ABSTRACT
Infection with murine leukemia virus (MLV) TR1.3 or the related molecular construct W102G causes severe neuropathology in vivo. Infection is causally linked to the development of extensive syncytia in brain capillary endothelial cells (BCEC). These viruses also induce cell fusion of murine cell lines, such as SC-1 and NIH 3T3, which are otherwise resistant to MLV-induced syncytium formation. Although the virulence of these viruses maps within the env gene, the mechanism of fusion enhancement is not fully determined. To this end, we examined the capacity of the syncytium-inducing (SI) TR1.3 and W102G MLVs to overcome the fusion inhibitory activity inherent in the full-length Env cytoplasmic tail. These studies showed that the TR1.3 and W102G Envs did not induce premature cleavage of p2E, nor did they override p2E fusion inhibition. Indeed, in the presence of mutations that disrupt p2E function, the TR1.3 and W102G Envs significantly increased the extent of cell fusion compared to that with the non-syncytium-inducing MLV FB29. Surprisingly, we also observed that TR1.3 and W102G Envs failed to elicit syncytium formation in these in vitro assays. Coexpression of gag-pol with env restored syncytium formation, and accordingly, mutations within gag-pol were used to examine the minimal functional requirements for the SI phenotype. The results indicate that both gag-dependent particle budding and cleavage of p2E are required to activate the SI phenotype of TR1.3 and W102G viruses. Collectively, these data suggest that the TR1.3 and W102G viruses induce cell fusion by the fusion-from-without pathway.
The prevention and management of infections by human retroviruses are challenging public health problems. A more complete understanding of the mechanisms of retroviral pathogenesis offers the potential to develop new therapies that target defined components of retrovirus-associated disease. The appearance of syncytium-inducing (SI) strains of human immunodeficiency virus (HIV) in vivo is thought to precede the T-cell inflection point and has been correlated with the progression of HIV infection to AIDS (7, 25, 31). Although receptor tropism has been linked to the SI phenotype of HIV viruses (16, 26, 45), the defining biological properties of SI retroviruses, in contrast to non-syncytium-inducing (NSI) retroviruses, remain unknown. To this end, we developed a facile murine model system of retrovirus-induced syncytium formation and pathogenesis.
The TR1.3 murine leukemia virus (MLV) is a member of the Friend MLV family. This virus possesses the distinctive ability to induce rapid and severe neurological disease associated with hemorrhagic stroke (37, 38). The onset of disease parallels the appearance of large multinucleated syncytia exclusively in brain capillary endothelial cells (BCEC) (37). Similarly, although primary endothelial cell cultures derived from the brain, kidney, liver, and lung are equivalently infected by TR1.3, cell fusion occurs only in BCEC (28). The aggressive fusion potential of TR1.3 is also evident in cell culture lines; unlike most MLVs, TR1.3 induces syncytium formation in SC-1 and NIH 3T3 murine fibroblast cells (10, 39).
The genetic sequence of MLV TR1.3 shares extensive identity with that of the NSI Friend MLV FB29, thus enabling mapping of disease and the SI phenotype to a tryptophan-to-glycine conversion at position 102 in the N-terminal region of the envelope protein (39). This single W102G conversion is both necessary and sufficient to transform FB29 into an SI virus, and the corresponding G102W reversion in the TR1.3 virus background completely abrogates the SI phenotype (35). The phenotype of the W102G virus is indistinguishable from that of TR1.3 in all aspects of virus tropism and disease kinetics, severity, location, and pathology (39). Therefore, comparison of TR1.3 and W102G viruses to FB29 provides a powerful tool with which to probe the biological mechanisms that regulate syncytium formation in vitro and in vivo.
MLV Env consists of the surface (SU) subunit gp70 tethered to the transmembrane (TM) subunit p15E. Together, these two subunits regulate receptor binding and membrane fusion. The first 230 residues of gp70 span the receptor binding domain of Env, which interacts with the multiple-membrane-spanning protein mCAT-1 (6, 11, 21). The crystal structure for the receptor binding domain predicts a potential receptor binding pocket composed of amino acids 84 and 86 and the tryptophan at position 102, which is mutated in the TR1.3 and W102G viruses (15). Paradoxically, despite an enhanced fusion effect, TR1.3 and W102G viruses have dramatically reduced receptor affinity compared to that of FB29 (34).
The N-terminal hydrophobic fusion peptide and two stretches of predicted amphipathic helices in p15E are essential for fusion (13, 49, 55). In contrast, the cytoplasmic tail of p15E inhibits fusion (41, 43). During the course of virus replication, the C-terminal 16 amino acids of p15E are cleaved from Env by the viral protease, activating the fusion potential of the protein (18, 46). The resulting lower-molecular-weight TM component is termed p12E, and the solubilized cleavage fragment is termed p2E. Truncations or mutations within the cytoplasmic tail of p15E can disrupt this inhibitory activity, triggering fusion in cells that are normally resistant to MLV-induced syncytium formation (41, 43, 47). Regulation of fusion activity by the Env cytoplasmic tail is a broadly conserved feature of retroviruses (8). Normally, p15E cleavage does not occur until during or after the budding process (36). In this way, MLV avoids the cytopathic effects of cell fusion and matures into a fully fusion-competent virus only after full egress from an infected cell.
The gag genes of retroviruses encode the proteins that promote budding of virions from the cell surface. Mutations within gag can have profound effects on the ability of viruses to assemble at the plasma membrane and/or the ability of progeny virions to bud from the cell (4, 19, 20, 30, 53). For example, the late-domain Gag sequence PPPY is a critical mediator of the interaction between viral and cellular proteins during particle release (17, 52): disruption or deletion of the PPPY sequence perturbs the budding process and results in the accumulation of virions on the cell surface (52, 53). The critical timing of p15E cleavage and corresponding acquisition of Env fusion capacity with respect to the events of particle assembly and budding indicates that mutations within gag have the potential to regulate Env-mediated membrane fusion (12).
In the studies presented here, the fusion pathway and mechanism of TR1.3 and W102G viral pathogenesis were examined in the context of p2E fusion regulatory activity, fusion kinetics, and synergy with Gag protein expression and function.
MATERIALS AND METHODS
Cells and viruses.The cell lines 293T/17 (CRL-11268) and NIH 3T3 (CRL-1658) were obtained from ATCC (Manassas, VA). 293T cells expressing mCAT-1 and primary endothelial cell line cultures were described previously (10).
Pseudotyped virus was produced by triple transfection of the appropriate envelope expression vector, pHIT60, and pHIT111. 293T/17 cells were grown in a 10-mm Primaria dish (BD Biosciences, San Jose, CA) to approximately 80% confluence. Cells were then transfected with 4 μg of each plasmid, using the Lipofectamine transfection reagent (Invitrogen, Carlsbad, CA) following the manufacturer's protocol at a concentration of 2.5 μl per μg of DNA in serum-free Dulbecco's modified Eagle's medium. Virus supernatants were harvested at 48 h posttransfection.
Constructs.The virus envelopes from the FB29, TR1.3, and W102G viruses were cloned from pcDNA3.1 into the pHIT123 vector, using PmeI restriction sites (34). For construction of the P2E+ and p2E− mutants, standard mutagenesis was performed using pHIT123 envelope expression vectors and a QuikChange mutagenesis kit (Stratagene) per the manufacturer's instructions. Primers used to introduce the leucine-to-arginine mutation at position 628 of the envelope (P2E+; GTAGTCCAGGCTAGAGTCCTGACTCAA and the complementary sequence) and a stop codon at position 629 of the envelope (p2E−; ATCTCAGTAGTCTGACAGGCTTTAGTC) were synthesized and purified by high-performance liquid chromatography. The full envelope gene sequence was verified for each construct to confirm that no spontaneous mutations arose during the procedure. The PAAA Gag-Pol construct was a kind gift of Paul Bates (University of Pennsylvania). The P150L and R119C/P133L Gag-Pol constructs were a kind gift of Mary Collins (University College London).
Western blot analysis.Cellular lysates were harvested using NP-40 lysis buffer (1% NP-40, 150 mM NaCl, 50 mM Tris, pH 7.8, 10% mammalian protease inhibitor cocktail [Roche]). Lysates were kept on ice for 30 min and then spun down in a refrigerated centrifuge for 30 min at 13,000 rpm. Clarified supernatants were harvested and frozen. NuPAGE gels (Invitrogen) were used to separate samples prior to transfer to nitrocellulose membranes. Primary antibodies for analysis were goat anti-gp70 (VR-1521AS-Gt; ATCC) and rabbit anti-p15E (a kind gift of John Elder, Scripps Research Institute) (18, 44), used at a dilution of 1:2,000. The appropriate horseradish peroxidase-conjugated secondary antibodies were used at a concentration of 1:50,000 (Jackson Immunoresearch, West Grove, PA).
FACS analysis.Fluorescence-activated cell sorter (FACS) analysis of surface Env expression was carried out essentially as described previously (34). Antibodies used for detection were anti-gp70 polyclonal antiserum (ATCC) and mouse monoclonal antibodies (MAbs) 48 and 573, kindly provided by Leonard Evans (NIAID, Rocky Mountain Laboratories) (9).
Fusion assays.NIH 3T3 or 293T cells were seeded in a six-well plate and grown to a density of approximately 75%. Envelope constructs were introduced by transfection of 2 μg DNA/well, using the Lipofectamine transfection reagent (Invitrogen). In the case of NIH 3T3 cells, fixation with 1% methylene blue in methanol was used to quantify fusion at 24 to 48 h posttransfection. In the case of 293T cell transfection, cells were detached 24 h after transfection and used to overlay a subconfluent layer of NIH 3T3 cells in a six-well plate, and methylene blue staining was performed 24 h after the overlay. Syncytia were defined as those cells containing five or more nuclei. The total number of nuclei in syncytial cells was then counted. Five random fields were counted from each well of triplicate samples, using a 10× objective.
Fusion kinetic analysis.293T cells and NIH 3T3 cells were labeled using PKH67 and PKH26 lipophilic dyes (Sigma-Aldrich, St. Louis, MO), respectively, following the manufacturer's protocol. NIH 3T3 cells were seeded into a T75 flask, and 293T cells were seeded into six-well plates. At a density of ∼75%, the 293T cells were transfected with 2 μg of Env expression plasmid, using Lipofectamine. After 24 h, the NIH 3T3 cells were detached with trypsin and the 293T cells were detached using enzyme-free cell dissociation buffer (Invitrogen). All cells were spun down and resuspended at a density of 3 × 105 cells/ml. Equal volumes of Env-expressing PKH67-labeled 293T cells and naïve PKH26-labeled 3T3 cells were then mixed, and 3 × 105 cells were seeded into 3 wells of multiple 48-well plates. Cells were immediately spun down onto the plate for 5 min at 1,500 rpm and placed at 37°C for between 10 and 70 min. Fusion was terminated by fixation in cold 4% paraformaldehyde. A Zeiss confocal microscope was used to acquire three representative images of each well of triplicate samples at each time point (10, 25, 40, 55, and 70 min). Analysis of dye colocalization was performed using Zeiss LSM510META, version 3.2, software. Dye colocalization was determined for each acquired image, and the average values were used for analysis.
RESULTS
p2E cleavage is not altered in TR1.3 Env.The p2E segment of the p15E TM domain is a critical regulator of Env fusion activity. Expression in cells of an Env protein that lacks the p2E C-terminal TM segment induces massive syncytium formation, yet the virus particles produced from these cells are less infectious than wild-type virions (27, 43). The similarity of this phenotype to that of MLV TR1.3 led us to examine whether premature cleavage of the p2E cytoplasmic tail during TR1.3 virus production might account for the SI phenotype. To investigate this possibility, Env and Gag-Pol expression plasmids were cotransfected into 293T cells, and the presence of both immature and mature viruses was examined in cell lysates and supernatants, respectively, by Western blotting. The distribution of p15E and the p12E TM domain cleavage product was visualized using polyclonal rabbit sera specific for epitopes shared in both the p15E and p12E proteins. As shown in Fig. 1, cleavage of the p15E cytoplasmic tail to yield p12E did not differ in magnitude between the FB29, TR1.3, and W102G viruses. Lysates from cells transfected with either TR1.3, W102G, or FB29 Env exhibited only a single band, consistent with the p15E protein found in immature virions. Supernatants containing mature virions displayed bands consistent with the presence of p12E protein and, to a lesser degree, p15E protein. Complete cleavage of p15E to p12E was not observed in any of the viral supernatants.
TR1.3 does not prematurely cleave p15E. Western blot analysis of p15E and p12E was performed with cell lysates or cell-free supernatants from FB29-, TR1.3-, W102G virus-, or mock-infected cell cultures. 293T cells were cotransfected with FB29, TR1.3, or W102G Env expression plasmid and the pHIT60 Gag-Pol expression plasmid, supernatants were harvested after 48 h, equal-volume lysates or cell-free supernatants were clarified by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and the presence of envelope cleavage products was detected using a pan-specific polyclonal rabbit serum directed against p15E.
W102G mutation enhances cell fusion in fusion-activated Env.Another possible explanation for the SI phenotype of TR1.3 is that the presence of the W102G Env mutation may bypass the inhibition of fusion activity resident in the full-length cytoplasmic tail (27, 47). To test this hypothesis, two new sets of constructs were generated from FB29, TR1.3, and W102G Env expression plasmids. The first set of constructs, termed p2E+ constructs, possess a mutation in the cytoplasmic tail which blocks p2E cleavage but nevertheless enhances fusion, presumably through disruption of p2E inhibitory function (27, 47). The second set of constructs, termed p2E− constructs, contain stop codon insertions that terminate p15E translation immediately before the p2E coding region, which also results in enhanced fusion activity. As shown in Fig. 2, representative Western blot analysis of p2E+ and p2E-FB29 virus supernatants confirmed that the Envp2E+ construct produced a stable p15E product that was uncleaved by viral protease, whereas the p2E− construct produced a protein with a migration pattern identical to that of the mature, cleaved p12E protein found in wild-type virus. Migration patterns identical to those shown here were also observed for p2E+ and p2E− mutations in the TR1.3 and W102G Env backgrounds (data not shown).
Impact of p2E+ and p2E− mutations on p15E cleavage. Western blot analysis was performed to detect gp70, p15E, and p12E cleavage in the P2E+ and P2E− envelope mutants. 293T cells were cotransfected with FB29, FB29p2E+, or FB29p2E− Env expression plasmid and the pHIT60 Gag-Pol expression plasmid. Supernatants were harvested at 48 h posttransfection and subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and the presence of gp70, p15E, and p12E was detected by a polyclonal antibody to either gp70 (top) or p15E (bottom).
To evaluate the possibility that the enhanced fusion induced by the W102G substitution results from a loss of fusion inhibition by p2E, we tested the effects of the p2E+ and p2E− mutations in the TR1.3 and W102G backgrounds on syncytium formation. If this hypothesis is correct, insertion of p2E− Env into the W102G background would not be expected to further increase cell fusion. In contrast, if the W102G effect is independent of p2E function, then fusion should be augmented significantly in association with either the p2E+ or p2E− construct.
We first measured the impact on syncytium formation of wild-type, p2E+, and p2E− Env expression vectors transfected into NIH 3T3 cells. Surprisingly, when expressed alone, neither the TR1.3 nor W102G wild-type Env construct induced cell fusion in NIH 3T3 cells (Fig. 3A), although both TR1.3 and W102G Envs were previously shown to be highly fusogenic in a cell-cell fusion assay driven by Env overexpression in vaccinia virus (10). The reason for this difference was fully examined and is described below, but this result also created the experimental advantage of a low background for testing the impact of p2E+ and p2E− mutations on fusion potential. As shown in Fig. 3A, introduction of the p2E+ mutation enhanced the fusion of each MLV Env, but more significantly in the TR1.3 and W102G backgrounds than with FB29. NIH 3T3 cells expressing either the TR1.3p2E+ or W102Gp2E+ construct fused at a magnitude 10 to 20 times greater than cells expressing FB29p2E+ (P < 0.05). Similar results were observed using the p2E− constructs: a fivefold increase in cell fusion was exhibited by the TR1.3p2E− and W102Gp2E− constructs relative to that with FB29p2E− (P < 0.05) (Fig. 3A). These differences were observed from the onset of fusion at 24 h until 48 h. After this point, extensive cytopathology prohibited further analysis.
W102G virus-mediated cell fusion is independent of p15E cleavage. (A) Cell fusion of NIH 3T3 cells following transient transfection with either wild-type, p2E+, or p2E− FB29 (dark bars), TR1.3 (striped bars), and W102G (light bars) Env expression plasmids. Results are presented as the total number (log scale) of nuclei in syncytia per field (10× objective; average of five random fields per well). Each assay was performed in triplicate, and averages for a representative assay are shown. This assay was repeated four times with similar results. (B) Cell fusion of NIH 3T3 cells following overlay with receptor-negative 293T cells transiently transfected with either wild-type, p2E+, or p2E− FB29 (dark bars), TR1.3 (striped bars), or W102G (light bars) Env expression plasmids. Each assay was performed in triplicate. This assay was repeated twice with similar findings. Asterisks indicate statistically significant differences (P < 0.05) between FB29 and TR1.3 Envs or FB29 and W102G Envs, as determined by the unpaired Student t test.
To confirm these results, we also evaluated the fusion potential of these constructs in a second MLV fusion assay. In this instance, 293T cells transfected with wild-type, p2E+, and p2E− constructs were used to overlay NIH 3T3 target cells, and nuclei in syncytia were counted following fixation and counterstaining. As shown in Fig. 3B, the results of these studies were identical to those reported above: wild-type TR1.3 and W102G Envs did not induce cell fusion above background levels, both TR1.3p2E+ and W102Gp2E+ showed a greater ability to induce cell fusion than did FB29p2E+ (P < 0.05), and both TR1.3p2E− and W102Gp2E− induced greater fusion than did FB29p2E− (P < 0.05).
Since p2E− Env constructs produce a prematurely truncated Env protein, we verified that surface expression of Env was not significantly altered by this mutation. The comparison of Env expression by wild-type and p2E− constructs was done by FACS analysis 24 h after transfection of 293T cells. As shown in Table 1, mean channel fluorescence was statistically indistinguishable for each Env, as measured by gp70 polyclonal antibody binding. Moreover, the same result was obtained using two conformationally dependent monoclonal antibodies (MAb 573 and MAb 48) that recognize epitopes within SU. Thus, Env expression levels as well as gross structural integrity appear to be conserved among these Env constructs.
Env expression is comparable between p2E− mutants and wild-type Enva
To determine the effects of the p2E+ and p2E− mutations on transduction efficiency, pseudotyped viral particles carrying a LacZ reporter gene were produced for each wild-type, p2E+, and p2E− Env in combination with the FB29, TR1.3, and W102G backgrounds. Transduction assays performed on NIH 3T3 cells showed that the p2E+ and p2E− mutations in Env uniformly reduced the transduction efficiency by >1 log (data not shown). This was similarly observed for FB29, TR1.3, and W102G Envs and was consistent with previous studies using Moloney MLV Env (27).
To determine the impact of the W102G substitution on syncytium formation and fusion kinetics, we next employed a modified version of the cell fusion overlay assay. In this assay, 293T cells were transfected with either the FB29p2E−, TR1.3p2E−, or W102Gp2E− Env construct or a vector control. These cells were then labeled with the green lipophilic dye PKH67. Green-labeled 293T effector cells were then mixed with NIH 3T3 target cells differentially labeled with the red dye PKH26. After being mixed in suspension, target and effector cells were spun down into 96-well plates and incubated at 37°C for 25 to 70 min to initiate fusion. As shown in both Fig. 4A and B, TR1.3p2E− and W102Gp2E− Env expression triggered the rapid appearance of syncytia, which were uniformly visible by 25 min. In contrast, the cell fusion events induced by the FB29p2E− Env were not evident at 25 min and accrued more slowly. After 70 min of incubation, fusion was evident in all cultures that contained Env effector cells, but both the magnitude of fusion and the size of syncytia were consistently greater with TR1.3p2E− and W102Gp2E− Envs than with FB29p2E− Env. Taken together, these experiments show that the presence of the W102G mutation augments Env-induced cell fusion by a mechanism independent of the regulation of fusion by the p2E element. Importantly, these data also show that expression of TR1.3 or W102G Env alone is not sufficient to drive cell fusion in NIH 3T3 cells.
W102G mutation enhances the kinetics of cell fusion. Cell fusion of NIH 3T3 cells (red) was examined following overlay with receptor-negative 293T cells (green) transiently transfected with either FB29, TR1.3, W102G, or a vector control. The magnitude of cell fusion (yellow) over time (70 min) was determined by sequential image analysis using a Zeiss confocal microscope and LSM510META, version 3.2, image software. (A) Confocal microscope images 25 and 70 min after cell mixing. Magnification, ×10. Arrows indicate representative areas of syncytium formation. (B) Quantitative analysis of cell fusion over time. Black boxes, TR1.3; triangles, W102G mutant; circles, FB29; smooth gray line, empty plasmid. The graph represents data from three experiments performed in triplicate.
Virus budding and p2E cleavage reconstitute the TR1.3 SI phenotype.The finding that expression of the TR1.3 and W102G Envs following transient transfection did not induce fusion in NIH 3T3 cells, which are otherwise highly susceptible to cell fusion during TR1.3 or W102G virus replication, may provide important clues to the mechanism of syncytium formation. The two predominant models for virus-induced cell fusion are fusion from within (FFWI) and fusion from without (FFWO). FFWI occurs when Env expressed on the surface of one cell mediates fusion with an adjacent receptor-bearing cell. FFWO occurs when a virion simultaneously engages two adjacent receptor-bearing cells, thereby fusing the two cell membranes. Although there have been indications that the FFWI pathway may regulate in vivo syncytium formation (1, 5), there is no definitive evidence for or against either pathway as a primary mechanism of disease.
The failure to observe syncytium formation following expression of TR1.3 or W102G Env in NIH 3T3 target cells suggests that FFWI might not be the primary mechanism of syncytium formation mediated by TR1.3; indeed, if TR1.3-induced fusion requires replication-competent virus, then an FFWO pathway may be responsible. To explore this possibility, we examined the capacity of Gag to rescue syncytium formation in our fusion assays. Env plasmids from TR1.3, W102G, or FB29 virus were cotransfected into NIH 3T3 cells with either an empty vector or a gag-pol expression vector (pHIT60). As shown in Fig. 5A, Gag-Pol expression rescued the ability of TR1.3 and W102G Envs to form syncytia in NIH 3T3 cells but had no impact on fusion in the presence of FB29 Env.
Gag-Pol coexpression is required for W102G Env-induced cell fusion. (A) Impact of wild-type Gag-Pol on Env-induced cell fusion. NIH 3T3 cells were cotransfected with either FB29 (dark bars), TR1.3 (striped bars), or W102G (light bars) Env and with wild-type Gag-Pol expression vectors. The amount of fusion was determined by counting the total number of nuclei in syncytia (five random fields; 10× objective) at 48 h posttransfection. Data shown are averages for triplicate wells. (B) Impact of altered Gag-Pol constructs on Env-induced cell fusion. The assays were performed as described for panel A, using the Gag-Pol expression vectors expressing Gag-Pol PAAA, Gag-Pol P150L, and Gag-Pol R119C/P150L. Each assay was performed in triplicate. This assay was repeated twice with similar findings. Asterisks indicate statistically significant differences (P < 0.05) between FB29 and TR1.3 Envs or FB29 and W102G Envs, as determined by the unpaired Student t test.
To test whether a fully functional Gag protein was necessary for the generation of syncytia, a panel of Gag-Pol mutants (Gag-Pol PAAA, Gag-Pol P150L, and Gag-Pol R119C/P133L) was used to assess the minimum requirements for restoration of the SI phenotype. The Gag-Pol PAAA mutant contains a series of alanine substitutions within the Gag p12 late domain. Mutations disrupting this domain were shown to reduce virion release and to impair p2E cleavage (53). The Gag-Pol P150L mutant also displays diminished but detectable levels of virion release. The double mutant Gag-Pol R119C/P133L completely abrogates virion release (3, 4, 53).
In contrast to the effect of wild-type Gag-Pol, coexpression of TR1.3 or W102G Env with either Gag-Pol PAAA or Gag-Pol R119C/P133L failed to induce syncytium formation (Fig. 5B). Cotransfection of TR1.3 and W102G Envs with Gag-Pol P150L significantly, but only partially, reconstituted the SI phenotype of TR1.3 and W102G Envs relative to that with wild-type Gag-Pol (Fig. 5B). As expected, none of these Gag-Pol constructs induced syncytium formation when coexpressed with FB29 Env.
Differences in the ability of Gag-Pol mutants to facilitate virus budding and particle release were next investigated to explore the relationship between cell fusion and Gag expression. To compare the levels of virus particle release following expression of each Gag-Pol construct, we measured the relative amounts of reverse transcriptase activity in cell-free supernatants from these cultures. As shown in Fig. 6A, both Gag-Pol PAAA and Gag-Pol P150L displayed measurable but reduced levels of reverse transcriptase activity compared to wild-type Gag-Pol, whereas, as reported previously, Gag-Pol R119C/P133L did not facilitate particle formation above the background (2-4).
Formation of mature viral particles is required for W102G Env-induced cell fusion. (A) Formation of virus particles with either wild-type or altered Gag-Pol. 293T cells were cotransfected with FB29 or W102G Env and the indicated Gag-Pol construct. Levels of budding were approximated by measurement of 32P incorporation in a reverse transcriptase assay (39). Each sample was run in triplicate, with averages shown. Asterisks indicate statistically significant differences compared to pcDNA3.1-transfected cells, as determined by the unpaired Student t test. (B) Western blot analysis of p15E and p12E in viral supernatants from 293T cells cotransfected with FB29 Env and the indicated Gag-Pol construct. Samples were prepared and analyzed as described in the legend to Fig. 1, with the exception of the PAAA sample, which was concentrated 10-fold using a microcentrifuge concentrator prior to gel loading. This blot is representative of three independent experiments.
Since particle release alone did not distinguish Gag-Pol P150L from Gag-Pol PAAA, we subsequently examined the cleavage of p15E within the virions formed following coexpression of these constructs with FB29 or W102G Env. Previous studies suggested that Gag-Pol PAAA might have a deficiency in p15E cleavage (52, 53). As shown in Fig. 6B, virions produced with the P150L mutation cleaved p2E, but at a reduced level compared to that of the wild type. Conversely, there was no p15E cleavage in PAAA mutant-containing virions; this was verified in an overexposure of this blot (not shown) and in two subsequent identical analyses.
DISCUSSION
The work presented in this paper provides several important insights into the mechanism of cell fusion induced by TR1.3 and, more specifically, the biological effect of the W102G substitution in Env, as follows: (i) the TR1.3 and W102G viruses do not stimulate premature cleavage of p15E, (ii) they do not override the inhibitory activity of the Env cytoplasmic tail on fusion, (iii) they display accelerated kinetics of MLV Env-mediated membrane fusion, and (iv) they acquire their SI phenotype only following budding and cleavage of p2E. Collectively, these data suggest that the TR1.3 virus induces cell fusion primarily by the FFWO pathway. Previous work from our group showed that NIH 3T3 cells and primary BCEC undergo FFWO when exposed to high concentrations of TR1.3 or W102G virus but not FB29 virus, supporting the conclusions of the results presented here (10, 28).
The observation that TR1.3 and W102G Envs alone did not induce cell fusion (Fig. 3) provided an important tool for this analysis. In previous FFWI assays, which showed enhanced cell fusion by these Envs relative to that by FB29 Env on NIH 3T3 cells, we employed vaccinia virus-driven Env expression in effector cells to achieve exceptionally high levels of Env and T7 polymerase expression (10). Although these assays are commonly utilized in analyzing virus fusion, envelope expression levels can influence syncytium formation, as seen recently in a simian immunodeficiency virus system as well as with other viruses (29, 40). Furthermore, vaccinia virus infection has numerous other effects on host cells which may alter the requirements for cell fusion. Indeed, the NSI FB29 Env also caused cell fusion in vaccinia virus-based assays, albeit to a lesser extent than that by TR1.3 or W102G Env (10). These findings led us to investigate the role of Gag-Pol in TR1.3 and W102G Env-mediated cell fusion.
Restriction of fusion activity in wild-type virus is overcome by the cleavage of p2E from the cytoplasmic tail after budding occurs. Here we show that the dependence of cell fusion on p2E cleavage is maintained in the TR1.3 and W102G viruses and, additionally, that fusion is dependent on the release of virions from the cell surface. These studies were enabled by the observation that while neither TR1.3 nor W102G Env enhanced cell fusion when expressed in isolation, cell fusion was rescued by coexpression of wild-type Gag-Pol (Fig. 5A).
Mutations in Gag that reduced the level of budding virus were used to further clarify the mechanism of MLV-induced cell fusion. Previous studies showed that the P150L and R119C/P133L mutations (2), as well as a mutation similar to PAAA (52), inhibited or eliminated processing of the Gag-Pol polyprotein. Verification that each of these mutations induced a similar block of Gag-Pol processing in our system was conducted by Western blot analysis (data not shown). Coexpression of mutated Gag-Pol constructs with TR1.3 or W102G Env uniformly reduced the level of syncytium formation (Fig. 5B and 6). This was most extensive with the Gag-Pol R119C/P133L mutant. The relative importance of virion release and p2E cleavage to TR1.3 syncytium formation was illustrated by examining the effects of the Gag-Pol PAAA and P150L constructs, both of which enabled virus budding, albeit at reduced levels. Coexpression of the Gag-Pol PAAA construct failed to elicit either p2E cleavage or syncytium formation. In contrast, coexpression of the P150L construct both stimulated p2E cleavage and enabled syncytium formation. From these results, we conclude that optimal cell fusion with either TR1.3 or W102G virus requires both the release of virus particles and cleavage of the p2E element, which supports FFWO as a mechanism of MLV pathogenesis.
Regulation of Env-mediated fusion by the Env cytoplasmic tail and by interaction between the cytoplasmic tail and Gag is a conserved feature of retroviruses (8). Although the cytoplasmic tail of HIV Env gp41 is not cleaved, truncations of this tail transformed an NSI virus into an SI virus, and synthetic extensions of this tail inhibit cell fusion (22, 33, 51). Truncations in the cytoplasmic tail of human T-cell leukemia virus can similarly enable syncytium formation by human T-cell leukemia virus Env in cell lines normally resistant to this cytopathic effect (24). Interactions between the gp41 cytoplasmic tail and Gag inhibit HIV fusion activity in the same manner that p2E inhibits MLV fusion activity (33, 50). Only following cleavage of Gag by the viral protease does HIV Env acquire its full fusion capability. This cleavage requirement is circumvented by truncation of the gp41 cytoplasmic tail (24). Taken together with the data presented here, this raises the possibility that FFWO may occur in vivo with HIV and may play an important role in the pathogenesis of this virus.
In both the current and previous studies, TR1.3 and W102G Envs were intrinsically more fusogenic than FB29 Env, although Env expression levels and stability on virus particles were equivalent (34). In previous studies, we also demonstrated that cell fusion is dependent on the level of virus receptor (10); thus, both Env and receptor expression levels mediate cell fusion potential. This finding was consistent across multiple assays (direct transfection and target overlay) and using several mutations (P2E+ and p2E−) to activate Env fusion (Fig. 3). Together, these observations prove that the W102G mutation enhances fusion through a mechanism independent from the regulation of fusion by p2E. Nevertheless, the TR1.3 and W102G viruses retain an absolute requirement for a functional TM domain. Two mutations previously shown to render TM essentially nonfusogenic (T461P and R553Q) in the Moloney MLV background were tested in the corresponding locations in the FB29, TR1.3, and W102G backgrounds (amino acids 471 and 563, respectively) (54, 55). In both cell fusion assays and transduction assays, the W102G mutation failed to restore detectable fusion activity (data not shown). Thus, while the W102G mutant can complement fusion activity triggered by changes in TM, this Env remains bound to the functionality of TM to achieve membrane fusion.
Studies on HIV and simian immunodeficiency virus have revealed that increased receptor binding is a correlate of the SI phenotype. A higher affinity for CD4 has been associated with increased syncytium formation in the HXB2 HIV Env (48); similarly, the highly pathogenic SI simian-human immunodeficiency virus (SHIV) strain KB9 has increased affinity for its coreceptor relative to the parental NSI strain SHIV 89.6 (14, 23). Paradoxically, TR1.3 and W102G Envs have a lower affinity for the mCAT-1 receptor molecule than that of wild-type FB29 Env (34). Data in this study shed light on this paradox in that TR1.3 and W102G Env-mediated fusion occurs more rapidly than FB29 Env-mediated fusion (Fig. 4). This suggests that receptor binding is not uniformly the rate-limiting step in retrovirus-induced cell fusion. Melikyan et al. also concluded that binding is not a rate-limiting step in MLV fusion by showing that prebinding of cells expressing a truncated Env to target cells did not noticeably accelerate the fusion process (32). The data presented here are also supported by the findings of Reeves et al., who showed that Env fusion kinetics but not receptor binding correlate with the extent of fusion induced by HIV Env (42).
TR1.3 and W102G Envs markedly enhance fusion both in vitro and in vivo. Until now, the relative importance of the FFWO and FFWI pathways in retroviral pathogenesis has remained unclear. Here we provide evidence that syncytium formation restricted to the FFWO pathway can result in dramatic pathogenesis in a model system of retrovirus-associated disease. While we cannot rule out the possibility that FFWI may also play a role in retrovirus fusion and pathogenesis, these studies highlight the importance and biological relevance of the FFWO pathway.
FOOTNOTES
- Received 16 April 2007.
- Accepted 11 July 2007.
- Copyright © 2007 American Society for Microbiology
