G541R within the 4070A TM Protein Regulates Fusion in Murine Leukemia Viruses

Abstract

The mutation G541R within the ectodomain of TM was isolated in three independent chimeric enveloped murine leukemia virus (MuLV) viral populations originally impaired in viral passage and in wild-type 4070A. Isolation of G541R in multiple populations suggested it played a critical role in viral envelope function. Using a viral vector system, the observed effects of the G541R mutation within MuLV envelope proteins were pleiotropic and included effects on the regulation of SU-TM interactions and membrane fusion. G541R suppresses enhanced cell-cell fusion events attributable to the absence of the R-peptide yet does not adversely affect virus titers. The ability to suppress cell-cell fusion is dependent on the presence of the C terminus of the amphotropic 4070A SU protein. Within the wild-type 4070A envelope background, the mutation results in a decreased level of Env at the cell surface that is mirrored in the virion. The TM mutation alters recognition of the SU C terminus by a monoclonal antibody, suggestive of an altered conformation. The presence of G541R allowed the virus to achieve a balance between cytopathogenicity and replication and restored productive viral entry.

Pathogenic viruses are frequently associated with induced cell-cell fusion. Membrane fusion, however, is a requisite step in the productive entry of enveloped viruses. The spatial and temporal control of virus-induced membrane fusion is therefore critical to understand.

Murine leukemia viruses (MuLVs) have adopted various strategies to control virus-cell and cell-cell fusion. Several key elements within the envelope (Env) surface (SU) and transmembrane (TM) proteins have been identified. Mutation of SU N-terminal amino acid H41 (numbering from the signal peptide) in Moloney-MuLV (M-MuLV) or H36 (numbering from the signal peptide) in amphotropic 4070A MuLV abolishes virus-cell and cell-cell fusion events (2, 19, 48). Elements within the proline-rich region (PRR) mediate conformational changes in SU and TM after receptor binding and affect syncytium formation (17). The fusion peptide, which is inserted into the host cell membrane during viral entry, is located at the N terminus of TM. By analogy to the influenza virus protein hemagglutinin, this domain is presumed to be buried in early conformations of the viral envelope (Env) protein, suggesting that its exposure is a controlled event. The 16 C-terminal amino acids of TM, named the R-peptide, act as a negative regulator of fusion. Removal of the R-peptide by the viral-encoded protease (Pr) activates viral fusion (38, 39).

Variations in the interaction of the viral Env with the host-cell receptor affect viral infectivity and fusion. Low expression levels of hemagglutinin (8, 9) and human immunodeficiency virus type 1 (HIV-1) envelope proteins (20) have been shown to down-regulate fusion. For HIV-1, it has been proposed that threshold amounts of receptor and coreceptor molecules are required to initiate fusion (15, 37). In nonmurine cells with high levels of receptor, both wild-type (WT) ecotropic MuLV and a variant, TR1.3, are fusogenic. However, under conditions where receptor expression is limited, only the TR1.3 variant is fusogenic (7). This TR1.3 neuropathogenic MuLV encodes the point mutation W102G (30). W102 forms a receptor-binding pocket with S84 and D86 within the receptor-binding domain (RBD) of Fr57-MuLV (10). Thus, mutations affecting Env-receptor binding can alter cell-cell fusion.

In the present study, the role of a point mutation, G541R, in regulating fusion is analyzed. G541 is located within the TM ectodomain, N-terminal to the intramolecular cysteine bond. This position was found mutated in one WT 4070A and three independent populations of chimeric viruses (EA6 and EA7) (31). In chimeric Env proteins bearing the ecotropic RBD and an amphotropic TM protein, G541R suppressed syncytia with rat XC cells and syncytia induced by the absence of R-peptide in NIH 3T3 cells. Reconstruction of G541R within virus lacking R-peptide improved virus viability and virus titer. The G541R substitution in the TM ectodomain leads to decreased levels of amphotropic Env expression at the cell surface and on the virus. These studies provide insights into the mechanism utilized by MuLV to balance cytopathogenicity and virulence and preserve its intrinsic entry characteristics.

MATERIALS AND METHODS

Cell lines and maintenance.

The generation and maintenance of canine D17/pJET and D17/gag/pol cell lines have been previously described (29). NIH 3T3 and rat XC cells were maintained as previously described (29).

Plasmids and DNA manipulations.

Chimeric proviral DNA clones (31) and plasmid pNCA-C (13) were as previously described. pNCA-C is an infectious provirus that expresses the gag, pol, and env gene products of the ecotropic Mo-MuLV. Plasmids pHIT 123, pHIT 456, and pHIT 111, expressing the Mo-MuLV Env, 4070A Env, and lacZ genes, respectively, were previously described (29, 44, 45). The pHIT 456/R⁻ construct was generated by replacement of the 172-bp ClaI-NheI fragment from pHIT 123/R⁻ (45).

Second-site changes detected in EA6 viral populations were reconstructed within the EA6 proviral backbone. The single mutations N261I and E311V and the double mutation N261I/E311V were each cloned as previously described (28). The double mutant N261I/G541R was generated by the ligation of the EA6 parental SfiI-ClaI 8,945-bp fragment with the 782-bp HindIII-ClaI and SfiI-HindIII 1,360-bp fragments from the appropriate EA6 populations. The single mutant G541R was generated by the ligation of the 782-bp HindIII-ClaI fragment from the N261I/G541R mutant with the 1,360-bp SfiI-HindIII and 8,945-bp SfiI-ClaI fragments from the EA6 parental DNA. The triple mutant N261I/E311V/G541R was generated by the ligation of the 782-bp HindIII-ClaI fragment from the N261I/G541R clone, the 1,360-bp SfiI-HindIII fragment from the N261I/E311V clone, and the 8,945-bp SfiI-ClaI fragment from parental EA6 DNA.

G541R was reconstructed into pHIT 456 by ligation of the 527-bp XhoI-HindIII fragment from 4070A-pNCA-Am (13), the 953-bp HindIII-NheI fragment from pNCA-C/EA6/G541R, and the 7,800-bp XhoI-NheI fragment from pHIT 456. The G541R/R-minus clone was constructed by the ligation of the 1,308-bp XhoI-ClaI fragment from pHIT 456/G541R with the ClaI-NheI R⁻ fragment and the 7,800-bp XhoI-NheI fragment from pHIT 456. The parental EA6 and mutants were introduced into pHIT 123 by exchanging the Mo-MuLV 2,100-bp PmlI-NheI fragment. The corresponding R-minus constructs were made by combining the PmlI-ClaI fragment from each respective env clone with the 400-bp ClaI-NheI fragment derived from pHIT123/R⁻ and the 2,160-bp PmlI-NheI fragment from pHIT123.

G552R was introduced into the Mo-MuLV envelope by an overlapping PCR mutagenesis scheme using 2.5 U of Pfu polymerase (Stratagene). The 650-bp HpaI-ClaI fragment from the G552R PCR product was substituted into pNCA-C (13) to generate pNCA-C/G552R. The 2,160-bp PmlI-NheI fragment of pNCA-C/G552R was substituted into pHIT 123 to generate pHIT 123/G552R. The 1,944-bp PmlI-ClaI fragment of the pNCA-C/G552R clone was substituted into pHIT 123 to generate the pHIT 123/G552/R⁻ clone. The mutation N261I was introduced into the E311V/pNCA-C clone containing a SalI site at bp 6745 to 6750 (28) by replacement of the 927-bp PmlI-DraIII fragment with the same fragment from the EA6/N261I/pNCA-C clone (28). The triple combination N261I, E311V, and G552R was introduced into pHIT 123 by replacement of the 1,000-bp PmlI-SalI fragment of either the G552R/pHIT clone or the G552R/pHIT/R⁻ clone with the same fragment from the N261I/E311V/pNCA-C clone. DNAs were purified by CsCl₂ banding prior to use in transfections.

Transfections.

Transfections of Env constructs, within the plasmid pHIT 123 backbone, into D17/gag-pol cells were performed as previously described for D17 cells using Lipofectamine reagents (Invitrogen) (29). For virus titers, pHIT 111 was cotransfected along with each env plasmid. NIH 3T3 cells were transfected using Lipofectamine in the presence of medium containing 10% serum and 0.1 mM nonessential amino acids (Invitrogen) per the manufacturer's directions. At 24 h posttransfection, cells were treated with 10 mM sodium butyrate for 6 to 8 h to enhance expression levels of Env (44). Sodium butyrate was subsequently replaced with cell maintenance medium.

Titer determinations of envelope proteins.

Transiently expressed virus was collected overnight between 36 and 48 h posttransfection, passed through an acrodisc 0.45-μm-pore-size filter, applied to D17/pJET cells in the presence of 8 μg of Polybrene/ml for 2 h at 37°C, and subsequently replaced with cell maintenance medium. At 48 h postinfection, cells were stained for lacZ expression as previously described (31).

Syncytium assays.

At 48 h posttransfection of Env expression plasmids (pHIT 123 based), D17-based env-expressing cells were UV irradiated for 30 s and overlaid with rat XC cells at 37°C as previously described (14). Syncytia containing 2 nuclei or >4 nuclei were quantitated in a microscopic field having an area of 13.86 mm²; data below in Tables 1 to 4 represent the mean ± standard deviation for syncytia in at least 10 fields for each independent assay.

TABLE 1.

Viral titers and 3T3 syncytium characteristics of EA6/R⁻ Env

Env	Titera	No. of syncytiab
		Assay 1		Assay 2
		2 nuclei	>4 nuclei	2 nuclei	>4 nuclei
EA6/R−
Parent	0	0.3 ± 0.5	0	0.3 ± 0.5	0.4 ± 0.5
N2611	<10	0.2 ± 0.6	2.1 ± 1.8	0.1 ±0.3	1.9 ± 1.1
E311V	15	0	5.8 ± 1.6	0.1 ± 0.3	5.9 ± 1.5
G541R	40	0	0.5 ± 0.8	0.1 ± 0.3	0.7 ± 1.1
N2611/E311V	(1.1 ± 0.2) × 103	0	14.7 ± 3.1	0	11.5 ± 2.4
N261I/G541R	ND	0	0	0.2 ± 0.4	0.4 ± 0.5
E311V/G541R	4.3 × 102*	0.1 ± 0.3	1.3 ± 1.5	0.3 ± 0.5	1.5 ± 1.9
N261I/E311V/G541R	(5.0 ± 2.0) × 103	0.1 ± 0.3	0.7 ± 0.6	0.1 ± 0.3	1.6 ± 1.1
M-MuLV/R+	(8.1 ± 2.0) × 104	0	0	0	0.1 ± 0.3
M-MuLV/R−	(4.8 ± 8.5) × 103	0	24 ± 3.4	0	26.3 + 4.5

^aDNAs encoding the lacz marker and each Env were transiently expressed in D17/gag-pol cells. At 48 h posttransfection, virus was collected and applied to D17/pJET cells. Viral titers are based on the number of D17/pJET cells scoring positive for β-galactosidase expression 48 h postinfection. Values represented as mean ± standard deviations (SD) are based on at least three independent determinations; values without SD represent mean of two independent determinations. *, value represents a mean of two independent determinations of 2.0 × 10² and 6.5 × 10². ND, not determined.

^bAt 48 h posttransfection of DNA encoding each Env into NIH 3T3 cells, syncytia were stained with hematoxylin and counted. Values represent means ± SD of 10 fields of 13.86 mm² of syncytia containing 2 nuclei or >4 nuclei.

TABLE 4.

Viral titers and syncytium characteristics of M-MuLV Env's

Env	Titera	No. of syncytiab
		2 nuclei	>4 nuclei
R+
M-MuLV	(9.0 ± 4.9) × 104	0	44.1 ± 11.1
M-MuLV/G552R	(5.6 ± 2.4) × 104	0	56.1 ± 6.5
M-MuLV/N261I/E311V/G552R	4.0 × 104*	ND	ND
R−
M-MuLV	(5.4 ± 2.8) × 104	0	38.3 ± 6.2
M-MuLV/G552R	(5.1 ± 2.7) × 104	0	40.9 ± 4.3
M-MuLV/N261I/E311V/G552R	2.9 × 104*	ND	ND

^aDNAs encoding the lacz marker and each Env were transiently expressed in D17/gag-pol cells. At 48 h posttransfection, virus was collected and applied to D17/pJET cells. Viral titers are based on the number of D17/pJET cells scoring positive for β-galactosidase expression 48 h postinfection. Values represent mean ± standard deviation (SD) of at least three independent determinations. *, value represents a mean of two independent determinations. ND, not determined.

^bAt 48 h posttransfection of each Env, D17/gag-pol producer cells were UV irradiated and cocultured with rat XC cells. Syncytia were stained with hematoxylin and counted. Values represent mean ± SD of 20 fields of 13.86 mm² containing 2 nuclei and >4 nuclei.

Radiolabeling, cell surface labeling, and Env analysis.

Envelope proteins were transiently expressed in D17/gag-pol cells by the Lipofectamine transfection protocol described above. Virus-associated envelope proteins were analyzed by metabolic labeling and immunoprecipitation as previously described (13, 14). Virus corresponding to one-quarter of a 100-mm plate was immunoprecipitated with 12 μl of antibody 80S-019 (anti-SU antiserum; Microbiological Associates). Following overnight labeling with Tran³⁵S-label, Env-producing cells were incubated with 0.5 mg of EZ-Link NHS-SS-biotin [sulfosuccinimidyl-2-(biotinamido) ethyl-1,3-dithiopropionate; Pierce] in 1 ml of phosphate-buffered saline (PBS) for 30 min at room temperature. The surface-labeling reactions were quenched by the subsequent addition of 10 ml of Dulbecco's modified Eagle's medium (DMEM) containing 10% heat-inactivated calf serum. Cells were then washed three times with PBS and lysed in 1 ml of phosphate lysis buffer. Lysates were clarified by centrifugation at 16,000 × g for 5 min. SU and SU-associated TM was immunoprecipitated with 12 μl of anti-SU polyclonal antibody 80S-019 and 50 μl of immobilized protein G (50% slurry in PLB) (Pierce) overnight at 4°C. Protein G-Env beads were washed three times with PLB. One-half of this immunoprecipitation was analyzed for total cellular proteins. Biotinylated Env, eluted by boiling in 20 μl of 10% sodium dodecyl sulfate (SDS), was subsequently complexed with 20 μl of immobilized streptavidin (50% slurry in PLB) (Pierce) for 4 h at 4°C. Biotinylated Env-streptavidin complexes were subsequently washed with PLB. Env associated with either protein G or streptavidin beads was recovered by boiling in SDS-polyacrylamide gel electrophoresis (PAGE) sample buffer containing 2% β-mercaptoethanol. Samples were applied to an SDS-12% PAGE; ³⁵S signals were enhanced by fluorography prior to exposure to film.

Virus was alternatively purified by centrifugation of 1-ml aliquots of culture supernatants through 0.2 ml of 20% sucrose in 10 mM Tris-HCl (pH 7.5), 1 mM EDTA, 50 mM NaCl (TEN) at 16,000 × g (46). Viral proteins were analyzed by Western blotting subsequent to SDS-PAGE analysis and visualized by chemiluminescence (Pierce). The resultant polyvinylidene difluoride (PVDF) membranes were cut into two pieces, consisting of proteins migrating as 31 kDa and higher molecular masses and 31 kDa and lower molecular masses. PVDF membranes with viral proteins of 31 kDa and higher molecular masses were incubated with antibodies to the SU Env protein. Duplicate panels of the same viral preparations were generated on a single gel to allow for analysis using two independent SU-specific antibodies. Posttransfer, the PVDF filter was cut vertically to separate the two panels and horizontally using the 45-kDa protein standard for analysis of anti-SU versus anti-CA antibodies. Viral SU was either detected with SU monoclonal antibody 83A25 (11) and secondary antibody goat anti-rat immunoglobulin G (IgG; Pierce; diluted 1:100,000) or using SU polyclonal 80S-019 (Microbiological Associates) and secondary rabbit anti-goat IgG (Pierce; diluted 1:100,000). PVDF membranes with viral proteins of 31 kDa and lower were incubated with CA polyclonal antibody (079-287) and secondary antibody goat anti-rabbit IgG (Pierce; diluted 1:100,000).

RESULTS

Summary of viral isolates containing G541R within the TM protein.

The Env protein on the retroviral particle is maintained in a metastable conformation through coordination of multiple domains within the SU and TM proteins. Key regulators of the Env conformation have been identified by the characterization of compensatory mutations generated through in vivo passage of chimeric ecotropic and amphotropic Env proteins (22, 29, 31). The amphotropic 4070A residue G541 was altered to an arginine in multiple independent chimeric Env isolate EA6 and EA7 (28) and WT 4070A (29) viral populations. The key features of these infectious viruses are summarized in Fig. 1. G541 is within the 4070A TM ectodomain between the fusion peptide and the CX₆CC motif, which is involved in covalent SU/TM interaction (35). There is high conservation of sequence between the ecotropic M-MuLV and the amphotropic 4070A isolates within this region. Passage of WT 4070A virus in canine D17 cells resulted in the selection of G100R (21) and G541R. G541R was also isolated after passage of two ecotropic and amphotropic chimeric Env-bearing junctions within the SU C terminus (EA6 and EA7). For EA6, G541R was isolated upon secondary passage of virus bearing the mutations N261I and E311V (28). G541R was also isolated in the secondary passage of EA7 chimeras bearing deletion of S39/P40 as well as an independent EA7/V421 M isolate (28).

FIG. 1.

Schematic of viral populations isolated with the G541R TM mutation. Viral populations of 4070A and chimeric EA6 and EA7 were isolated with second-site changes in the TM ectodomain residue G541R after multiple passages through canine D17 and D17/pJET cells. Amino acid numbering begins with the initiation methionine. Chimeric junctions 6 and 7 are as previously described (31); amino acid junctions correspond to the ecotropic residue and amphotropic residue. Additional second-site changes isolated in these populations are noted. For comparison, a schematic of M-MuLV is shown at the top and denotes the arrangement of variable domains VRA, VRC, VRB, and the PRR within the SU protein, as well as the positions of the two cysteine residues which comprise the CWLC motif. The amino acid sequence in the ectodomain of TM in the vicinity of 4070A residue G541 is noted, and it is conserved between M-MuLV and 4070A MuLV.

Due to its localization within TM, the role of G541R in regulating fusion was investigated. For MuLVs, two systems have been used to characterize cell-cell fusion. With the first system, envelope constructs lacking the R-peptide are activated for fusion and syncytia are visible in NIH 3T3 cells. The second system utilizes rat XC cells, which induce syncytia in the presence of ecotropic MuLV but not amphotropic MuLV (27, 41, 43). The effect of G541R was analyzed in each of these systems.

G541R affects cell-cell fusion events.

Removal of the R-peptide from the TM protein has been shown to activate fusion for both ecotropic and amphotropic MuLV envelope proteins (38, 39). Enhanced cell-cell fusion can be observed when R-minus Env proteins are directly expressed in NIH 3T3 cells. WT M-MuLV does not induce syncytia in NIH 3T3 cells (Fig. 2C), while very large syncytia are observed for the R-minus M-MuLV control (Fig. 2B). The expression of EA6 chimeric Env's bearing the G541R mutation within an R-minus backbone led to the visualization of a remarkable phenotype, the suppression of syncytia. EA6 Env bearing the N261I/E311V mutations within the R-minus backbone resulted in large multinucleated cells (Fig. 2E). In contrast, EA6/R⁻ virus bearing N261I/E311V plus G541R resulted in no syncytia, even in the absence of R-peptide (Fig. 2F).

FIG. 2.

Suppressive effects of G541R in the R-minus Env backbone. DNA corresponding to R⁻ Env was transiently transfected in NIH 3T3 cells. At 48 h posttransfection, syncytia were visualized by hematoxylin staining. (A to C) Mock and WT M-MuLV controls. (D to F) Syncytia for Env mutants in the EA6/R⁻ backbone. In all panels (A to F), virus titers (lacz⁺ IFU) of M-MuLV controls and R⁻ Env's transiently expressed in D17/gag-pol cells, representative of one assay, are shown at the top of each syncytia panel.

Table 1 provides a quantitative analysis of the syncytia and virus titers of the complete panel of EA6/R⁻ chimeras bearing N261I, E311V, and G541R mutations individually and in combination. The Env proteins were expressed within a three-plasmid vector system (pHIT [44]) and thus represent a single round of infection. Virus titer was determined by the transfer and expression of the lacZ gene from pHIT 111. For the M-MuLV/R⁻ control, virus titers of 4.8 × 10³ infectious units (IFU)/ml were obtained with 24 large multinucleated syncytia (>4 nuclei) visible per 13.86-mm² field. These syncytia were not observed with the WT M-MuLV (R⁺). In contrast, no transfer of the lacZ gene was observed with the parental EA6/R⁻ virus. Incorporation of each of the single mutations N261I, E311V, and G541R within the EA6/R-minus constructs yielded titers on the order of 10¹ IFU/ml. A low but detectable level of syncytia was observed with EA6/R⁻ viruses bearing either N261I (average, 2/field) or E311V (5.8/field). Cooperation between mutations was observed. The combination of N261I plus E311V within EA6/R⁻ yielded virus titers of 1.1 × 10³ IFU/ml with 14.7 and 11.5 syncytia (>4 nuclei), respectively, in two independent assays. This corresponds to levels of syncytia between 44 and 61% of that observed with the M-MuLV/R⁻ virus. The combination of N261I/E311V/G541R in EA6/R⁻ yielded the highest titer, 5.0 ± 2.0 × 10³ IFU/ml. Remarkably, EA6/R⁻ virus bearing N261I/E311V/G541R showed a marked reduction of syncytia, even in the absence of the R-peptide (0.7 to 1.6 syncytia/field).

The combination of E311V plus G541R within EA6/R⁻ resulted in a modest titer (10² IFU/ml); the presence of G541R again decreased the size and number of the syncytia present compared to E311V alone.

G541R suppresses syncytia with rat XC cells.

The EA6 chimera encodes the ecotropic RBD, a key determinant for formation of syncytia with XC cells (31). Chimeric viral Env's (R⁺) were transiently expressed in a D17/gag-pol cell line and analyzed for cell-cell fusion following coculturing with rat XC cells. Table 2 summarizes resultant virus titers (in IFU per milliliter) and quantification of syncytia. The control WT M-MuLV Env results in numerous large, multinucleated cells within each field (between 23.7 and 31.2). Parental EA6 exhibits a low number of syncytia, consistent with a lack of detectable virus titer.

TABLE 2.

Viral titers and XC syncytium characteristics of EA6/R⁺ Env's

Env	Titera	No. of syncytiab
		Assay 1		Assay 2
		2 nuclei	>4 nuclei	2 nuclei	>4 nuclei
EA6
Parent	0*	0.3 ± 0.7	1.9 ± 1.6	0.2 ± 0.4	3.1 ± 1.6
N261I	<10*	0	2.1 ± 1.2	0	2.2 ± 1.2
E311V	(2.8 ± 2) × 103	0	15.2 ± 4.2	0	11.1 ± 2.2
G541R	(2.3 ± 1.2) × 103	0	0	0	0.2 ± 0.4
N261I/E311V	(1.2 ± 0.2) × 104	0	28.3 ± 10.3	0	28.0 ± 4.8
N261I/E311V/G541R	(2.4 ± 0.2) × 104	0.1 ± 0.3	0.1 ± 0.3	0	0
M-MuLV	(8.0 ± 6.0) × 104	0	31.2 ± 4.3	0	23.7 ± 4.2

^aDNAs encoding the lacZ marker and each Env were transiently expressed in D17/gag-pol cells. At 48 h posttransfection, virus was collected and applied to D17/pJET cells. Viral titers are based on the number of D17/pJET cells scoring positive for β-galactosidase expression 48 h postinfection; values represent means ± standard deviations (SD) for at least three independent determinations.

^bAt 48 h posttransfection of each Env, D17/gag-pol producer cells were UV irradiated and cocultured with rat XC cells. Syncytia were stained with hematoxylin and counted. Values represent the mean number of syncytia ± SD containing either 2 nuclei or >4 nuclei in 10 fields, each with an area of 13.86 mm².

The single replacement N261I results in limited cell-cell fusion events with background levels of transduction of the lacZ gene (in IFU per milliliter). Replacement of the single mutation E311V results in a pronounced number of syncytia and an improved virus titer of 2.8 × 10³ IFU/ml relative to the parental EA6 control. The double combination of N261I and E311V mutations produces large multinucleated syncytia in the presence of XC cells similar in quantity to that observed for WT M-MuLV, with a concomitant increase in titer to 1.2 × 10⁴ IFU/ml (28). Incorporation of G541R into EA6/N261I/E311V inhibited the syncytium-positive phenotype while, interestingly, it had a titer of 2.4 × 10⁴ IFU/ml. Despite improving the titer to 2.3 × 10³ IFU/ml in comparison to the parental EA6 titer, incorporation of G541R alone into EA6 did not yield syncytia with XC cells. Similar results were seen when these chimeric Env proteins were expressed in D17 cells (data not shown). This is consistent with previous observations that syncytia with XC cells do not require R-peptide cleavage, which is catalyzed by the viral protease encoded by the pol gene (45). Thus, the presence of G541R was capable of suppressing syncytium formation for fusion-activated Env constructs lacking the R-peptide, as well as with XC cells for chimeric Env containing the R-peptide. In combination, the titer and the syncytia results indicate G541R acts at the cellular level to regulate fusion events and yield productive viral infection.

G541R suppresses fusion for the WT 4070A amphotropic envelope.

The mutation G541R was previously also isolated in a WT 4070A MuLV population passaged through D17 cells (29). Syncytium assays were performed with this WT variant in order to determine how universal the syncytium-suppressive effect might be. Syncytia are not normally observed for the WT amphotropic Env in an XC cocultivation assay, necessitating the analysis of this mutation in an R⁻ backbone. Analyses of syncytia on XC and NIH 3T3 cells are summarized in Table 3. 4070A Env lacking the R-peptide produced syncytia with both XC as well as NIH 3T3 cells. Inclusion of G541R into 4070A Env/R⁻ decreased the number of syncytia 6-fold when cocultivated with rat XC cells and 6- to 10-fold within NIH 3T3 cells. Control experiments of M-MuLV Env with XC cells yielded massive syncytia (data not shown). While G541R improved the titers of the chimeric EA6 (Tables 1 and 2) and EA7 viruses (1.7 × 10³ [28]), substitution of G541R in the WT 4070A backbone decreased virus titers by two- to fourfold in two independent determinations (Table 3).

TABLE 3.

Viral titers and syncytium characteristics of 4070A Env's

Env	Titera	No. of syncytiab in:
		XCb		NIH 3T3c
		2 nuclei	>4 nuclei	Assay 1		Assay 2
				2 nuclei	>4 nuclei	2 nuclei	>4 nuclei
R+
4070A	6.1 × 103	0	0	0.2 ± 0.4	1.2 ± 0.9	0.3 ± 0.5	0.8 ± 1.0
4070A/G541R	2.6 × 103	0	0	0	0	0	0
R−
4070A	6.7 × 103	0	15.8 ± 3.1	15.2 ± 4.2	11.7 ± 1.7	0.6 ± 0.7	12.5 ± 2.0
4070A/G541R	1.9 × 103	0	2.5 ± 1.1	15.2 ± 4.2	1.4 ± 1.1	0.1 ± 0.3	2.1 ± 1.3

^aDNAs encoding the lacz marker and each Env were transiently expressed in D17/gag-pol cells. At 48 h posttransfection, virus was collected and applied to D17 cells. Viral titers are based on the number of D17 cells scoring positive for β-galactosidase expression 48 h postinfection. Values represent the average of two independent determinations.

^bAt 48 h posttransfection of each Env, D17/gag-pol producer cells were UV irradiated and cocultured with rat XC cells. Syncytia were stained with hematoxylin and counted. Values represent the mean ± standard deviation (SD) of 20 fields of 13.86 mm² containing 2 nuclei and >4 nuclei.

^cAt 48 h posttransfection of each env DNA into NIH 3T3 cells, syncytia were stained with hematoxylin and counted. Values represent the mean ± SD of 10 fields of 13.86 mm² of syncytia containing 2 nuclei and >4 nuclei.

G541R affects cell surface expression levels and subsequent viral expression.

The effect of amphotropic residue G541R on the viral Env protein was examined at the cellular level. The viral Env proteins bearing the G541R substitution were transiently expressed and radiolabeled overnight with ³⁵S. Cells were subsequently incubated with a membrane-impermeable biotinylating agent. Cell lysates were divided into two fractions. In the first, the Env proteins were immunoprecipitated with antibodies to the SU protein, yielding the total cellular population. TM protein that is tightly associated with SU can coimmunoprecipitate with SU. The second fraction was immunoprecipitated with anti-SU antibody and then purified on streptavidin-agarose to isolate the surface biotinylated Env protein. This represents the fraction of the Env protein that is on the cell surface.

Figure 3A represents the analysis of comparative fractions from the cell (left panel) and the cell surface (center panel). Envelopes within the chimeric EA6 series that were analyzed included the syncytium-positive N261I/E311V mutant and the syncytium-negative N261I/E311V/G541R mutant. In addition, the 4070A Env in the presence and absence of G541R as well as M-MuLV Env in the presence and absence of R-peptide were analyzed. The level of expression of precursor SU-TM for all the constructs was similar in the total cellular extract. Heterogeneity in the SU corresponding to proteolytic processing and glycosylation was visible with the WT controls M-MuLV and 4070A Env (lanes 6 and 8, respectively). This is supported by the presence of TM (p15E) coimmunoprecipitating in the fraction. For the M-MuLV/R⁻ Env, the corresponding p12E protein was observed (lane 5). In comparison with the control samples, the level of TM protein associated with the EA6 virus (lane 4) as well as the 4070A/G541R virus (lane 7) was greatly decreased. An overexposure of the region containing the TM proteins from the total cellular protein was included on the bottom of the left panel. For the EA6/N261I/E311V virus, a low level of p15E TM protein was visible (lane 2). This correlates with the increased titer and syncytia associated with incorporating these two mutations within EA6. The level of TM was diminished in EA6/N261I/E311V/G541R, relative to that in the N261I/E311V double combination (lane 3). In the parental EA6, a protein product in between the p15E and p12E protein was visible (lane 4). This product is observed as a minor species in the control M-MuLV/R⁺ Env sample (lane 6).

FIG. 3.

Levels of expression of Env's. DNA encoding different Env's was transiently expressed in D17/gag-pol cells. At 24 h posttransfection, cells were metabolically labeled overnight with Tran³⁵S-label. (A) Cells were subsequently incubated with 1 ml of PBS containing 0.5 mg of EZ-Link NHS-SS-biotin for 30 min at room temperature. Biotinylation reactions were quenched by the addition of 10 ml of DMEM containing 10% calf serum. Cells were lysed in 1 ml of PLB and processed as described in Materials and Methods. The total cellular fraction (left) was immunoprecipitated with anti-SU antibody 80S-019. The cell surface fraction (middle) was immunoprecipitated with anti-SU antibody 80S-019, followed by purification on avidin-agarose beads. (B) The viral fraction was isolated following an overnight labeling with Tran³⁵S-label. Culture supernatant was centrifuged at 35,000 rpm in an SW41 rotor for 4 h to concentrate viral particles. Viral particles were resuspended in PLB, and Env was immunoprecipitated with anti-SU antibody 80S-019. The sample corresponding to one-fourth of a 100-mm plate was analyzed by SDS-PAGE.

Analysis of the cell surface-associated fraction proved informative. As predicted, the WT M-MuLV, M-MuLV/R⁻, and 4070A Env proteins were readily detected on the cell surface, and the corresponding TM protein coimmunoprecipitated with SU (Fig. 3A, lanes 14, 13, and 16, respectively). However, incorporation of G541R into the 4070A protein decreased the level of SU on the cell surface, with a marked decrease in the presence of TM (lane 15). For the EA6 molecules, a low level of heterogeneously sized SU proteins was detected with the parental EA6 (lane 12) and EA6 N261V/E311V (lane 10) constructs. The presence of p15E TM associated with EA6/N261V/E311V was maintained on the cell surface (lane 10). Similar to that observed for 4070A, incorporation of G541R into EA6/N261V/E311V decreased the level of surface SU and TM proteins (lane 11). These results are intriguing in light of the fact that the presence of G541R maintained the titer of EA6/N261V/E311V.

Viral SU was analyzed by immunoprecipitation using the same antibody (80S-019) as in the cell surface experiment (Fig. 3B). While appreciable amounts of virus-associated Env were present on the EA6 parental (lane 20) and the EA6/N261I/E311V (lane 18) isolates, little or no Env was detected in the EA6/N261I/E311V/G541R viral isolate (lane 19). Consistent with the lower detection of cell surface expression, there was a decreased level of virus-associated 4070A Env bearing G541R (lane 23) compared to that in the WT 4070A control (lane 24).

Several mechanisms can explain why a TM mutation decreases the level of virion and cell surface-associated SU protein. Processing of the SU-TM precursor protein may be inhibited. However, pulse-chase experiments of both the WT 4070A and the G541R-substituted 4070A Env's indicated the presence of SU 2 h postchase (data not shown). A blockage in the transport to the cell surface can occur. Alternatively, the G541R TM mutation can induce a conformation in SU that alters the recognition by the antibody utilized in the immunoprecipitation. It is possible that several of these pathways are working in concert. The question of differential antibody recognition was further analyzed below.

The TM mutation G541R abolishes recognition of the amphotropic SU C terminus by a C-terminal-specific monoclonal antibody.

Altered cell surface and virus expression levels implied a possible alteration in SU-TM interactions. Figure 4 shows the results of probing Western blots of virus-associated G541R Env's with two different SU antibodies. In the top panel (A), viral preparations of M-MuLV and 4070A were compared to EA6/N261I/E311V/G541R. Samples from the same transient-expression experiment were replicated on a single PAGE, separated, and subsequently visualized with either the SU monoclonal 83A25 (11) or the SU polyclonal 80S-019 (Microbiological Associates). By Western blotting, antibody 83A25 did not recognize either the M-MuLV or EA6 viral SU. This is consistent with recent results in which reduced SU is not detectable on Western blotting using the SU antibody 83A25 (5). Viral SU was detected, however, in all Env's by the SU polyclonal 80S-019. The level of EA6/N261I/E311V/G541R Env detected on viral particles was lower than that of the control M-MuLV or 4070A-MuLV Env. Since the antibody reacted with both the ecotropic and amphotropic SU, these results indicate that the level of EA6 N261I/E311V/G541R on the virus is decreased.

FIG. 4.

Effect of G541R on antibody recognition. Viral Env's were transiently expressed in D17/gag-pol cells. At 48 h posttransfection, viral supernatants were collected, filtered through an acrodisc 0.45-μm-pore-size filter, and centrifuged at 16,000 × g through a 20% sucrose cushion for 45 min. Viral pellets were resuspended in SDS-PAGE sample buffer with 2% β-mercaptoethanol and separated on an SDS-10% PAGE. Proteins were transferred onto PVDF and subsequently visualized by immunostaining. (A) Analysis of N261I/E311V/G541R/EA6 compared to M-MuLV and 4070A-MuLV Env's. The upper left panel denotes immunostaining with SU primary antibody 83A25 (11) followed by secondary antibody goat anti-rat IgG-horseradish peroxidase (Pierce) and chemiluminescence (Pierce); the upper right panel represents immunostaining with SU polyclonal 80S-019 (Microbiological Associates), followed by secondary antibody rabbit anti-goat IgG (Pierce) and chemiluminescence. Lower panels represent analysis of CA protein using primary antibody 79S-287 (Microbiological Associates), secondary antibody rabbit anti-goat IgG (Pierce), and chemiluminescence. (B) Analysis of G541R⁻ 4070A Env's in comparison to WT M-MuLV, R-minus Mo-MuLV, and 4070A-MuLV Env's. The left panel represents results following probing with SU monoclonal 83A25. The Western blot was then incubated with SDS and β-mercaptoethanol in Tris-HCl (pH 6.8) buffer at 55°C for 30 min to remove bound antibodies. The blot was subsequently incubated with primary antibody 80S-019 (right panel). Secondary antibodies and visualization are as noted for panel A.

Viral preparations of M-MuLV and 4070A were compared to 4070A/G541R Env's with and without the R-peptide (Fig. 4B). Similar to the results seen in Fig. 4A, M-MuLV was not recognized by the monoclonal antibody 83A25 by Western analysis, while the WT 4070A Env's in the presence or absence of R-peptide were detected. Interestingly, SU monoclonal 83A25 did not recognize each 4070A Env containing the TM G541R substitution. The same blot was subsequently reprobed with the SU polyclonal antibody 80S-019. Both the WT 4070A and G541R-4070A viral envelopes were recognized, as were the M-MuLV Env controls. Therefore, the absence of reactivity with 83A25 is not due to the absence of protein. Less SU is present in the R-minus G541R-substituted 4070A Env than the R-minus 4070A control. The lower amount of viral SU in the G541R-substituted Env is consistent with expression levels in the biotinylation and virus immunoprecipitation experiments (Fig. 3). In the case of the R⁺ Env's, no difference in expression levels could be detected between the G541R Env and the WT 4070A control. This may be due to the denaturing conditions required for stripping the blot. In all other experiments (total of three), the level of Env bearing G541R in the 4070A TM domain was decreased compared with the WT 4070A isolate. The lack of recognition of all the G541R-substituted Env's by the SU monoclonal antibody 83A25 suggests that G541R alters SU C-terminus conformation. The fact that this is observed after SDS-PAGE indicates the stable conformations of the SU are greatly influenced by the TM protein. Additional experiments of intracellular 4070A Env bearing G541R have indicated that the SU-TM precursor protein could be recognized by 80S-019 and an antibody generated against the R-peptide (a gift of John Elder, Scripps Institute), but it was not recognized by the monoclonal antibody 42-114 (34) (data not shown). This supports the idea that G541R alters the conformation of TM; however, disruption of the epitope recognized by the monoclonal antibody 42-114 cannot be ruled out.

Analysis of G552 in M-MuLV.

4070A G541R is localized within a highly conserved region of the ectodomain of TM, three amino acids N-terminal to the first cysteine residue (Fig. 1). The homologous position within the ecotropic M-MuLV is G552. The ability of G552 to regulate syncytia within M-MuLV was investigated. The ecotropic M-MuLV TM protein (G552) was mutated to an arginine. Results of an XC cocultivation assay following transient expression are summarized in Table 4. Virus titer was determined by the transfer of the lacZ gene following infection into D17/pJET cells. No appreciable effect was observed on either the virus titer or production of syncytia for either the R⁺ or R⁻ M-MuLV Env's bearing the G552R substitution in comparison to results with the wild-type M-MuLV counterparts. The assembly of the SU/TM complexes within amphotropic and ecotropic Env therefore displays subtle differences.

The effects of incorporating G552R were further analyzed within an M-MuLV backbone bearing the N261I and E311V mutations. The N261I and E311V mutations were located within the ecotropic domain of the EA6 chimera. The results presented in Fig. 3 indicated that the presence of these two mutations within EA6 increased the proteolytic processing of the TM protein and the association of SU and TM by coimmunoprecipitation. The two mutations acted cooperatively to increase virus titer (Table 1), kinetics of viral passage (data not shown), and induction of syncytia (Table 1). These mutations were selected in response to a chimeric junction in the SU C terminus. It was therefore of interest whether these stabilizing mutations in the first two-thirds of the ecotropic SU would respond in the presence of the equivalent G552R mutation in the context of an intact ecotropic SU C terminus and TM. M-MuLV Env bearing the substitutions N261V/E311V/G552R was generated and analyzed for expression for surface as well as viral protein expression (Fig. 5). Surprisingly, the presence of G552R within the M-MuLV N261V/E311V construct had the opposite effect as within an amphotropic TM/SU C terminus. The G552R mutation increased the surface expression of SU and stabilized the coimmunoprecipitation of TM (Fig. 5, Surface, lane 2). This cell surface-expressed SU/TM complex, however, was not efficiently packaged within virion particles (Fig. 5, Total Supernatant, lane 2). Analysis of viral proteins released into the medium indicated a small but detectable decrease in SU proteins. In addition, the level of TM protein that coimmunoprecipitated with SU was also decreased. Analysis of virus titers indicated that the virus titer of M-MuLV N261V/E311V/G552R was 3.7- to 12-fold decreased from that of the WT M-MuLV virus (Table 4).

FIG. 5.

Effect of homologous substitution in the M-MuLV TM protein. DNA encoding different Env's was transiently expressed in D17/gag-pol cells. At 24 h posttransfection, cells were metabolically labeled overnight with Tran³⁵S-label. Cells were subsequently incubated with 1 ml of PBS containing 0.5 mg of EZ-Link NHS-SS-biotin for 30 min at room temperature. Biotinylation reactions were quenched by the addition of 10 ml of DMEM containing 10% calf serum. Cells were lysed in 1 ml of PLB and processed as described in Materials and Methods. The total cellular fraction (left) was immunoprecipitated with anti-SU antibody 80S-019. The cell surface fraction (right) was immunoprecipitated with anti-SU antibody 80S-019, followed by purification on avidin-agarose beads. Molecular mass markers were obtained from Bio-Rad. Lane 1, mock; lane 2, M-MuLV N261I/E311V/G552R; lane 3, M-MuLV; lane 4, M-MuLV/R⁻.

These results indicate that the Env protein with increased SU-TM interaction was not efficiently packaged into virions, and virions containing this heterodimer had a decreased virus titer. In contrast, when G541R was expressed within 4070A and EA6 chimeras bearing the amphotropic TM, the association of SU and TM was decreased, the levels of heterodimer on the cell and viral surface were diminished, and the virus titer was improved. These results indicate that the Gly residue within the ectodomain of TM plays a critical role in the association with SU and the regulation of membrane fusion.

DISCUSSION

The amphotropic TM mutation G541R was isolated in multiple infectious viral populations in combination with other second-site changes upon transient expression within a replication-competent provirus (28). The effects of the G541R mutation were pleiotropic. Within the EA6 chimera and WT 4070A Env, expression of G541R resulted in decreased cell surface expression of SU/TM protein, decreased coimmunoprecipitation of TM with SU, and altered antibody recognition of the SU protein by monoclonal antibodies. Within the M-MuLV Env, mutation of the equivalent Gly, G552R, resulted in stabilization of the SU-TM heterodimer on the cell surface. These results indicate this Gly plays a role both in the regulation of membrane fusion and in the association of TM with SU.

Cell surface biotinylation experiments revealed that within EA6 and WT 4070A, G541R altered the levels of SU and TM present on the cell and viral surfaces. The low level of Env bearing the G541R substitution was not attributable to shedding into the culture medium. Comparison of total SU in the medium versus virus-associated SU after pelleting through a sucrose gradient did not reveal release of 4070A/G541R versus the WT 4070A Env (data not shown). Mutational studies of various amino acids within the homologous ectodomain regions of TM in HIV-1 and feline leukemia virus led to the phenotype of decreased levels of cell-cell fusion events (6, 26). For feline leukemia virus, this phenotype was attributed to inhibition of protein processing. For HIV, mutation of the equivalent position, W596, resulted in loss of gp120 association.

Cells with decreased levels of G541R-Env at the cell surface were limited for cell-cell fusion relative to the controls. The effect of Env density on cell-cell fusion was previously studied with influenza virus (8, 9), where low hemagglutinin cell surface density correlated with a longer lag period before the onset of cell-cell fusion. For HIV-1 (20), the level of cellular expression of the Env proteins (gp120/gp41) affects the efficiency of cell-cell fusion events (20) and suggests that a conserved mechanism to limit cell-cell fusion events among retroviruses is via the level of Env expression. Since syncytium formation results in host cell death, this strategy of limiting cell-cell fusion events would ultimately be beneficial for the virus in maintaining a symbiotic relationship with its host.

Although the number of cell-cell fusion events is dictated in part by the cell surface density of the Env protein, virus-cell fusion is not adversely affected. For M-MuLV, efficient infection requires only a minimal level of Env protein on the virus (1). Viral isolates such as EA6 N261I/E331V/G541R, with decreased cell-cell fusion, maintained high virus titer. G541R was additionally isolated in D17 cell-derived 4070A in combination with G100R within VRA. G100R enhances binding and entry kinetics of amphotropic 4070A viral particles to multiple cell types with varying levels of Pit2 receptor (21). If stronger binding signifies better signaling, one strategy to restore the spatial and temporal control of fusion would be to decrease the level of Env protein, through the acquisition of G541R. The net effect of decreased surface expression in response to stronger signaling has led to MuLV variants that preserve the intrinsic nature of the fusion process.

Residue G541 is the first glycine in a double-Gly motif conserved among retroviral TM proteins. Based on homology modeling with the M-MuLV TM ectodomain crystal structure, G541 is N-terminal to the TM intramolecular disulfide bond and helps to form a reverse turn between the coiled-coil motif and a short alpha helix (12). Covalent SU-TM interactions are mediated through a third TM cysteine with the C terminus of SU (35, 42). The TM disulfide-bonded loop-chain reversal region of HIV is a critical SU contact site (6, 26). Immunological data presented in this paper suggest that G541R interacts with the C terminus of SU near or at the epitope of the SU monoclonal antibody 83A25. Chimera and polymorphism mapping localize the 83A25 epitope to a position inclusive of SU C-terminal amino acids 416 to 449 (M-MuLV numbering, including the signal sequence) (5, 40), while saturation mutagenesis studies (40) further localize its epitope to M-MuLV Env amino acids 428 to 431. G541R was also isolated in chimeric EA7 Env viral populations (Fig. 1) that bear a C-terminal SU junction close in primary sequence to the proposed monoclonal antibody 83A25 epitope (5, 40). G541R restored viral viability compared to the parental EA7 virus and additionally suppressed a unique cold-sensitive phenotype when expressed in cis with changes in either the N or C terminus of SU (28). In total, these observations suggest amino acids within the ectodomain of TM, particularly glycine 541, are critical determinants of the stability of SU-TM interactions (14) and conformation of the SU C terminus.

The C terminus of SU plays an important role in the stability of SU-TM interactions (14, 33, 36) and the achievement of virus-cell fusion (3, 4, 16, 18, 42). Rearrangement of the SU-TM disulfide bond is proposed to facilitate structural conformational changes required for fusion (42). A viral entry model proposes that alternate contacts between SU and TM form upon receptor binding (3) which lead to fusion, such as that between the SU N terminus and a conserved disulfide loop in the C terminus of SU (16). By altering the conformation of the C terminus, the G541R mutation may alter the critical SU N- and C-terminal interactions and lead to delayed fusion kinetics.

Mutations to the cysteine-proximal region of the human T-cell leukemia virus type 1 TM protein resulted in either unchanged or increased levels of Env with a concomitant decreased level of cell-cell fusion (25) and led to a proposal that the double-Gly motif regulates the transition of TM from the prehairpin stage to the six-helix intermediate. In an analogous role, MuLV G541R could be acting as a control point for the amphotropic Env to regulate the conversion of TM from the prehairpin intermediate to the six-helix bundle.

Despite their C-terminal SU and TM homologies, the ecotropic and amphotropic MuLV Env's have adapted subtle distinct properties to bind their individual receptors and transmit the post-receptor-binding signals. Amphotropic MuLV transduction is dependent on high levels of receptor (24, 32), and the 4070A PRR has less fusogenic potential relative to that of M-MuLV (17). The substitution of the analogous G552R in the WT M-MuLV has no effect on cell-cell fusion events, suggesting the mutation was specifically adapted for viruses with the amphotropic C terminus. Interestingly, passage of a converse chimera containing the 4070A RBD and the ecotropic SU C terminus and TM (IVAE4 [22]) resulted in the selection of the G552E mutation in addition to G100R and L435I (23). The M-MuLV G552R substitution (G541R in 4070A), along with K550 and K558, generates a large positive surface area at the top of the hairpin structure within the TM ectodomain (12). The acidic G552E residue would disrupt the positive charge surface, suggesting that this charge cluster may play a regulatory role. The proximity of L435 to junction 7 and the proposed recognition site for SU monoclonal antibody 83A25 further supports the concept of an interaction between these localized regions of SU and TM. By mutagenesis and subsequent compensatory mutations acquired by in vivo evolution, the intrinsic Env characteristics required for optimal signal transmission (47) and entry of 4070A and M-MuLV were maintained.