same as the pH optima for the parent fusion peptide (H2) mutants. The results suggest that although the H2 and H10/A4 sites may affect membrane fusion independently, a possible interaction between these two sites cannot be ruled out.     

The cleavage of p62, the precursor of E2 and E3, is an early and continuous event in Semliki Forest virus-infectedAedes albopictus cells

Summary The cleavage of p62 of Semliki Forest virus (SFV) in C6/36 (Aedes albopictus) cells was investigated by pulse-chase labeling experiments and analysis of the sugar side chain of E₁ using endoglycosidases. Similar to vertebrates, E₁, E₂, and p62 are transported as complexes in C6/36 cells. This observation allows the use of E₁ as a positional marker for the transport and processing of E₂ and p62. The oligosaccharide on the viral spike E₁ protein was modified first to an Endo-D-sensitive (35 min) and then to an Endo-H-resistant structure (55 min), whereas the oligosaccharides of p62 remained sensitive towards Endo-H the whole time.E₂ could be detected already at 10–20 min post synthesis, suggesting that p62 cleavage starts early, probably before the protein has been transported to the Golgi apparatus. This is in contrast to the cleavage taking place later mainly near the plasma membrane of higher eukaryotes.The spike proteins finally appeared in extracellular virions after about 70–90 min post synthesis.     

Crosslink analysis of N-terminal, C-terminal, and N/B determining regions of the Moloney murine leukemia virus capsid protein

To analyze contacts made by Moloney murine leukemia virus (M-MuLV) capsid (CA) proteins in immature and mature virus particles, we have employed a cysteine-specific crosslinking approach that permits the identification of retroviral Gag protein interactions at particular residues. For analysis, single cysteine creation mutations were made in the context of protease-deficient or protease-competent parental constructs. Cysteine creation mutations were chosen near the N- and C-termini of CA and at a site adjacent to the M-MuLV Glu-Ala Fv1 N/B host range determination sequence. Analysis of immature virions showed that PrGag proteins were crosslinked at C-terminal CA residues to form dimers while crosslinking of particle-associated N-terminal and N/B region mutant proteins did not yield dimers, but showed evidence of linking to an unknown 140- to 160-kDa partner. Analysis of mature virions demonstrated that both N- and C-terminal CA residues participated in dimer formation, suggesting that processed CA N- and C-termini are free to establish interprotein associations. Interestingly, N/B region mutant residues in mature virus particles did not crosslink to form dimers, but showed a novel crosslinked band, consistent with an interaction between the N/B tropism determining region and a cellular protein of 45-55 kDa.     

Enzymatic depalmitoylation of viral glycoproteins with acyl-protein thioesterase 1 in vitro.

Many glycoproteins of enveloped viruses as well as cellular proteins are covalently modified with fatty acids. Palmitoylation is often reversible, but the enzymology of this hydrophobic protein modification is not understood. Recently a cytosolic enzyme designated acyl-protein thioesterase 1 (APT1) was purified, which depalmitoylates several cellular proteins. Since hitherto no transmembrane proteins have been tested as substrates for APT1 we have investigated whether palmitoylated viral membrane glycoproteins can be deacylated by use of this enzyme. Recombinant APT1 was purified from Escherichia coli, and depalmitoylation of [3H]palmitate-labeled glycoproteins present in virus particles was measured by SDS-PAGE, fluorography, and scanning densitometry. We find that APT1 causes rapid and almost complete cleavage of fatty acids from the G-protein of vesicular stomatitis virus, hemagglutinin proteins of influenza A and C virus, and E2 of Semliki Forest virus (SFV). In contrast, E1 of SFV is largely resistant against APT1 activity. This substrate specificity of APT1 was also observed using microsomes prepared from SFV-infected cells. Our data emphasize the potential of APT1 as a tool for functional analysis of protein-bound fatty acids.     

Vaccinia virus temperature-sensitive mutants in the A28 gene produce non-infectious virions that bind to cells but are defective in entry

The vaccinia virus temperature-sensitive mutations Cts6 and Cts9 were mapped by marker rescue and DNA sequencing to the A28 gene. Cts6 and Cts9 contain an identical 2-bp deletion truncating the A28 protein and removing the fourth conserved cysteine near the C-terminus. Cts9 mutant virions produced at 40 degrees C were non-infectious and unable to cause cytopathic effect. However, the mutant A28 protein localized to purified mature virions (MV) at 31 degrees C and 40 degrees C. MV of Cts9 produced at 40 degrees C bound to cells but did not enter cells. Low pH treatment of Cts9-infected cells at 18 h p.i. failed to produce fusion from within at 40 degrees C, but gave fusion at 31 degrees C. Adsorption of Cts9 mutant virions to cells followed by low pH treatment showed a defect in fusion from without. The Cts9 phenotype suggests that the A28 protein is involved in both virus entry and cell-cell fusion, and supports the linkage between the two processes.     

Cross-linking of the spike glycoproteins in Semliki Forest virus with dimethylsuberimidate

The bifunctional reagent dimethylsuberimidate was used to cross-link the proteins in: (1) intact Semliki Forest virus (SFV); (2) SFV containing a small amount of the detergent Triton X-100; (3) detergent released SFV membranes; (4) membranes solubilized with Triton X-100 into lipoprotein-detergent complexes; and (5) membranes solubilized into lipid-free glycoprotein-detergent complexes. Analyses of the cross-linked glycoprotein polymers by SDS gel electrophoresis showed that the two 50,000 molecular weight (MW) glycoproteins E1 and E2 of the SFV membrane were preferentially cross-linked into dimers in all preparations. This suggests that oligomeric glycoprotein units, containing two 50,000 MW glycoproteins, are present in the viral membrane and that these remain associated with each other when solubilized with Triton X-100.     

Low pH-dependent Sindbis virus-induced fusion of BHK cells: differences between strains correlate with amino acid changes in the E1 glycoprotein

Expression of alphavirus glycoproteins on the surface of infected cells leads to cell fusion after exposure to acidic pH. Two strains of Sindbis virus, AR339 (SV) and neuroadapted Sindbis virus (NSV), which differ in virulence for weanling mice, were found to differ in pH-dependent fusion. BHK-21 cells infected with SV fused maximally after shifting to pH 5.4, whereas cells infected with NSV required a lower pH, pH 4.8, for maximal fusion. No difference was noted in the optimal pH for agglutination of goose erythrocytes (5.75 for both viruses). To determine the molecular basis for the difference in fusion a series of recombinant viruses was constructed using a cDNA clone of Sindbis virus from which infectious RNA can be transcribed in vitro. Cells infected with a recombinant virus that had the SV E1 and NSV E2 genes had a fusion response curve as a function of pH like SV, while cells infected with recombinant virus with the NSV E1 and SV E2 genes fused like NSV. The E1 glycoproteins of SV and NSV differ at two positions: Val-72 in SV is Ala in NSV (a change near the putative fusion site), and Gly-313 in SV is Asp in NSV. Recombinant viruses which had Val-72 (SV) and Asp-313 (NSV) or Ala-72 (NSV) and Gly-313 (SV) had a lowered pH of fusion like NSV suggesting that both positions participate in determining some aspect of the conformational change in the E1-E2 heterodimer associated with pH-dependent fusion.