Linkage of Reduced Receptor Affinity and Superinfection to Pathogenesis of TR1.3 Murine Leukemia Virus

Abstract

TR1.3 is a Friend murine leukemia virus (MLV) that induces selective syncytium induction (SI) of brain capillary endothelial cells (BCEC), intracerebral hemorrhage, and death. Syncytium induction by TR1.3 has been mapped to a single tryptophan-to-glycine conversion at position 102 of the envelope glycoprotein (Env102). The mechanism of SI by TR1.3 was examined here in comparison to the non-syncytium-inducing, nonpathogenic MLV FB29, which displays an identical BCEC tropism. Envelope protein expression and stability on both infected cells and viral particles were not statistically different for TR1.3 and FB29. However, affinity measurements derived using purified envelope receptor binding domain (RBD) revealed a reduction of >1 log in the K_D of TR1.3 RBD relative to FB29 RBD. Whole-virus particles pseudotyped with TR1.3 Env similarly displayed a markedly reduced binding avidity compared to FB29-pseudotyped viral particles. Lastly, decreased receptor affinity of TR1.3 Env correlated with the failure to block superinfection following acute and chronic infection by TR1.3. These results definitively show that acquisition of a SI phenotype can be directly linked to amino acid changes in retroviral Env that decrease receptor affinity, thereby emphasizing the importance of events downstream of receptor binding in the cell fusion process and pathology.

The ecotropic Friend murine leukemia virus (MLV) TR1.3 was previously defined as an acutely cytopathic, syncytium-inducing (SI) variant of the noncytopathic, non-syncytium-inducing (NSI) MLV FB29. When inoculated into neonatal BALB/c mice, TR1.3 infected and caused syncytium formation in brain capillary endothelial cells (34-36). These cytopathic effects paralleled a breakdown of the blood brain barrier and hemorrhagic stroke formation. The molecular basis of both the SI phenotype and in vivo neurologic disease was previously mapped to a tryptophan (W)-to-glycine (G) substitution at amino acid position Env 102 (37). Introduction of the W102G substitution into the NSI FB29 molecular clone yielded an SI pathogenic phenotype that was identical to that seen with TR1.3. The mechanism(s) whereby mutations at Env 102 mediate SI and disease is unknown.

The MLV Env consists of surface and transmembrane subdomains. Amino acids 1 to 236 of SU encompass the putative receptor binding domain (RBD) (10). Analysis of the RBD crystal structure indicates that amino acid 102 lies at the base of the predicted receptor binding pocket (13). Interaction between RBD and the ecotropic receptor murine cationic amino acid transporter 1 (mCAT-1) is essential for MLV membrane fusion and entry and for syncytium formation (1, 22, 53). Previous studies assessed the impact of multiple amino acid changes in Env on expression, binding, and viral entry in MLV with related genetic backgrounds (11, 33). These studies showed that nonconservative substitutions of glycine or threonine for tryptophan at Env 102 altered receptor binding and transduction efficiency. However, unlike with TR1.3 and W102G, Env substitutions on these backgrounds did not result in the SI phenotype (11, 33).

Following MLV infection, Env downmodulates the level of available receptor on the cell surface (16, 31). This process, known as superinfection interference, prevents a virally infected cell from undergoing additional rounds of infection by the same or related viruses that utilize the same cellular receptor (44, 50, 55). In rare circumstances, failure of MLV Env to establish superinfection interference is linked to either low Env expression levels and/or diminished mCAT-1 binding (3, 4, 49).

Important pathological consequences can result from retroviral superinfection. Superinfection can lead to the accumulation of unintegrated linear DNA in the cytoplasm (51); by virtue of the similarity between unintegrated viral DNA and damaged DNA, this may trigger apoptosis induction in superinfected cells (9, 51). Failure to block superinfection has been linked to the in vivo cytotoxicity of mink cell focus-forming virus M13 and the neuropathogenic MLV Moloney ts1 (51, 56).

Murine leukemia viruses are facile tools for investigating the mechanisms of retroviral pathogenesis (26). Several correlates of increased fusion activity have been attributed to Env including increased Env expression (24, 25, 38, 45), increased receptor affinity (12, 15, 54), and decreased stability of the association between Env and the viral core (27-30). We describe here a link between diminished receptor affinity, superinfection, and selective SI in MLV TR1.3.

MATERIALS AND METHODS

Cells.

The cells used in this study were SC-1 (American Type Culture Collection, Manassas, Va.) and 293T cells (American Type Culture Collection). Cell lines were maintained in Dulbecco's modified Eagle medium (DMEM) supplemented with 1% penicillin-streptomycin, 10 mM glutamine, and 10% fetal bovine serum (Invitrogen, Carlsbad, CA). 293T cells stably expressing hemagglutinin-tagged mCAT-1 were individually cloned after selection in 20-μg/ml blasticidin HCl and maintained in 10-μg/ml blasticidin HCl-supplemented medium (Invitrogen). A23 cells and the 99 cell lines from the T31 radiation hybrid panel were a generous gift of Susan Ross (University of Pennsylvania). With the exception of the A23 cells (maintained in DMEM), these cells were maintained in medium supplemented with hypoxanthine-aminopterin-thymidine (Sigma, St. Louis, MO).

Plasmids.

The MCHA plasmid containing a C-terminal hemagglutinin-tagged mCAT-1 gene was generated based on previously published methods (11). The vector pBacPAK8 (BD Biosciences) was used for construction of the RBD expression vectors pBacPAKFB and pBacPAKTR containing the RBD sequence (amino acids 1 to 236) in frame with a six-His tag, followed by a stop codon using the primers 5′GCGGGATCCAGCAGCACCCGGCTCCAGC and 3′CGCGAATTCCTAGTGGTGGTGGTGGTGGTGGAAAGTAAGCCCTGGGTC. Sequence fidelity was confirmed by DNA sequencing.

Production and purification of RBD protein.

Using pBacPAKFB and pBacPAKTR constructs, we followed previously published methodology to generate baculovirus clones expressing RBD proteins (47). Clones were selected for high levels of expression on the basis of gp70 antibody reactivity on Western blots. Two successive rounds of virus cloning yielded a virus supernatant suitable for expansion into viral stock. Production of RBD was initiated by infection of 4 × 10⁹ Sf9 cells in 1 liter of supplemented Grace's medium at a multiplicity of infection (MOI) of 4. Cell viability was monitored daily, and protein harvest commenced when viability declined to 85%. Following dialysis, RBD-containing retentate was incubated overnight at 4°C with 0.5 ml of Talon Superflow metal affinity resin (BD Biosciences, Franklin Lakes, NJ). RBD was eluted using an imidazole gradient (10 to 250 mM). Fractions were screened for the presence of RBD by Western blot analysis and Coomassie staining. Molecular mass cutoff filter devices (15,000 kDa) were used to remove imidazole from the RBD preps (Fisher Scientific, Hampton, NH).

RBD binding assay.

RBD proteins were labeled using carrier-free Na¹²⁵I (Perkin-Elmer, Wellesley, MA) and Iodobeads (Pierce) per the manufacturer's instructions. Typically, 25 μg of RBD was labeled using 250 μCi of Na¹²⁵I. Labeled protein was separated from free ¹²⁵I with NAP 10 columns preloaded with Sephadex G-25 (Amersham Biosciences, Piscataway, N.J.). For equilibrium dissociation constant (K_D) determination, MCHA2 cells and 293T cells were seeded in polystyrene tubes at 2 × 10⁵ to 3 × 10⁵ cells/tube in 100 μl of binding buffer (50 mM HEPES [pH 7.4]-150 mM NaCl-1 mM CaCl₂-5 mM MgCl₂). Dilutions of RBD were made in binding buffer and added in equal volumes to tubes containing MCHA2 cells and 293T cells. After 1 h of binding at 24°C, reactions were stopped by the addition of 20 volumes of ice-cold assay buffer. Following three washes in ice-cold binding buffer, cells were lysed, and bound ¹²⁵I was measured with a gamma counter. Specific binding in saturation analysis was determined by subtraction of background (¹²⁵I bound to 293T cells at each RBD concentration) from the ¹²⁵I bound to MCHA2 cells at the same concentration. From the raw data, K_D measurements and binding curves were generated using GraphPad Prism 4.0 software nonlinear regression analysis (GraphPad Software, San Diego, CA).

For the detection of surface receptor using ¹²⁵I-labeled FB29 RBD, SC-1 cells were first seeded in two fluorescence-activated cell sorter (FACS) tubes per sample. Cells were spun down for 5 min at 1,200 rpm and then resuspended in either 500 μl of pH 3.0 citrate buffer (400 mM sodium citrate, 10 mM KCl, 135 mM NaCl) or 500 μl of phosphate-buffered saline (PBS). After 60 s, 8 volumes of complete DMEM supplemented with 75 mM HEPES (pH 7.5) was added to each tube to restore normal pH. Cells were then spun down and resuspended at a concentration of 10⁶ cells/ml in binding buffer. For each sample, 100 μl of cells was added to FACS tubes in triplicate. Radiolabeled FB29 RBD was diluted in binding buffer and added to each tube to give a final volume of 150 μl/sample and a final concentration of 21.6 nM. After a 30-min incubation at room temperature, the samples were washed five times in ice-cold wash buffer (PBS supplemented with 1% bovine serum albumin). Samples were then lysed in wash buffer supplemented with 1% Triton X-100 and read in a gamma counter.

Virus production.

FB29, TR1.3, and W102G viruses were made as previously described (34) and purified by ultracentrifugation on a 20% sucrose cushion at 19,500 rpm for 2 h (6). Virus titers were determined by a modified XC cell plaque assay (43). For use in generating Env-pseudotyped viruses, molecular clones of FB29 (48), TR1.3, and W102G viruses were digested with AscI and BsaAI to isolate the entire Env gene. DNA fragments were cloned into pcDNAI at the EcoRV site (Invitrogen). Correct orientation was confirmed by diagnostic restriction enzyme digestions. Pseudotyped viruses were generated by triple transfection as described previously (2, 33). For viral interference assays, 10⁵ cells were plated and infected overnight with pseudotyped viruses at an MOI of 5 in 8-μg/ml Polybrene. Forty-eight hours postinfection, an X-galactosidase assay was performed and positive beta-galactosidase units were counted (40).

Superinfection analysis.

SC-1 cells (10⁶) were seeded into a T150 cell culture flask (Corning, Cambridge, MA) in 24 ml of DMEM and incubated with 1.5 × 10⁷ PFU of primary infecting virus (MOI = 15) in 8 μg of Polybrene (Sigma)/ml. Cells were incubated with primary virus for 24 h to 1 month, as indicated in Results, and then the medium was replaced by 24 ml of DMEM containing 1.5 × 10⁷ PFU of secondary virus (MOI = 15). Biological cloning of superinfected SC-1 cells was achieved by first diluting cells to the concentration of 8 cells per 1 ml. A total of 100 μl of this cell suspension was seeded into 96-well plates (Corning, Cambridge, MA) for a final cell concentration of 0.8 cells per well and incubated for up to 2 months in culture. Single-cell clones were expanded into 24-well plates and then into T75 cell culture flasks (Corning, Cambridge, MA). Cultures were maintained until 100% confluent, trypsinized, and washed in 1× PBS (Invitrogen). One half of the cells was used for DNA isolation, the other half was used to repeat a second round of biological cloning. After a second round of biological cloning, DNA was isolated (QIAamp Tissue kit; QIAGEN, Santa Clarita, CA) for PCR analysis.

Amplification of viral genes by PCR.

Cellular DNA (500 ng) was amplified through 30 cycles of PCR with 5 U of Taq (Promega, Madison, WI) and 300 mM of sense and antisense primers in PCR buffer consisting of 1.5 mM MgCl₂ (Promega, Madison, WI), 0.2 mM deoxynucleoside triphosphates (Pharmacia, Piscataway, NJ), and 2.0% formamide in case of TR1.3 amplification. Each cycle consisted of denaturation at 94°C for 30 s, annealing at 60°C for 1 min, elongation at 72°C for 45 s for TR1.3 or FB29 amplification, and denaturation at 94°C for 30 s. Reactions were performed with following primers. For TR1.3, the sense TGGAAGCCCTCCTCTC and antisense CTTCGGACAGGGTCAA primers were used; for FB29, the sense GCCCCCTATTCCTCGC and antisense, CTTCGGACAGGGTCAG primers were used. The PCR was highly sensitive, allowing the detection of 0.5 ng of specific DNA (data not shown). This represents the detection of viral DNA present in 1 cell in 187,500 total cells used in each superinfection experiment.

Immunofluorescence and immunoblotting assays.

A total of 2 × 10⁵ cells were harvested at each time point 12, 24, and 48 h postinfection and immunostained for gp70 expression. FACS analysis of gp70 expression using gp70-specific goat serum (American Type Culture Collection) has been described previously (19). For immunoblotting, virally infected cell lysates were harvested at 24, 48, and 72 h postinfection in 50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 0.02% sodium azide, 100-μg/ml phenylmethylsulfonyl fluoride, 1-μg/ml aprotinin, 1-μg/ml leupeptin, and 1% Triton X-100. Purified viral particles were resuspended in DMEM, lysed, and mixed with sodium dodecyl sulfate-polyacrylamide gel electrophoresis sample buffer before sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis. After transfer to nitrocellulose membrane, blots were probed with goat anti-gp70 polyclonal antibody (1:2,000) for 1 h at room temperature. For quantitative Western blotting, blots were incubated in protein A conjugated to ¹²⁵I diluted to 0.1 μCi in TBST (1 M Tris [pH 7.5] and 9% NaCl) with 0.1% Tween-20 and 1% bovine serum albumin for 1 h at room temperature. Gels were then washed in TBST (three washes for 10 min each) and dried. Quantitation of blots was performed by STORM image quantification.

RESULTS

Env gp70 expression and stability.

Syncytium formation and superinfection are controlled by the levels of both Env expression and receptor availability. Flow cytometry and quantitative Western blot analysis were used here to determine the levels of Env expression on FB29- and TR1.3-infected SC-1 cells. Infected SC-1 cells (MOI = 15) were harvested at 12, 24, and 48 h postinfection, immunostained for Env gp70 expression, and analyzed by FACS. As shown in Fig. 1A, the surface expression of gp70 was lower on TR1.3-infected cells than on FB29-infected cells at 12 and 24 h postinfection; however, at 48 h postinfection, the level of gp70 was indistinguishable on FB29 and TR1.3-infected cells. In agreement with previous results (8), TR1.3-induced syncytium formation of SC-1 cells was first observed at 24 h and increased at 48 h postinfection (data not shown). To control for variation in the extent of virus infection by different MLV isolates, Env levels were also determined by comparing the ratio of gp70 to p30 (Gag) expression. As shown in Fig. 1B, quantitative Western blots of whole-cell lysates from cultures infected with FB29, TR1.3, or W102G showed no significant difference in the gp70:p30 ratio among MLV isolates from 24 to 72 h postinfection.

FIG. 1.

Time course of gp70 expression on FB29- and TR1.3-infected SC-1 cells. (A) Flow cytometric analysis of gp70 surface expression. SC-1 cells were infected with FB29 or TR1.3 viruses at an MOI of 15 and harvested at 12, 24, and 48 h postinfection. The cells were stained using a primary polyclonal antiserum directed against MLV gp70 and a fluorescein isothiocyanate-conjugated secondary antibody. Representative FACS plots are shown for 12-h (top left), 24-h (top right), and 48-h (bottom left) time points. Solid line, FB29-infected cells; dotted line, TR1.3-infected cells; shaded curve, uninfected cells. (B) Quantitative immunoblot determination of gp70:p30 expression in cell lysates. SC-1 cells were infected as described above, and cell lysates were harvested at the times indicated. The ratio of gp70 to p30 was calculated by Western blot detection and STORM image quantification. Each ratio represents three independent experiments, shown as the mean and standard error of the mean. No statistically significant differences were found between the samples at any time, as determined by Student's t test (P > 0.05).

To determine the stability of gp70 association within MLV particles, particle integrity was examined following purification by ultracentrifugation through a 20% sucrose gradient, which was previously shown to disrupt unstable Env proteins (29, 30). As shown in Fig. 2, the ratio of gp70 to p30 was not statistically different between FB29, TR1.3, and W102G isolates following this procedure.

FIG. 2.

Quantitative immunoblot analysis of gp70:p30 expression on virus. (A) Immunoblot of gp70 and p30 expression in FB29, TR1.3, and W102G virus particles. Shown is a representative blot of viral particles purified by ultracentrifugation through a 20% sucrose cushion. Viral proteins were detected with polyclonal antibodies and quantification of bound antibody was achieved using ¹²⁵I-conjugated protein A. (B) STORM image quantification of the data shown in panel A was used to determine the ratio of gp70:p30 in immunoblots. These data are representative of six independent experiments. The ratios were not statistically different as determined by the unpaired Student's t test, (P > 0.05).

MLV RBD affinity for mCAT-1.

Alterations in receptor affinity have previously been associated with phenotypic changes to retrovirus-induced syncytium formation and superinfection interference (3, 4, 12). The binding constants of TR1.3 and FB29 Env to mCAT-1 were determined using purified Env RBD (amino acids 1 to 236) prepared in a baculovirus expression system. A histidine tag was added to the C terminus of each RBD to enable purification. In each instance, the RBD eluted from affinity columns as a doublet of approximately 32 to 34 kDa and was free of significant protein contaminants, as shown by a representative Gelcode Blue protein staining of FB29 RBD (data not shown).Treatment of RBD with PNGase F prior to gel loading yielded a single Env band of approximately 30 kDa, indicating that the doublet represented differential N-linked glycosylation of Env (data not shown). Western blot analysis using gp70-specific antibody confirmed the integrity of each RBD (data not shown).

The purified RBD was radiolabeled using carrier-free Na¹²⁵I and Iodobeads to allow for quantitative analysis of receptor binding. These studies were performed using 293T cells that stably express a high level of mCAT-1 (MCHA2 cell line). Both saturation binding assays and cold-competition assays were employed to make K_D determinations for each RBD. Figure 3A and B show representative saturation binding experiments for FB29 and TR1.3 RBD, respectively. Various concentrations of radiolabeled RBD were added to 2 × 10⁵ MCHA-2 cells or to an equal number of receptor-negative 293T control cells. After incubation at room temperature for 60 min, reactions were stopped by the addition of 20 volumes of ice-cold assay buffer. Reaction mixtures were then subjected to repeated washes, followed by cell lysis and detection of bound radioligand in a gamma counter. Specific binding of each RBD was used to calculate K_D by nonlinear regression analysis. As shown in Table 1, the binding affinity of TR1.3 RBD was 14-fold lower than that of FB29 RBD.

FIG. 3.

K_D determination for FB29 and TR1.3 RBD. Results are presented as the number of ¹²⁵I-labeled RBD counts per minute specifically bound to MCHA-2 cells after subtraction of background binding to receptor-negative 293T cells. (A and B) Comparison of saturation binding of FB29 and TR1.3 RBD, respectively. (C and D) Comparison of binding of FB29 and TR1.3 RBD, respectively, in the presence of a concentration gradient of homotypic, unlabeled RBD.

TABLE 1.

Summary of RBD binding experiments

Expt and virus	KD (nM)a
Saturation binding
FB29	4.22 ± 1.08
TR1.3	57.37 ± 26.35
Cold competition
FB29	1.04 ± 0.09
TR1.3	20.45 ± 0.17

^aCalculations were made using GraphPad Prism analysis software and are based on the average of at least two independent experiments.

To complement these studies, K_D measurements were also made through the use of cold competition experiments. In these experiments, radiolabeled RBD was used at a concentration approximately equal to the K_D, as determined through saturation binding experiments. As shown in Fig. 3C and D, competition from increasing concentrations of unlabeled, homotypic RBD resulted in the loss of labeled RBD bound to the target cells. The results of cold competition analysis mirrored those seen in saturation binding: the binding affinity of TR1.3 RBD was 20-fold lower than that of FB29 RBD.

Analysis of virus particle binding.

Following the determination that monomeric TR1.3 RBD had a lower binding affinity for mCAT-1 than did FB29, studies were undertaken to examine whether this difference would impact the multivalent interactions of intact virions with receptor-expressing cells. A modified FACS-based virus-to-cell binding assay was utilized to evaluate these events. Virus particles were purified through a sucrose cushion, and then particle concentration was determined by electron microscopy and standardized by reverse transcriptase activity. Virus binding to cells was then measured by anti-gp70 immunofluorescence and FACS analysis. As shown in Fig. 4, the relative binding avidity of FB29 was markedly greater than either TR1.3 or W102G on SC-1 cells. Similar results were obtained in binding experiments using NIH 3T3 cells or 293T cells transfected with mCAT-1, and there was no detectable difference between TR1.3, W102G, and FB29 whole-virus binding to 293T cells which lack mCAT-1 expression (data not shown). Although it was not possible to calculate binding constants in these experiments because saturation binding of virus particles could not be achieved, these results reinforce the RBD analysis shown above.

FIG. 4.

Concentration-dependent binding of MLV, FB29, TR1.3, and W102G to SC-1 cells. Whole-virus binding was assayed over a range of virus concentrations, as determined by reverse transcriptase activity. FB29 (squares), TR1.3 (circles), and W102G (diamonds) virus binding to SC-1 cells (counts per minute) was carried out at 37°C for 30 min and measured by anti-gp70 immunofluorescence with a FACScan.

Receptor availability on TR1.3-infected cells.

The diminution of available surface receptor by Env is an important component in controlling retrovirus cytopathology and syncytium formation. In this regard, the reduction in relative binding of TR1.3 for mCAT-1 could attenuate the ability of Env to modulate receptor levels following infection. To investigate this possibility, SC-1 cells were first infected with replication-competent FB29, TR1.3, or W102G at a saturating MOI and passaged a minimum of five times. Direct measurements of available surface receptor were then made through binding of high-affinity radiolabeled FB29 RBD to each cell line. As shown in Fig. 5, uninfected SC-1 cells possessed 68,323 receptor molecules per cell, assuming a 1:1 binding stoichiometry. Identical analysis of SC-1 cells infected with FB29, TR1.3, or W102G demonstrated that all of these chronically infected cell lines displayed a significantly reduced capacity for binding FB29 RBD. Indeed, the binding to chronically infected cell lines was indistinguishable from that of receptor-negative 293T cells. Because of the low levels of receptor, we were unable to statistically distinguish differences in receptor availability between FB29, TR1.3, and W102G chronically infected cell cultures.

FIG. 5.

Receptor availability in uninfected and MLV chronically infected cells. Receptor availability was determined by measuring the binding of purified ¹²⁵I-labeled FB29 RBD at saturating conditions to cells as indicated. The data shown indicate the number of receptors per cell, based on the specific activity of bound RBD. The asterisk indicates a statistically significant (P < 0.05) difference in binding between infected and uninfected cells by the unpaired Student t test. These data are representative of four independent experiments.

Use of mCAT-1 receptor by FB29 and TR1.3 MLV.

Previous interference studies have shown that mCAT-1 functions as the sole ecotropic MLV receptor. To verify the absence of alternate receptors orentry factors for TR1.3, the T31 mouse-hamster radiation hybrid panel was used to map the chromosomal determinants of entry in the mouse genome independently for TR1.3 and FB29 (17, 41, 42). Each cell line within the panel contains the full hamster genome and approximately 25% of the mouse genome. The composition of mouse genomic fragments present within each cell line has been determined to allow for linkage analysis of phenotypic data (18, 32, 42, 52). Pseudotyped virus incorporating FB29 or TR1.3 Env with a beta-galactosidase reporter gene was used to transduce each cell line in the panel. Cell lines were scored positive or negative based on the presence of colonies following X-galactosidase staining. Neither FB29 nor TR1.3 pseudotyped virus was capable of transducing the parental hamster cell line A23, and the pattern of transduction of FB29 and TR1.3 pseudotyped virus within the hybrid panel was identical. By scoring each cell line with a 1 for positive, 0 for negative, and 2 for not applicable, the following data set was generated: 11011001010011010111000000000100000000011010010010010 0000110 0001002100100011000110000000000001000100. These data were then analyzed by Jackson Laboratories for linkage analysis (http://www.jax.org/resources/documents/cmdata/). Thesedata returned a logarithm of odds scores that linked transduction to expressed sequence tags AI447493 and M26687, which encode sequences from the mCAT-1 gene. Thus, as reported for other MLVs, these results indicate that TR1.3 and FB29 Env exclusively use the mCAT-1 receptor for transduction in mouse cells.

Superinfection susceptibility following TR1.3 infection.

Although the results shown in the previous sections failed to distinguish differences in receptor availability following chronic infection of SC-1 cells by FB29 or TR1.3 MLV, radioligand binding assays may have a threshold of detection which obscures subtle differences in receptor availability. The superinfection interference assay provides an extremely sensitive and perhaps more biologically relevant method for monitoring receptor availability by using productive infection as a measure of available surface receptor.

Superinfection by FB29 and TR1.3 was evaluated by virus-specific primers designed to distinguish unique nucleotide differences within the Env genes of these viruses. As shown in Fig. 6A, TR1.3-specific primers amplified a 604-bp fragment specifically from TR1.3- but not FB29-infected SC-1 genomic DNA (lanes 1 and 2); FB29-specific primers amplified a 194-bp fragment specifically from FB29- but not TR1.3-infected SC-1 genomic DNA (lanes 3 and 4); and neither primer set amplified genomic DNA isolated from uninfected SC-1 cells (lanes 5 and 6).

FIG. 6.

Superinfection in TR1.3- and FB29-infected cell cultures. (A) Specificity of primers used for PCRs. DNA isolated from TR1.3-infected (TR), FB29-infected (FB), or uninfected (SC) cells was incubated with PCR primer specifically directed to either TR or FB virus DNA. (B) Superinfection of TR1.3 chronically infected SC-1 cells 30 days postinfection. DNA from TR1.3-infected cells (30 days postinfection) incubated with FB29 (TR/FB) and FB29-infected cells (30 days postinfection) incubated with TR1.3 (FB/TR) were analyzed for superinfection by PCR analysis at 48 h using TR1.3- and FB29-specific primers. (C) Superinfection analysis of TR1.3-infected clonal lines. Representative samples from one (clones 5, 6, 8, and 11) or two (clones 5.1, 5.3, 8.4, and 11.6) rounds of biological cloning of TR1.3-infected SC-1 cells secondarily incubated with FB29. DNA from each population was analyzed by PCR using FB29 and TR1.3-specific primers.

Superinfection by TR1.3 and FB29 was examined using these primers over a 24-h to 1-month time course of FB29 or TR1.3 infection. To establish primary infection, SC-1 cells were first incubated with either FB29 or TR1.3 at an MOI of 15. Superinfection challenge was then conducted at 24 h, 7 days, 14 days, or 30 days by the addition of TR1.3 virus to FB29-infected cultures or reciprocally FB29 virus to TR1.3-infected cultures. Analysis of these cultures by PCR, as described above, yielded identical results at each time point of primary infection and secondary challenge; FB29 was able to superinfect TR1.3-infected cultures at all time points tested, while TR1.3 was unable to superinfect FB29-infected cultures at any time point. A representative example of these results, showing FB29 superinfection after 30 days of TR1.3 chronic infection, is presented in Fig. 6B. Both TR1.3 (Fig. 6B, lane 1) and FB29 (lane 2) DNA was detected in TR1.3 cultures secondarily incubated with FB29. In contrast, only FB29 DNA was detected in FB29 cultures secondarily incubated with TR1.3 (lanes 3 and 4). To eliminate the possibility that superinfection of TR1.3 cultures by FB29 was the result of infection of parental cells that escaped primary TR1.3 infection, secondary challenge and PCR analysis were repeated with cells obtained following one or two rounds of biological cloning from TR1.3-infected primary cultures. Representative results from eight separate TR1.3-infected clonal isolates, shown in Fig. 6C, demonstrated that FB29 superinfection of TR1.3-infected cells was uniformly observed.

Superinfection frequency and specificity.

A second viral interference assay using pseudotyped particles was next used to quantify the frequency and range of superinfection susceptibility. FB29 or TR1.3 Env-pseudotyped virus was evaluated for superinfection of SC-1 cells chronically infected (five passages) with either FB29 or TR1.3 or W102G. This analysis complements the PCR-based methodology described above in that one may additionally evaluate superinfection by the primary virus, i.e., FB29 superinfection of FB29 primary infected cells and/or TR1.3 superinfection of TR1.3 primary infected cells. As shown in Table 2, FB29 chronically infected cells were not superinfected by FB29 or TR1.3 Env-pseudotyped virus. Conversely, TR1.3 and W102G chronically infected cells were superinfected by FB29 Env-pseudotyped viruses, although they were not superinfected by TR1.3 pseudotyped viruses.

TABLE 2.

Viral interference assay monitoring pseudotyped virus infection

Primary infectiona	MLV Env pseudotyped virusb	LacZ expressionc
FB29	Mock	0 ± 0
TR1.3	Mock	0 ± 0
W102G	Mock	0 ± 0
FB29	FB29	0 ± 0
TR1.3	FB29	200 ± 108*
W102G	FB29	18 ± 7*
FB29	TR1.3	0 ± 0
TR1.3	TR1.3	0 ± 0
W102G	TR1.3	0 ± 0

^aVirus used for chronic infection of SC-1 cells.

^bPseudotyped virus with a LacZ transgene was generated with the indicated Env construct. Mock, absence of Env construct. Superinfection challenge of the primary infected cells was conducted with an MOI of 5 in the presence of 8-μg/ml polybrene.

^cLacZ expression was based on counting beta-galactosidase-positive cells in a 24-well plate with approximately 10⁵ cells per well. This result is representative of five experiments. Student's t test was performed to compare TR1.3 and W102G chronically infected SC-1 cells superinfected by FB29 Env pseudotyped virus versus FB29 chronically infected SC-1 cells superinfected by FB29 Env pseudotyped virus. An asterisk indicates a P value of <0.05.

DISCUSSION

The studies presented here definitively link the pathogenic TR1.3 Friend MLV virus Env to a 14- to 20-fold reduction in receptor binding affinity and the loss of superinfection interference, in addition to syncytium formation. This is the first description of a direct linkage between an Env variant that reduces receptor affinity and yet increases syncytium formation in a retroviral envelope protein.

Previous work has shown a correlation between Env expression and syncytium formation following specific amino acid changes in simian immunodeficiency virus envelope and varicella-zoster virus gH glycoprotein (24, 25, 38, 45). Additional studies using amphotropic MLV gp70 demonstrated an inverse relationship between Env expression levels and superinfection (49). The data presented in the manuscript show that TR1.3 and W102G Env expression on infected cells was the same as or less than that of FB29. Importantly, FB29- and TR1.3-infected SC-1 cells expressed the same levels of gp70 by 48 to 72 h postinfection by both FACS analysis and quantitative Western blotting (Fig. 1), which are the maximal points of TR1.3-induced syncytium formation in vitro. The biphasic expression profile of TR1.3 Env, not seen with FB29 Env, may be due to a failure of this virus to uniformly infect every cell in the culture at the MOI used in these experiments.

Several laboratories have previously identified and characterized mutations in the Env SU proline-rich and C-terminal regions that increase cell fusion activity (27-30). Each of these mutations destabilized the association between gp70 and p15E, which is believed to augment the SI phenotype. The shedding of Env from those virus particles was experimentally demonstrated by a reduction in the gp70:p30 ratio relative to the wild-type virus, following centrifugation of the virions through a sucrose cushion. Importantly, we failed to detect changes in the gp70:p30 ratio of TR1.3, W102G, and FB29 following identical treatment, as shown in Fig. 2. This observation indicates that the mechanism for enhanced cell fusion seen in TR1.3 is distinct from those Env which acquire an SI phenotype through mutations that destabilize the gp70-p15E complex.

Previous studies of the interaction of ecotropic MLV with their receptors have characterized the binding of a variety of MLV envelopes and RBD to mCAT-1. When NIH 3T3 cells and crude membrane preparations were used, both purified Rauscher MLV Env and purified Moloney MLV RBD were found to have equilibrium constants between 2 and 8 nM (5, 7, 14, 20, 39). The data presented here for FB29 RBD at 1 to 4 nM compare favorably to these studies, showing conservation in receptor binding constants across ecotropic strains of MLV. Binding of the purified N-terminal domain of Friend clone 57 SU was also studied on Xenopus oocytes expressing mCAT-1, with a K_D of 55 nM (10); subtle differences in assay performance, target cells, and RBD protein preparation may account for this difference. Importantly, for each of these MLVs, the W102G Env substitution did not result in syncytium formation (11).

Increased receptor affinity has been reported to be a correlate of conversion to SI phenotype for multiple viruses. The highly pathogenic simian-human immunodeficiency virus SHIV-KB9 (SI) has been shown to have a higher affinity for its coreceptor than its nonpathogenic counterpart SHIV-89.6 (NSI) (12, 21). HIV Env affinity for CD4 has also been shown to correlate with the SI phenotype independently of Env expression levels (54). In the TR1.3 MLV system, analysis of the impact of the TR1.3 and W102G mutations within Env on mCAT-1 binding yielded quite different results. The relative binding affinity of TR1.3 RBD to mCAT-1 was found to be 14- to 20-fold lower than that of FB29 RBD (Table 1), and the relative avidity of intact TR1.3 and W102G viruses for cells expressing mCAT-1 was also markedly diminished, as shown in Fig. 4.

In view of the TR1.3 binding data, surface mCAT-1 expression during viral infection was assessed through binding of FB29 RBD to chronically infected SC-1 cells and by measurement of virus superinfection. Measurements of available surface receptor by RBD binding showed that cells chronically infected by FB29 or TR1.3 expressed lower receptor levels than uninfected cells but failed to distinguish differences in receptor levels between FB29 and TR1.3 (Fig. 5). In contrast, more-sensitive PCR or pseudotyped virus analyses of superinfection revealed that SC-1 cells chronically infected by TR1.3, but not FB29, could be superinfected (Fig. 6). PCR-based results showed that TR1.3-infected cells could be superinfected by FB29 at all time points tested after primary infection. In contrast, FB29-infected cells were not superinfected by TR1.3 at any point. These results suggest that differential modulation of mCAT-1 by TR1.3 Env, as contrasted to FB29 Env, occurs early after viral entry and is sustained throughout infection.

Transduction experiments additionally demonstrated that TR1.3 and W102G chronically infected cells were susceptible to transduction by FB29 Env-pseudotyped viruses, but not TR1.3 Env-pseudotyped viruses; therefore, interference to superinfection was unidirectional (Table 2). There is ample evidence that the entry of retroviruses requires multiple Env-receptor complexes (46). Work comparing wild-type and mutant CCR5 molecules studied across a range of expression levels suggests that for HIV a decreased receptor affinity corresponds to an increase in the number of surface receptors required to support virus entry (23). Application of these observations to MLV suggests that receptor downregulation following TR1.3 or W102G infection may reduce the level of surface receptor below the threshold required for the lower-affinity TR1.3 Env, while maintaining enough surface receptor to support entry of higher-affinity FB29 Env-bearing particles. This explanation is also supported by previous studies in our laboratory, which showed that SI by TR1.3 in vitro is critically dependent on the levels of receptor (8).

Studies presented here establish a unique relationship between decreased receptor affinity and conversion to SI phenotype in a naturally occurring neuropathogenic MLV. The appreciation of the distinction between receptor binding and virus or cell fusion suggests that steps subsequent to binding in the fusion process may be equally important to the initial binding events in determining the fusion potential of retroviral Env.