Envelope Determinants for Dual-Receptor Specificit
http://www.100md.com
病菌学杂志 2006年第4期
Program in Molecular and Cellular Biology
Department of Microbiology, University of Washington
Division of Human Biology, Fred Hutchinson Cancer Research Center, Seattle, Washington
ABSTRACT
Gammaretroviruses, including the subgroups A, B, and C of feline leukemia virus (FeLV), use a multiple-membrane-spanning transport protein as a receptor. In some cases, such as FeLV-T, a nonclassical receptor that includes both a transport protein (Pit1) and a soluble cofactor (FeLIX) is required for entry. To define which regions confer specificity to classical versus nonclassical receptor pathways, we engineered mutations found in either FeLV-A/T or FeLV-T, individually and in combination, into the backbone of the transmissible form of the virus, FeLV-A. The receptor specificities of these viruses were tested by measuring infection and binding to cells expressing the FeLV-A receptor or the FeLV-T receptors. FeLV-A receptor specificity was maintained when changes at amino acid position 6, 7, or 8 of the mature envelope glycoprotein were introduced, although differences in infection efficiency were observed. When these N-terminal mutations were introduced together with a C-terminal 4-amino-acid insertion and an adjacent amino acid change, the resulting viruses acquired FeLV-T receptor specificity. Additionally, a WL change at amino acid position 378, although not required, enhanced infectivity for some viruses. Thus, we have found that determinants in the N and C termini of the envelope surface unit can direct entry via the nonclassical FeLV-T receptor pathway. The region that has been defined as the receptor binding domain of gammaretroviral envelope proteins determined entry via the FeLV-A receptor independently of the presence of the N- and C-terminal FeLV-T receptor determinants.
INTRODUCTION
Receptors for gammaretroviruses are multiple-membrane-spanning proteins involved in small-molecule or ion transport (31, 37). For most retroviruses, it appears that these receptors are both necessary and sufficient for virus binding and entry and are referred to here as the "classical" receptors. The T-cell-tropic feline leukemia virus (FeLV-T) was the first identified example of a naturally occurring gammaretrovirus that requires two host proteins for entry into the cell, Pit1 and FeLIX (1). These two proteins, the membrane-spanning receptor Pit1 and the soluble host protein FeLIX, constitute what we call a "nonclassical" receptor. More recently, it has been shown that other naturally arising retroviruses, mink cell focus-forming murine leukemia virus and porcine endogenous retroviruses, can use a nonclassical receptor for entry, although they may also infect cells using classical receptors (22, 41).
Studies of gammaretroviruses have indicated that the classical receptor determinants of the envelope protein lie in the N-terminal region of the envelope surface unit (SU) in variable region A (VRA) and VRB. This domain was first identified in murine leukemia virus (MuLV) and has been defined as a region approximately encompassing amino acid positions 50 to 190 in the mature SU, depending on the strain (3, 8, 29). In MuLV, as well as in related gammaretroviruses such as FeLV, VRA is the primary determinant for receptor specificity while VRB serves as a secondary determinant (8, 10, 36). Because the N terminus of the SU encompassing VRA and VRB is sufficient to bind specifically to cognate receptors (3, 7, 15), it is considered the receptor binding domain (RBD) in gammaretroviruses.
Initial infection of cats by subgroup A FeLV (FeLV-A), the transmissible form of the virus, occurs via the recently identified FeLV-A receptor (referred to here as FLVAR) (R. Mendoza, M. Anderson, and J. Overbaugh, submitted for publication). Once infection is established, FeLV-A can undergo changes to give rise to the other FeLV subgroups with distinct receptor usage (31): FeLV-B, which uses the related phosphate transporters Pit1 and Pit2 (2, 39); FeLV-C, which uses the heme transporter FLVCR (32, 33, 38); and FeLV-T, which requires both Pit1 and FeLIX (1, 31).
FeLV-A and FeLV-T clones (61E and 61C, respectively) were originally isolated from a single cat with FeLV-feline AIDS-induced immunodeficiency (30), and they have approximately 96% identity (12, 30). Using chimeras, the T-cell specificity and cytopathicity of FeLV-T-61C were mapped to two regions within the envelope SU: amino acid changes within the N-terminal 109 amino acids, including an N-terminal deletion within the VRA of RBD, and a C-terminal 6-amino-acid insertion and adjacent amino acid change relative to FeLV-A (12).
FeLV-81T is a clone isolated from a cat inoculated with FeLV-A-61E that also causes T-cell killing (34). FeLV-81T can use both the FeLV-A and the FeLV-T receptors, making it dual-receptor specific (reference 34 and unpublished observations). Interestingly, FeLV-A/T-81T contains a 4-amino-acid insertion in the exact location where the C-terminal 6-amino-acid insertion in FeLV-T-61C is found. Chimeras containing the N terminus of the SU from FeLV-A-61E and the C terminus of the SU (including the insertion) from FeLV-81T were replication defective, but replication was restored by a single mutation at amino acid position 7 of a tissue culture-adapted virus, EE(EQPT3)E (14). Thus, all T-cell-tropic clones that we have characterized to date—FeLV-T-61C, FeLV-A/T-81T, and EE(EQPT3)E—contain an N-terminal amino acid change (at position 6, 7, or 8) and either a 6- or 4-amino-acid insertion (and adjacent amino acid change) in an identical location in the C terminus(30, 34). These changes outside of the RBD are associated with T-cell tropism in vitro, suggesting that they may also confer specificity for the FeLV-T receptor complex because T cells express high levels of Pit1 and FeLIX (20).
The N-terminal changes associated with the T-cell tropism of FeLV-T are in a highly conserved PHQ motif in gammaretroviruses (3, 24). In MuLV, substitutions or deletions of the histidine within the motif at amino acid position 8 (H8) result in defective viruses (4, 21, 24). However, these viruses can still bind to their cognate receptors, signifying a role for MuLV H8 in postbinding events (3, 4, 21, 42). These defective MuLVs can be rescued by the trans addition of soluble RBD fragments that contain the wild-type histidine (5, 6, 21, 24). The laboratory-engineered soluble RBDs that rescue MuLVs lacking the critical H8 resemble the soluble cofactors of nonclassical receptors that function in natural infections, such as FeLIX (1). This suggests that there may be common mechanisms of entry underlying these laboratory-derived MuLV mutants and circulating variants such as FeLV-T. However, because naturally arising FeLV-T variants acquire multiple mutations throughout the envelope sequence that may contribute to their fitness in the host, it is difficult to predict whether the PHQ domain in FeLV-T serves the same function as it does for MuLV.
In this study, we sought to dissect the determinants for dual-receptor specificity—the ability to use both a classical receptor (FLVAR) and a nonclassical receptor (Pit1 and FeLIX)—by examining the effects of mutations at the N terminus (amino acid positions 6, 7, and 8), the C-terminal insertion and adjacent amino acid change, and the VRA within the context of either the FeLV-A or FeLV-C envelope SU backbone on binding and infection with various FeLV receptors. In this manner, we have identified the determinants for classical and nonclassical entry of a gammaretrovirus, and we show that these determinants can be independent of each other.
MATERIALS AND METHODS
Cell culture. The MDTF murine fibroblast cell line has been described previously (19), as has the MDTF-FePit1 line (2). Construction of the MDTF cell line expressing the receptor for FeLV-A will be described elsewhere (Mendoza, Anderson, and Overbaugh, submitted). The receptor for FeLV-C (FLVCR) was introduced into MDTF cells by transduction using virus particles packaging pMSCV(FLVCR)neo (32). Particles were generated in 293T human embryonic kidney (HEK) cells using transient transfection with the following plasmids: amphotropic MuLV envelope (SV-A-MLV-env [18]), a construct encoding FeLV gag and pol (61E-LTR- gag-pol [35]), and pMSCV(FLVCR)neo (32), using the transduction and selection method described previously (2). All cell lines, including MDTF murine fibroblasts and 293T HEK fibroblasts, were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum, 100 U penicillin/ml, 100 μg streptomycin/ml, 0.25 mg amphotericin fungicide/ml, and 2 mM L-glutamine (complete DMEM). The MDTF-FePit1, MDTF-FLVAR, and MDTF-FLVCR cell lines were maintained in complete DMEM with 1.0 mg G418/ml (Geneticin; Gibco, Carlsbad, CA).
Construction of mutant virus clones. Envelope expression clones, in which the FeLV-A-61E envelope gene (35) or the FeLV-T-61C envelope gene (20) are expressed in the pcDNA3.1 zeo(–) vector, have been described previously. An analogous clone used to express the FeLV-A/T-81T (clone 109) envelope was made by the same method. Briefly, the EET109E plasmid (14) was digested with XhoI and SacII, and an approximately 1.5-kb fragment was isolated from an agarose gel, purified, and ligated to a 5-kb fragment of pcDNA3.1-61Eenv (35) that had been digested with the same restriction enzymes and purified in the same manner.
The amino acid modifications 7P, 6D, 6P-8M, ins (HD at amino acid position 350 adjacent to a GESQ insertion), and 378L in the mature envelope SU in the 61Eenv/pcDNA3.1 Zeo(–) plasmid (35) were introduced by site-directed mutagenesis (Quick-Change Mutagenesis; Invitrogen). Sequences of the mutagenic primers are available upon request. All modified plasmid constructs were verified by sequencing. Standard subcloning techniques were used to create chimeras expressing mutations at both the N terminus (7P, 6D, or 6P-8M) and C terminus (ins or ins-378L). Modified FeLV RBDs were cloned into the pcDNA3.1 Zeo(–) expression vector using a 5' XhoI site and a 3' Bsu361 site, as previously described (35).
A FeLV-C-FSC envelope expression clone consisting of the FeLV-A-61E backbone and containing the FeLV-C-FSC RBD was made using the XhoI and SacII sites in a manner similar to the construction of the FeLV-A/T-81T envelope expression clone described above. Analogous chimeric envelopes were constructed that expressed one of the two N-terminal changes in combination with one of the two C-terminal mutations described above. As before, a 5' XhoI site and a 3' Bsu361 site flanking the FeLV RBD sequence in pcDNA3.1 Zeo(–) were used. However, due to an additional Bsu36I site in the RBD of FeLV-C-FSC, site-directed mutagenesis (Quick-Change Mutagenesis; Invitrogen) was used to remove the Bsu361 site internal to the FeLV-C-FSC RBD, resulting in a silent mutation (primer sequences available upon request).
The 61E7P--ins-378L construct, containing a 6-amino-acid deletion similar to one found in FeLV-T-61C, was made in the context of the 61E7P-ins-378L mutant backbone using overlap extension PCR (17). Sequences of the mutagenic primers are available upon request. The first set of primers amplified a fragment 5' of the deletion, and the second set of primers amplified a fragment 3' of the deletion. In the second round of PCR, the 5' fragment and 3' fragment were combined to prime each other and form a 1.5-kb fragment, which was cloned back into the full-length FeLV-A-61E envelope expression vector (35) using the XhoI and SacII sites as described above.
Virus production and infection assays. To maintain mutations of interest and avoid introduction of new ones, we created pseudotyped viruses that were limited to a single round of replication. Virus was made by calcium phosphate transfection of 293T cells (Mammalian Transfection Kit; Stratagene, La Jolla, CA) with three plasmids: 4 μg of FeLV-61E-gag-pol (61E-LTR--gag-pol [35]), 4 μg of MuLV-derived reporter gene that expresses -galactosidase (pRT43.2Tnlsgal-1 [40]), and 10 μg of FeLV-envelope construct. One day post-transfection, the cells were washed three times in phosphate-buffered saline, and 7 ml complete DMEM was added. Cell-free virus supernatant was collected at 2 days post-transfection and filtered through a 0.2-μm filter, aliquoted, and stored at –80°C.
Viral infectivity was assayed using a single-cycle infection assay described previously (2). Cells were seeded at 2 x 104 cells per well of a 24-well dish and incubated for 24 h, after which virus and complete DMEM containing 4 μg Polybrene per ml were added to a total volume of 1 ml per well. Two days postinfection, cells were fixed and stained for -galactosidase activity, and -galactosidase-positive blue foci were scored as previously described (14).
In some cases, we added 100 μl of conditioned medium containing soluble cofactors to cultured cells just prior to adding virus. Soluble cofactors were generated, as previously described, by calcium phosphate transfection of 293T cells with plasmids encoding either FeLIX (20) or a soluble RBD derived from FeLV-A-61E envelope that terminated at amino acid 231, analogous to FeLIX (61E sRBD; Anderson, Cheng, and Overbaugh, unpublished data). The expression of FeLIX and 61E sRBD was verified by Western blot analysis (data not shown). For these infection studies, we used fluorescence-activated cell-sorted MDTF-FLVAR cells that expressed high levels of FLVAR, as described elsewhere (Cheng, Mendoza, and Overbaugh, unpublished data) to ensure that receptor levels were not limiting.
Virus ultracentrifugation. One milliliter of virus supernatant (made as described above) was layered over 750 μl of a solution of 20% sucrose in TNE (10 mM Tris, 150 mM NaCl, and 1 mM EDTA). The sample volume was brought up to 5 ml and balanced with complete DMEM in 13- by 51-mm polyallomer tubes (Beckman, Palo Alto, CA), and it was subsequently ultracentrifuged for 2 h at 24,000 rpm at 4°C using an Sw55 rotor in an Optima-L 90K Ultracentrifuge (Beckman Coulter, Palo Alto, CA). When the spin was completed, supernatants were discarded by inverting the tubes, followed by draining for 10 min. Samples were then resuspended in 20 μl of standard Laemmli loading buffer for Western blot analysis.
Immunoprecipitation and Western blot analysis. Two milliliters of each virus supernatant was precleared for three hours at 4°C using 300 μl of a slurry containing Protein A beads (Sigma, St. Louis, MO). Samples were centrifuged at 14,000 rpm at 4°C, and the supernatant was immunoprecipitated for 2.5 h at 4°C using 50 μl of Protein A beads conjugated to the goat anti-FeLV polyclonal antiserum 77S (Quality Biotech, Inc., Resource Lab, Camden, NJ) (11). Protein-bound beads were then washed three times with a solution of 0.1% NP-40 and 0.1% Triton X-100 in phosphate-buffered saline and resuspended in 20 μl of sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis loading buffer, and the entire sample was resolved on an SDS-10% polyacrylamide gel. Ultracentrifuged virus samples (prepared as described above) were boiled and resolved on an SDS-10% polyacrylamide gel, followed by wet transfer to a polyvinylidene difluoride blotting membrane (Millipore, Bedford, MA).
Western blot analysis was performed using the anti-FeLV gp70 monoclonal C11D8 antibody (Custom Monoclonal Antibodies, Sacramento, CA) (13) and a horseradish peroxidase-conjugated goat anti-mouse secondary antibody (Bio-Rad, Hercules, CA). Bound antibodies were detected by chemiluminescence (Amersham, Piscataway, NJ).
Binding assay. Binding of virus to cells was performed as previously described (20). Briefly, 2 x 105 cells were incubated with 1 ml of cell-free supernatant and 100 μl of ultracentrifuge-purified virus in a 2-ml total volume for 45 min at 37°C. Incubations with the primary antibody, anti-FeLV gp70 monoclonal C11D8 or anti-hemagglutinin (HA) monoclonal HA.11 (Covance, Berkeley, CA), and the secondary antibody, r-phycoerythrin-conjugated goat anti-mouse (DAKO, Carpinteria, CA), were performed as described previously, except that antibody incubation volumes were at half the previous levels (20).
RESULTS
Individual effects of a C-terminal insertion or N-terminal mutations on FeLV-A and FeLV-T receptor specificity. FeLV-T-61C and the dual-receptor-specific FeLV-A/T-81T contain either a 6- or 4-amino-acid insertion and an adjacent amino acid change in an identical location in the C terminus of the envelope SU (30, 34). We engineered a virus envelope containing only the insertion and adjacent amino acid change in the FeLV-A-61E backbone (61Eins). In order to assess receptor specificity, we tested a pseudotyped virus expressing the mutant envelope in a single-cycle infection assay. Virus was introduced to cell lines that express either the FeLV-A receptor, MDTF-FLVAR (Mendoza, Anderson, and Overbaugh, submitted), or the FeLV-T receptors, MDTF-FePit1, in the presence of FeLIX (henceforth referred to as MDTF-FePit1/FeLIX for simplicity) (Fig. 1). We observed that 61Eins could infect MDTF-FLVAR cells despite a 2-log10 reduction in titer compared to FeLV-A-61E (Fig. 1). We could also detect a level of binding of 61Eins to MDTF-FLVAR comparable to that of the FeLV-A-61E parental virus (Fig. 1). Importantly, this virus did not infect MDTF-FePit1/FeLIX cells. Thus, viruses with the C-terminal insertion and adjacent amino acid change maintained specificity for the FeLV-A receptor but were unable to infect via the FeLV-T receptor.
All FeLV-T and dual-receptor-specific FeLV-A/T variants characterized to date have an N-terminal change of one or more amino acids at positions 6, 7, and 8 in the mature SU relative to FeLV-A-61E. These include a single amino acid change at position 6 (HD) found in the FeLV-A/T-81T variant (34), changes at both positions 6 and 8 (6HP, 8IM) found in the prototype FeLV-T-61C variant (30), or a change at position 7 (QP) found in a tissue culture-derived FeLV-A/T variant, EE(EQPT3)E (14). We used site-directed mutagenesis to introduce these mutations individually into the FeLV-A-61E envelope backbone (Fig. 1). The results of these assays revealed that all of the pseudotyped viruses with N-terminal mutations retained the ability to infect cells expressing the FeLV-A receptor, although we did observe that infectivity was reduced by approximately 3 log10 units compared to FeLV-A-61E (Fig. 1).
Binding of virus to MDTF-FLVAR cells was detected with pseudotyped viruses made with envelope clones containing 6D and 7P, although binding was reduced relative to the controls, FeLV-A-61E and FeLV-A/T-81T (Fig. 1). We could not detect binding of virus encoding an envelope with changes at positions 6P and 8M to MDTF-FLVAR at the volume of supernatant tested. However, we could detect a subtle but reproducible shift in fluorescence intensity when we tested higher volumes of this supernatant (data not shown), consistent with its observed ability to infect MDTF-FLVAR cells. As a further test of binding of these mutants, we also examined binding of the SU monomer (data not shown). Again, we observed a reduction in binding with SUs encoding the 7P or 6D change relative to 61E, although binding of the mutant SUs could be more readily detected with the monomeric SU protein than with whole virus. As a control, we also examined MuLV SUs with changes in the PHQ domain; as shown previously, there was no obvious effect of these mutations on binding of MuLV SU to cells expressing the cognate receptor (3).
Whereas FeLV-A/T-81T could infect MDTF-FePit1/FeLIX cells at titers on the order of 106 IU/ml, none of the pseudoviruses containing envelopes with N-terminal changes were able to infect these cells (Fig. 1). We could not detect binding of any pseudoviruses to MDTF-FePit1/FeLIX cells (data not shown), which is consistent with the fact that we have not been able to detect binding of any FeLV-T virus to the FeLV-T receptor complex by our flow cytometric assay (20). Importantly, these data suggest that the N-terminal changes alone have no effect on receptor specificity, although they may reduce infectivity to some extent.
Infection of FeLV-A particles with N-terminal mutations in the envelope via the FLVAR is not enhanced by the addition of soluble FeLV-A RBD fragments. To address whether infection of envelopes containing the N-terminal changes could be enhanced by the addition of soluble FeLV-A RBD fragments, we measured the infectious titers with and without soluble FeLV-A RBD fragments. The titers of 61E6D and 61E7P viruses were approximately 103 IU/ml (Fig. 1 and 2). The addition of soluble 61E RBD resulted in 61E6D or 61E7P viruses giving titers on the order of 102 IU/ml (Fig. 2). The titer of 61E7P-ins-378L virus was on the order of 2 x 105 IU/ml (Fig. 1 and 2), and the addition of soluble 61E RBD fragments resulted in an average titer of 4 x 105 IU/ml (Fig. 2). These data suggest that the addition of the soluble 61E RBD fragments did not increase the titer of viruses containing an N-terminal change alone.
A combination of an N-terminal change and the C-terminal insertion is sufficient to render FeLV-A-61E dualtropic. All of the viruses that had a combination of both the N-terminal change (either at position 6, 7, or 8) and the C-terminal change (a 4-amino-acid insertion with adjacent amino acid change) were able to infect both MDTF-FLVAR cells and MDTF-FePit1/FeLIX cells (Fig. 3A). The titer of each virus pseudotyped with these envelopes on MDTF-FLVAR was reduced by 3 to 4 log10 units relative to FeLV-A-61E, with the pseudovirus expressing the 61E6D-ins envelope showing the greatest reduction. Interestingly, although this virus showed the lowest infectivity on MDTF-FLVAR, 61E6D-ins also exhibited a greater level of binding to the same cells than the other two variants, 61E7P-ins and 61E6P8M-ins (Fig. 3A). This suggests that differences in receptor binding alone are not sufficient to account for the differences in infectivity between viral envelopes. All of the viruses with both an N-terminal change and C-terminal insertion were able to infect MDTF-FePit1/FeLIX cells with titers ranging from 103 to 105 IU/ml. In comparison, the controls FeLV-A/T-81T and FeLV-T-61C had titers of about 106 IU/ml on MDTF-FePit1/FeLIX (Fig. 3A). These data suggest that a small insertion and a total of two amino acid changes—the first at position 6, 7, or 8 of the envelope and the second adjacent to the insertion—are sufficient to extend the receptor specificity of an FeLV-A envelope to the FeLV-T receptor and thereby confer dual-receptor specificity.
A WL change at amino acid position 378 may enhance infection via FeLV-A and FeLV-T receptors. In two independently isolated FeLV-T viruses, the tissue-culture-derived EE(EQPT3)E and the natural isolate FeLV-T-61C, we observed a common mutation at amino acid position 378 in the envelope SU from tryptophan (W) to leucine (L) (12, 14, 30, 34). Introduction of this mutation, alone (61E378L) or together with the insertion (61Eins-378L), virtually abolished infectivity (data not shown). However, when the 378L amino acid change was introduced into envelope clones that had both an N-terminal change and a C-terminal insertion with adjacent amino acid change (61E6D-ins-378L, 61E7P-ins-378L, and 61E6P8M-ins-378L), virus titers on MDTF-FePit1/FeLIX cells were higher by a half log10 unit on average (Fig. 3B). This difference was consistently observed in five independent experiments (data not shown) and suggests that the 378L change can enhance infection via the FeLV-T receptors. The titer on MDTF-FLVAR cells for 61E7P-ins-378L increased about 2 log10 units relative to 61E7P-ins, the clone differing only at 378L (Fig. 3B). This result was also reproducible in five independent experiments (data not shown). However, we did not observe marked differences in fluorescence intensity as a measure of virus binding to MDTF-FLVAR cells with or without the 378L amino acid change (Fig. 3B). We interpret these data to mean that while the 378L amino acid change is not a critical determinant of receptor specificity, it may enhance the infectivity of the viruses for the FeLV-T receptor and, in some cases, for the FeLV-A receptor.
The mutated envelope SU is virus-associated. To test whether the engineered envelope SU constructs were expressed and processed, we performed Western blot analysis on lysates of virus-producing 293T cells and on cell-free virus supernatants. For envelopes with any one of the three N-terminal substitutions (61E6D, 61E7P, or 61E6P8M), we observed both unprocessed envelope gp85 and processed envelope gp70 in lysates, as we expected (Fig. 4A). We also observed gp70 in cell-free supernatants, albeit lower than in supernatants of FeLV-A-61E (Fig. 4B). In cells expressing any of the envelopes containing the C-terminal insertion, there was a very low to undetectable level of processed envelope SU, although levels of unprocessed envelope were not dramatically affected (Fig. 4). On the basis of previous pulse-chase studies of FeLV-A/T variants (14), we interpret these data to indicate that processing is impaired by the presence of the insertion. We could detect the presence of envelope in virus supernatants for most of the mutant viruses, including some unprocessed envelope, which has been reported to be a particular issue with viral pseudotypes generated by transient transfection (16).
We also noted that the added presence of the 378L mutation appeared to increase the amount of envelope detected in the supernatant, although in some cases the increase appeared to be gp85 (Fig. 4). Although some virus supernatants—most strikingly 61E7P and 61E7P-ins—expressed considerably less envelope gp70 than did FeLV-A-61E as measured by Western blot analysis, they were able to infect cells expressing either the FeLV-A or FeLV-T receptor (Fig. 3 and 4).
The presence of gp85 in cell-free supernatants could be due to the presence of contaminating non-virion-associated envelope in cell-free supernatants. To address whether virus purified of free envelope could bind to cells expressing receptors, we purified a subset of virus supernatants by ultracentrifugation through a 20% sucrose cushion to remove any non-virion-associated envelope protein. In this case, we used a panel of viruses containing the 7P mutation, plus or minus the C-terminal insertion and 378L mutation because they had very low levels of gp70 in unpurified virus supernatants (Fig. 4). We measured these purified virions by Western analysis and found, despite differences in expression observed in cell lysates and supernatants, that purified virions of all mutants contain approximately comparable amounts of processed gp70 (Fig. 5A). These purified viruses showed binding properties similar to the virus supernatants that were not purified (Fig. 5B).
To verify that what we detected by Western blotting was gp70 and not gp85, we analyzed 61E and 61E7P along with a control envelope previously shown to be defective for processing [EE(ET3)E] (14) (Fig. 5C). This analysis confirmed that the protein present in purified viruses such as 61E7P was primarily the processed SU protein. In at least six replicates of this assay, we detected gp70 associated with purified viruses for all the viruses examined (data not shown). However, we have found that some mutants—in particular 61Eins and 61Eins-378L (data not shown)—sometimes contained smaller amounts of envelope, even after virion purification. However, despite variations in envelope detection, the relative infectivity between independent virus preparations did not vary (data not shown). Virus purification also did not alter infectivity compared to unpurified virus. This suggests that the differences in virus titers observed in our mutants are not explained simply by differences in envelope expression.
A deletion within the VRA of the RBD abolishes classical receptor specificity, leading to the loss of dual tropism. We have shown that both N- and C-terminal changes are required for dual-receptor specificity for FeLV-A and FeLV-T receptors in the context of FeLV-A-61E. These changes are present in the FeLV-T-61C variant, yet it cannot infect via the FeLV-A receptor. Because the FeLV-T-61C clone also contains a 6-amino-acid deletion within the VRA relative to the FeLV-A-61E and FeLV-A/T viruses, we hypothesized that this deletion was responsible for the loss of specificity for the FeLV-A receptor. We tested this by engineering the 6-amino-acid deletion into the 61E7P-ins-378L clone, which best represented a dualtropic virus of our panel of mutants. The resulting clone, 61E7P--ins-378L, was tested for envelope expression and binding to MDTF-FLVAR cells (Fig. 6). This envelope was expressed in cell-free supernatants at levels comparable to 61E7P-ins-378L, but it failed to infect or bind to MDTF-FLVAR cells (Fig. 6). When we tested infection of MDTF-FePit1/FeLIX cells, we found that both 61E7P-ins-378L and 61E7P--ins-378L could infect with a titer of 105 IU/ml, which was comparable to the titer of the control FeLV-T-61C (Fig. 6). This suggests that the deletion found in FeLV-T-61C is sufficient to abolish binding and infection via the FeLV-A receptor, and it also implies that an intact RBD (without a 6-amino-acid deletion) is required for entry via a classical receptor pathway. However, even when the RBD is disrupted, infection via the nonclassical pathway of Pit1 and FeLIX is maintained, suggesting that the two determinants can be distinct and independent.
The classical receptor specificity of a dualtropic virus can be altered by exchange of the RBD. In order to further test the independence of the determinants of the nonclassical FeLV-T receptor pathway from those of the classical receptor pathway, we replaced the RBD of FeLV-A-61E with the RBD of the subgroup FeLV-C prototype variant, FSC, in the presence of the N- and C-terminal changes that confer FeLV-T receptor specificity. The endogenous murine FLVCR can facilitate infection of FeLV-C-FSC (unpublished observations), and we observed a virus titer of less than 103 IU/ml on MDTF and MDTF-FePit1 cells (Fig. 7A and data not shown). The stable addition of feline FLVCR into MDTF cells (MDTF-FLVCR) allows a virus containing an envelope with the RBD from FeLV-C-FSC in the backbone of FeLV-A-61E (61ESRBD) to infect with a titer of 105 IU/ml (Fig. 7A). This level of infection of MDTF-FLVCR cells is reduced to background levels (103 IU/ml) in the presence of the C-terminal insertion (61ESRBD-ins), and it is abolished in the presence of both the C-terminal insertion and the 378L amino acid change (61ESRBD-ins-378L). However, the addition of an HD change at amino acid position 6 to an envelope encoding the C-terminal insertion can permit infection of MDTF-FePit1/FeLIX at a titer of about 105 IU/ml. Moreover, a QP change at amino acid position 7 along with the C-terminal insertion allows specificity for both the FeLV-C and FeLV-T receptors in the range of approximately 104 to 105 IU/ml. It is interesting that, when paired with the C-terminal insertion and adjacent amino acid change, the N-terminal changes tested were not equivalent in the ability to maintain FeLV-C receptor specificity, although both were able to confer FeLV-T receptor specificity.
All of these mutants exhibited comparable increases in fluorescence intensity as measured by flow cytometry when bound to MDTF cells expressing the receptor for FeLV-C (MDTF-FLVCR) compared to MDTF cells alone (Fig. 7B). In addition, all of the mutant envelopes were expressed, as visualized by Western blot analysis of lysates from transfected 293T cells, although there was some reduction in gp70 (Fig. 7C).
DISCUSSION
The goal of this study was to identify the minimal determinants required to allow a virus to use both the classical FeLV-A receptor and the nonclassical, two-component FeLV-T receptor complex. We found that the determinants for FeLV-T receptor specificity are located outside of the classical receptor binding domain and include at least two SU domains: first, amino acid changes at the N terminus at positions 6, 7, and/or 8, and second, a 4-amino-acid insertion and adjacent amino acid change within the C terminus. We also found that these nonclassical determinants are largely distinct and separable from the RBD, which specifies recognition and entry via the classical receptor pathway used by FeLV-A and FeLV-C (Fig. 8).
Viruses containing envelopes with N-terminal changes at amino acid positions 6, 7, and 8 (where the conserved histidine of the PHQ motif is at position 6 in FeLV) maintain specificity for the FeLV-A receptor, although the efficiency of infection was reduced compared to FeLV-A-61E. These findings reveal some differences between FeLV and MuLV because in MuLV the conserved histidine of the PHQ motif at position 8 (H8) appears to have a much more pronounced effect on infectivity, and in some cases infection could not be detected (3, 24, 42). As a result, MuLV viruses with disruptions of H8 have been described as defective. However, they can bind efficiently to their cognate receptors and they can also be rescued in trans by the addition of a wild-type soluble RBD fragment, in apparent similarity to naturally occurring FeLV-T (4, 5, 21, 24). Here we show that, FeLVs containing N-terminal changes, including the analogous H6, did not acquire the ability to infect via the nonclassical receptor pathway of Pit1 and FeLIX. Furthermore, infection of MDTF-FLVAR by FeLVs with N-terminal changes was not enhanced after the addition of soluble FeLV-A RBD, as was seen with the defective MuLVs. Thus, there may be some differences in the role of the conserved histidine at position 6 in the FeLV SU in receptor specificity compared to the H8 in the MuLV SU. These differences may depend in part on sequence context, as we have found that the infectivity of some of the mutants that encode both N- and C-terminal changes can be increased with the addition of 61E sRBD (Cheng, Anderson, and Overbaugh, unpublished data).
For FeLV, both N- and C-terminal changes are required to render a dual-receptor-specific phenotype; viruses expressing an envelope with the insertion with an adjacent amino acid change and any one of the three N-terminal changes can infect cells using both the FeLV-A and the FeLV-T receptors. Although there was variability in the amount of envelope detected in cell-free supernatants among the different mutant viruses, we saw no obvious correlation between differences in either expression or binding and relative infectivity. Thus, the impact of the N-terminal mutation and C-terminal insertion on infectivity appears not to be simply the result of effects on steady-state expression levels of the envelope SU. The reduced infectivity of constructs containing the N- and C-terminal changes on MDTF-FePit1/FeLIX compared to both FeLV-T-61C and FeLV-A/T-81T suggests that other changes may enhance infection via the FeLV-T receptor complex. For example, viruses with the 378L change had slightly increased titers in MDTF-FePit1/FeLIX in some cases. Thus, while the N-terminal changes at positions 6, 7, and 8 and the C-terminal insertion are together sufficient to confer the dual-receptor phenotype, changes at other residues may increase infectivity for one or both receptors.
In order to test whether the nonclassical determinants are distinct from the classical RBD determinants for FeLV receptor specificity, we engineered a 6-amino-acid deletion within the VRA of RBD present in FeLV-T-61C into an envelope also containing N- and C-terminal modifications. This envelope could neither bind MDTF-FLVAR cells nor infect them, presumably due to the disruption of the RBD determinants, yet it maintained the ability to infect MDTF-FePit1/FeLIX cells. As a further test, we substituted the RBD from FeLV-C-FSC into the FeLV-A envelope SU backbone containing different N-terminal changes along with the C-terminal insertion. We found that envelopes with one combination of N- and C-terminal changes had specificity for both FeLV-C and FeLV-T receptors, again indicating the independence of the nonclassical from the classical receptor determinants. However, envelopes with another combination of N- and C-terminal changes could infect using the FeLV-T receptors but not the FeLV-C receptor, indicating that the particular combination of N- and C-terminal changes may be important for some interactions with the classical FeLV-C receptor and that the different combinations may not elicit the same changes. All of the mutant envelopes containing the FeLV-C-FSC RBD were capable of binding MDTF-FLVCR cells regardless of receptor specificity, suggesting that the effects of N- and C-terminal mutations were on events after binding, such as fusion or entry.
Critical interactions between the N and C termini of SU have been implicated in the classical receptor pathway (21, 23, 25, 26, 28). A model has emerged from various studies in which the gammaretroviral RBD binds to the cognate receptor, signals to a region of the C terminus via the proline-rich region, and induces a conformational change in which residues in the N and C termini interact and trigger the dissociation of the SU from TM and allow the initiation of fusion (4, 21, 26). We found changes in the same regions of the N and C termini of the FeLV SU to be the determinants of infection via the nonclassical FeLV-T receptor pathway (14), perhaps indicating that naturally occurring viruses, such as FeLV-T, that use the nonclassical receptors may exist in a conformation different from those that use a classical receptor. These models for gammaretroviruses remain speculative; however, they are somewhat similar to models that have been proposed for some CD4-independent human immunodeficiency virus/simian immunodeficiency virus envelopes, in which the native structure of CD4-independent SUs are thought to exist in a conformation that is induced in CD4-dependent viruses upon SU binding to CD4 (9, 27). In FeLV, the changes required for nonclassical receptor use can maintain the integrity of the RBD and entry via a classical receptor (such as FLVAR or FLVCR), resulting in dual-receptor-specific variants. Thus, these envelopes that are capable of utilizing two distinct receptor pathways may shed light on the sequential events required for fusion and entry of gammaretroviruses.
ACKNOWLEDGMENTS
We thank Gretchen Strauch, Ramon Mendoza, Mario Pineda, and Colin Pritchard for comments on the manuscript, as well as Brannon Orton, Ramon Mendoza, and Nick Rabena for technical assistance.
This work was supported by NIH grant CA 51080. H.H.C. was supported by a fellowship from the Paul Allen Foundation.
REFERENCES
Anderson, M. M., A. S. Lauring, C. C. Burns, and J. Overbaugh. 2000. Identification of a cellular cofactor required for infection by feline leukemia virus. Science 287:1828-1830.
Anderson, M. M., A. S. Lauring, S. Robertson, C. Dirks, and J. Overbaugh. 2001. Feline Pit2 functions as a receptor for subgroup B feline leukemia viruses. J. Virol. 75:10563-10572.
Bae, Y., S. M. Kingsman, and A. J. Kingsman. 1997. Functional dissection of the Moloney murine leukemia virus envelope protein gp70. J. Virol. 71:2092-2099.
Barnett, A. L., and J. M. Cunningham. 2001. Receptor binding transforms the surface subunit of the mammalian C-type retrovirus envelope protein from an inhibitor to an activator of fusion. J. Virol. 75:9096-9105.
Barnett, A. L., R. A. Davey, and J. M. Cunningham. 2001. Modular organization of the Friend murine leukemia virus envelope protein underlies the mechanism of infection. Proc. Natl. Acad. Sci. USA 98:4113-4118.
Barnett, A. L., D. L. Wensel, W. Li, D. Fass, and J. M. Cunningham. 2003. Structure and mechanism of a coreceptor for infection by a pathogenic feline retrovirus. J. Virol. 77:2717-2729.
Battini, J. L., O. Danos, and J. M. Heard. 1995. Receptor-binding domain of murine leukemia virus envelope glycoproteins. J. Virol. 69:713-719.
Battini, J. L., J. M. Heard, and O. Danos. 1992. Receptor choice determinants in the envelope glycoproteins of amphotropic, xenotropic, and polytropic murine leukemia viruses. J. Virol. 66:1468-1475.
Bhattacharya, J., P. J. Peters, and P. R. Clapham. 2003. CD4-independent infection of HIV and SIV: implications for envelope conformation and cell tropism in vivo. AIDS 17(Suppl. 4):S35-S43.
Boomer, S., M. Eiden, C. C. Burns, and J. Overbaugh. 1997. Three distinct envelope domains, variably present in subgroup B feline leukemia virus recombinants, mediate Pit1 and Pit2 receptor recognition. J. Virol. 71:8116-8123.
Burns, C. C., M. L. Poss, E. Thomas, and J. Overbaugh. 1995. Mutations within a putative cysteine loop of the transmembrane protein of an attenuated immunodeficiency-inducing feline leukemia virus variant inhibit envelope protein processing. J. Virol. 69:2126-2132.
Donahue, P. R., S. L. Quackenbush, M. V. Gallo, C. M. C. deNoronha, J. Overbaugh, E. A. Hoover, and J. I. Mullins. 1991. Viral genetic determinants of T-cell killing and immunodeficiency disease induction by the feline leukemia virus FeLV-FAIDS. J. Virol. 65:4461-4469.
Grant, C. K., B. J. Ernisse, O. Jarrett, and F. R. Jones. 1983. Feline leukemia virus envelope gp70 of subgroups B and C defined by monoclonal antibodies with cytotoxic and neutralizing functions. J. Immunol. 131:3042-3048.
Gwynn, S. R., F. C. Hankenson, A. S. Lauring, J. L. Rohn, and J. Overbaugh. 2000. Feline leukemia virus envelope sequences that affect T-cell tropism and syncytium formation are not part of known receptor-binding domains. J. Virol. 74:5754-5761.
Heard, J. M., and O. Danos. 1991. An amino-terminal fragment of the Friend murine leukemia virus envelope glycoprotein binds the ecotropic receptor. J. Virol. 65:4026-4032.
Herrera, C., P. J. Klasse, E. Michael, S. Kake, K. Barnes, C. W. Kibler, L. Campbell-Gardener, Z. Si, J. Sodroski, J. P. Moore, and S. Beddows. 2005. The impact of envelope glycoprotein cleavage on the antigenicity, infectivity, and neutralization sensitivity of Env-pseudotyped human immunodeficiency virus type 1 particles. Virology 338:154-172.
Ho, S. N., H. D. Hunt, R. M. Horton, J. K. Pullen, and L. R. Pease. 1989. Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene 77:51-59.
Landau, N. R., K. A. Page, and D. R. Littman. 1991. Pseudotyping with human T-cell leukemia virus type I broadens the human immunodeficiency virus host range. J. Virol. 65:162-169.
Lander, M. R., and S. K. Chattopadhyay. 1984. A Mus dunni cell line that lacks sequences closely related to endogenous murine leukemia viruses and can be infected by ecotropic, amphotropic, xenotropic, and mink cell focus-forming viruses. J. Virol. 52:695-698.
Lauring, A. S., M. M. Anderson, and J. Overbaugh. 2001. Specificity in receptor usage by T-cell-tropic feline leukemia viruses: implications for the in vivo tropism of immunodeficiency-inducing variants. J. Virol. 75:8888-8898.
Lavillette, D., B. Boson, S. J. Russell, and F. L. Cosset. 2001. Activation of membrane fusion by murine leukemia viruses is controlled in cis or in trans by interactions between the receptor-binding domain and a conserved disulfide loop of the carboxy terminus of the surface glycoprotein. J. Virol. 75:3685-3695.
Lavillette, D., and D. Kabat. 2004. Porcine endogenous retroviruses infect cells lacking cognate receptors by an alternative pathway: implications for retrovirus evolution and xenotransplantation. J. Virol. 78:8868-8877.
Lavillette, D., A. Ruggieri, B. Boson, M. Maurice, and F. L. Cosset. 2002. Relationship between SU subdomains that regulate the receptor-mediated transition from the native (fusion-inhibited) to the fusion-active conformation of the murine leukemia virus glycoprotein. J. Virol. 76:9673-9685.
Lavillette, D., A. Ruggieri, S. J. Russell, and F. L. Cosset. 2000. Activation of a cell entry pathway common to type C mammalian retroviruses by soluble envelope fragments. J. Virol. 74:295-304.
Lu, C. W., and M. J. Roth. 2001. Functional characterization of the N termini of murine leukemia virus envelope proteins. J. Virol. 75:4357-4366.
Lu, C. W., and M. J. Roth. 2003. Functional interaction between the N- and C-terminal domains of murine leukemia virus surface envelope protein. Virology 310:130-140.
Myszka, D. G., R. W. Sweet, P. Hensley, M. Brigham-Burke, P. D. Kwong, W. A. Hendrickson, R. Wyatt, J. Sodroski, and M. L. Doyle. 2000. Energetics of the HIV gp120-CD4 binding reaction. Proc. Natl. Acad. Sci. USA 97:9026-9031.
O'Reilly, L., and M. J. Roth. 2003. Identification of conformational and cold-sensitive mutations in the MuLV envelope protein. Virology 312:337-349.
Ott, D., and A. Rein. 1992. Basis for receptor specificity of nonecotropic murine leukemia virus surface glycoprotein gp70SU. J. Virol. 66:4632-4638.
Overbaugh, J., P. R. Donahue, S. L. Quackenbush, E. A. Hoover, and J. I. Mullins. 1988. Molecular cloning of a feline leukemia virus that induces fatal immunodeficiency disease in cats. Science 239:906-910.
Overbaugh, J., A. D. Miller, and M. V. Eiden. 2001. Receptors and entry cofactors for retroviruses include single and multiple transmembrane-spanning proteins as well as newly described glycophosphatidylinositol-anchored and secreted proteins. Microbiol. Mol. Biol. Rev. 65:371-389.
Quigley, J. G., C. C. Burns, M. M. Anderson, E. D. Lynch, K. M. Sabo, J. Overbaugh, and J. L. Abkowitz. 2000. Cloning of the cellular receptor for feline leukemia virus subgroup C (FeLV-C), a retrovirus that induces red cell aplasia. Blood 95:1093-1099.
Quigley, J. G., Z. Yang, M. T. Worthington, J. D. Phillips, K. M. Sabo, D. E. Sabath, C. L. Berg, S. Sassa, B. L. Wood, and J. L. Abkowitz. 2004. Identification of a human heme exporter that is essential for erythropoiesis. Cell 118:757-766.
Rohn, J. L., M. S. Moser, S. R. Gwynn, D. N. Baldwin, and J. Overbaugh. 1998. In vivo evolution of a novel, syncytium-inducing and cytopathic feline leukemia virus variant. J. Virol. 72:2686-2696.
Sugai, J., M. Eiden, M. M. Anderson, N. Van Hoeven, C. D. Meiering, and J. Overbaugh. 2001. Identification of envelope determinants of feline leukemia virus subgroup B that permit infection and gene transfer to cells expressing human Pit1 or Pit2. J. Virol. 75:6841-6849.
Tailor, C. S., and D. Kabat. 1997. Variable regions A and B in the envelope glycoproteins of feline leukemia virus subgroup B and amphotropic murine leukemia virus interact with discrete receptor domains. J. Virol. 71:9383-9391.
Tailor, C. S., D. Lavillette, M. Marin, and D. Kabat. 2003. Cell surface receptors for gammaretroviruses. Curr. Top. Microbiol. Immunol. 281:29-106.
Tailor, C. S., B. J. Willett, and D. Kabat. 1999. A putative cell surface receptor for anemia-inducing feline leukemia virus subgroup C is a member of a transporter superfamily. J. Virol. 73:6500-6505.
Takeuchi, Y., R. G. Vile, G. Simpson, B. O'Hara, M. K. Collins, and R. A. Weiss. 1992. Feline leukemia virus subgroup B uses the same cell surface receptor as gibbon ape leukemia virus. J. Virol. 66:1219-1222.
Ting, Y. T., C. A. Wilson, K. B. Farrell, G. J. Chaudry, and M. V. Eiden. 1998. Simian sarcoma-associated virus fails to infect Chinese hamster cells despite the presence of functional gibbon ape leukemia virus receptors. J. Virol. 72:9453-9458.
Wensel, D. L., W. Li, and J. M. Cunningham. 2003. A virus-virus interaction circumvents the virus receptor requirement for infection by pathogenic retroviruses. J. Virol. 77:3460-3469.
Zavorotinskaya, T., and L. M. Albritton. 1999. Suppression of a fusion defect by second site mutations in the ecotropic murine leukemia virus surface protein. J. Virol. 73:5034-5042.(Heather H. Cheng, Maria M)
Department of Microbiology, University of Washington
Division of Human Biology, Fred Hutchinson Cancer Research Center, Seattle, Washington
ABSTRACT
Gammaretroviruses, including the subgroups A, B, and C of feline leukemia virus (FeLV), use a multiple-membrane-spanning transport protein as a receptor. In some cases, such as FeLV-T, a nonclassical receptor that includes both a transport protein (Pit1) and a soluble cofactor (FeLIX) is required for entry. To define which regions confer specificity to classical versus nonclassical receptor pathways, we engineered mutations found in either FeLV-A/T or FeLV-T, individually and in combination, into the backbone of the transmissible form of the virus, FeLV-A. The receptor specificities of these viruses were tested by measuring infection and binding to cells expressing the FeLV-A receptor or the FeLV-T receptors. FeLV-A receptor specificity was maintained when changes at amino acid position 6, 7, or 8 of the mature envelope glycoprotein were introduced, although differences in infection efficiency were observed. When these N-terminal mutations were introduced together with a C-terminal 4-amino-acid insertion and an adjacent amino acid change, the resulting viruses acquired FeLV-T receptor specificity. Additionally, a WL change at amino acid position 378, although not required, enhanced infectivity for some viruses. Thus, we have found that determinants in the N and C termini of the envelope surface unit can direct entry via the nonclassical FeLV-T receptor pathway. The region that has been defined as the receptor binding domain of gammaretroviral envelope proteins determined entry via the FeLV-A receptor independently of the presence of the N- and C-terminal FeLV-T receptor determinants.
INTRODUCTION
Receptors for gammaretroviruses are multiple-membrane-spanning proteins involved in small-molecule or ion transport (31, 37). For most retroviruses, it appears that these receptors are both necessary and sufficient for virus binding and entry and are referred to here as the "classical" receptors. The T-cell-tropic feline leukemia virus (FeLV-T) was the first identified example of a naturally occurring gammaretrovirus that requires two host proteins for entry into the cell, Pit1 and FeLIX (1). These two proteins, the membrane-spanning receptor Pit1 and the soluble host protein FeLIX, constitute what we call a "nonclassical" receptor. More recently, it has been shown that other naturally arising retroviruses, mink cell focus-forming murine leukemia virus and porcine endogenous retroviruses, can use a nonclassical receptor for entry, although they may also infect cells using classical receptors (22, 41).
Studies of gammaretroviruses have indicated that the classical receptor determinants of the envelope protein lie in the N-terminal region of the envelope surface unit (SU) in variable region A (VRA) and VRB. This domain was first identified in murine leukemia virus (MuLV) and has been defined as a region approximately encompassing amino acid positions 50 to 190 in the mature SU, depending on the strain (3, 8, 29). In MuLV, as well as in related gammaretroviruses such as FeLV, VRA is the primary determinant for receptor specificity while VRB serves as a secondary determinant (8, 10, 36). Because the N terminus of the SU encompassing VRA and VRB is sufficient to bind specifically to cognate receptors (3, 7, 15), it is considered the receptor binding domain (RBD) in gammaretroviruses.
Initial infection of cats by subgroup A FeLV (FeLV-A), the transmissible form of the virus, occurs via the recently identified FeLV-A receptor (referred to here as FLVAR) (R. Mendoza, M. Anderson, and J. Overbaugh, submitted for publication). Once infection is established, FeLV-A can undergo changes to give rise to the other FeLV subgroups with distinct receptor usage (31): FeLV-B, which uses the related phosphate transporters Pit1 and Pit2 (2, 39); FeLV-C, which uses the heme transporter FLVCR (32, 33, 38); and FeLV-T, which requires both Pit1 and FeLIX (1, 31).
FeLV-A and FeLV-T clones (61E and 61C, respectively) were originally isolated from a single cat with FeLV-feline AIDS-induced immunodeficiency (30), and they have approximately 96% identity (12, 30). Using chimeras, the T-cell specificity and cytopathicity of FeLV-T-61C were mapped to two regions within the envelope SU: amino acid changes within the N-terminal 109 amino acids, including an N-terminal deletion within the VRA of RBD, and a C-terminal 6-amino-acid insertion and adjacent amino acid change relative to FeLV-A (12).
FeLV-81T is a clone isolated from a cat inoculated with FeLV-A-61E that also causes T-cell killing (34). FeLV-81T can use both the FeLV-A and the FeLV-T receptors, making it dual-receptor specific (reference 34 and unpublished observations). Interestingly, FeLV-A/T-81T contains a 4-amino-acid insertion in the exact location where the C-terminal 6-amino-acid insertion in FeLV-T-61C is found. Chimeras containing the N terminus of the SU from FeLV-A-61E and the C terminus of the SU (including the insertion) from FeLV-81T were replication defective, but replication was restored by a single mutation at amino acid position 7 of a tissue culture-adapted virus, EE(EQPT3)E (14). Thus, all T-cell-tropic clones that we have characterized to date—FeLV-T-61C, FeLV-A/T-81T, and EE(EQPT3)E—contain an N-terminal amino acid change (at position 6, 7, or 8) and either a 6- or 4-amino-acid insertion (and adjacent amino acid change) in an identical location in the C terminus(30, 34). These changes outside of the RBD are associated with T-cell tropism in vitro, suggesting that they may also confer specificity for the FeLV-T receptor complex because T cells express high levels of Pit1 and FeLIX (20).
The N-terminal changes associated with the T-cell tropism of FeLV-T are in a highly conserved PHQ motif in gammaretroviruses (3, 24). In MuLV, substitutions or deletions of the histidine within the motif at amino acid position 8 (H8) result in defective viruses (4, 21, 24). However, these viruses can still bind to their cognate receptors, signifying a role for MuLV H8 in postbinding events (3, 4, 21, 42). These defective MuLVs can be rescued by the trans addition of soluble RBD fragments that contain the wild-type histidine (5, 6, 21, 24). The laboratory-engineered soluble RBDs that rescue MuLVs lacking the critical H8 resemble the soluble cofactors of nonclassical receptors that function in natural infections, such as FeLIX (1). This suggests that there may be common mechanisms of entry underlying these laboratory-derived MuLV mutants and circulating variants such as FeLV-T. However, because naturally arising FeLV-T variants acquire multiple mutations throughout the envelope sequence that may contribute to their fitness in the host, it is difficult to predict whether the PHQ domain in FeLV-T serves the same function as it does for MuLV.
In this study, we sought to dissect the determinants for dual-receptor specificity—the ability to use both a classical receptor (FLVAR) and a nonclassical receptor (Pit1 and FeLIX)—by examining the effects of mutations at the N terminus (amino acid positions 6, 7, and 8), the C-terminal insertion and adjacent amino acid change, and the VRA within the context of either the FeLV-A or FeLV-C envelope SU backbone on binding and infection with various FeLV receptors. In this manner, we have identified the determinants for classical and nonclassical entry of a gammaretrovirus, and we show that these determinants can be independent of each other.
MATERIALS AND METHODS
Cell culture. The MDTF murine fibroblast cell line has been described previously (19), as has the MDTF-FePit1 line (2). Construction of the MDTF cell line expressing the receptor for FeLV-A will be described elsewhere (Mendoza, Anderson, and Overbaugh, submitted). The receptor for FeLV-C (FLVCR) was introduced into MDTF cells by transduction using virus particles packaging pMSCV(FLVCR)neo (32). Particles were generated in 293T human embryonic kidney (HEK) cells using transient transfection with the following plasmids: amphotropic MuLV envelope (SV-A-MLV-env [18]), a construct encoding FeLV gag and pol (61E-LTR- gag-pol [35]), and pMSCV(FLVCR)neo (32), using the transduction and selection method described previously (2). All cell lines, including MDTF murine fibroblasts and 293T HEK fibroblasts, were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum, 100 U penicillin/ml, 100 μg streptomycin/ml, 0.25 mg amphotericin fungicide/ml, and 2 mM L-glutamine (complete DMEM). The MDTF-FePit1, MDTF-FLVAR, and MDTF-FLVCR cell lines were maintained in complete DMEM with 1.0 mg G418/ml (Geneticin; Gibco, Carlsbad, CA).
Construction of mutant virus clones. Envelope expression clones, in which the FeLV-A-61E envelope gene (35) or the FeLV-T-61C envelope gene (20) are expressed in the pcDNA3.1 zeo(–) vector, have been described previously. An analogous clone used to express the FeLV-A/T-81T (clone 109) envelope was made by the same method. Briefly, the EET109E plasmid (14) was digested with XhoI and SacII, and an approximately 1.5-kb fragment was isolated from an agarose gel, purified, and ligated to a 5-kb fragment of pcDNA3.1-61Eenv (35) that had been digested with the same restriction enzymes and purified in the same manner.
The amino acid modifications 7P, 6D, 6P-8M, ins (HD at amino acid position 350 adjacent to a GESQ insertion), and 378L in the mature envelope SU in the 61Eenv/pcDNA3.1 Zeo(–) plasmid (35) were introduced by site-directed mutagenesis (Quick-Change Mutagenesis; Invitrogen). Sequences of the mutagenic primers are available upon request. All modified plasmid constructs were verified by sequencing. Standard subcloning techniques were used to create chimeras expressing mutations at both the N terminus (7P, 6D, or 6P-8M) and C terminus (ins or ins-378L). Modified FeLV RBDs were cloned into the pcDNA3.1 Zeo(–) expression vector using a 5' XhoI site and a 3' Bsu361 site, as previously described (35).
A FeLV-C-FSC envelope expression clone consisting of the FeLV-A-61E backbone and containing the FeLV-C-FSC RBD was made using the XhoI and SacII sites in a manner similar to the construction of the FeLV-A/T-81T envelope expression clone described above. Analogous chimeric envelopes were constructed that expressed one of the two N-terminal changes in combination with one of the two C-terminal mutations described above. As before, a 5' XhoI site and a 3' Bsu361 site flanking the FeLV RBD sequence in pcDNA3.1 Zeo(–) were used. However, due to an additional Bsu36I site in the RBD of FeLV-C-FSC, site-directed mutagenesis (Quick-Change Mutagenesis; Invitrogen) was used to remove the Bsu361 site internal to the FeLV-C-FSC RBD, resulting in a silent mutation (primer sequences available upon request).
The 61E7P--ins-378L construct, containing a 6-amino-acid deletion similar to one found in FeLV-T-61C, was made in the context of the 61E7P-ins-378L mutant backbone using overlap extension PCR (17). Sequences of the mutagenic primers are available upon request. The first set of primers amplified a fragment 5' of the deletion, and the second set of primers amplified a fragment 3' of the deletion. In the second round of PCR, the 5' fragment and 3' fragment were combined to prime each other and form a 1.5-kb fragment, which was cloned back into the full-length FeLV-A-61E envelope expression vector (35) using the XhoI and SacII sites as described above.
Virus production and infection assays. To maintain mutations of interest and avoid introduction of new ones, we created pseudotyped viruses that were limited to a single round of replication. Virus was made by calcium phosphate transfection of 293T cells (Mammalian Transfection Kit; Stratagene, La Jolla, CA) with three plasmids: 4 μg of FeLV-61E-gag-pol (61E-LTR--gag-pol [35]), 4 μg of MuLV-derived reporter gene that expresses -galactosidase (pRT43.2Tnlsgal-1 [40]), and 10 μg of FeLV-envelope construct. One day post-transfection, the cells were washed three times in phosphate-buffered saline, and 7 ml complete DMEM was added. Cell-free virus supernatant was collected at 2 days post-transfection and filtered through a 0.2-μm filter, aliquoted, and stored at –80°C.
Viral infectivity was assayed using a single-cycle infection assay described previously (2). Cells were seeded at 2 x 104 cells per well of a 24-well dish and incubated for 24 h, after which virus and complete DMEM containing 4 μg Polybrene per ml were added to a total volume of 1 ml per well. Two days postinfection, cells were fixed and stained for -galactosidase activity, and -galactosidase-positive blue foci were scored as previously described (14).
In some cases, we added 100 μl of conditioned medium containing soluble cofactors to cultured cells just prior to adding virus. Soluble cofactors were generated, as previously described, by calcium phosphate transfection of 293T cells with plasmids encoding either FeLIX (20) or a soluble RBD derived from FeLV-A-61E envelope that terminated at amino acid 231, analogous to FeLIX (61E sRBD; Anderson, Cheng, and Overbaugh, unpublished data). The expression of FeLIX and 61E sRBD was verified by Western blot analysis (data not shown). For these infection studies, we used fluorescence-activated cell-sorted MDTF-FLVAR cells that expressed high levels of FLVAR, as described elsewhere (Cheng, Mendoza, and Overbaugh, unpublished data) to ensure that receptor levels were not limiting.
Virus ultracentrifugation. One milliliter of virus supernatant (made as described above) was layered over 750 μl of a solution of 20% sucrose in TNE (10 mM Tris, 150 mM NaCl, and 1 mM EDTA). The sample volume was brought up to 5 ml and balanced with complete DMEM in 13- by 51-mm polyallomer tubes (Beckman, Palo Alto, CA), and it was subsequently ultracentrifuged for 2 h at 24,000 rpm at 4°C using an Sw55 rotor in an Optima-L 90K Ultracentrifuge (Beckman Coulter, Palo Alto, CA). When the spin was completed, supernatants were discarded by inverting the tubes, followed by draining for 10 min. Samples were then resuspended in 20 μl of standard Laemmli loading buffer for Western blot analysis.
Immunoprecipitation and Western blot analysis. Two milliliters of each virus supernatant was precleared for three hours at 4°C using 300 μl of a slurry containing Protein A beads (Sigma, St. Louis, MO). Samples were centrifuged at 14,000 rpm at 4°C, and the supernatant was immunoprecipitated for 2.5 h at 4°C using 50 μl of Protein A beads conjugated to the goat anti-FeLV polyclonal antiserum 77S (Quality Biotech, Inc., Resource Lab, Camden, NJ) (11). Protein-bound beads were then washed three times with a solution of 0.1% NP-40 and 0.1% Triton X-100 in phosphate-buffered saline and resuspended in 20 μl of sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis loading buffer, and the entire sample was resolved on an SDS-10% polyacrylamide gel. Ultracentrifuged virus samples (prepared as described above) were boiled and resolved on an SDS-10% polyacrylamide gel, followed by wet transfer to a polyvinylidene difluoride blotting membrane (Millipore, Bedford, MA).
Western blot analysis was performed using the anti-FeLV gp70 monoclonal C11D8 antibody (Custom Monoclonal Antibodies, Sacramento, CA) (13) and a horseradish peroxidase-conjugated goat anti-mouse secondary antibody (Bio-Rad, Hercules, CA). Bound antibodies were detected by chemiluminescence (Amersham, Piscataway, NJ).
Binding assay. Binding of virus to cells was performed as previously described (20). Briefly, 2 x 105 cells were incubated with 1 ml of cell-free supernatant and 100 μl of ultracentrifuge-purified virus in a 2-ml total volume for 45 min at 37°C. Incubations with the primary antibody, anti-FeLV gp70 monoclonal C11D8 or anti-hemagglutinin (HA) monoclonal HA.11 (Covance, Berkeley, CA), and the secondary antibody, r-phycoerythrin-conjugated goat anti-mouse (DAKO, Carpinteria, CA), were performed as described previously, except that antibody incubation volumes were at half the previous levels (20).
RESULTS
Individual effects of a C-terminal insertion or N-terminal mutations on FeLV-A and FeLV-T receptor specificity. FeLV-T-61C and the dual-receptor-specific FeLV-A/T-81T contain either a 6- or 4-amino-acid insertion and an adjacent amino acid change in an identical location in the C terminus of the envelope SU (30, 34). We engineered a virus envelope containing only the insertion and adjacent amino acid change in the FeLV-A-61E backbone (61Eins). In order to assess receptor specificity, we tested a pseudotyped virus expressing the mutant envelope in a single-cycle infection assay. Virus was introduced to cell lines that express either the FeLV-A receptor, MDTF-FLVAR (Mendoza, Anderson, and Overbaugh, submitted), or the FeLV-T receptors, MDTF-FePit1, in the presence of FeLIX (henceforth referred to as MDTF-FePit1/FeLIX for simplicity) (Fig. 1). We observed that 61Eins could infect MDTF-FLVAR cells despite a 2-log10 reduction in titer compared to FeLV-A-61E (Fig. 1). We could also detect a level of binding of 61Eins to MDTF-FLVAR comparable to that of the FeLV-A-61E parental virus (Fig. 1). Importantly, this virus did not infect MDTF-FePit1/FeLIX cells. Thus, viruses with the C-terminal insertion and adjacent amino acid change maintained specificity for the FeLV-A receptor but were unable to infect via the FeLV-T receptor.
All FeLV-T and dual-receptor-specific FeLV-A/T variants characterized to date have an N-terminal change of one or more amino acids at positions 6, 7, and 8 in the mature SU relative to FeLV-A-61E. These include a single amino acid change at position 6 (HD) found in the FeLV-A/T-81T variant (34), changes at both positions 6 and 8 (6HP, 8IM) found in the prototype FeLV-T-61C variant (30), or a change at position 7 (QP) found in a tissue culture-derived FeLV-A/T variant, EE(EQPT3)E (14). We used site-directed mutagenesis to introduce these mutations individually into the FeLV-A-61E envelope backbone (Fig. 1). The results of these assays revealed that all of the pseudotyped viruses with N-terminal mutations retained the ability to infect cells expressing the FeLV-A receptor, although we did observe that infectivity was reduced by approximately 3 log10 units compared to FeLV-A-61E (Fig. 1).
Binding of virus to MDTF-FLVAR cells was detected with pseudotyped viruses made with envelope clones containing 6D and 7P, although binding was reduced relative to the controls, FeLV-A-61E and FeLV-A/T-81T (Fig. 1). We could not detect binding of virus encoding an envelope with changes at positions 6P and 8M to MDTF-FLVAR at the volume of supernatant tested. However, we could detect a subtle but reproducible shift in fluorescence intensity when we tested higher volumes of this supernatant (data not shown), consistent with its observed ability to infect MDTF-FLVAR cells. As a further test of binding of these mutants, we also examined binding of the SU monomer (data not shown). Again, we observed a reduction in binding with SUs encoding the 7P or 6D change relative to 61E, although binding of the mutant SUs could be more readily detected with the monomeric SU protein than with whole virus. As a control, we also examined MuLV SUs with changes in the PHQ domain; as shown previously, there was no obvious effect of these mutations on binding of MuLV SU to cells expressing the cognate receptor (3).
Whereas FeLV-A/T-81T could infect MDTF-FePit1/FeLIX cells at titers on the order of 106 IU/ml, none of the pseudoviruses containing envelopes with N-terminal changes were able to infect these cells (Fig. 1). We could not detect binding of any pseudoviruses to MDTF-FePit1/FeLIX cells (data not shown), which is consistent with the fact that we have not been able to detect binding of any FeLV-T virus to the FeLV-T receptor complex by our flow cytometric assay (20). Importantly, these data suggest that the N-terminal changes alone have no effect on receptor specificity, although they may reduce infectivity to some extent.
Infection of FeLV-A particles with N-terminal mutations in the envelope via the FLVAR is not enhanced by the addition of soluble FeLV-A RBD fragments. To address whether infection of envelopes containing the N-terminal changes could be enhanced by the addition of soluble FeLV-A RBD fragments, we measured the infectious titers with and without soluble FeLV-A RBD fragments. The titers of 61E6D and 61E7P viruses were approximately 103 IU/ml (Fig. 1 and 2). The addition of soluble 61E RBD resulted in 61E6D or 61E7P viruses giving titers on the order of 102 IU/ml (Fig. 2). The titer of 61E7P-ins-378L virus was on the order of 2 x 105 IU/ml (Fig. 1 and 2), and the addition of soluble 61E RBD fragments resulted in an average titer of 4 x 105 IU/ml (Fig. 2). These data suggest that the addition of the soluble 61E RBD fragments did not increase the titer of viruses containing an N-terminal change alone.
A combination of an N-terminal change and the C-terminal insertion is sufficient to render FeLV-A-61E dualtropic. All of the viruses that had a combination of both the N-terminal change (either at position 6, 7, or 8) and the C-terminal change (a 4-amino-acid insertion with adjacent amino acid change) were able to infect both MDTF-FLVAR cells and MDTF-FePit1/FeLIX cells (Fig. 3A). The titer of each virus pseudotyped with these envelopes on MDTF-FLVAR was reduced by 3 to 4 log10 units relative to FeLV-A-61E, with the pseudovirus expressing the 61E6D-ins envelope showing the greatest reduction. Interestingly, although this virus showed the lowest infectivity on MDTF-FLVAR, 61E6D-ins also exhibited a greater level of binding to the same cells than the other two variants, 61E7P-ins and 61E6P8M-ins (Fig. 3A). This suggests that differences in receptor binding alone are not sufficient to account for the differences in infectivity between viral envelopes. All of the viruses with both an N-terminal change and C-terminal insertion were able to infect MDTF-FePit1/FeLIX cells with titers ranging from 103 to 105 IU/ml. In comparison, the controls FeLV-A/T-81T and FeLV-T-61C had titers of about 106 IU/ml on MDTF-FePit1/FeLIX (Fig. 3A). These data suggest that a small insertion and a total of two amino acid changes—the first at position 6, 7, or 8 of the envelope and the second adjacent to the insertion—are sufficient to extend the receptor specificity of an FeLV-A envelope to the FeLV-T receptor and thereby confer dual-receptor specificity.
A WL change at amino acid position 378 may enhance infection via FeLV-A and FeLV-T receptors. In two independently isolated FeLV-T viruses, the tissue-culture-derived EE(EQPT3)E and the natural isolate FeLV-T-61C, we observed a common mutation at amino acid position 378 in the envelope SU from tryptophan (W) to leucine (L) (12, 14, 30, 34). Introduction of this mutation, alone (61E378L) or together with the insertion (61Eins-378L), virtually abolished infectivity (data not shown). However, when the 378L amino acid change was introduced into envelope clones that had both an N-terminal change and a C-terminal insertion with adjacent amino acid change (61E6D-ins-378L, 61E7P-ins-378L, and 61E6P8M-ins-378L), virus titers on MDTF-FePit1/FeLIX cells were higher by a half log10 unit on average (Fig. 3B). This difference was consistently observed in five independent experiments (data not shown) and suggests that the 378L change can enhance infection via the FeLV-T receptors. The titer on MDTF-FLVAR cells for 61E7P-ins-378L increased about 2 log10 units relative to 61E7P-ins, the clone differing only at 378L (Fig. 3B). This result was also reproducible in five independent experiments (data not shown). However, we did not observe marked differences in fluorescence intensity as a measure of virus binding to MDTF-FLVAR cells with or without the 378L amino acid change (Fig. 3B). We interpret these data to mean that while the 378L amino acid change is not a critical determinant of receptor specificity, it may enhance the infectivity of the viruses for the FeLV-T receptor and, in some cases, for the FeLV-A receptor.
The mutated envelope SU is virus-associated. To test whether the engineered envelope SU constructs were expressed and processed, we performed Western blot analysis on lysates of virus-producing 293T cells and on cell-free virus supernatants. For envelopes with any one of the three N-terminal substitutions (61E6D, 61E7P, or 61E6P8M), we observed both unprocessed envelope gp85 and processed envelope gp70 in lysates, as we expected (Fig. 4A). We also observed gp70 in cell-free supernatants, albeit lower than in supernatants of FeLV-A-61E (Fig. 4B). In cells expressing any of the envelopes containing the C-terminal insertion, there was a very low to undetectable level of processed envelope SU, although levels of unprocessed envelope were not dramatically affected (Fig. 4). On the basis of previous pulse-chase studies of FeLV-A/T variants (14), we interpret these data to indicate that processing is impaired by the presence of the insertion. We could detect the presence of envelope in virus supernatants for most of the mutant viruses, including some unprocessed envelope, which has been reported to be a particular issue with viral pseudotypes generated by transient transfection (16).
We also noted that the added presence of the 378L mutation appeared to increase the amount of envelope detected in the supernatant, although in some cases the increase appeared to be gp85 (Fig. 4). Although some virus supernatants—most strikingly 61E7P and 61E7P-ins—expressed considerably less envelope gp70 than did FeLV-A-61E as measured by Western blot analysis, they were able to infect cells expressing either the FeLV-A or FeLV-T receptor (Fig. 3 and 4).
The presence of gp85 in cell-free supernatants could be due to the presence of contaminating non-virion-associated envelope in cell-free supernatants. To address whether virus purified of free envelope could bind to cells expressing receptors, we purified a subset of virus supernatants by ultracentrifugation through a 20% sucrose cushion to remove any non-virion-associated envelope protein. In this case, we used a panel of viruses containing the 7P mutation, plus or minus the C-terminal insertion and 378L mutation because they had very low levels of gp70 in unpurified virus supernatants (Fig. 4). We measured these purified virions by Western analysis and found, despite differences in expression observed in cell lysates and supernatants, that purified virions of all mutants contain approximately comparable amounts of processed gp70 (Fig. 5A). These purified viruses showed binding properties similar to the virus supernatants that were not purified (Fig. 5B).
To verify that what we detected by Western blotting was gp70 and not gp85, we analyzed 61E and 61E7P along with a control envelope previously shown to be defective for processing [EE(ET3)E] (14) (Fig. 5C). This analysis confirmed that the protein present in purified viruses such as 61E7P was primarily the processed SU protein. In at least six replicates of this assay, we detected gp70 associated with purified viruses for all the viruses examined (data not shown). However, we have found that some mutants—in particular 61Eins and 61Eins-378L (data not shown)—sometimes contained smaller amounts of envelope, even after virion purification. However, despite variations in envelope detection, the relative infectivity between independent virus preparations did not vary (data not shown). Virus purification also did not alter infectivity compared to unpurified virus. This suggests that the differences in virus titers observed in our mutants are not explained simply by differences in envelope expression.
A deletion within the VRA of the RBD abolishes classical receptor specificity, leading to the loss of dual tropism. We have shown that both N- and C-terminal changes are required for dual-receptor specificity for FeLV-A and FeLV-T receptors in the context of FeLV-A-61E. These changes are present in the FeLV-T-61C variant, yet it cannot infect via the FeLV-A receptor. Because the FeLV-T-61C clone also contains a 6-amino-acid deletion within the VRA relative to the FeLV-A-61E and FeLV-A/T viruses, we hypothesized that this deletion was responsible for the loss of specificity for the FeLV-A receptor. We tested this by engineering the 6-amino-acid deletion into the 61E7P-ins-378L clone, which best represented a dualtropic virus of our panel of mutants. The resulting clone, 61E7P--ins-378L, was tested for envelope expression and binding to MDTF-FLVAR cells (Fig. 6). This envelope was expressed in cell-free supernatants at levels comparable to 61E7P-ins-378L, but it failed to infect or bind to MDTF-FLVAR cells (Fig. 6). When we tested infection of MDTF-FePit1/FeLIX cells, we found that both 61E7P-ins-378L and 61E7P--ins-378L could infect with a titer of 105 IU/ml, which was comparable to the titer of the control FeLV-T-61C (Fig. 6). This suggests that the deletion found in FeLV-T-61C is sufficient to abolish binding and infection via the FeLV-A receptor, and it also implies that an intact RBD (without a 6-amino-acid deletion) is required for entry via a classical receptor pathway. However, even when the RBD is disrupted, infection via the nonclassical pathway of Pit1 and FeLIX is maintained, suggesting that the two determinants can be distinct and independent.
The classical receptor specificity of a dualtropic virus can be altered by exchange of the RBD. In order to further test the independence of the determinants of the nonclassical FeLV-T receptor pathway from those of the classical receptor pathway, we replaced the RBD of FeLV-A-61E with the RBD of the subgroup FeLV-C prototype variant, FSC, in the presence of the N- and C-terminal changes that confer FeLV-T receptor specificity. The endogenous murine FLVCR can facilitate infection of FeLV-C-FSC (unpublished observations), and we observed a virus titer of less than 103 IU/ml on MDTF and MDTF-FePit1 cells (Fig. 7A and data not shown). The stable addition of feline FLVCR into MDTF cells (MDTF-FLVCR) allows a virus containing an envelope with the RBD from FeLV-C-FSC in the backbone of FeLV-A-61E (61ESRBD) to infect with a titer of 105 IU/ml (Fig. 7A). This level of infection of MDTF-FLVCR cells is reduced to background levels (103 IU/ml) in the presence of the C-terminal insertion (61ESRBD-ins), and it is abolished in the presence of both the C-terminal insertion and the 378L amino acid change (61ESRBD-ins-378L). However, the addition of an HD change at amino acid position 6 to an envelope encoding the C-terminal insertion can permit infection of MDTF-FePit1/FeLIX at a titer of about 105 IU/ml. Moreover, a QP change at amino acid position 7 along with the C-terminal insertion allows specificity for both the FeLV-C and FeLV-T receptors in the range of approximately 104 to 105 IU/ml. It is interesting that, when paired with the C-terminal insertion and adjacent amino acid change, the N-terminal changes tested were not equivalent in the ability to maintain FeLV-C receptor specificity, although both were able to confer FeLV-T receptor specificity.
All of these mutants exhibited comparable increases in fluorescence intensity as measured by flow cytometry when bound to MDTF cells expressing the receptor for FeLV-C (MDTF-FLVCR) compared to MDTF cells alone (Fig. 7B). In addition, all of the mutant envelopes were expressed, as visualized by Western blot analysis of lysates from transfected 293T cells, although there was some reduction in gp70 (Fig. 7C).
DISCUSSION
The goal of this study was to identify the minimal determinants required to allow a virus to use both the classical FeLV-A receptor and the nonclassical, two-component FeLV-T receptor complex. We found that the determinants for FeLV-T receptor specificity are located outside of the classical receptor binding domain and include at least two SU domains: first, amino acid changes at the N terminus at positions 6, 7, and/or 8, and second, a 4-amino-acid insertion and adjacent amino acid change within the C terminus. We also found that these nonclassical determinants are largely distinct and separable from the RBD, which specifies recognition and entry via the classical receptor pathway used by FeLV-A and FeLV-C (Fig. 8).
Viruses containing envelopes with N-terminal changes at amino acid positions 6, 7, and 8 (where the conserved histidine of the PHQ motif is at position 6 in FeLV) maintain specificity for the FeLV-A receptor, although the efficiency of infection was reduced compared to FeLV-A-61E. These findings reveal some differences between FeLV and MuLV because in MuLV the conserved histidine of the PHQ motif at position 8 (H8) appears to have a much more pronounced effect on infectivity, and in some cases infection could not be detected (3, 24, 42). As a result, MuLV viruses with disruptions of H8 have been described as defective. However, they can bind efficiently to their cognate receptors and they can also be rescued in trans by the addition of a wild-type soluble RBD fragment, in apparent similarity to naturally occurring FeLV-T (4, 5, 21, 24). Here we show that, FeLVs containing N-terminal changes, including the analogous H6, did not acquire the ability to infect via the nonclassical receptor pathway of Pit1 and FeLIX. Furthermore, infection of MDTF-FLVAR by FeLVs with N-terminal changes was not enhanced after the addition of soluble FeLV-A RBD, as was seen with the defective MuLVs. Thus, there may be some differences in the role of the conserved histidine at position 6 in the FeLV SU in receptor specificity compared to the H8 in the MuLV SU. These differences may depend in part on sequence context, as we have found that the infectivity of some of the mutants that encode both N- and C-terminal changes can be increased with the addition of 61E sRBD (Cheng, Anderson, and Overbaugh, unpublished data).
For FeLV, both N- and C-terminal changes are required to render a dual-receptor-specific phenotype; viruses expressing an envelope with the insertion with an adjacent amino acid change and any one of the three N-terminal changes can infect cells using both the FeLV-A and the FeLV-T receptors. Although there was variability in the amount of envelope detected in cell-free supernatants among the different mutant viruses, we saw no obvious correlation between differences in either expression or binding and relative infectivity. Thus, the impact of the N-terminal mutation and C-terminal insertion on infectivity appears not to be simply the result of effects on steady-state expression levels of the envelope SU. The reduced infectivity of constructs containing the N- and C-terminal changes on MDTF-FePit1/FeLIX compared to both FeLV-T-61C and FeLV-A/T-81T suggests that other changes may enhance infection via the FeLV-T receptor complex. For example, viruses with the 378L change had slightly increased titers in MDTF-FePit1/FeLIX in some cases. Thus, while the N-terminal changes at positions 6, 7, and 8 and the C-terminal insertion are together sufficient to confer the dual-receptor phenotype, changes at other residues may increase infectivity for one or both receptors.
In order to test whether the nonclassical determinants are distinct from the classical RBD determinants for FeLV receptor specificity, we engineered a 6-amino-acid deletion within the VRA of RBD present in FeLV-T-61C into an envelope also containing N- and C-terminal modifications. This envelope could neither bind MDTF-FLVAR cells nor infect them, presumably due to the disruption of the RBD determinants, yet it maintained the ability to infect MDTF-FePit1/FeLIX cells. As a further test, we substituted the RBD from FeLV-C-FSC into the FeLV-A envelope SU backbone containing different N-terminal changes along with the C-terminal insertion. We found that envelopes with one combination of N- and C-terminal changes had specificity for both FeLV-C and FeLV-T receptors, again indicating the independence of the nonclassical from the classical receptor determinants. However, envelopes with another combination of N- and C-terminal changes could infect using the FeLV-T receptors but not the FeLV-C receptor, indicating that the particular combination of N- and C-terminal changes may be important for some interactions with the classical FeLV-C receptor and that the different combinations may not elicit the same changes. All of the mutant envelopes containing the FeLV-C-FSC RBD were capable of binding MDTF-FLVCR cells regardless of receptor specificity, suggesting that the effects of N- and C-terminal mutations were on events after binding, such as fusion or entry.
Critical interactions between the N and C termini of SU have been implicated in the classical receptor pathway (21, 23, 25, 26, 28). A model has emerged from various studies in which the gammaretroviral RBD binds to the cognate receptor, signals to a region of the C terminus via the proline-rich region, and induces a conformational change in which residues in the N and C termini interact and trigger the dissociation of the SU from TM and allow the initiation of fusion (4, 21, 26). We found changes in the same regions of the N and C termini of the FeLV SU to be the determinants of infection via the nonclassical FeLV-T receptor pathway (14), perhaps indicating that naturally occurring viruses, such as FeLV-T, that use the nonclassical receptors may exist in a conformation different from those that use a classical receptor. These models for gammaretroviruses remain speculative; however, they are somewhat similar to models that have been proposed for some CD4-independent human immunodeficiency virus/simian immunodeficiency virus envelopes, in which the native structure of CD4-independent SUs are thought to exist in a conformation that is induced in CD4-dependent viruses upon SU binding to CD4 (9, 27). In FeLV, the changes required for nonclassical receptor use can maintain the integrity of the RBD and entry via a classical receptor (such as FLVAR or FLVCR), resulting in dual-receptor-specific variants. Thus, these envelopes that are capable of utilizing two distinct receptor pathways may shed light on the sequential events required for fusion and entry of gammaretroviruses.
ACKNOWLEDGMENTS
We thank Gretchen Strauch, Ramon Mendoza, Mario Pineda, and Colin Pritchard for comments on the manuscript, as well as Brannon Orton, Ramon Mendoza, and Nick Rabena for technical assistance.
This work was supported by NIH grant CA 51080. H.H.C. was supported by a fellowship from the Paul Allen Foundation.
REFERENCES
Anderson, M. M., A. S. Lauring, C. C. Burns, and J. Overbaugh. 2000. Identification of a cellular cofactor required for infection by feline leukemia virus. Science 287:1828-1830.
Anderson, M. M., A. S. Lauring, S. Robertson, C. Dirks, and J. Overbaugh. 2001. Feline Pit2 functions as a receptor for subgroup B feline leukemia viruses. J. Virol. 75:10563-10572.
Bae, Y., S. M. Kingsman, and A. J. Kingsman. 1997. Functional dissection of the Moloney murine leukemia virus envelope protein gp70. J. Virol. 71:2092-2099.
Barnett, A. L., and J. M. Cunningham. 2001. Receptor binding transforms the surface subunit of the mammalian C-type retrovirus envelope protein from an inhibitor to an activator of fusion. J. Virol. 75:9096-9105.
Barnett, A. L., R. A. Davey, and J. M. Cunningham. 2001. Modular organization of the Friend murine leukemia virus envelope protein underlies the mechanism of infection. Proc. Natl. Acad. Sci. USA 98:4113-4118.
Barnett, A. L., D. L. Wensel, W. Li, D. Fass, and J. M. Cunningham. 2003. Structure and mechanism of a coreceptor for infection by a pathogenic feline retrovirus. J. Virol. 77:2717-2729.
Battini, J. L., O. Danos, and J. M. Heard. 1995. Receptor-binding domain of murine leukemia virus envelope glycoproteins. J. Virol. 69:713-719.
Battini, J. L., J. M. Heard, and O. Danos. 1992. Receptor choice determinants in the envelope glycoproteins of amphotropic, xenotropic, and polytropic murine leukemia viruses. J. Virol. 66:1468-1475.
Bhattacharya, J., P. J. Peters, and P. R. Clapham. 2003. CD4-independent infection of HIV and SIV: implications for envelope conformation and cell tropism in vivo. AIDS 17(Suppl. 4):S35-S43.
Boomer, S., M. Eiden, C. C. Burns, and J. Overbaugh. 1997. Three distinct envelope domains, variably present in subgroup B feline leukemia virus recombinants, mediate Pit1 and Pit2 receptor recognition. J. Virol. 71:8116-8123.
Burns, C. C., M. L. Poss, E. Thomas, and J. Overbaugh. 1995. Mutations within a putative cysteine loop of the transmembrane protein of an attenuated immunodeficiency-inducing feline leukemia virus variant inhibit envelope protein processing. J. Virol. 69:2126-2132.
Donahue, P. R., S. L. Quackenbush, M. V. Gallo, C. M. C. deNoronha, J. Overbaugh, E. A. Hoover, and J. I. Mullins. 1991. Viral genetic determinants of T-cell killing and immunodeficiency disease induction by the feline leukemia virus FeLV-FAIDS. J. Virol. 65:4461-4469.
Grant, C. K., B. J. Ernisse, O. Jarrett, and F. R. Jones. 1983. Feline leukemia virus envelope gp70 of subgroups B and C defined by monoclonal antibodies with cytotoxic and neutralizing functions. J. Immunol. 131:3042-3048.
Gwynn, S. R., F. C. Hankenson, A. S. Lauring, J. L. Rohn, and J. Overbaugh. 2000. Feline leukemia virus envelope sequences that affect T-cell tropism and syncytium formation are not part of known receptor-binding domains. J. Virol. 74:5754-5761.
Heard, J. M., and O. Danos. 1991. An amino-terminal fragment of the Friend murine leukemia virus envelope glycoprotein binds the ecotropic receptor. J. Virol. 65:4026-4032.
Herrera, C., P. J. Klasse, E. Michael, S. Kake, K. Barnes, C. W. Kibler, L. Campbell-Gardener, Z. Si, J. Sodroski, J. P. Moore, and S. Beddows. 2005. The impact of envelope glycoprotein cleavage on the antigenicity, infectivity, and neutralization sensitivity of Env-pseudotyped human immunodeficiency virus type 1 particles. Virology 338:154-172.
Ho, S. N., H. D. Hunt, R. M. Horton, J. K. Pullen, and L. R. Pease. 1989. Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene 77:51-59.
Landau, N. R., K. A. Page, and D. R. Littman. 1991. Pseudotyping with human T-cell leukemia virus type I broadens the human immunodeficiency virus host range. J. Virol. 65:162-169.
Lander, M. R., and S. K. Chattopadhyay. 1984. A Mus dunni cell line that lacks sequences closely related to endogenous murine leukemia viruses and can be infected by ecotropic, amphotropic, xenotropic, and mink cell focus-forming viruses. J. Virol. 52:695-698.
Lauring, A. S., M. M. Anderson, and J. Overbaugh. 2001. Specificity in receptor usage by T-cell-tropic feline leukemia viruses: implications for the in vivo tropism of immunodeficiency-inducing variants. J. Virol. 75:8888-8898.
Lavillette, D., B. Boson, S. J. Russell, and F. L. Cosset. 2001. Activation of membrane fusion by murine leukemia viruses is controlled in cis or in trans by interactions between the receptor-binding domain and a conserved disulfide loop of the carboxy terminus of the surface glycoprotein. J. Virol. 75:3685-3695.
Lavillette, D., and D. Kabat. 2004. Porcine endogenous retroviruses infect cells lacking cognate receptors by an alternative pathway: implications for retrovirus evolution and xenotransplantation. J. Virol. 78:8868-8877.
Lavillette, D., A. Ruggieri, B. Boson, M. Maurice, and F. L. Cosset. 2002. Relationship between SU subdomains that regulate the receptor-mediated transition from the native (fusion-inhibited) to the fusion-active conformation of the murine leukemia virus glycoprotein. J. Virol. 76:9673-9685.
Lavillette, D., A. Ruggieri, S. J. Russell, and F. L. Cosset. 2000. Activation of a cell entry pathway common to type C mammalian retroviruses by soluble envelope fragments. J. Virol. 74:295-304.
Lu, C. W., and M. J. Roth. 2001. Functional characterization of the N termini of murine leukemia virus envelope proteins. J. Virol. 75:4357-4366.
Lu, C. W., and M. J. Roth. 2003. Functional interaction between the N- and C-terminal domains of murine leukemia virus surface envelope protein. Virology 310:130-140.
Myszka, D. G., R. W. Sweet, P. Hensley, M. Brigham-Burke, P. D. Kwong, W. A. Hendrickson, R. Wyatt, J. Sodroski, and M. L. Doyle. 2000. Energetics of the HIV gp120-CD4 binding reaction. Proc. Natl. Acad. Sci. USA 97:9026-9031.
O'Reilly, L., and M. J. Roth. 2003. Identification of conformational and cold-sensitive mutations in the MuLV envelope protein. Virology 312:337-349.
Ott, D., and A. Rein. 1992. Basis for receptor specificity of nonecotropic murine leukemia virus surface glycoprotein gp70SU. J. Virol. 66:4632-4638.
Overbaugh, J., P. R. Donahue, S. L. Quackenbush, E. A. Hoover, and J. I. Mullins. 1988. Molecular cloning of a feline leukemia virus that induces fatal immunodeficiency disease in cats. Science 239:906-910.
Overbaugh, J., A. D. Miller, and M. V. Eiden. 2001. Receptors and entry cofactors for retroviruses include single and multiple transmembrane-spanning proteins as well as newly described glycophosphatidylinositol-anchored and secreted proteins. Microbiol. Mol. Biol. Rev. 65:371-389.
Quigley, J. G., C. C. Burns, M. M. Anderson, E. D. Lynch, K. M. Sabo, J. Overbaugh, and J. L. Abkowitz. 2000. Cloning of the cellular receptor for feline leukemia virus subgroup C (FeLV-C), a retrovirus that induces red cell aplasia. Blood 95:1093-1099.
Quigley, J. G., Z. Yang, M. T. Worthington, J. D. Phillips, K. M. Sabo, D. E. Sabath, C. L. Berg, S. Sassa, B. L. Wood, and J. L. Abkowitz. 2004. Identification of a human heme exporter that is essential for erythropoiesis. Cell 118:757-766.
Rohn, J. L., M. S. Moser, S. R. Gwynn, D. N. Baldwin, and J. Overbaugh. 1998. In vivo evolution of a novel, syncytium-inducing and cytopathic feline leukemia virus variant. J. Virol. 72:2686-2696.
Sugai, J., M. Eiden, M. M. Anderson, N. Van Hoeven, C. D. Meiering, and J. Overbaugh. 2001. Identification of envelope determinants of feline leukemia virus subgroup B that permit infection and gene transfer to cells expressing human Pit1 or Pit2. J. Virol. 75:6841-6849.
Tailor, C. S., and D. Kabat. 1997. Variable regions A and B in the envelope glycoproteins of feline leukemia virus subgroup B and amphotropic murine leukemia virus interact with discrete receptor domains. J. Virol. 71:9383-9391.
Tailor, C. S., D. Lavillette, M. Marin, and D. Kabat. 2003. Cell surface receptors for gammaretroviruses. Curr. Top. Microbiol. Immunol. 281:29-106.
Tailor, C. S., B. J. Willett, and D. Kabat. 1999. A putative cell surface receptor for anemia-inducing feline leukemia virus subgroup C is a member of a transporter superfamily. J. Virol. 73:6500-6505.
Takeuchi, Y., R. G. Vile, G. Simpson, B. O'Hara, M. K. Collins, and R. A. Weiss. 1992. Feline leukemia virus subgroup B uses the same cell surface receptor as gibbon ape leukemia virus. J. Virol. 66:1219-1222.
Ting, Y. T., C. A. Wilson, K. B. Farrell, G. J. Chaudry, and M. V. Eiden. 1998. Simian sarcoma-associated virus fails to infect Chinese hamster cells despite the presence of functional gibbon ape leukemia virus receptors. J. Virol. 72:9453-9458.
Wensel, D. L., W. Li, and J. M. Cunningham. 2003. A virus-virus interaction circumvents the virus receptor requirement for infection by pathogenic retroviruses. J. Virol. 77:3460-3469.
Zavorotinskaya, T., and L. M. Albritton. 1999. Suppression of a fusion defect by second site mutations in the ecotropic murine leukemia virus surface protein. J. Virol. 73:5034-5042.(Heather H. Cheng, Maria M)