Functional Domains in the Retroviral Transmembrane Protein

Abstract

The envelope glycoproteins of the mammalian type C retroviruses consist of two subunits, a surface (SU) protein and a transmembrane (TM) protein. SU binds to the viral receptor and is thought to trigger conformational changes in the associated TM protein that ultimately lead to the fusion of viral and host cell membranes. For Moloney murine leukemia virus (MoMuLV), the envelope protein probably exists as a trimer. We have previously demonstrated that the coexpression of envelope proteins that are individually defective in either the SU or TM subunits can lead to functional complementation (Y. Zhao et al., J. Virol. 71:6967–6972, 1997). We have now extended these studies to investigate the abilities of a panel of fusion-defective TM mutants to complement each other. This analysis identified distinct complementation groups within TM, with implications for interactions between different regions of TM in the fusion process. In viral particles, the C-terminal 16 amino acids of the MoMuLV TM (the R peptide) are cleaved by the viral protease, resulting in an increased fusogenicity of the envelope protein. We have examined the consequences of R peptide cleavage for the different TM fusion mutants and have found that this enhancement of fusogenicity can only occur in cis to certain of the TM mutants. These results suggest that R peptide cleavage enhances the fusogenicity of the envelope protein by influencing the interaction of two distinct regions in the TM ectodomain.

The entry of enveloped viruses into cells requires the fusion of viral and cellular membranes in a process catalyzed by specific viral fusion proteins (41). Fusion is initiated by the binding of the fusion protein itself, or an associated protein, to a specific cellular receptor and involves a series of steps that includes the insertion of a stretch of hydrophobic amino acids (the fusion peptide) into the target cell membrane. The paradigm for viral fusion proteins is the influenza virus hemagglutinin (HA) protein (reviewed in reference 42). The HA₁ subunit binds to cell surface sialic acid residues, allowing the virus to be internalized into endosomes by receptor-mediated endocytosis. In the endosome, the low pH triggers conformational changes in HA₁ and the associated HA₂ subunits, leading to the translocation of the fusion peptide at the N terminus of HA₂ toward the target cell membrane (2). An important part of this structural reorganization is the recruitment of part of a heptad repeat sequence in HA₂ into a triple-stranded coiled coil (2, 3).

Several other viral fusion proteins have been shown to possess features in common with influenza virus HA. Hydrophobic fusion peptides have been identified at the N termini of the transmembrane components of the fusion proteins of the paramyxoviruses and several retroviruses (5, 13), while the avian retroviruses, the filoviruses, and the coronaviruses may also contain fusion peptides in their transmembrane proteins (5, 15). Heptad repeat sequences have been found adjacent to all of these fusion peptides (5, 15), and for human immunodeficiency virus type 1 (HIV-1) (6, 39) and murine leukemia virus (MuLV) (11), these heptad repeats form triple-stranded coiled coils when crystallized.

Not all viral fusion proteins are activated by low pH. The paramyxoviruses, coronaviruses, and some retroviruses can fuse at neutral pH and are able to mediate fusion at the cell surface (22, 24, 30, 40). However, conformational changes can be induced in the envelope proteins of HIV-1 (35) and avian sarcoma-leukosis virus (16) by exposure to soluble forms of their receptors, suggesting that structural rearrangements are likely to be a common step in viral fusion, even if triggered by different stimuli. In the MuLVs, a low pH step has been suggested to be a requirement for entry by Moloney MuLV (MoMuLV), because in certain cell lines viral entry is sensitive to lysosomotropic agents (24, 28). However, this sensitivity is cell type dependent, other closely related MuLV strains are not sensitive to such agents, and MuLV fusion proteins cannot be activated by exogenous low-pH treatments (24, 28). Therefore, the reported pH dependence of MoMuLV infection may occur at some step distal to the fusion process.

The MuLV envelope protein is initially translated as a precursor protein, Pr85, assembled into oligomers in the endoplasmic reticulum and proteolytically cleaved by a host protease into two subunits, the surface (SU) protein, gp70, and the transmembrane (TM) protein, p15E (8, 9). Crystallographic studies have now provided evidence that the oligomeric form of MoMuLV TM is a trimer (11). At or shortly after the time of virus budding, the TM is further processed by the viral protease to release a 16-amino-acid peptide, the R peptide, from the C terminus of the cytoplasmic tail. Both p15E and the processed form, p12E, coexist in the virion (17, 19).

R peptide cleavage of MuLV has profound effects on the fusogenicity of the envelope protein, promoting cell-cell fusion and syncytium formation in NIH 3T3 cells which are not fused by the full-length envelope protein (19, 31, 32). It is likely that the virus has adopted a regulatory mechanism to prevent the premature activation of fusion before the envelope protein is incorporated into a virion, which could be cytopathic to the host cell or interfere with the budding process. The fusogenicity of the envelope proteins of Mason-Pfizer monkey virus (1a) and equine infectious anemia virus (33) is also enhanced by the cleavage of their cytoplasmic tails, and artificial truncations of the cytoplasmic tails of HIV-1, HIV-2 and simian immunodeficiency virus (SIV) have also been shown to enhance fusogenicity (12, 27, 47).

We have previously demonstrated that mutant MoMuLV envelope proteins that are defective in either the SU or TM subunits can functionally complement each other when coexpressed, presumably through the formation of hetero-oligomers (45). This observation suggests that the binding signal from one SU monomer can trigger fusion by the associated TM proteins, even when its own TM subunit is defective. We have interpreted these data to indicate that cross talk can occur between monomers of the envelope protein complex. We have now extended those studies to include a panel of fusion-defective mutants that map to distinct predicted features of MoMuLV TM, in order to analyze functional interactions between different regions of the TM protein. In addition, we have compared the ability of R peptide cleavage to enhance fusion in cis to the different TM mutants. In this way, we have been able to identify discrete functional domains within TM, to analyze their interactions, and to suggest how R peptide cleavage may act to modulate the fusogenicity of the envelope protein.

MATERIALS AND METHODS

Envelope protein mutants and cell lines.

Point mutations of MoMuLV TM protein p15E were constructed in the envelope protein expression vector CEE+ (23), using an oligonucleotide-directed in vitro mutagenesis system (version 2.1; Amersham, Arlington Heights, Ill.). NIH 3T3, GP8, GPG4, and 293T cells were grown in Dulbecco modified Eagle medium (Core Facility, University of Southern California) supplemented with 10% fetal calf serum (HyClone, Logan, Utah) and 2 mM glutamine (Gibco-BRL, Grand Island, N.Y.); XC cells were maintained in Earle basal medium (Gibco-BRL) supplemented with 10% fetal calf serum and 2 mM glutamine. GP8 (26) is an NIH 3T3-derived cell line expressing MoMuLV gag-pol; GPG4 cells additionally contain the retroviral vector G1nBgSvNa (18).

Retroviral vector production and characterization.

Retroviral vectors were produced by transient transfection of 293T cells by calcium phosphate precipitation essentially as described previously (18, 36). The plasmids used were an MoMuLV gag-pol expression plasmid pHIT60 (36), the retroviral vector pCnBg, which expresses lacZ and neo (18), and an env expression plasmid. Ten micrograms of each plasmid was used per 10-cm-diameter dish of 293T cells; when two different env expression plasmids were cotransfected, 5 μg of each was used. Thirty-six hours posttransfection, the supernatants were harvested and filtered through 0.45-μm-pore-size filters. The protein content of virions partially purified through 20% sucrose was assessed by Western blot analysis as described previously (19). The ability of virions to bind to the ecotropic receptor expressed on NIH 3T3 cells was determined by a fluorescence-activated cell sorting-based assay as described previously (44).

The titer of each retroviral vector was determined by plating 3 × 10⁴ NIH 3T3 cells in 30-mm-diameter wells of six-well tissue culture plates and, 18 to 24 h later, replacing the medium with 1 ml of appropriately diluted supernatant containing Polybrene (8 μg/ml). Following overnight incubation, the cells were selected for neo expression by growth in G418 (600 μg/ml; Sigma, St. Louis, Mo.) for 9 days. G418-resistant colonies were counted after methylene blue staining.

Cell surface expression of envelope proteins.

The level of cell surface envelope protein was measured by fluorescence-activated cell sorting analysis of 293T cells transiently expressing the wild-type or mutant envelope proteins as described previously (19).

Cell-cell fusion assays.

The XC cocultivation cell fusion assay has been described previously (26). To measure syncytium formation in GPG4 cells, 2 × 10⁵ cells were plated in a 60-mm-diameter tissue culture dish and transfected the following day with 10 μg of envelope protein expression plasmid as described previously (45). Following overnight incubation, the precipitate was replaced with fresh medium, and the plates were stained with methylene blue 24 h later. Cells with more than four nuclei were scored as syncytia.

Coimmunoprecipitation.

The procedure is the same as described previously (45). Briefly, envelope protein expression plasmids were transiently transfected into 293T cells. The cells were labeled for 4 h in cell-labeling medium containing 100 μCi each of [³⁵S]methionine and [³⁵S]cysteine (Amersham), lysed on ice with immunoprecipitation buffer {25 mM Tris-HCl (pH 7.4), 200 mM NaCl, 6 mM 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate (CHAPS; Pierce, Rockford, Ill.)}, and centrifuged. The supernatant was immunoprecipitated with 4 μl of goat anti-gp70 antiserum (lot 79S656; Quality Biotech, Camden, N.J.) or 4 μl of rabbit anti-R peptide antiserum (kindly provided by John Elder, Scripps Institute), together with 20 μl of protein G-Sepharose (Sigma), and incubated overnight at 4°C. The immunoprecipitates were washed three times with immunoprecipitation buffer, resuspended in 2 × sodium dodecyl sulfate (SDS) gel loading buffer, and electrophoresed on SDS–8 to 16% polyacrylamide gels. The dried gels were exposed to BioMax MR film at −70°C.

RESULTS

Predicted structural features of MoMuLV TM.

The MoMuLV TM protein, p15E, is shown schematically in Fig. 1. The N terminus contains a hydrophobic stretch of amino acids that probably constitute a fusion peptide, followed by a region rich in glycine and threonine residues (20, 46). This region is followed by a heptad repeat sequence that has been shown to form a triple-stranded coiled coil when crystallized (11) and then a region containing three cysteine residues that is highly conserved in all retroviral TM proteins (13).

FIG. 1.

Predicted functional domains of MoMuLV TM protein p15E. (A) The MoMuLV TM protein extends from amino acid 437 at the SU-TM cleavage site to residue 632. The C-terminal 16 amino acids (the R peptide) are removed by the viral protease (arrow). Three regions predicted to form amphipathic α-helices, including a heptad repeat sequence (residues 483 to 521) that has been shown by crystallographic studies to form a triple-stranded coiled coil (11), are represented as shaded boxes. C, cysteine residue. The relative positions of the mutants used in this study are shown by asterisks. (B) Sequence of p15E protein and locations of mutants used in this study. The R peptide cleavage site is marked with an arrow.

Neural net algorithms (34) predict several α-helical regions in p15E. In the ectodomain, the heptad repeat sequence and a second, more downstream region are predicted to form α-helices, as is the transmembrane region and the first 16 amino acids of the cytoplasmic tail. While crystallographic studies using a peptide spanning residues 482 to 533 have confirmed the helical nature of the heptad repeat region (11), no structural information is available for the other predicted ectodomain helix between residues 539 and 561. However, a second, more membrane proximal α-helix has been shown to exist in the HIV-1 TM protein (6, 39), and additional α-helices are also predicted in the analogous regions of several other retroviral TM proteins (references 14 and 17 and data not shown). In the cytoplasmic tail, the first 16 amino acids are predicted to form an amphipathic helix (43) and a peptide corresponding to residues 601 to 616 adopts a helical conformation in the presence of lipids (34a). The final 16 amino acids of the cytoplasmic tail, the R peptide, are cleaved by the viral protease and regulate envelope protein fusogenicity (19, 31, 32).

Fusion defective mutants of MoMuLV map to distinct regions of TM.

A large number of point mutants of the MoMuLV TM protein were screened to identify those envelope proteins that were primarily defective in fusion (references 19 and 46 and data not shown). We define fusion mutants as envelope proteins that are normally processed and transported to the surface of the cell, are incorporated into virions, bind to the ecotropic MuLV receptor on NIH 3T3 cells but are unable to induce cell-cell fusion (syncytium formation) in XC cells, and are defective at promoting virus-cell fusion, as measured by the transduction of retroviral vectors.

We assembled a panel of six fusion mutants that mapped to the distinct regions of MoMuLV TM that we had identified (Table 1). At the N terminus, we chose mutant L445E in the hydrophobic core of the fusion peptide and mutant T461P in the GT-rich region that we have previously shown to be fusion defective (46). In the heptad repeat region, we used two substitutions of residue L493; mutant L493V is incorporated efficiently into viral particles, although it has a tendency to lose SU when pelleted through sucrose, whereas mutant L493R is less efficiently incorporated. Both mutants proteins are present at normal levels on the surface of transfected cells but are unable to induce syncytia. In the second predicted α-helical region in the ectodomain we identified mutant R553Q as being primarily fusion defective, and in the cytoplasmic tail we chose the deletion mutant del603-606 (19).

TABLE 1.

Properties of fusion-defective TM mutants

Envelope proteina	Cell surface expressionb	Binding to NIH 3T3 cellsb	Incorporation into virionsc	Cell-cell fusiond	Relative titere (CFU/ml)
None	0	0	−	−	0
CEE+ (wild type)	100	100	+++	+++	(3.1 ± 2.8) × 106
L445E	111	110	+++	−	1 ± 0.6
T461P	101	105	+++	−	3 ± 0.8
L493V	92	105	+++ (SU labile)	−	1 ± 0.4
L493R	94	106	+ (SU labile)	−	0
R553Q	97	106	+++	−	73 ± 14
del603-606	98	102	+++	−	45 ± 9

^aPoint mutations of MoMuLV TM protein were introduced into wild-type expression vector CEE+. 

^bLog absolute fluorescence value relative to CEE+. 

^cRelative amount of envelope protein in viral particles. Mutant L493V had normal levels of TM protein but lower amounts of SU; L493R had lower overall levels of TM also. 

^dNIH 3T3 cells transiently expressing envelope proteins were overlaid with XC cells; syncytia (four or more nuclei) were counted. Wild-type CEE+ produced 11 to 20 syncytia per field. 

^eRetroviral vectors were produced by transient transfection of 293T cells, and titers were determined on NIH 3T3 cells. The values shown are averaged from four transfections ± standard error. 

Coexpression of certain TM mutants results in transduction and defines distinct complementation groups.

We have previously demonstrated functional complementation between binding-defective SU mutants and fusion-defective TM mutants, as coexpression of the two envelope proteins rescues both cell-cell fusion and retroviral vector transduction (45). We now wished to extend those studies to determine whether different TM fusion mutants were able to complement each other and restore envelope protein function.

Retroviral vectors containing each of the individual TM mutant proteins were unable to transduce NIH 3T3 cells efficiently. However, the coexpression of certain combinations of TM mutants gave rise to virions that had greater abilities to transduce NIH 3T3 cells than either mutant alone (Table 2). This complementation did not occur for all combinations of TM mutants that we tested. Notably, the three mutants in the predicted helical regions (L493V, R553Q, and del603-606) were unable to complement each other, whereas all could be complemented by T461P. This result suggests that the helix group mutants form one complementation group that is distinct from T461P. L445E could also complement the helix group of mutants, although to a lesser extent than could T461P, but the combination of L445E and T461P did not result in complementation. These data suggest that the two N-terminal mutants, L445E and T461P, lie in the same complementation group, which is distinct from the helix group mutants.

TABLE 2.

Complementation for titer by coexpression of TM mutants

Envelope protein(s)a	Titer on NIH 3T3 cellsb (CFU/ml)	Complementation indexc
None	0
CEE+ (wild type)	(3.1 ± 1.2) × 106
L445E	1 ± 0.6
T461P	3 ± 0.8
L493V	1 ± 0.4
R553Q	73 ± 14
del603-606	45 ± 9
L445E + T461P	1 ± 0.4	0.5
L445E + L493V	5 ± 3	5
L445E + R553Q	475 ± 163	12.8
L445E + del603-606	375 ± 182	16.3
T461P + L493V	375 ± 182	188
T461P + R553Q	5,350 ± 2,327	141
T461P + del603-606	7,750 ± 1,949	323
L493V + R553Q	21 ± 7	0.6
L493V + del603-606	13 ± 4	0.6
R553Q + del603-606	105 ± 32	1.8

^aEnvelope protein expression plasmids were transfected separately or in combination into 293T cells, along with plasmids pHIT60 and pCnBg, in order to generate retroviral vectors. 

^bTiters of virions produced by transient transfection of 293T cells were determined on NIH 3T3 cells. The values shown are averaged from four independent transfections ± standard error. 

^cRatio of mean titer of coexpressed mutants a and b: (mean titer a + mean titer b)/2. 

These results were obtained by using a transient expression system to generate the retroviral vectors (36). To confirm that these data were not the result of an artifact of this system, we also examined the ability of mutants T461P and R553Q to complement each other when expressed in a stable producer cell line. We separately, or sequentially, introduced T461P and R553Q envelope proteins into GP8 cells (which express gag-pol). One of the envelope protein expression cassettes was based on the retroviral vector LXSN (25), which allowed the resulting supernatants to be titered for transfer of neo resistance. While stable GP8 cells expressing the T461P or R553Q proteins individually produced supernatants with titers of less than 20 CFU per ml, five of seven individual clones examined following selection for both envelope protein mutants produced supernatants with titers of 10³ to 10⁴ CFU/ml (data not shown), indicating functional complementation.

Hetero-oligomers form efficiently between different TM mutants.

We have previously argued that functional complementation between defective envelope proteins occurs through hetero-oligomer formation (45). We considered the possibility that the lack of complementation that we observed with certain combinations of TM mutants was due to inefficient hetero-oligomer formation. This was especially of concern as some of the mutants were located in regions of the TM that have been implicated in envelope protein oligomerization (4, 9, 10, 29, 38).

We have previously analyzed the ability of mutants to form hetero-oligomers by using a coimmunoprecipitation assay with antiserum that recognizes the R peptide (45). This assay takes advantage of the fact that MoMuLV envelope protein expressed in the absence of the viral protease will retain the R peptide and, additionally, that we can construct envelope proteins artificially truncated at the natural R peptide cleavage site. Plasmid CEE+ expresses full-length MoMuLV envelope protein, and plasmid CEETR expresses an R-peptide-truncated form of the protein.

The envelope protein expression plasmids CEE+ and CEETR were transfected individually or together in equal amounts into 293T cells, and cell lysates were immunoprecipitated with either anti-SU or anti-R peptide antiserum. The anti-SU antiserum could immunoprecipitate both the full-length (p15E) form of the TM protein expressed by CEE+ and the R-less (p12E) form expressed by CEETR (Fig. 2A). As expected, the anti-R peptide antiserum recognized envelope proteins from cells transfected only with CEE+ and not the R-less CEETR. However, when CEE+ and CEETR were transfected together, the anti-R peptide antiserum immunoprecipitated both the full-length p15E protein from CEE+ and also the R-less p12E form from CEETR (Fig. 2B). The ability of the anti-R peptide antiserum to bring down the R-less CEETR protein suggests a close physical association between these two TM proteins. Such an association cannot be achieved simply by mixing lysates from cells singly transfected with CEE+ or CEETR (45) but requires the coexpression of both proteins. This assay therefore demonstrates that the coexpression of two envelope proteins results in a close physical association, most probably through the formation of mixed oligomers.

FIG. 2.

Coimmunoprecipitation of full-length and R-less envelope proteins by anti-R peptide serum. Envelope proteins were transiently expressed in 293T cells and labeled with [³⁵S]Met and [³⁵S]Cys. Cells were lysed, and the supernatants divided into two aliquots and immunoprecipitated with either 5 μl of anti-SU antiserum or 5 μl of anti-R peptide antiserum. Full-length TM (p15E), R-peptide-truncated TM-R (p12E), and SU (gp70) proteins were resolved on SDS–8 to 16% polyacrylamide gels. (A) Both CEE+ (full-length) and CEETR (R-less) TM proteins could be immunoprecipitated by the anti-SU antiserum, whether transfected separately or together. (B) The TM-R protein expressed by CEETR was immunoprecipitated by the R-peptide antiserum only in the presence of the full-length CEE+ protein. The right-hand panel is a lighter exposure of the SU protein; Bkg is a background band. (C) Various combinations of full-length and R-less versions of the TM mutants were coexpressed and shown to form hetero-oligomers equally efficiently, as assessed by the immunoprecipitation of the R-less proteins by the anti-R peptide antiserum. Lane 1, L445E and T461P-TR; lane 2, L445E and L493V-TR; lane 3, L445E and R553Q-TR; lane 4, L445E and del603-606-TR; lane 5, T461P and L493V-TR; lane 6, T461P and R553Q-TR; lane 7, T461P and del603-606-TR; lane 8, L493V and R553Q-TR; lane 9, L493V and del603-606-TR; lane 10, R553Q and del603-606-TR; lane 11, CEE+ and CEETR.

We coexpressed various combinations of the TM mutants in full-length and R-less forms and performed immunoprecipitations using both the anti-SU and the anti-R peptide antisera. All of the mutants demonstrated equivalent abilities to oligomerize in this assay (Fig. 2C and data not shown). In particular, we noted that the helix group mutants were equally able to oligomerize with each other as with mutant T461P. This result indicates that the lack of complementation that we observed between the helix group mutants did not arise because of an inability to oligomerize efficiently.

Effect of R peptide cleavage on TM mutants.

We (19, 31) and others (32) have previously demonstrated that R peptide cleavage enhances the fusogenicity of the MoMuLV envelope protein, allowing syncytia to form when an R-less protein is expressed in NIH 3T3 cells. We were interested to determine whether R peptide truncation could in some way compensate for the defects in fusion of the various TM mutants. We therefore constructed R-less versions of the panel of TM mutants and assessed their ability to induce syncytia in GPG4 cells. These cells, which are derived from NIH 3T3 cells and express MoMuLV gag-pol, were chosen because they demonstrate a clear phenotypic difference between full-length and R-less envelope proteins (Table 3). R-less versions of the mutants L445E, T461P, L493R, and R553Q did not induce syncytia in GPG4 cells. However, R-less versions of mutants L493V and del603-606 were able to induce some syncytia, suggesting that these proteins were not as defective as the other mutants.

TABLE 3.

Syncytium formation in GPG4 cells

Envelope proteina	Syncytium formationb
None	−
CEE+ (wild type)	−
CEE-TR (R-less)	++++
L445E	−
L445E-TR	−
T461P	−
T461P-TR	−
L493V	−
L493V-TR	++
L493R	−
L493R-TR	−
R553Q	−
R553Q-TR	−
del603-606	−
del603-606-TR	+++

^aFull-length and R-peptide-truncated versions of wild-type and mutant envelope proteins were transiently expressed in GPG4 cells, and syncytia were counted 48 h posttransfection. 

^b++++, >20 syncytia per field; +++, 11 to 20 syncytia; ++, 6 to 10 syncytia; −, no syncytia. 

Nonreciprocal enhancement of fusogenicity in trans by R peptide truncation.

We have previously shown that the enhancement of fusogenicity mediated by R peptide truncation can occur in trans within a mixed envelope protein oligomer (45). The expression of an R-less form of a binding-defective SU mutant, construct D84K-TR (23), in NIH 3T3 cells will not give rise to syncytia, due to its inability to bind to the ecotropic receptor. However, the coexpression of D84K-TR with the full-length wild-type protein resulted in syncytia (45). We concluded from this study that in the hetero-oligomers that we presumed to form, the SU components contributed by the wild-type protein bound the complexes to the ecotropic receptor, while the R-less TM proteins contributed by the D84K-TR protein activated fusion in trans.

The fact that R peptide truncation enhances fusion in trans provided us with a tool with which to further investigate the relationships between the different TM mutants and, in addition, to investigate the mechanism of the R peptide cleavage enhancement of fusogenicity. Accordingly, we coexpressed R-less versions of the TM mutants with the wild-type envelope protein in GPG4 cells and assessed their ability to enhance fusion in trans. As a control, we also included the coexpression of D84K-TR with the wild-type protein. For these analyses, we did not use the R-less version of mutant del603-606, as this protein by itself produces a high background level of syncytia (Table 3). In addition, we used an R-less version of mutant L493R in preference to mutant L493V, as this construct gave no background levels of syncytia when expressed in GPG4 cells.

While neither the wild-type full-length protein, CEE+, or any of the R-less versions of the mutant proteins alone produced syncytia, the combinations of CEE+ with either D84K-TR or T461P-TR gave rise to syncytia (Table 4). In contrast, the coexpression of CEE+ with L445E-TR, L493R-TR, or R553Q-TR did not result in any syncytia. This analysis therefore revealed a difference between the two N-terminal mutants, L445E and T461P; while both mutants are fusion defective, even when R-less, L445E cannot form part of a fusogenic complex. This observation also agrees with the data from the titer complementation assays (Table 2), which revealed that the two mutants had different abilities to complement the helix group of TM mutants and indicates that L445E is a more severely defective fusion mutant than T461P.

TABLE 4.

R-less enhancement of fusogenicity in trans

Envelope proteina	Syncytium formationb
None	−
CEE+ (wild type)	−
CEE-TR (R-less)	++++
D84K-TR	−
L445E-TR	−
T461P-TR	−
L493R-TR	−
R553Q-TR	−
CEE+ + D84K-TR	+++
CEE+ + L445E-TR	−
CEE+ + T461P-TR	+++
CEE+ + L493R-TR	−
CEE+ + R553Q-TR	−
T461P-TR + L493R	+
T461P-TR + L493V	++
T461P-TR + R553Q	++
T461P-TR + del603-606	++
L493R-TR + T461P	−
R553Q-TR + T461P	−

^aFull-length and R-peptide truncated versions of wild-type and mutant envelope proteins were transiently expressed in GPG4 cells, and syncytia were counted 48 h posttransfection. 

^b++++, >20 syncytia per field; +++, 11 to 20 syncytia per field; ++, 6 to 10 syncytia; +, 1 to 5 syncytia; −, no syncytia. 

The titer complementation assay had demonstrated that T461P and the helix group mutants could complement each other for virus-cell fusion (titer). We were interested in examining whether T461P-TR could enhance fusogenicity in trans when expressed with the helix group mutants. In addition, we wished to determine whether any such complementation would be reciprocal. We therefore coexpressed T461P-TR with the helix group mutants L493R and R553Q and also coexpressed T461P with R-less versions of these two mutants (Table 4). This analysis revealed that T461P-TR could indeed enhance fusogenicity when expressed in combination with either L493R or R553Q. In contrast, R-less versions of L493R and R553Q could not enhance fusogenicity when coexpressed with T461P, which is in agreement with their inability to enhance fusion when coexpressed with the wild-type protein. The nonreciprocal nature of this effect demonstrates that the trans enhancement of fusogenicity mediated by R peptide cleavage has specific cis requirements in TM; while it is tolerant of the N-terminal mutant T461P, it cannot function in cis to mutants L493R and R553Q.

DISCUSSION

We have identified several fusion-defective mutants of MoMuLV envelope protein that appear to map to distinct functional regions of the TM protein. By coexpressing these mutants and looking for rescue of fusion ability or infectivity, we have assigned these mutants to two different complementation groups. Interestingly, mutants in the heptad repeat, a predicted α-helix in the TM ectodomain, and a predicted α-helix in the cytoplasmic tail all appear to be in the same group, suggesting a functional interaction between these three regions. In addition, we examined the effects of R peptide cleavage on mutants from the two complementation groups. Our data suggest that the enhancement of fusogenicity resulting from R peptide cleavage appears to work through an interaction with the two helical regions in the ectodomain of TM.

The fusion-defective mutants used in this study were located in five regions of the MoMuLV TM protein, defined by structural studies, computer modeling, and our previous mutational analyses (11, 19, 34, 46). At the N terminus, mutant L445E occurs in the presumed hydrophobic core of the N-terminal fusion peptide. The other N-terminal mutant, T461P, is situated in a GT-rich region, the analogous region of which in HIV-1 TM has been reported to be important for SU-TM interactions (21). The remaining mutants, L493V/R, R553Q, and del603-606, all occur in regions predicted to be amphipathic α-helices (the helix group). The two proposed helical regions in the ectodomain appear to be vital for the fusion process, as even the relatively conservative substitutions of L493V and R553Q resulted in severely defective proteins.

To examine the functional relationships between these regions of the TM protein, we used the fact that defective envelope proteins can in some cases complement each other to restore function. We coexpressed various combinations of the fusion-defective proteins on retroviral vector particles and looked for an ability to rescue viral titer. These analyses revealed that while mutant T461P could efficiently complement all of the helix group mutants and restore viral titer, it could not complement mutant L445E. Mutant L445E could also complement the helix group mutants, albeit at a lower level than T461P. Despite the demonstrated ability of the helix group mutants to be functionally complemented by both T461P and L445E, they were unable to complement each other. These three helical regions of the protein therefore form a distinct complementation group from L445E and T461P. Their inability to complement each other was not due to a lack of efficient hetero-oligomer formation, as coimmunoprecipitation analyses revealed that all of the mutants were able to oligomerize equally efficiently. Taken together, our data provide evidence for functional domains in the retroviral TM protein that may cooperate during the fusion process.

There are precedents for the interaction of helical regions in the ectodomains of viral fusion proteins. In influenza virus HA₂, the low-pH-induced conformational rearrangements extend the helical heptad repeat into a triple-stranded coiled coil that is supported at its base by an additional helical region (2). That a similar interaction may be involved in fusion mediated by retroviral TM proteins was initially proposed based on studies with fusion-inhibitory peptides derived from the HIV-1 TM. Peptides corresponding to two predicted helices in the ectodomain individually inhibited fusion but were found to sequester each other when both were present, suggesting an interaction (7). Such an interaction has now been confirmed by data from cocrystallized HIV-1 peptides spanning these two regions (6, 39). While there are no crystallographic data to support the presence of a helical region between MuLV residues 539 and 561, computer modeling also predicts similar helices in the TM proteins of Rous sarcoma virus and Mason-Pfizer monkey virus (data not shown).

The lack of complementation between del603-606 in the cytoplasmic tail and L493V or R553Q in the ectodomain suggests that the cytoplasmic tail can influence the ectodomain of the protein in a manner that involves these two helical regions. The ability of the cytoplasmic tail to modulate the function of the protein ectodomain is also indicated by the increase in fusogenicity that occurs upon R peptide cleavage. Similarly, truncation of the cytoplasmic tail of the SIV envelope protein has previously been shown to alter both the fusogenicity of the protein and the gross conformation of its ectodomain, as detected by altered susceptibility to biotinylation reagents (37).

If R peptide cleavage affects fusogenicity by influencing the interaction between two helical regions in the TM ectodomain, a predicted consequence would be that the ability of R peptide cleavage to enhance fusogenicity in trans within a mixed oligomer would not work in cis to either mutant L493R or R553Q, and indeed we have found this to be the case. In contrast, the trans enhancement of fusogenicity caused by R peptide truncation was unaffected by the mutation T461P. Furthermore, while the R-less form of T461P could enhance fusogenicity in trans when coexpressed with both mutants L493R and R553Q, the reciprocal arrangement with the R peptide truncation occurring on mutants L493R or R553Q did not lead to syncytia.

The mechanism of control of fusion by the R peptide is unknown. It is possible that the R peptide interacts with a fusion-inhibiting protein whose effect is relieved upon cleavage. An interaction with such a cellular protein could explain the apparent ability of the MuLV R peptide to regulate the fusogenicity of a truncated SIV envelope protein in the context of a chimeric envelope protein (43). However, such a model is at odds with the lack of trans dominance of the R peptide. The explanation that we currently favor is that R peptide cleavage removes a conformational constraint on the cytoplasmic tail. Possibly, the amphiphathic nature of the remainder of the cytoplasmic tail enhances a subsequent association with the membrane or promotes intermolecular interactions in the cytoplasmic tail within an envelope protein oligomer. Any conformational changes occurring in the tail could then be transmitted to the rest of the molecule. We have data to suggest that the structure of the cytoplasmic tail and transmembrane regions of MoMuLV TM can influence the strength of the interactions between the SU and TM subunits (1), and so it is possible that R peptide truncation leads to a reorganization of the ectodomain of the protein, perhaps facilitating SU-TM dissociation subsequent to receptor binding.

A common feature of viral fusion proteins is the adoption of a metastable state, primed for the transition to the fusogenic state upon exposure to the appropriate trigger. A major example of this could be the cleavage event between the SU and TM subunits that presumably positions the fusion peptide ready to be translocated toward the host cell membrane following the interaction with the receptor. It is possible that certain retroviruses use a second cleavage event in the cytoplasmic tail of the protein to allow an additional conformational change to further prime the protein. The entire fusion process for MoMuLV could therefore be viewed as a series of conformational changes or energy state transitions in the envelope protein, starting with SU-TM cleavage, followed by R peptide cleavage and ultimately triggered to a fusogenic state by the interaction with the viral receptor.