A Single Point Mutation Controls the Cholesterol Dependence of Semliki Forest Virus Entry and Exit

Membrane fusion and budding are key steps in the life cycle of all enveloped viruses. Semliki Forest virus (SFV) is an enveloped alphavirus that requires cellular membrane cholesterol for both membrane fusion and efficient exit of progeny virus from infected cells. We selected an SFV mutant, srf-3, that was strikingly independent of cholesterol for growth. This phenotype was conferred by a single amino acid change in the E1 spike protein subunit, proline 226 to serine, that increased the cholesterol independence of both srf-3 fusion and exit. The srf-3 mutant emphasizes the relationship between the role of cholesterol in membrane fusion and virus exit, and most significantly, identifies a novel spike protein region involved in the virus cholesterol requirement.     

Cholesterol is required in the exit pathway of Semliki Forest virus

The enveloped alphavirus Semliki Forest virus (SFV) infects cells via a membrane fusion reaction triggered by low pH. For fusion to occur cholesterol is required in the target membrane, as demonstrated both in in vitro fusion assays and in vivo for virus infection of a host cell. In this paper we examine the role of cholesterol in postfusion events in the SFV life cycle. Cholesterol-depleted insect cells were transfected with SFV RNA or infected at very high multiplicities to circumvent the fusion block caused by the absence of cholesterol. Under these conditions, the viral spike proteins were synthesized and transported to the site of p62 cleavage with normal kinetics. Surprisingly, the subsequent exit of virus particles was dramatically slowed compared to cholesterol-containing cells. The inhibition of virus production could be reversed by the addition of cholesterol to depleted cells. In contrast to results with SFV, no cholesterol requirement for virus exit was observed for the production of either the unrelated vesicular stomatitis virus or a cholesterol-independent SFV fusion mutant. Thus, cholesterol was only critical in the exit pathway of viruses that also require cholesterol for fusion. These results demonstrate a specific and unexpected lipid requirement in virus exit, and suggest that in addition to its role in fusion, cholesterol is involved in the assembly or budding of SFV.     

In vivo generation and characterization of a soluble form of the Semliki forest virus fusion protein

During infection of host cells, a number of enveloped animal viruses are known to produce soluble forms of viral membrane glycoproteins lacking the transmembrane domain. The roles of such soluble glycoproteins in viral life cycles are incompletely understood, but in several cases they are believed to modulate host immune response and viral pathogenesis. Semliki Forest virus (SFV) is an enveloped alphavirus that infects cells through low-pH-dependent fusion and buds from the plasma membrane. Fusion is mediated by the E1 subunit of the SFV spike protein. Previous studies described the in vivo generation of E1s, a truncated soluble form of E1, under conditions in which budding is inhibited in mammalian host cells. We have here examined the properties of E1s generation and the biological activity of E1s. E1s cleavage required spike protein transport out of the endoplasmic reticulum and was independent of virus infection. Cell surface E1 efficiently acted as a precursor for E1s. E1s generation was strongly pH dependent in BHK cells, with optimal cleavage at a pH of < or =7.0, conditions that inhibited the budding of SFV but not the budding of the rhabdovirus vesicular stomatitis virus. The pH dependence of E1s production and SFV budding was unaffected by the stability of the spike protein dimer but was a function of the host cell. Similar to the intact virus and in vitro-generated E1 ectodomain, treatment of E1s at low pH in the presence of target membranes triggered specific acid-dependent conformational changes. Thus, under a variety of conditions, SFV-infected cells can produce a soluble form of E1 that is biologically active.     

Spike protein-nucleocapsid interactions drive the budding of alphaviruses.

Semliki Forest virus (SFV) particles are released from infected cells by budding of nucleocapsids through plasma membrane regions that are modified by virus spike proteins. The budding process was studied with recombinant SFV genomes which lacked the nucleocapsid protein gene or, alternatively, the spike genes. No subviral particles were released from cells which expressed only the nucleocapsid protein or the spike proteins. Virus release was found to be strictly dependent on the coexpression of the nucleocapsid and the spike proteins. These results provide direct proof for the hypothesis that the alphavirus budding is driven by nucleocapsid-spike interactions. The importance of the viral 42S RNA for virus assembly and budding was investigated by using the heterologous vaccinia virus-T7 expression system for the synthesis of the SFV structural proteins. The results demonstrate that the viral genome is not absolutely required for formation of budding competent nucleocapsids, since small amounts of viruslike particles were assembled in the absence of 42S RNA.     

Mutations in the putative fusion peptide of Semliki Forest virus affect spike protein oligomerization and virus assembly.

The two transmembrane spike protein subunits of Semliki Forest virus (SFV) form a heterodimeric complex in the rough endoplasmic reticulum. This complex is then transported to the plasma membrane, where spike-nucleocapsid binding and virus budding take place. By using an infectious SFV clone, we have characterized the effects of mutations within the putative fusion peptide of the E1 spike subunit on spike protein dimerization and virus assembly. These mutations were previously demonstrated to block spike protein membrane fusion activity (G91D) or cause an acid shift in the pH threshold of fusion (G91A). During infection of BHK cells at 37 degrees C, virus spike proteins containing either mutation were efficiently produced and transported to the plasma membrane, where they associated with the nucleocapsid. However, the assembly of mutant spike proteins into mature virions was severely impaired and a cleaved soluble fragment of E1 was released into the medium. In contrast, incubation of mutant-infected cells at reduced temperature (28 degrees C) dramatically decreased E1 cleavage and permitted assembly of morphologically normal virus particles. Pulse-labeling studies showed that the critical period for 28 degrees C incubation was during virus assembly, not spike protein synthesis. Thus, mutations in the putative fusion peptide of SFV confer a strong and thermoreversible budding defect. The dimerization of the E1 spike protein subunit with E2 was analyzed by using either cells infected with virus mutants or mutant virus particles assembled at 28 degrees C. The altered-assembly phenotype of the G91D and G91A mutants correlated with decreased stability of the E1-E2 dimer.     

The cholesterol requirement for sindbis virus entry and exit and characterization of a spike protein region involved in cholesterol dependence

Semliki Forest virus (SFV) and Sindbis virus (SIN) are enveloped alphaviruses that enter cells via low-pH-triggered fusion in the endocytic pathway and exit by budding from the plasma membrane. Previous studies with cholesterol-depleted insect cells have shown that SFV requires cholesterol in the cell membrane for both virus fusion and efficient exit of progeny virus. An SFV mutant, srf-3, shows efficient fusion and exit in the absence of cholesterol due to a single point mutation in the E1 spike subunit, proline 226 to serine. We have here characterized the role of cholesterol in the entry and exit of SIN, an alphavirus quite distantly related to SFV. Growth, primary infection, fusion, and exit of SIN were all dramatically inhibited in cholesterol-depleted cells compared to control cells. Based on sequence differences within the E1 226 region between SFV, srf-3, and SIN, we constructed six SIN mutants with alterations within this region and characterized their cholesterol dependence. A SIN mutant, SGM, that had the srf-3 amino acid sequence from E1 position 224 to 235 showed increases of approximately 100-fold in infection and approximately 250-fold in fusion with cholesterol-depleted cells compared with infection and fusion of wild-type SIN. Pulse-chase analysis demonstrated that SGM exit from cholesterol-depleted cells was markedly more efficient than that of wild-type SIN. Thus, similar to SFV, SIN was cholesterol dependent for both virus entry and exit, and the cholesterol dependence of both steps could be modulated by sequences within the E1 226 region.     

The Fusion glycoprotein shell of Semliki Forest virus: an icosahedral assembly primed for fusogenic activation at endosomal pH

Semliki Forest virus (SFV) has been extensively studied as a model for analyzing entry of enveloped viruses into target cells. Here we describe the trace of the polypeptide chain of the SFV fusion glycoprotein, E1, derived from an electron density map at 3.5 A resolution and describe its interactions at the surface of the virus. E1 is unexpectedly similar to the flavivirus envelope protein, with three structural domains disposed in the same primary sequence arrangement. These results introduce a new class of membrane fusion proteins which display lateral interactions to induce the necessary curvature and direct budding of closed particles. The resulting surface protein lattice is primed to cause membrane fusion when exposed to the acidic environment of the endosome.     

Virus Maturation by Budding

Enveloped viruses mature by budding at cellular membranes. It has been generally thought that this process is driven by interactions between the viral transmembrane proteins and the internal virion components (core, capsid, or nucleocapsid). This model was particularly applicable to alphaviruses, which require both spike proteins and a nucleocapsid for budding. However, genetic studies have clearly shown that the retrovirus core protein, i.e., the Gag protein, is able to form enveloped particles by itself. Also, budding of negative-strand RNA viruses (rhabdoviruses, orthomyxoviruses, and paramyxoviruses) seems to be accomplished mainly by internal components, most probably the matrix protein, since the spike proteins are not absolutely required for budding of these viruses either. In contrast, budding of coronavirus particles can occur in the absence of the nucleocapsid and appears to require two membrane proteins only. Biochemical and structural data suggest that the proteins, which play a key role in budding, drive this process by forming a three-dimensional (cage-like) protein lattice at the surface of or within the membrane. Similarly, recent electron microscopic studies revealed that the alphavirus spike proteins are also engaged in extensive lateral interactions, forming a dense protein shell at the outer surface of the viral envelope. On the basis of these data, we propose that the budding of enveloped viruses in general is governed by lateral interactions between peripheral or integral membrane proteins. This new concept also provides answers to the question of how viral and cellular membrane proteins are sorted during budding. In addition, it has implications for the mechanism by which the virion is uncoated during virus entry.     

Role of cell surface spikes in alphavirus budding.

Alphaviruses mature by budding at cell surfaces. According to a prevailing hypothesis, the viral membrane protein, which is a heterodimeric protein unit, is transported to the plasma membrane (PM), where it awaits binding to the viral nucleocapsid (NC). This hypothesis predicts that the viral membrane protein heterodimers accumulate at the cell surface when expressed in the absence of NCs. We have tested this prediction by analyzing the spike protein expression phenotype of a Semliki Forest virus (SFV) variant which contains a capsid gene deletion. We found that viral membrane protein heterodimers were formed and transported to the cell surface normally. However, instead of accumulating at the PM as expected, the membrane proteins were rapidly degraded. In the case of the E1 subunit, degradation resulted in the release of a soluble E1 fragment into the medium. The fact that this pathway of protein degradation is mostly inhibited during wild-type virus infection suggests that viral membrane proteins are very efficiently captured by NCs into budding complexes and that normally no sizeable pool of free membrane protein complexes exists at the PM.     

Biochemical consequences of a mutation that controls the cholesterol dependence of Semliki Forest virus fusion

The enveloped alphavirus Semliki Forest virus (SFV) infects cells via a low-pH-triggered membrane fusion reaction that requires cholesterol and sphingolipid in the target membrane. Cholesterol-depleted insect cells are highly resistant to alphavirus infection and were used to select srf-3, an SFV mutant that is approximately 100-fold less cholesterol dependent for infection due to a single amino acid change in the E1 spike subunit, proline 226 to serine. Sensitive lipid-mixing assays here demonstrated that the in vitro fusion of srf-3 and wild-type (wt) virus with cholesterol-containing liposomes had comparable kinetics, activation energies, and sphingolipid dependence. In contrast, srf-3 fusion with sterol-free liposomes was significantly more efficient than that of wt virus. Thus, the srf-3 mutation does not affect its general fusion properties with purified lipid bilayers but causes a marked and specific reduction in cholesterol dependence. Upon exposure to low pH, the E1 spike subunit undergoes distinct conformational changes, resulting in the exposure of an acid conformation-specific epitope and formation of an E1 homotrimer. These conformational changes were strongly cholesterol and sphingolipid dependent for wt SFV and strikingly less cholesterol dependent for srf-3. Our results thus demonstrate the functional importance of fusogenic E1 conformational changes in the control of SFV cholesterol dependence.     

Budding of enveloped viruses from the plasma membrane

Many enveloped viruses are released from infected cells by maturing and budding at the plasma membrane. During this process, viral core components are incorporated into membrane vesicles that contain viral transmembrane proteins, termed ‘spike’ proteins. For many years these spike proteins, which are required for infectivity, were believed to be incorporated into virions via a direct interaction between their cytoplasmic domains and viral core components. More recent evidence shows that, while such direct interactions drive budding of alphaviruses, this may not be the case for negative strand RNA viruses and retroviruses. These viruses can bud particles in the absence of spike proteins, using only viral core components to drive the process. In some cases the spike proteins, without the viral core, can be released as virus-like particles. Optimal budding and release may, therefore, depend on a ‘push-and-pull’ concerted action of core and spike, where oligomerization of both components plays a crucial role.     

Redistribution of cellular and herpes simplex virus proteins from the trans-golgi network to cell junctions without enveloped capsids

Herpes simplex virus (HSV) and other alphaherpesviruses assemble enveloped virions in the trans-Golgi network (TGN) or endosomes. Enveloped particles are formed when capsids bud into TGN/endosomes and virus particles are subsequently ferried to the plasma membrane in TGN-derived vesicles. Little is known about the last stages of virus egress from the TGN/endosomes to cell surfaces except that the HSV directs transport of nascent virions to specific cell surface domains, i.e., epithelial cell junctions. Previously, we showed that HSV glycoprotein gE/gI accumulates extensively in the TGN at early times after infection and also when expressed without other viral proteins. At late times of infection, gE/gI and a cellular membrane protein, TGN46, were redistributed from the TGN to epithelial cell junctions. We show here that gE/gI and a second glycoprotein, gB, TGN46, and another cellular protein, carboxypeptidase D, all moved to cell junctions after infection with an HSV mutant unable to produce cytoplasmic capsids. This redistribution did not involve L particles. In contrast to TGN membrane proteins, several cellular proteins that normally adhere to the cytoplasmic face of TGN, Golgi, and endosomal membranes remained primarily dispersed throughout the cytoplasm. Therefore, cellular and viral membrane TGN proteins move to cell junctions at late times of HSV infection when the production of enveloped particles is blocked. This is consistent with the hypothesis that there are late HSV proteins that reorganize or redistribute TGN/endosomal compartments to promote virus egress and cell-to-cell spread.     

Cholesterol is required for infection by Semliki Forest virus

Semliki Forest virus (SFV) and many other enveloped animal viruses enter cells by a membrane fusion reaction triggered by the low pH within the endocytic pathway. In vitro, SFV fusion requires cholesterol in the target membrane, but the role of cholesterol in vivo is unknown. In this paper, the infection pathway of SFV was studied in mammalian and inset cells substantially depleted of sterol. Cholesterol-depleted cells were unaltered in their ability to bind, internalize, and acidify virus, but were blocked in SFV fusion and subsequent virus replication. Depleted cells could be infected by the cholesterol-independent vesicular stomatitis virus, which also enters cells via endocytosis and low pH-mediated fusion. The block in SFV infection was specifically reversed by cholesterol but not by cholestenone, which lacks the critical 3 beta-hydroxyl group. Cholesterol thus is central in the infection pathway of SFV, and may act in vivo to modulate infection by SFV and other pathogens.     

In vitro mutagenesis of a full-length cDNA clone of Semliki Forest virus: the small 6,000-molecular-weight membrane protein modulates virus release.

We report on the construction of a full-length cDNA clone of Semliki Forest virus (SFV). By placing the cDNA under the SP6 promoter, infectious RNA can be produced in vitro and used to transfect cells to initiate virus infection. To achieve efficient transfections, a new protocol for electroporation of RNA was developed. This method gave up to 500-fold improvement over the traditional DEAE-dextran transfection procedure. Since virtually 100% of the cells can be transfected by electroporation, this method is a useful tool for detailed biochemical studies of null mutations of SFV that abolish production of infections virus particles. We used the cDNA clone of SFV to study what effects a deletion of the 6,000-molecular-weight membrane protein (6K membrane protein) had on virus replication. The small 6K protein is part of the structural precursor molecule (C-p62-6K-E1) of the virus. Our results conclusively show that the 6K protein is not needed for the heterodimerization of the p62 and E1 spike membrane proteins in the endoplasmic reticulum, nor is it needed for their transport out to the cell surface. The absence of the 6K protein did, however, result in a dramatic reduction in virus release, suggesting that the protein exerts its function late in the assembly pathway, possibly during virus budding.     

Silencing the morphogenesis of rotavirus

The morphogenesis of rotaviruses follows a unique pathway in which immature double-layered particles (DLPs) assembled in the cytoplasm bud across the membrane of the endoplasmic reticulum (ER), acquiring during this process a transient lipid membrane which is modified with the ER resident viral glycoproteins NSP4 and VP7; these enveloped particles also contain VP4. As the particles move towards the interior of the ER cisternae, the transient lipid membrane and the nonstructural protein NSP4 are lost, while the virus surface proteins VP4 and VP7 rearrange to form the outermost virus protein layer, yielding mature infectious triple-layered particles (TLPs). In this work, we have characterized the role of NSP4 and VP7 in rotavirus morphogenesis by silencing the expression of both glycoproteins through RNA interference. Silencing the expression of either NSP4 or VP7 reduced the yield of viral progeny by 75 to 80%, although the underlying mechanism of this reduction was different in each case. Blocking the synthesis of NSP4 affected the intracellular accumulation and the cellular distribution of several viral proteins, and little or no virus particles (neither DLPs nor TLPs) were assembled. VP7 silencing, in contrast, did not affect the expression or distribution of other viral proteins, but in its absence, enveloped particles accumulated within the lumen of the ER, and no mature infectious virus was produced. Altogether, these results indicate that during a viral infection, NSP4 serves as a receptor for DLPs on the ER membrane and drives the budding of these particles into the ER lumen, while VP7 is required for removing the lipid envelope during the final step of virus morphogenesis.     

Role of spike protein conformational changes in fusion of Semliki Forest virus.

The alphavirus Semliki Forest virus (SFV) and a number of other enveloped animal viruses infect cells via a membrane fusion reaction triggered by the low pH within endocytic vesicles. In addition to having a low pH requirement, SFV fusion and infection are also strictly dependent on the presence of cholesterol in the host cell membrane. A number of conformational changes in the SFV spike protein occur following low-pH treatment, including dissociation of the E1-E2 dimer, conformational changes in the E1 and E2 subunits, and oligomerization of E1 to a homotrimer. To allow the ordering of these events, we have compared the kinetics of these conformational changes with those of fusion, using pH treatment near the fusion threshold and low-temperature incubation to slow the fusion reaction. Dimer dissociation, the E1 conformational change, and E1 trimerization all occur prior to the mixing of virus and cell membranes. Studies of cells incubated at 20 degrees C showed that as with virus fusion, E1 trimerization occurred in the endosome before transport to lysosomes. However, unlike the strictly cholesterol-dependent membrane fusion reaction, the E1 homotrimer was produced in vivo during virus uptake by cholesterol-depleted cells or in vitro by low-pH treatment of virus in the presence of artificial liposomes with or without cholesterol. Purified, lipid-free spike protein rosettes were assayed to determine the requirement for virus membrane cholesterol in E1 homotrimer formation. Spike protein rosettes were found to undergo E1 oligomerization upon exposure to low pH and target liposomes and showed an enhancement of oligomerization with cholesterol-containing membranes. The E1 homotrimer may represent a perfusion complex that requires cholesterol to carry out the final coalescence of the viral and target membranes.     

An epitope of the Semliki Forest virus fusion protein exposed during virus-membrane fusion

Semliki Forest virus (SFV) is an enveloped alphavirus that infects cells via a membrane fusion reaction triggered by acidic pH in the endocytic pathway. Fusion is mediated by the spike protein E1 subunit, an integral membrane protein that contains the viral fusion peptide and forms a stable homotrimer during fusion. We have characterized four monoclonal antibodies (MAbs) specific for the acid conformation of E1. These MAbs did not inhibit fusion, suggesting that they bind to an E1 region different from the fusion peptide. Competition analyses demonstrated that all four MAbs bound to spatially related sites on acid-treated virions or isolated spike proteins. To map the binding site, we selected for virus mutants resistant to one of the MAbs, E1a-1. One virus isolate, SFV 4-2, showed reduced binding of three acid-specific MAbs including E1a-1, while its binding of one acid-specific MAb as well as non-acid-specific MAbs to E1 and E2 was unchanged. The SFV 4-2 mutant was fully infectious, formed the E1 homotrimer, and had the wild-type pH dependence of infection. Sequence analysis demonstrated that the relevant mutation in SFV 4-2 was a change of E1 glycine 157 to arginine (G157R). Decreased binding of MAb E1a-1 was observed under a wide range of assay conditions, strongly suggesting that the E1 G157R mutation directly affects the MAb binding site. These data thus localize an E1 region that is normally hidden in the neutral pH structure and becomes exposed as part of the reorganization of the spike protein to its fusion-active conformation.     

A statistical-thermodynamic model of viral budding

We present a simple statistical thermodynamic model for budding of viral nucleocapsids at the cell membrane. The membrane is modeled as a flexible lipid bilayer embedding linker (spike) proteins, which serve to anchor and thus wrap the membrane around the viral capsids. The free energy of a single bud is expressed as a sum of the bending energy of its membrane coat, the spike-mediated capsid-membrane adhesion energy, and the line energy associated with the bud's rim, all depending on the extent of wrapping (i.e., bud size), and density of spikes in the curved membrane. This self-energy is incorporated into a simple free energy functional for the many-bud system, allowing for different spike densities, and hence entropy, in the curved (budding) and planar membrane regions, as well as for the configurational entropy of the polydisperse bud population. The equilibrium spike densities in the coexisting, curved and planar, membrane regions are calculated as a function of the membrane bending energy and the spike-mediated adhesion energy, for different spike and nucleocapsid concentrations in the membrane plane, as well as for several values of the bud's rim energy. We show that complete budding (full wrapping of nucleocapsids) can only take place if the adhesion energy exceeds a certain, critical, bending free energy. Whenever budding takes place, the spike density in the mature virions is saturated, i.e., all spike adhesion sites are occupied. The rim energy plays an important role in determining the size distribution of buds. The fraction of fully wrapped buds increases as this energy increases, resulting eventually in an all-or-nothing mechanism, whereby nucleocapsids at the plasma membrane are either fully enveloped or completely naked (just touching the membrane). We also find that at low concentrations all capsids arriving at the membrane get tightly and fully enveloped. Beyond a certain concentration, corresponding approximately to a stoichiometric spike/capsid ratio, newly arriving capsids cannot be fully wrapped; i.e., the budding yield decreases.     

Ubiquitin depletion and dominant-negative VPS4 inhibit rhabdovirus budding without affecting alphavirus budding

The budding reactions of a number of enveloped viruses use the cellular machinery involved in the formation of the luminal vesicles of endosomal multivesicular bodies (MVB). Budding of these viruses is dependent on the presence of specific late-domain motifs in membrane-associated viral proteins. Such budding reactions usually involve ubiquitin and are blocked by expression of an ATPase-deficient form of VPS4, a cellular AAA+ ATPase believed to be required late in the MVB pathway for the disassembly/release of the MVB machinery. Here we examined the role of the MVB pathway in the budding of the late-domain-containing rhabdovirus vesicular stomatitis virus (VSV) and the alphavirus Semliki Forest virus (SFV). We tested early and late steps in the MVB pathway by depleting ubiquitin with the proteasome inhibitor MG-132 and by using cell lines inducibly expressing VPS4A or VPS4B protein. As previously shown, VSV budding was strongly dependent on ubiquitin. In contrast to the findings of previous studies with VPS4A, expression of ATPase-deficient mutants of either VPS4A or VPS4B inhibited VSV budding. Inhibition by VPS4 required the presence of the PPPY late domain on the VSV matrix protein and resulted in the accumulation of nonreleased VSV particles at the plasma membrane. In contrast, SFV budding was independent of both ubiquitin and the activity of VPS4, perhaps reflecting the important role of the highly organized envelope protein lattice during alphavirus budding.     

Differential cholesterol binding by class II fusion proteins determines membrane fusion properties

The class II fusion proteins of the alphaviruses and flaviviruses mediate virus infection by driving the fusion of the virus membrane with that of the cell. These fusion proteins are triggered by low pH, and their structures are strikingly similar in both the prefusion dimer and the postfusion homotrimer conformations. Here we have compared cholesterol interactions during membrane fusion by these two groups of viruses. Using cholesterol-depleted insect cells, we showed that fusion and infection by the alphaviruses Semliki Forest virus (SFV) and Sindbis virus were strongly promoted by cholesterol, with similar sterol dependence in laboratory and field isolates and in viruses passaged in tissue culture. The E1 fusion protein from SFV bound cholesterol, as detected by labeling with photocholesterol and by cholesterol extraction studies. In contrast, fusion and infection by numerous strains of the flavivirus dengue virus (DV) and by yellow fever virus 17D were cholesterol independent, and the DV fusion protein did not show significant cholesterol binding. SFV E1 is the first virus fusion protein demonstrated to directly bind cholesterol. Taken together, our results reveal important functional differences conferred by the cholesterol-binding properties of class II fusion proteins.