Residues 21 to 38 located internally in the TM subunit of the ASLV envelope glycoprotein are believed to encode the fusion peptide. Several lines of evidence suggest definition of this region as the putative FPD of the ASLV Env. First, hydropathy plot analysis of the TM subunit of ASLV implicates this region to be the FPD and this domain is highly conserved among ASLV, as is generally found for FPDs (Fig. ). Importantly, the proposed FPD is juxtaposed to the heptad repeat region, as has been observed for a number of other viral FPDs (9
). In addition, this region shares characteristics with other putative internal FPDs (49
). Finally, recent mutational analysis demonstrated that alteration of Val31 to Glu within the putative FPD of a soluble form of EnvA impairs receptor-induced liposome binding, believed to represent the initial step of glycoprotein-mediated membrane fusion (25
Mutational analysis of envelope glycoproteins has been instrumental in the identification and functional characterization of a number of viral FPDs. To define the role of the putative ASLV FPD in viral infection, an extensive number of conservative point mutations and deletion and insertion mutations were introduced into the ASLV EnvA FPD, and the mutant FPDs assayed for infectivity, receptor binding activity, and incorporation into MLV virions. Of the 22 mutants that were processed and incorporated efficiently into MLV virions, only one point mutation and two insertion mutations significantly affected ASLV envelope function, decreasing titers by 10-fold or more. None of these three mutations decreased receptor binding, supporting the idea that they directly affect membrane fusion mediated by the ASLV envelope glycoprotein.
An interesting finding of this study is that most of the point mutations introduced into the ASLV FPD had little effect on viral entry. This is in sharp contrast to mutational analysis of other viral FPDs. Conservative alterations to the putative FPDs of vesicular stomatitis virus G (18
), Semliki Forest virus E1 (32
), and influenza virus HA (46
) glycoproteins resulted in a decrease in fusion activity. Conversely, individual Gly-to-Ala mutations in the simian virus 5 F protein putative FPD increased the fusogenic activity of this glycoprotein (29
) by increasing the kinetics of fusion (1
). One possibility for the lack of ASLV EnvA mutant phenotype may be that the single-cycle infection assay is not sensitive enough to detect subtle differences in infectivity between FPD mutant Myc-EnvA glycoproteins. However, using a similar assay to analyze the Moloney MLV putative FPD, numerous infection defective mutants, including glycoproteins containing conservative substitutions, were detected (52
), indicating that a single-cycle infection assay is suitable for this type of analysis. In addition, Hernandez and White showed that nonconservative substitutions to the subtype A ASLV putative FPD significantly impaired glycoprotein function, using an infectivity assay nearly identical to the one employed here (26
Our observation that numerous single point mutations within the putative FPD have little effect on ASLV EnvA function may suggest that no individual amino acid in this region is critical for EnvA membrane fusion activity. A centrally located proline in the FPD is a common feature of internal viral FPDs (49
) and may allow a bend, or kink, in the proposed amphipathic helix. Previous mutational analysis showed that alterations to the centrally located proline of the Semliki Forest virus FPD in the E1 subunit resulted in retention of this subunit in the rough endoplasmic reticulum (32
). Furthermore, when Pro127 in the vesicular stomatitis virus G protein putative FPD was change to Asp, Leu, or Gly, cell-cell fusion was dramatically affected (18
). These results may imply a requirement for this residue in internal FPDs. Similarly, our data demonstrate that altering the central proline in the ASLV FPD to Val (P29V) impaired viral infectivity. In contrast, a mutant in which the ASLV FPD Pro29 was changed to Gly (P29G) retained 25% of wild-type envelope function. One hypothesis to explain this observation is that the functional requirement in this region is not specific for proline but rather for an ability to bend or flex near the center of the internal FPD. Thus, the smaller glycine residue is tolerated in place of the proline residue while a bulky valine residue, which should impede bending, is not tolerated. This hypothesis is supported by the observation that three additional mutants in this region, P29G/G30P, P29[A]G30, and G30P, all of which should support bending in the middle of the FPD (Fig. ), have infectious titers similar to that of wild-type EnvA.
Peptide studies employing the FPDs of several viral glycoproteins have yielded significant information on the structure of amino-terminal FPDs in a lipid environment. These studies demonstrated that for human immunodeficiency virus type 1, simian immunodeficiency virus, influenza virus, and Newcastle disease virus, peptides corresponding to the FPD adopt an amphipathic helix conformation (7
) and insert into the lipid bilayer at an oblique orientation (7
). In many cases, these peptides cause liposome leakage (13
) as well as liposome fusion (13
), suggesting that they are biologically active. These results, in addition to photoaffinity labeling studies of the influenza virus HA FPD (23
), have led to generalization of the amphipathic helix model to other viral fusion peptides including ASLV (49
). In order to determine if this model aptly describes the structure of the putative ASLV FPD during membrane fusion, insertion mutants that would affect the size of the proposed hydrophobic patch were produced and analyzed. The results suggest that, for ASLV, this region does not conform to an amphipathic helix. Indeed, for the ASLV mutants analyzed, there is no correlation between the size of the hydrophobic patch and infectivity (Fig. ; Table ). Mutants A24[A]S25 and P29[A]G30, both of which disrupt most of the proposed hydrophobic patch, have a modest effect and no effect on viral titer, respectively. This is in sharp contrast to the severely impaired mutant, A34[A]Q35, which retains most of the hydrophobic patch. The observation that the insertion mutations near the carboxy terminus of the putative ASLV FPD have a more dramatic effect on viral infectivity may suggest that the orientation of the FPD with respect to the downstream heptad repeat domain must be preserved. Alternatively, during the fusion process, the FPD may need to interact with other regions of Env. By inserting an alanine residue at or near the carboxy terminus, the register of the FPD may be altered such that intrasubunit interactions required in steps of the membrane fusion process other than FPD insertion are impeded.
It seems likely that internal FPDs may have functional requirements distinct from those of the more thoroughly studied amino-terminal FPDs. Internal FPDs may not adopt an amphipathic helix conformation due to constraints asserted by the location of this domain within the membrane-anchoring subunit. While the internal FPD of PH-30, the sperm fusion protein, has been modeled as an amphipathic helix (4
), Muga et al. have demonstrated that a peptide corresponding to this domain does not have α-helical properties when inserted into a membrane (38
). The studies presented here suggest that the ASLV FPD may also not be an amphipathic α-helix.
The observation that the mutants F23W, P29V, A24[A]S25, A34[A]Q35, and L37[A]R38 diminish viral entry without affecting receptor binding is consistent with the hypothesis that these mutations perturb membrane fusion. However, additional experiments analyzing receptor-induced structural rearrangements, receptor-triggered liposome binding, and development of a quantitative cell-cell fusion assay will be required to more precisely define the mechanistic effects of these mutations on glycoprotein-mediated membrane fusion.