Wild Mouse Variants of Envelope Genes of Xenotropic/Polytropic Mouse Gammaretroviruses and Their XPR1 Receptors Elucidate Receptor Determinants of Virus Entry

Abstract

Mouse xenotropic and polytropic leukemia viruses (XMVs and PMVs) are closely related gammaretroviruses that use the XPR1 receptor for entry. To identify amino acid residues in XPR1 important for virus entry, we tested mouse cells derived from evolutionarily divergent species for susceptibility to prototypical PMVs, XMVs, and the wild mouse isolate CasE#1. CasE#1 has a variant XMV/PMV host range, and sequence analysis of the CasE#1 env gene identifies segments related to PMVs and XMVs. Cells from the Asian mouse species Mus pahari show a unique pattern of susceptibility to these three viruses; these cells are susceptible to XMVs and CasE#1 but are resistant to PMVs, whereas NIH 3T3 cells show the reciprocal pattern, susceptibility to only PMVs. The M. pahari XPR1 gene differs from that of NIH 3T3 in the two extracellular loops (ECLs) previously shown to mediate virus entry (M. Marin, C. S. Tailor, A. Nouri, S. L. Kozak, and D. Kabat, J. Virol. 73:9362-9368, 1999, and N. S. Van Hoeven and A. D. Miller, Retrovirology 2:76, 2005). Using transfected hamster cells expressing chimeric and mutated XPR1s, we demonstrated that the susceptibility differences between NIH 3T3 and M. pahari cells are receptor mediated, that PMV entry requires residues in ECL3, that the CasE#1 entry determinant is in ECL4, and that determinants for XMV entry are in both ECL3 and ECL4. Additional substitutions in ECL3 and ECL4 modulate virus susceptibility and suggest that ECL3 and ECL4 may contribute to the formation of a single virus receptor site. The position of M. pahari at the base of the Mus phylogenetic tree indicates that XPR1-mediated susceptibility to XMVs is the ancestral type in this genus and that the phenotypic variants of mouse XPR1 likely arose in conjunction with exposure to gammaretrovirus infections and coevolutionary adaptations in the viral envelope.

The mouse gammaretroviruses can be classified into several host range groups based on receptor usage. The xenotropic and polytropic mouse virus (XMV and PMV, respectively) subgroups were initially described as distinct host range groups based on their ability to infect mouse cells (11, 14, 22). However, XMVs and PMVs utilize the same receptor, XPR1, and the observed host range differences are due to sequence polymorphisms in both receptor and viral envelopes. Among mouse species, three allelic variants of Xpr1 have been described: the Xpr1ⁿ receptor found in common strains of the laboratory mouse mediates entry of PMVs but not XMVs, the Xpr1^sxv receptor variant found in many wild mouse species, such as Mus spretus, mediates entry of XMVs as well as PMVs (17), and the Xpr1^c gene of M. castaneus is defective for entry of both virus types (26). In addition to susceptibility differences due to receptor polymorphism, in some mice XPR1 is blocked by expression of endogenous interfering PMVs (the Rmcf resistance gene of DBA/2 mice) (15) or endogenous XMVs (the Rmcf2 resistance gene of M. castaneus) (25, 36). Cells of many other mammalian species, such as human and mink, generally are susceptible to infection by XMVs, and many also are susceptible to PMVs (8).

The XPR1 receptor has been characterized as a transmembrane protein of unknown function, although it shows homology to yeast genes involved in signal transduction and phosphate transport (4, 32, 37). XPR1 has eight predicted transmembrane domains, with the greatest sequence divergence in its fourth extracellular loop (ECL), ECL4. Mutagenesis identified two critical amino acids, K500 in ECL3 and T582 in ECL4, for XMV entry (26). Chimeras made between human XPR1 and hamster XPR1 confirm that XPR1 has two receptor determinants that independently mediate entry of XMVs in ECL3 and ECL4 and identified a receptor determinant for PMVs in the ECL4 of human XPR1 (33). The viral sequences critical for XPR1 receptor binding have not been identified, although analysis of Env chimeras indicates that the primary determinants for this specificity are in the N-terminal 118 amino acids containing the first variable domain, VRA (3), and a recent study (2) identified two residues in the C terminus of the Env receptor binding domain (RBD) needed for utilization of the human and mink XPR1 genes by the XMV/PMV isolate SL3-2.

Both PMVs and XMVs are present as endogenous copies in the laboratory mouse genome (28), but these proviruses generally are restricted to different taxonomic groups of wild mouse species (7, 19). While XMV proviruses can be nondefective and capable of producing infectious virus (18), infectious PMVs are generated only after recombination between endogenous defective PMVs and replicating mouse-tropic murine leukemia viruses (MLVs) (16). Also, XMVs are not known to be pathogenic in mice, whereas PMVs are associated with neoplastic transformation. Some studies also suggest that PMVs and XMVs differ in their binding to XPR1, and that this difference may be responsible for the failure of PMVs to establish superinfection interference, thus contributing to their pathogenicity (26, 35). This failure to establish superinfection interference also accounts for the fact that PMVs, but not XMVs, produce distinctive cytopathic effects in mink lung cells (14, 39).

The coevolution of the XPR1 receptor and the XMV/PMV host range group is important in the natural history of pathogenic gammaretroviruses and represents an area of investigation that should illuminate questions related to trans-species transmission or epizoonosis. In this study, we screened cells from mouse species representing each of the four subgenera in the genus Mus for susceptibility to XMVs/PMVs. We describe a novel infectivity phenotype associated with the XPR1 gene of the Asian mouse species M. pahari of the subgenus Coelomys, and we identify a novel XMV/PMV host range subgroup defined by the wild mouse isolate CasE#1 (8). We use these novel receptor and virus variants to define sites within the mouse XPR1 gene that affect receptor function and illuminate the natural history of this class of viruses in the genus Mus.

MATERIALS AND METHODS

Mice.

M. pahari mice, a randomly bred line of an Asian wild mouse species, originally were obtained from M. Potter (National Cancer Institute, Bethesda, MD). NFS/N mice originally were obtained from the Small Animal Section, NIH (Bethesda, MD). A congenic strain of NFS/N was developed in our laboratory: NFS/N-Sxv/Sxv (N/Sxv mice), carrying the Xpr1^sxv locus of M. musculus domesticus (formerly M. praetextus). Two breeding lines of M. castaneus, CAST/Rp and CAST/EiJ, were obtained from R. Elliott (Roswell Park, Buffalo, NY) and The Jackson Laboratory (Bar Harbor, ME), respectively.

Viruses, cells, and virus assays.

Viruses originally obtained from J. Hartley (National Institute of Allergy and Infectious Diseases [NIAID], Bethesda, MD) included the Friend dual-tropic PMV FrMCF, NFS Th-1 XMV, and the isolate CasE#1 (also known as Cas. E No. 1) (8). CAST-X is a xenotropic MLV isolated in our laboratory from the spleen of a CAST/EiJ mouse.

Susceptibility to PMV or XMV was tested in cultures of tail biopsy tissue prepared as described by Lander and colleagues (21) and in various cell lines, including M. dunni (20), NIH 3T3, mink Mv-1-Lu (CCL64), canine kidney cells (MDCK), Tu-1-Lu bat lung (ATCC CCL88), and MA139 ferret (obtained from J. Hartley). Cell lines derived from tail fibroblasts of the wild mouse species M. pahari, M. setulosis, and M. platythrix were obtained from J. Rodgers (Baylor College of Medicine, Houston, TX). Embryo fibroblasts were prepared from the progeny of crosses between CAST/Rp and NFS/N mice that were homozygous for Xpr1^c. These cells were used to establish permanent cell lines by the 3T3 method (1); this line is termed NXPR-C. Susceptibility to virus infection was quantitated as follows. Cells were infected with dilutions of XMV/PMV stocks in the presence of polybrene (4 μg/ml; Aldrich, Milwaukee, WI). After 4 to 5 days, cultures were UV irradiated and overlaid with 6 × 10⁵ mink S⁺ L⁻ cells (29). Foci were counted 6 to 7 days later.

Cloning the Xpr1 gene of M. pahari and generation of mutants and chimeras.

Total RNA was extracted from M. pahari and M. castaneus cells and from NIH 3T3 cells using TriReagent (Molecular Research Center, Cincinnati, OH). cDNA was synthesized using the SuperScript II reverse transcriptase kit (Invitrogen, Carlsbad, CA). PCR was performed to amplify the Xpr1 coding regions using 2 μl of cDNA as the template in a 50-μl PCR containing 2.5 U of PfuUltra high-fidelity DNA polymerase (Stratagene, La Jolla, CA) and 0.2 μM of the primers Fxpr1 (5′-CACCATGAAGTTCGCCGAGCACCTCTC) and Rxpr1 (5′-AGTGTTAGCTTCGTCATCTGTGTC), designed using GenBank accession no. NM_011273. The PCR was carried out in a GeneAmp PCR system 9700 machine (PE Applied Biosystems, Foster City, CA). After denaturation at 94°C for 5 min, the reaction proceeded with 35 cycles of the following: denaturation at 94°C for 30 s, annealing at 55°C for 30 s, and extension at 68°C for 4 min. The 2.1-kb PCR product from M. pahari and NIH 3T3 was purified with a QIAquick gel extraction kit (QIAGEN Sciences, Germantown, MD), cloned into the expression vector pcDNA 3.1D/V5-His-TOPO (Invitrogen), and sequenced. The NIH 3T3 Xpr1 sequence was identical to that reported previously (GenBank accession no. NM_011273) (26). The PCR product of the M. castaneus Xpr1 was partially sequenced; the ECL3 and ECL4 regions were identical to those reported previously (GenBank accession no. AF131102) (26).

Eight variants of the NIH 3T3 Xpr1 gene were generated. Three of these variants, Pah3, Pah4, and Pah3/4, have the ECL3 and/or ECL4 segment(s) of the M. pahari gene. Pah3 was produced by cloning a 0.5-kb Van91I-PshAI fragment containing the M. pahari ECL3 into the corresponding position of the NIH 3T3 Xpr1 gene.

To generate Pah4, site-directed mutagenesis (Stratagene) was used to make three amino acid substitutions in the NIH 3T3 gene using three pairs of primers. The first pair (F1, 5′-ATCTCTATTACTGCTACAACGTTTAAGCCTCATGTTGG; R1, 5′-CCAACATGAGGCTTAAACGTTGTAGCAGTAATAGAGAT) was used to insert threonine (Δ582T). The second primer pair (F2, 5′-AAGCCTCATGTTGGGGACATCATTGCTACTG; R2, 5′-CAGTAGCAATGATGTCCCCAACATGAGGCTT) was used to make the mutation N590D, and the third pair (F3, 5′-CCAAATCTCTATTACTGTTACAACGTTTAAGCCTC; R3, 5′-GAGGCTTAAACGTTGTAACAGTAATAGAGATTTGG) was used to make the substitution A581V. The high-fidelity polymerase PfuUltra (Stratagene) was used for all of these mutagenesis PCRs. After denaturation at 94°C for 5 min, 16 cycles were run with a 30-s denaturation at 94°C, a 1-min annealing step at 55°C, and a 12-min extension step at 68°C.

Pah3/4 contains both ECL3 and ECL4 from the M. pahari gene. It was made by replacing the 0.5-kb Van91I-PshAI ECL3-containing fragment of the Pah4 mutant with the corresponding fragment from the M. pahari Xpr1 clone.

Three additional Xpr1 variants containing one or two of the three mutations used to construct Pah4 were generated by mutagenesis PCR using the primers described above. These mutants were ECL4-1, with Δ582T; ECL4-2, with Δ582T and N590D; and ECL4-3, with A581V and Δ582T.

Mutagenesis PCR also was used to generate the mutants ECL3-1 (E500K) and ECL3/4-1 (E500K, Δ582T, and N590D). The primer pair (3F, 5′-CTTTACAGCACTCACAAAGAACAAAATCAC; 3R, 5′-GTGATTTTGTTCTTTGTGAGTGCTGTAAAG) was used to make the E500K substitution.

All chimeras and mutants generated for this study were confirmed by sequencing.

The recombinant plasmids were transfected into the E36 Chinese hamster cells (12). Stable transfectants were selected with Geneticin (830 μg/ml), and the expression of the Xpr1 variants was confirmed by Western analysis. Proteins were extracted from transfected cells with mammalian protein extraction reagent (Pierce, Rockford, IL). The expression vector used for XPR1 inserts a V5 epitope at the C terminus; XPR1 expression was detected in Western blots by using anti-V5 antibody (Invitrogen), followed by using goat anti-mouse immunoglobulin G conjugated with horseradish peroxidase (Invitrogen). After being washed, the membrane was incubated with chemiluminescence agents (PerkinElmer, Boston, MA) for 1 min and exposed to film. The membrane was then stripped with Restore Western blot stripping buffer (Pierce) for 20 min and was incubated with mouse anti-α-tubulin (Sigma, St. Louis, MO) and goat anti-mouse immunoglobulin G conjugated with horseradish peroxidase (Invitrogen), and then it was exposed to film.

Cloning and sequencing of the CasE#1 and CAST-X env genes.

RNA was extracted from a culture of CasE#1 virus-infected mink cells. The viral env gene was amplified by reverse transcription-PCR (RT-PCR) using forward primer NZB-F1 (5′-AGACGGCATCTCTGCGTGG) and reverse primer Xr-Z (5′-CAGCTAGCTTGCTAAGCCTTATGGTGG) based on GenBank accession no. K02730. A PCR product of 2.1 kb was isolated, cloned into pCR2.1-TOPO, and sequenced.

A 0.9-kb segment of the CAST-X env gene was amplified by RT-PCR from RNA of CAST-X virus-infected mink cells using forward primer 5′-GGATCCACGCCGCTCACGTA and reverse primer 5′-TGTCTCCCGTCCCAGGTTGT. The PCR product was purified using the QIAquick gel extraction kit (QIAGEN Sciences) and sequenced.

Pseudotype assay.

PMV and XMV LacZ pseudotype viruses were generated by transfection of human TELCeB6 cells (9) with expression vectors for the MCF247 PMV Env, pCRUCM, or a xenotropic variant of pCRUCM (3) containing the env gene of NZB-IU-6 (27). This XMV Env vector was generated by replacing the env-containing SalI-NheI segment of pCRUCM with the corresponding segment of the XMV clone. Transfected TELCeB6 cells produce viral particles harboring the MFGnlslacZ retroviral vector.

LacZ pseudotypes also were generated for CasE#1, for additional xenotropic isolates CAST-X and NFS Th-1, and for the PMV FrMCF by virus infection of the packaging cell line GP2-293 (Clontech, Mountain View, CA) that had been transfected by J. Silver (NIAID, Bethesda, MD) with pCL-MFG-LacZ (Imgenex, San Diego, CA) along with pMSCVpuro (Clontech). Supernatants of the virus-infected cells contained a mixture of infectious virus and LacZ pseudotypes.

Cells infected with the LacZ pseudotypes were fixed 32 to 48 h after infection with 0.4% glutaraldehyde and were stained to reveal the presence of β-galactosidase activity using as substrate 5-bromo-4-chloro-indolyl-β-d-galactosidaside (2 mg/ml; ICN Biomedicals, Aurora, OH).

Nucleotide sequence accession numbers.

The sequence of the M. pahari Xpr1 gene has been deposited in GenBank under accession no. EF606903. Sequences of the envelope genes of CAST-X and CasE#1 were deposited under nos. EF606902 and EF606901.

RESULTS

env sequence variation among XMV/PMV gammaretroviruses.

We selected several host range variants of XMV/PMV gammaretroviruses to define receptor determinants for entry into mouse cells. MCF247 and Friend mink cell focus-forming viruses were used as representative PMVs with broad polytropic host ranges, i.e., they infect mouse cells and cells of heterologous species. NZB-IU-6, NFS Th-1, and CAST-X are XMVs isolated from NZB/B1NJ, NFS/N, and M. castaneus mice, respectively, and are noninfectious for cells from laboratory mice such as NIH 3T3 mice. The third virus type, CasE#1, was isolated from a wild mouse trapped in Lake Casitas, CA (8); this mouse population may have acquired mouse gammaretroviruses from Asian mice (19). CasE#1 is not clearly classed as XMV or PMV based on its reactivity with XMV- or PMV-specific monoclonal antibodies and its host range in heterologous cells. Like PMVs, CasE#1 induces foci in mink lung cells, and while this virus also resembles PMVs in its pattern of nonreciprocal interference with XMVs in rat cells, the interference of CasE#1 with prototypical PMVs is not fully reciprocal (8).

The env genes of the CAST-X and CasE#1 isolates were sequenced, and the results were compared to the sequences of prototypical PMVs and XMVs. Within the RBD there are differences at 38 sites in the set of five viruses shown in Fig. 1. CAST-X shows 98% amino acid identity to the previously sequenced XMVs NZB-IU-6 (termed NZB-9-1) and NFS Th-1 with five amino acid substitutions in the RBD, all of which are conservative changes. At the sites that distinguish CasE#1 from the XMVs or PMVs, CasE#1 shares 7 residues with one or both PMVs and 14 residues with XMVs, and it has 10 unique substitutions. Like XMVs, CasE#1 has a 4-amino-acid insertion in VRA. These differences are clustered such that, despite the closer general similarity of CasE#1 to XMVs than to PMVs, there is a region that more closely resembles PMVs, beginning at the C terminus of VRA. The novel CasE#1 substitutions not found in either PMVs or XMVs define segments in the middle of VRA and in the C-terminal third of the RBD.

FIG. 1.

Comparison of the deduced amino acid sequences of the RBD region of the viral env gene of CasE#1 and the XMVs and PMVs used for infection. Variable regions VRA, VRB, and VRC are indicated with bars. Sequences for MCF247, FrMCF, and NZB-9-1 were previously determined (GenBank accession nos. K02727, X01679, and K02730, respectively).

Virus infectivity.

Viruses were used to infect subconfluent cultures of cells from laboratory mouse strains and from various wild mouse species. The mouse cells included cells with the three previously identified allelic variants of Xpr1: Xpr1ⁿ, Xpr1^sxv, and Xpr1^c. Additionally, we infected cells of several heterologous species: mink, dog, and bat (Table 1).

TABLE 1.

Virus titers of XMV/PMV gammaretroviruses on mouse cells and cells of heterologous species

Cell typeb	Mouse Xpr1 allele	Log10 virus titera
		PMV	CasE#1	XMV
NIH 3T3	Xpr1n	2.7	<0	<0
M. dunni	Xpr1sxv	5.2	5.7	4.5
M. pahari	Xpr1p	<0	2.3	2.5
NXPR-C	Xpr1c	<0	<0	1.8
Mv-1-Lu mink		4.7	4.9	5.2
MDCK canine kidney		<0	<0	3.9
Tu-1-Lu bat lung		<0	3.8	4.0

^aMeasured as the number of focus-forming units in 100 μl of dilutions of virus stocks. Foci were counted 5 to 7 days after infected cultures were UV irradiated and overlaid with mink S⁺ L⁻ cells. M. pahari cells were tested as cultured cells from tail biopsies using an M. pahari cell line. Titers represent the averages from three different experiments, using FrMCF as the PMV and CAST-X as the XMV.

^bThe first four cell types listed are of mouse origin. NXPR-C is a permanent line derived from embryo fibroblasts of a mouse carrying Xpr1^c on the NFS/N genetic background.

The three viruses, the PMV FrMCF, CasE#1, and the XMV CAST-X, identified three distinct infectivity phenotypes among mouse cells with the previously described Xpr1 variants. As expected, mouse cells with the laboratory mouse allele of the Xpr1 receptor, Xpr1ⁿ, were susceptible to infection only by PMVs. M. dunni cells, with the wild mouse Sxv allele (17, 20), were susceptible to all three viruses. Cells carrying the Xpr1^c allele of M. castaneus were not susceptible to infection with PMV, as previously shown (24); these cells also were not susceptible to infection by CasE#1, but they were susceptible to XMV infection. This XMV susceptibility was unexpected, because a previous study had reported resistance of M. castaneus cells to XMV based on the failure of Xpr1^c to function as a viral receptor in transfected hamster cells (26). Sequence analysis identified no differences in ECL3 and ECL4 between our Xpr1^c sequence and the published sequence (26) (GenBank accession no. AF131102).

Cells from nonrodent species also showed differential susceptibility to these three viruses (Fig. 1). These results are consistent with the previous demonstration that CasE#1 is not clearly PMV or XMV in host range; this virus resembles PMVs in its infectivity in MDCK and NXPR-C cells but resembles XMVs in NIH 3T3 and bat cells (8).

We also tested cells of the Asian wild mouse species M. pahari (Gairdner's shrewmouse) for virus susceptibility. These cells differed from all other mouse cells in that they restricted PMVs but supported the replication of XMVs and CasE#1 (Table 1). Focus-forming titers of these two viruses on M. pahari cells generally were somewhat lower than those on fully susceptible mink cells or on M. dunni cells (Table 1). This type of resistance to specific isolates of a single host range group previously has been reported for ecotropic gammaretroviruses; such resistance has been attributed to altered receptor glycosylation, because treatment with inhibitors of glycosylation restored virus infectivity (10). The Xpr1 gene has several potential glycosylation sites, including two in putative ECLs (26). M. pahari cells were treated with tunicamycin (0.05 μg/ml) prior to virus infection. Tunicamycin treatment did not increase the efficiency of infection with XMVs and did not result in detectable PMV replication (data not shown), suggesting that glycosylation plays no role in virus resistance in these cells.

To determine if the unusual pattern of susceptibility of M. pahari cells is entry related, the cells were infected with LacZ pseudotypes with PMV (FrMCF), XMV (CAST-X), and CasE#1 envelopes (Table 2). Results for all cells tested replicated the results of virus infectivity tests (Table 1), although the titers observed for the viral pseudotypes in M. pahari were higher than those of the corresponding viruses in Table 1, suggesting there is a postentry restriction of virus replication in these cells. M. pahari cells were susceptible to XMV and CasE#1 pseudotypes, but no LacZ-positive cells were identified after infection with PMV pseudotypes. In addition, E36 Chinese hamster cells defined a novel susceptibility phenotype to PMV FrMCF, CasE#1, and the XMV CAST-X. These cells were resistant to all three viruses, although trace levels of XMV infection were detected.

TABLE 2.

Titers of LacZ pseudotypes of XMVs and PMVs in mouse cells and cells of heterologous species

Cell typeb	Log10 LacZ pseudotype titera
	PMV	CasE#1	XMV
NIH 3T3	3.0	<0	<0
N/SXV	4.5	2.3	3.5
M. pahari	<0	>5.0	>5.0
NXPR-C	<0	<0	4.5
MA139 ferret	5.0	4.0	5.0
E36 Chinese hamster	<0	<0	1.3
M. platythrix	4.0	4.4	4.0
M. setulosis	1.4	4.0	4.0

^aMeasured as the number of LacZ-positive cells in cultures infected with 100 μl of virus dilutions, using FrMCF as the PMV and CAST-X as the XMV.

^bAll but the ferret and hamster cells are of mouse origin. NXPR-C and N/SXV are embryo fibroblasts derived from mice carrying Xpr1^c and Xpr1^sxv, respectively, on the NFS/N genetic background.

Because M. pahari cells presented a susceptibility type not previously identified for mice, we also tested several cells from other Mus species that, like M. pahari, are only distantly related to the European mouse-derived laboratory strains and the more commonly studied wild mouse species (Table 2). M. platythrix (subgenus Pyromys) showed an Sxv-like susceptibility to all three viruses, but M. setulosis (subgenus Nannomys) cells resembled M. pahari cells in their resistance to PMVs.

Sequence analysis of the M. pahari Xpr1 receptor.

We amplified the M. pahari Xpr1 receptor, Xpr1^p, by RT-PCR and cloned it into the expression vector pcDNA 3.1D/V5-His-TOPO. Alignment of the sequenced Xpr1^p gene with reference to that of NIH 3T3 mice is shown in Fig. 2. Xpr1^p is 98.7% identical to Xpr1ⁿ, and all of the differences are within the two regions known to contain determinants for virus entry, ECL3 and ECL4. In these two domains, Xpr1^p contains both residues known to be critical for XMV entry: T582 in ECL4 and K500 in ECL3. However, relative to Xpr1ⁿ, Xpr1^p also has additional substitutions in both of these ECLs. ECL4 contains the novel substitution A581V and the substitution N590D, which also is present in the M. dunni gene (26). Sequence differences between Xpr1^p and Xpr1ⁿ in the much larger ECL3 are primarily at the carboxy terminus; Xpr1^p lacks the putative glycosylation site at N503 and has four additional novel substitutions, E436G, S444T, T507Y, and V508K.

FIG. 2.

Comparison of the deduced amino acid sequences of the Xpr1 genes of M. pahari and NIH 3T3. Bars are used to indicate the predicted locations of ECL3 and ECL4.

We transfected Xpr1^p into Chinese hamster E36 cells to determine if this gene is responsible for the unique M. pahari pattern of virus susceptibility. E36 cells are not susceptible to CasE#1 or PMV and are very poorly susceptible to infection by XMVs. Infection of the transfected cells with virus (data not shown) or with LacZ pseudotypes reproduced the susceptibility pattern of the M. pahari cells (Fig. 3A; Table 2).

FIG. 3.

(A) Susceptibility of cells with different Xpr1 receptors to LacZ pseudotypes of XMV, PMV, and CasE#1. Receptor genes from NIH 3T3 cells (Xpr1ⁿ) and M. pahari cells (Xpr1^p) were tested along with the indicated chimeras in transfected E36 hamster cells. MCF247 was used to generate the PMV pseudotype, and NFS Th-1 was used for the XMV pseudotype; experiments using CAST-X and FrMCF gave comparable results (data not shown). Titers represent the average results from three experiments and are given as the number of LacZ-positive cells/100 μl. Arrows indicate the restriction fragment used to make the ECL3 substitutions. Because the NIH 3T3 and M. pahari ECL4 regions differ by only three amino acids, site-specific mutagenesis was used to introduce these substitutions into the NIH 3T3 gene. (B) Western analysis of E36 cells expressing XPR1 constructs. XPR1 expression was detected by using an anti-V5 antibody; the V5 epitope was added to the C terminus of Xpr1 in the expression constructs.

Functional analysis of XPR1 chimeras and mutated variants.

To identify determinants for XMV/PMV entry, we generated XPR1 chimeras between Xpr1ⁿ and Xpr1^p. Because previous reports (26, 33) had identified receptor determinants in the putative ECLs ECL3 and ECL4, and because there were no amino acid differences between these XPR1 genes outside of these ECLs (Fig. 2), chimeras were constructed to evaluate the contribution of these two XPR1 segments to virus susceptibility. The chimeras were constructed using Xpr1ⁿ as the backbone (Fig. 3A), and the constructs were transfected into E36 cells. Western analysis of the transfected cells identified a protein of about 70 kDa in all transfectants (Fig. 3B).

Cells expressing chimeric Xpr1 genes were infected with virus or with virus LacZ pseudotypes of all three host range types. The three chimeras showed three different susceptibility phenotypes, as shown for pseudotype infections in Fig. 3A. Chimera Pah3 with the Xpr1^p ECL3 was resistant to all three viruses, Pah4 with Xpr1^p ECL4 was susceptible to all three viruses, and Pah3/4 with both Xpr1^p ECL domains was susceptible to XMVs and CasE#1 but not to PMV. Susceptibility to PMVs thus is associated with the presence of the Xpr1ⁿ ECL3. Susceptibility to CasE#1 is concordant with susceptibility to XMVs, and this susceptibility correlated with the presence of the Xpr1^p ECL4. The Pah3 chimera with the Xpr1^p ECL3 did not mediate XMV infection, despite the fact that Pah3 contains the ECL3 critical residue K500. This suggests that K500 alone is not sufficient for XMV infection.

Specific mutations then were introduced into Xpr1ⁿ to determine the contributions of various residues to receptor function. These mutations included substitutions at the two sites shown previously (26) to be critical for XMV susceptibility: E500K in ECL3 and Δ582T in ECL4. The introduction of either of these mutations into the NIH 3T3 receptor produced receptors capable of mediating infection by XMVs (Fig. 4), as was shown previously for Δ582T (26). Only one of these two XMV entry determinants, however, was capable of supporting infection by CasE#1: Δ582T in ECL4. This result is consistent with two different models of XPR1-mediated virus entry. If ECL3 and ECL4 carry independent entry determinants for XMVs, this result shows that, in tests with CasE#1, these two separate determinants are not functionally equivalent. Alternatively, if ECL3 and ECL4 together contribute to the formation of a single virus receptor site, changes in either loop might be expected to alter receptor specificity. This second model, requiring both ECLs for efficient receptor function, is consistent with the observed difference between Pah3/4 and Pah4 in the efficiency of CasE#1 infection. The Pah3/4 chimera is substantially more susceptible to this virus than Pah4 (Fig. 3A); although the M. pahari ECL3 does not function as a CasE#1 receptor in Pah3, the combination of ECL3 and ECL4 in Pah3/4 produces a more efficient receptor for this virus than ECL4 alone. This model also explains the observed reduced titer for CasE#1 in the Δ582T mutant, ECL4-1; this single mutation produces a minimally functional receptor (Fig. 4). These observations together suggest that CasE#1 entry mainly is affected by ECL4, but that some combination of the Xpr1^p ECL3 and ECL4 regions produces a receptor that can more effectively mediate virus entry.

FIG. 4.

Susceptibility of NIH 3T3, M. pahari, and E36 Chinese hamster cells expressing XPR1 mutants to LacZ pseudotypes of PMV, XMV, and CasE#1. NZB-IU-6 was used as the XMV, and MCF247 was used as the PMV. Titers are reported as the number of LacZ-positive cells in 100 μl and represent the averages from four experiments.

To further define the genetic basis for CasE#1 infectivity, we made several additional substitutions together with Δ582T at sites that distinguish Xpr1^p from Xpr1ⁿ: E500K, A581V, and N590D (Fig. 4). All of these mutants were susceptible to XMV and PMV. All of these mutants also were susceptible to CasE#1, although one, ECL4-3, showed reduced susceptibility to this virus. These results together suggest that Δ582T is critical for CasE#1 and that the substitutions A581V and N590D may enhance the efficiency of CasE#1 entry.

DISCUSSION

We identified receptor determinants in XPR1 that modulate virus entry by using a novel receptor variant from the Asian mouse species M. pahari, along with XMV and PMV with three distinct host ranges. We showed that two ECLs of the XPR1 protein contribute to receptor function, and that infectivity of the three viruses used here, PMV, XMV, and CasE#1, relies on different determinants in XPR1. Entry determinants for PMV were identified in ECL3 of the mouse XPR1, for XMV in ECL3 and ECL4, and for CasE#1 in ECL4.

Previous studies had implicated two XPR1 ECLs in virus entry and identified two specific amino acids, either one of which could mediate entry of XMVs: K500 in ECL3 and T582 in ECL4 (26, 33). While addition of either of these residues to the XMV-resistant NIH 3T3 Xpr1 generates a functional XMV receptor, our data show that these two mutant receptors are not functionally equivalent; the Δ582T mutation generates a receptor for CasE#1, but the E500K mutation does not. Also, the functionality of these two critical receptor determinants clearly is modulated by other residues. Thus, while all XPR1 mutants with Δ582T are susceptible to infection with XMV and CasE#1, substitutions at additional sites in ECL3 or ECL4 can alter susceptibility to CasE#1. Similarly, the presence of K500 is not sufficient for XMV infection, as this residue is present in XMV-resistant hamster cell XPR1 (26) as well as the resistant Pah3 chimera that contains the M. pahari ECL3.

These observations suggest that ECL3 and ECL4 both contribute to the interaction with viral Env rather than specifying separate entry determinants as previously suggested (26, 33). The Pah4 chimera with Xpr1^p ECL4 is an efficient receptor for XMV but is less efficient for CasE#1. The presence of ECL3 together with ECL4 in the Pah3/4 chimera generates a receptor that is very efficient for both viruses. All of the known gammaretrovirus receptors have multiple transmembrane domains, and the suggestion that residues in two XPR1 loops are needed for receptor function is consistent with the fact that all of these other receptors, with the exception of the mCAT-1 ecotropic MLV receptor, require multiple domains for receptor function, as discussed by Brown et al. (6).

In contrast to results of studies on XMV entry, identification of the residues that specify the PMV receptor determinant in XPR1 remains elusive. In fact, our identification of a PMV receptor determinant in the mouse NIH 3T3 ECL3 is at odds with the results of Van Hoeven and Miller (33), who identified a PMV entry determinant in ECL4 of the human XPR1. One possible explanation for this discrepancy is the major sequence differences in the receptor-determining regions of the human and mouse XPR1 orthologues. Of the 13 ECL4 residues, 6 differ in human and mouse cells, and the remaining 7 sites are conserved in most sequenced XPR1 genes regardless of virus susceptibility (26; Y. Yan and C. A. Kozak, unpublished data). ECL3 is considerably larger than ECL4 (88 amino acids), with significant sequence variation near the XMV receptor determinant K500; this variation includes a potential glycosylation site at N503 in Xpr1ⁿ that is absent from the human and M. pahari genes. Glycosylation sites frequently are associated with virus receptor sites and may have a regulatory role in their function (31). These observed XPR1 sequence differences are likely to produce substantially different human and mouse XPR1 receptor structures, and these differences may alter the relative contributions of the critical residues for PMV entry in these orthologues. PMVs may be more sensitive to minor sequence changes, as illustrated by the fact that XMV-susceptible cells of heterologous species are not always PMV susceptible. It also has been suggested that PMV-receptor interactions are weaker, as they only inefficiently induce superinfection interference (26, 35, 39).

Evolutionary origin of Xpr1 variants and coevolution of XMV/PMV env variants.

The four functionally distinct mouse XPR1 receptors described here were identified in specific taxonomic groups of wild mice species (Fig. 5). The appearance of these four receptor phenotypes can be examined in relation to the appearance and spread of XMVs and PMVs in these populations to shed light on the adaptive coevolution of receptor and virus (31). The genus Mus is thought to have originated on the Indian subcontinent about 7.5 million years ago, and successive expansions/radiations have produced at least 40 species in four subgenera. Three of the four Xpr1 variants are found in the extensively studied Eurasian subgenus Mus (17, 24). The XPR1 variant with the broadest host range susceptibility phenotype, Xpr1^sxv, is the most widely distributed and is found in the older species of the Mus subgenus as well as in M. platythix, subgenus Pyromys. This distribution suggests a long period of stasis that ended after exposure to infectious XMVs and PMVs that led to the acquisition of endogenous XMV/PMV-related sequences. The variant receptor alleles, Xpr1^c and Xpr1ⁿ, appeared in recently diverged Mus species that have acquired endogenous env copies of the viruses that use these variant receptors. Thus, XMV-susceptible M. castaneus carries multiple XMV env genes, whereas PMV-susceptible M. domesticus carries multiple PMV env genes (19). It is likely that the mutations restricting XPR1 function in these mice initially provided a survival advantage in the face of endemic gammaretrovirus infection, but they may have also resulted in the selection of viral variants with altered receptor specificities. For the European mice, the Xpr1ⁿ mutation may have driven the evolution of the PMV Env variants that target this novel XPR1.

FIG. 5.

Schematic representation of the evolution of Mus. This evolutionary tree is based largely on the synthetic tree of Boursot and colleagues as well as Guénet and Bonhomme (5, 13) and is derived from morphological characteristics, DNA sequence data, and analysis of karyotypic rearrangements (23, 30, 34). All species from M. domesticus through M. cervicolor are of the subgenus Mus. The most recent node of the tree represents the house mouse M. musculus complex; M. molossinus is a natural hybrid of M. castaneus and M. musculus (38). Among the other, older species in the Mus subgenus is M. dunni, part of the M. terricolor group. The species that have been tested for susceptibility to PMVs and XMVs are marked by asterisks, and the susceptibility type is indicated using designations for the four known functional variants of Xpr1. For species that carry endogenous PMV/XMV env genes, the predominant type is indicated (19). Mya, millions of years ago.

At the base of the Mus phylogenetic tree, the relationships among the four subgenera have not been decisively resolved, although most recent studies argue for the basal placement of Coelomys (23, 30, 33). This placement suggests that the XPR1 receptor variant identified in the Coelomys species M. pahari and shared by the Nannomys species M. setulosis is the ancestral type in Mus and may be found in other, older Muridae.

For the XMV/PMV gammaretroviruses, the coevolution of virus and host clearly has resulted in exceptional sequence diversity and functional plasticity, as illustrated by CasE#1. We identified several receptor mutations that modulate CasE#1 entry but have no effect on XMV and PMV infection. Although this virus uses the M. pahari XMV T582 determinant, it does not use the XMV K500 determinant. The CasE#1 Env sequence, like its host range, resembles that of both XMVs and PMVs; further analysis of this and other unusual naturally occurring XMV/PMV variants may help identify the determinants of receptor specificity.