|Home | About | Journals | Submit | Contact Us | Français|
Genetic conflicts between retroviruses and their receptors result in the evolution of novel host entry restrictions and novel virus envelopes, and such variants can influence trans-species transmission. We screened rodents and other mammals for sequence variation in the Xpr1 receptor for the mouse xenotropic or polytropic mouse leukemia viruses (X-MLVs or P-MLVs, respectively) of the gammaretrovirus family and for susceptibility to mouse-derived X/P-MLVs and to XMRV (xenotropic murine leukemia virus-related virus), an X-MLV-like virus isolated from humans with prostate cancer and chronic fatigue syndrome. We identified multiple distinct susceptibility phenotypes; these include the four known Xpr1 variants in Mus and a novel fifth Xpr1 gene found in Mus molossinus and Mus musculus. We describe the geographic and species distribution of the Mus Xpr1 variants but failed to find the X-MLV-restrictive laboratory mouse allele in any wild mouse. We used mutagenesis and phylogenetic analysis to evaluate the functional contributions made by constrained, variable, and deleted residues. Rodent Xpr1 is under positive selection, indicating a history of host-pathogen conflicts; several codons under selection have known roles in virus entry. All non-Mus mammals are susceptible to mouse X-MLVs, but some restrict other members of the X/P-MLV family, and the resistance of hamster and gerbil cells to XMRV indicates that XMRV has unique receptor requirements. We show that the hypervariable fourth extracellular XPR1 loop (ECL4) contains three evolutionarily constrained residues that do not contribute to receptor function, we identify two novel residues important for virus entry (I579 and T583), and we describe a unique pattern of ECL4 variation in the three virus-restrictive Xpr1 variants found in MLV-infected house mice; these mice carry different deletions in ECL4, suggesting either that these sites or loop size affects receptor function.
The XPR1 receptor mediates entry for the mouse leukemia viruses (MLVs) with xenotropic and polytropic host ranges (X-MLVs and P-MLVs, respectively). X-MLVs and P-MLVs can be isolated from laboratory mice and are capable of infecting cells of nonrodent species; these viruses are distinguished by the ability of P-MLVs, but not X-MLVs, to infect cells of the laboratory mouse and by the cytopathic and leukemogenic properties of P-MLVs, also termed MCF MLVs (mink cell focus-inducing MLVs) (11, 16, 24). XPR1 is also the receptor for several wild mouse isolates with an atypical host range (6, 48, 49) and for the recently described virus XMRV (xenotropic murine leukemia virus-related virus) (8), isolated from human patients with prostate cancer or chronic fatigue syndrome (27, 37, 43). Studies on the XPR1 receptor have identified residues critical for virus entry and described functionally distinct variants of XPR1 in human and rodent species that differ in their abilities to mediate entry of various virus isolates (18, 29, 31, 48, 49).
In Mus, four receptor variants of Xpr1 are found in different taxonomic groups. Xpr1n was originally described in strains of the laboratory mouse (1, 41, 51), which are largely derived from Mus domesticus (50). Xpr1c was identified in the Asian species Mus castaneus (29, 31); Xpr1p is in the Asian species Mus pahari (48); and Xpr1sxv was found in several Eurasian species (18, 31). These variants are distinguished by their differential susceptibilities to prototype X-MLVs and P-MLVs as well as to two wild mouse isolates, CasE#1 and Cz524 (49); only Xpr1sxv encodes a receptor that is fully permissive for all isolates. The host range differences of these various virus isolates are due to sequence polymorphisms in both receptor and viral envelope genes.
The various mouse X/P-MLV isolates and the humanized XMRV define six different tropism patterns based on infectivity on rodent cells carrying Xpr1 variants (49). These tropisms distinguish the two wild mouse isolates, CasE#1 and Cz524, and identify two P-MLV host range subgroups and two X-MLV/XMRV subgroups. Specific XPR1 residues responsible for entry of these viruses have been identified by analysis of rodent Xpr1 variants and mutants. These receptor determinants lie in two of the four predicted extracellular loops (ECLs) of Xpr1, ECL3 and ECL4 (31, 44, 48, 49). Two critical amino acids have been defined for X-MLV entry: K500 in ECL3 and T582 in ECL4 (31). These two receptor determinants independently produce X-MLV receptors but are not functionally equivalent, as the Δ582Τ insertion into Xpr1n generates a receptor for CasE#1, but the E500K substitution does not (48). Sensitivity to the six tropism subgroups is further modulated by specific substitutions at ECL3 residues 500, 507, and 508 (49). The sequence variations that distinguish the rodent XPR1 receptors can result in subtle differences in the efficiency of virus infection or complete resistance to specific X/P-MLVs.
The characterization of host genes that effect and/or block entry has obvious importance for a broader understanding of how viruses spread in natural populations and are transmitted to new hosts and how those populations adapt to retrovirus infections. The four house mouse species carry endogenous retroviruses (ERVs) for X-MLVs and P-MLVs (XMVs and PMVs, respectively) (3, 20, 42), and three of these species harbor infectious X-MLVs (4, 19, 48, 49). Restrictive variants of the XPR1 receptor have evolved in these virus-infected mice, along with the virus envelope (env) variants that define the tropism subgroups. We thus sought to examine the evolution of Xpr1 in rodent species, and we extended this functional and sequence analysis to nonrodent species for two reasons. First, identification of XMRV in several human patient cohorts (27, 37, 43), the recent detection of P-MLV-related sequences in patients and blood donors (26), and the multiple instances of transspecies transmission of mouse gammaretroviruses (33) support an effort to describe factors that mediate or modulate virus entry in these species. Second, analysis of nonrodent species with novel patterns of virus restriction may uncover different or additional entry determinants. In the present study, we characterized 49 mice of different species or from different geographic locations and 24 other mammalian species for sequence and functional variants of the Xpr1 receptor. We identified a novel 5th functional Xpr1 variation in Mus, showed that restrictive XPR1 receptors in the three MLV-infected house mouse lineages have different deletions in ECL4, demonstrated that XPR1 is under positive selection, identified novel virus restriction phenotypes in nonrodent species, and demonstrated that XMRV relies on unique entry determinants.
CAST-X is a xenotropic MLV isolated in our laboratory from the spleen of a CAST/EiJ mouse (48). Cz524 is a novel MLV isolated from the spleen of a CZECHII/EiJ mouse 2 months after inoculation with Moloney MLV (MoMLV) (49). The human xenotropic murine leukemia virus-related virus, XMRV (8), was kindly provided by R. Silverman (Cleveland Clinic, Cleveland, OH). AKR6, CasE#1, FrMCF, HIX MoMCF, and amphotropic A-MLV 4070A were originally obtained from J. Hartley (NIAID, Bethesda, MD).
Susceptibility to X/P-MLVs was tested in the cells from the following sources or cell lines: Mus dunni (21), ferret MA139 and goat lung cells obtained from J. Hartley (NIAID, Bethesda, MD), human 293 (CRL-1573), mink Mv-1-Lu (ATCC CCL64), Rat-2 (CRL-1764), gerbil GeLu (CCL-100), guinea pig JH4 clone 1 (CCL-158), dog MDCK (CCL-34), bat Tb-1-Lu (CCL-88), African green monkey kidney COS-1 (CRL-1650), cat CRFK (CCL-94), rabbit SIRC (CCL-60), buffalo IMR30 (obtained from M. Eiden; NIMH, Bethesda, MD), armadillo DNI.Tr (CRL-6009), Chinese hamster lines E36 (14) and Lec8 (CRL-1737), and E36 hamster cells transfected with Xpr1 variants (48, 49). Embryo fibroblasts homozygous for Xpr1c were described previously (48); tail fibroblasts were prepared from wild mouse-derived strains MOLD/EiJ, MOLG/EiJ, and CZECHII/EiJ as described previously (22).
DNA and RNA were isolated from animals, from cell lines that were developed from wild mice and wild mouse-derived breeding colonies or inbred strains (see Table S1 in the supplemental material), and from the mammalian cells listed above. Many wild-derived mice were obtained from M. Potter (NCI, Bethesda, MD). Mice trapped in California were provided by S. Rasheed (University of Southern California, Los Angeles). CAST/Rp mice were obtained from R. Elliott (Roswell Park Cancer Institute, Buffalo, NY). Cells from some wild mouse species were obtained from J. Rodgers (Baylor College of Medicine, Houston, TX). Mice or DNA samples from M. castaneus (CAST/EiJ), M. domesticus (PERA/EiJ, LEWES/EiJ, and WSB/EiJ), M. musculus (CZECHI/EiJ and CZECHII/EiJ), and various inbred lines derived from Mus molossinus were obtained from The Jackson Laboratory (Bar Harbor, ME). DNA samples from M. musculus (VEJ, and BRNO, Viborg, Belgrade) were obtained from S. Chattopadhyay and H. Morse (NIAID, Bethesda, MD). An African pygmy mouse sample was provided by H. Leathers (Philadelphia, PA). DNA samples from wild-trapped European M. domesticus mice were provided by M. Nachman (University of Arizona, Tucson).
A set of DNAs from African pygmy mouse DNA and other Muridae were obtained from Y. Cole and P. D'Eustachio (Departments of Biochemistry and Medicine, NYU, New York); the pygmy mice were classed into five species of subgenus Nannomys on the basis of skeletal features by J. T. Marshall (Smithsonian Natural History Museum, Washington, DC), and the other wild-trapped African Muridae (Arvicanthis niloticus, Lemniscomys striatus, Hybomys univittatus, Uranomys ruddi, Lophuromys sikapusi, Lophuromys flavopunctatus, Mastomys natalensis, Mastomys fumatus, Mastomys stella, and Grammomys dolichurus) were typed using standards at the NY Natural History Museum.
Mouse taxa are identified by genus and species, or where needed for clarification, by genus, subgenus, and species.
Viral pseudotypes carrying the LacZ reporter were generated for the various X/P-MLVs by infecting GP2-293 cells transfected with pCL-MFG-LacZ as described previously (49). Cells were tested for susceptibility by infection with appropriate dilutions of these pseudotype virus stocks in the presence of 4 to 8 μg/ml Polybrene. One day after infection, cells were fixed with 0.4% glutaraldehyde and assayed for β-galactosidase activity, using as a substrate 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal; 2 mg/ml) (ICN Biomedicals, Aurora, Ohio). Infectious titers were expressed as the number of blue cells per 50 μl of virus supernatant.
The full-length Xpr1 genes from Mus spicilegus, Mus shortridgei, M. domesticus (WSB), M. musculus (CZECHI/EiJ), Mus minutoides, goat, cat, ferret, rabbit, gerbil, buffalo, armadillo, dog, African green monkey (AGMK), and bat cells were amplified by reverse transcription (RT)-PCR, cloned into pCR2.1-TOPO, and sequenced. Primers for amplification were 5′-ATGAAGTTCGCCGAGCACCTCTC and 5′-AGTGTTAGCTTCGTCATCTGTGTC.
Xpr1 exon13 sequences, containing ECL4, were amplified from DNA of the rodents and other mammals listed below in Fig. Fig.11 and and22 by using primers Ex13F, 5′-GCCTATTACTACTGTGCC and Ex13R, 5′-CGGAAAACCTCAAGGGGG. Xpr1 exons 10 to 12, which contain ECL3, were amplified from many of the same DNAs using the following two sets of primer pairs: Ex10-11F (5′-CGAGTATTTACTGCTCCCTTCC) and Ex10-11R (5′-GCAGGGATGGCAAAGTCCAGG) and Ex12F (5′-GGAGTTCCTGGAGAGAGAGTGG) and Ex12R (5′-GACCCCAGTCCATCTTGACATCC). All fragments were cloned into pCR2.1-TOPO before sequencing.
Five novel mutant variants of the Xpr1 gene were generated by mutagenesis PCR with the QuikChange II XL site-directed mutagenesis kit (Stratagene, La Jolla, CA), using as templates previously described clones of Xpr1n and Xpr1sxv (48). Mutations were produced using the following primers and their reverse complements: T583K (5′-CCCAACATGAGGCTTAAACTTTGTAGCAGTAATAGAG), I579Δ (5′-GACTATCCAAATCTCTACTGCTACAACGTTTAAGC), S578A (5′-GGACTATCCAAATCGCCATTACTGCTACAACG), T580A (5′-ATGAGGCTTAAACGTTGTAGCGGCAATAGAGATTTGGATAGTC), and G589A (5′-CGTTTAAGCCTCATGTTGCCGACATCATTGCTACTG). All mutants were confirmed by sequencing.
The recombinant plasmids were transfected into E36 Chinese hamster cells. 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 M-PER 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 using anti-V5 antibody (Invitrogen, Carlsbad, CA) followed by goat anti-mouse IgG conjugated with horseradish peroxidase (HRP) (Invitrogen). The membrane was then stripped and incubated with mouse anti-α-tubulin (Sigma, St. Louis, MO) and goat anti-mouse IgG conjugated with HRP (Invitrogen).
Two data sets were analyzed: one for the rodent ECL3-4 sequences and one for the rodent ECL4 sequences. DNA sequences were aligned using MUSCLE (9) and improved manually. Kimura two-parameter distance-based neighbor-joining phylogenies were generated for each set by PHYLIP (version 3.68) (10) and were corrected for closer correspondence to the appropriate consensus phylogeny of murid rodents (38) and mice (28, 45).
The codeml program of the PAML4 package (52) was used for maximum likelihood analysis of codon evolution (2). Both lineage-specific and codon-specific analyses were performed. In the lineage-specific selection analyses, the free-ratio model (codon model = 1) was used to calculate branch-specific rates of nonsynoymous/synonymous evolutionary changes (dN/dS). In this model, each branch is assumed to have a specific dN/dS ratio. The likelihood of the phylogeny under this model was tested against the likelihood of the phylogeny under the model of one uniform dN/dS ratio across all branches (codon model 0) using a likelihood ratio test (LRT). The significance of the LRT value was assessed using a chi-square distribution with 37 degrees of freedom for the ECL3/4 analysis and 53 degrees of freedom for the ECL4 analysis.
Selection acting on Xpr1 codons was analyzed using two models of equilibrium codon frequencies and four models of codon selection. The two codon frequency models used were the F3x4 model (codon frequencies estimated from the nucleotide frequencies in the data at each codon site) and the F61 (codon table) model (frequencies of each of the 61 non-stop codons estimated from the data). The codon selection models were two neutral/negative selection models (M1 and M7) which were compared against corresponding positive selection models which included a category for dN/dS > 1 (M2 and M8, respectively). The significance of this additional codon selection category was assessed using LRTs of the phylogeny likelihoods under the neutral and positive selection models. Significance of the test statistics was calculated using a chi-square distribution with 2 degrees of freedom. The posterior probabilities of individual codons experiencing dN/dS > 1 was calculated using the naïve empirical Bayes (NEB) and Bayes empirical Bayes (BEB) algorithms (53).
The sequences of the Xpr1 genes have been deposited in GenBank under accession no. HQ022826 to HQ022857.
Mouse XPR1 is encoded by a gene on chromosome 1 with 15 exons over 140 kb. We screened 45 DNAs from species in all four Mus subgenera and mice trapped in different geographic locations for sequence variations in exon 13 (Fig. (Fig.11 and and2;2; see Table S1 in the supplemental material). Exon 13 encodes the Xpr1 ECL4, a highly variable segment that contains receptor determinants (31). We also tested cells from available taxa for virus susceptibility.
ECL4 has only 13 amino acid residues, but deletions or replacement mutations in Mus Xpr1 genes affect 8 codon sites (Fig. (Fig.1).1). These mutations define six receptor variants, four of which represent previously described alleles that differ in their ability to mediate entry of the various members of the X/P-MLV family of viruses (18, 29, 31, 48). One of the two novel sequence variants, here termed Xpr1m, has a virus restrictive phenotype. A sixth sequence variant in the African pygmy mouse species M. tenellus was also identified (Fig. (Fig.1),1), but it was not characterized for receptor function. The species and geographic distribution of these sequence variants in Mus and the relationship between Xpr1 variation and exposure to X/P-MLV infection can be described as follows.
The largest number of Xpr1 sequences, 27, had an Xpr1sxv-like ECL4 (Fig. (Fig.2).2). The sxv receptor variant is fully permissive for all host range groups of X/P-MLVs (18, 49), and the 27 mice that carry this sequence include species in 3 of the 4 Mus subgenera (Fig. (Fig.22 and and3).3). We also identified this ECL4 sequence in other Murinae: rat and five wild-trapped mice in four species of two African genera, Mastomys and Grammomys (Fig. (Fig.2).2). The conservation of the sxv-like ECL4 sequence is consistent with the observation that most species of Mus as well as rat cells are permissive for X/P-MLV infection (6, 18, 23, 35). Surprisingly, this sequence was also found in all mice classed as M. domesticus (see below); these included six wild-trapped and three wild-derived lines of M. domesticus mice of Western Europe and North Africa, as well as three wild-caught California house mice and breeding lines of wild-derived mice from the Delmarva (Delaware-Maryland-Virginia) Peninsula and from South America.
The previously described variant Xpr1p restricts P-MLVs and Cz524, but not X-MLVs or CasE#1, and was originally described in M. Coelomys pahari (48). We also identified this variant in one additional Southeast Asian species, M. Pyromys shortridgei, although the other Pyromys species tested, M. Pyromys saxicola, carries Xpr1sxv (Fig. (Fig.2).2). Thus, Mus species of two subgenera carry a restrictive XPR1 receptor, while other Asian, palearctic, and African species, as well as other species of Murinae, carry the permissive sxv variant. Previous Southern blot analysis of genomic DNA from one of the two species carrying the Xpr1p variant, M. pahari, failed to identify X-MLV or P-MLV env-related endogenous copies (19). The absence of endogenous copies indicative of past virus exposure that might have selected for receptor mutations restricting entry suggests either that the evolutionary pressures resulting in the replacement mutations fixed in this species are unrelated to XPR1 receptor function, or these substitutions evolved in conjunction with an undiscovered retrovirus infection.
There are four species of house mice in three major lineages (Fig. (Fig.3).3). It is these house mouse species that uniquely carry endogenous X/PMV env sequences (3, 20, 42), and three different restrictive Xpr1 variants are found in these mice (Fig. (Fig.2)2) (1, 18, 29, 41, 51). Xpr1c restricts all viruses but X-MLV and was originally identified in M. castaneus (29) (Fig. (Fig.2).2). The sequence of this gene is identical to Xpr1sxv, except for a five-codon deletion in ECL4 (31). M. castaneus mice are found in Southeast Asia and carry dozens of XMV env genes, some of which are capable of producing infectious virus (4, 20, 48), at least one of which is associated with interference-mediated resistance to infection (47). Three laboratory breeding lines of M. castaneus, CAST/EiJ, CAST/Rp, and CAST/Ncr, were found to carry Xpr1c (Fig. (Fig.2).2). Xpr1c was not, however, found in any of the three lines of the Japanese mouse M. molossinus (Fig. (Fig.2),2), which is a naturally occurring hybrid of M. castaneus and M. musculus (54) (Fig. (Fig.33).
A novel ECL4 sequence variant was identified in two house mouse species, the Japanese mouse M. molossinus and M. musculus, which has a range extending from eastern Europe to the Pacific. Both species carry dozens of copies of XMVs and few copies of PMV env genes (20, 42), and infectious virus has been isolated from both species (4, 19, 25, 49). The full-length Xpr1m sequence is distinguishable from Xpr1sxv at two sites, both of which are in ECL4: a deletion, I579Δ, and a substitution, T583K. To characterize receptor function of this novel variant, we infected cultured tail cells from three wild-derived strains carrying this gene: MOLG/DnJ, CZECHII/EiJ, and MOLF/EiJ. As shown for MOLG/DnJ (Fig. (Fig.4),4), these cells were susceptible to infection by amphotropic MLV but were not infected by P-MLVs or wild mouse X/P-MLVs and showed only low levels of susceptibility to CAST-X X-MLV. Such a restrictive phenotype can be due to the presence of interfering env genes as well as receptor polymorphism (47; reviewed in reference 39), so Xpr1m was cloned and expressed in E36 Chinese hamster cells that show only trace levels of X-MLV susceptibility. Transfectants showed increased susceptibility to CAST-X compared to E36 cells and showed no susceptibility to other X/P-MLVs, indicating that this susceptibility profile is receptor mediated (Fig. (Fig.44).
The mutations found in Xpr1m, I579Δ and T583K, were individually introduced into Xpr1sxv. In transfected cells, the Xpr1 clone with T583K showed a 100-fold reduction in susceptibility to X-MLVs and wild mouse X/P-MLVs relative to M. dunni cells but a 10,000-fold decrease in susceptibility to P-MLVs (Fig. (Fig.4).4). Cells expressing Xpr1-I579Δ showed poor but reproducible susceptibility to CAST-X and, at best, trace levels of infectivity with all other X/P-MLVs. This suggests that both mutations contribute to the virus resistance of this XPR1 variant, with the deletion having a greater effect on restriction. These results also confirm that the presence of K500 in ECL3 is not sufficient for efficient receptor function, consistent with the previous suggestion that the receptor interface requires residues in both ECLs (48).
The Xpr1n variant was originally cloned from laboratory mouse NIH 3T3 cells (1, 41, 51), which, like cells of various inbred laboratory strains, are resistant to infection by X-MLVs. The common strains of laboratory mice are a mosaic of three house mouse species, M. musculus, M. domesticus, and M. castaneus, with the largest genomic contribution coming from M. domesticus (50). The expectation that we would therefore find Xpr1n in M. domesticus was also suggested by the fact that restriction of X-MLVs by Xpr1n is consistent with the absence of XMV ERVs in M. domesticus (20). Surprisingly, however, all nine M. domesticus mice of Europe and North Africa carry Xpr1sxv rather than Xpr1n, as do wild-trapped or wild-derived mice from Peru and the East and West coasts of the United States (Fig. (Fig.2).2). Thus, if the Xpr1n allele exists in wild mouse species, it has a limited distribution.
Previous mutational analysis of Xpr1n determined that residues K500 and T582 function as independent receptor determinants for X-MLV and Cz524, but only T582 also produces a receptor for CasE#1. Because Xpr1n also restricts XMRV, we tested Xpr1n mutants carrying Δ582T or E500K for sensitivity to XMRV to determine if either or both of these residues are important for XMRV entry. Both transfectants were infected by XMRV, although XMRV infectivity was less efficient in the Xpr1n-Δ582T mutant, with LacZ virus titers reduced 1 to 2 logs relative to the E500K mutant, although the cells were equivalently infectible by X-MLV (Fig. (Fig.55).
The substantial polymorphism seen in the Xpr1 ECL4 of Mus is also characteristic of the 16 additional sequenced Xpr1 genes of other rodent species (Fig. (Fig.1).1). ECL4 variability is due largely to replacement mutations; however, the three house mouse Xpr1 alleles Xpr1c, Xpr1m, and Xpr1n, and only these three alleles, carry deletions in this region of XPR1 (Fig. (Fig.1).1). These three alleles all encode restrictive XPR1 receptors that restrict two or more viruses in this family, and the appearance of these variants roughly coincides with the acquisition of MLV ERVs (Fig. (Fig.3).3). The hypervariability of this region in functional receptors suggests that XPR1 can tolerate multiple ECL4 replacement mutations, and the fact that three different deletions lie in this region in virus-positive mice suggests either that the six deleted residues are important for receptor function or, alternatively, that altering the size of this ECL may be an efficient way to disable receptor function in mice exposed to endemic retrovirus infection.
The additional sequence variants identified in other murid rodents were not tested for receptor function. Instead, we used a phylogenetic approach to try to identify residues showing evidence of genetic conflicts. Host genes involved in antagonistic interactions with pathogens can be identified by sequence comparisons that reveal positive selection through an excess of nonsynonymous substitutions. To identify evidence of a possible antiviral role for Xpr1 throughout Mus evolution, and to identify possible sites of virus interaction, we analyzed two sets of rodent Xpr1 genes: a set of 28 sequences of the 138-bp exon that includes ECL4, and a 543-bp segment from 19 rodents that encodes ECL3 and ECL4 (see Fig. S1 and S2 in the supplemental material).
Taxonomy-based phylogenetic trees were generated for both sets of sequences. Values of dN/dS along each branch were calculated by using the free-ratio model of PAML. A dN/dS value of >1 suggests that positive selection has acted along that lineage. For the ECL3/4 DNA set, one branch shows clear positive selection, while none of the others had more than three inferred nonsynonymous changes when there were no synonymous changes (see Fig. S3 in the supplemental material); for the ECL4 set, only one branch had more than 3 nonsynonymous changes when there were zero synonymous changes (not shown).
Likelihood ratio tests of the ECL3/4 data set indicate that Xpr1 has a significant probability of having experienced positive selection under the F3x4 model (Fig. (Fig.66 A), whereas ratios for the ECL4 data set were below the level of significance (not shown). We calculated posterior probabilities of positive selection at individual codon sites in the ECL3/4 data set using two methods in the PAML suite of programs: the Bayes empirical Bayes (BEB) calculation and the naïve empirical Bayes calculation (NEB) (see Table S2 in the supplemental material). NEB identified nine codons under positive selection with posterior probability of >0.95; four of these sites were also identified by the more conservative BEB (Fig. (Fig.6B).6B). Six of these selected codons are in ECL3. Two of these ECL3 codons, 434 and 442, lie near the 5′end of this ECL, but the others (codons 502, 503, 507, 508) are clustered at the 3′ end in a region known to harbor entry determinants; positively selected residues T507 and V508 were previously identified as critical for entry of X/P-MLVs (49). The other ECL3 residues under positive selection have not been tested as receptor determinants, but 503 marks a potential glycosylation site in some XPR1 variants. The NEB analysis identified three codons in ECL4 under positive selection (Fig. (Fig.6B),6B), although only one of these, 584, approached significance using BEB. Of these four codons, 583, as indicated in the Xpr1m analysis, affects P-MLV entry. Also, two of these selected ECL4 codons are deleted in Xpr1c, and the deletion of one of the two adjacent Thr residues at 582 and 583 in Xpr1n eliminates an X-MLV receptor determinant (31). The fact that most of the amino acid sites identified as being under positive selection are at or near sites implicated in receptor function suggests that this gene may have experienced positive selection in Mus evolution because of an antiviral role.
Because of known examples of the transmission of mouse gammaretroviruses to other species, including humans (8, 26, 27, 33, 37), we extended our analysis of the functional and sequence variation of the Xpr1 receptor to other mammals with special emphasis on species that serve as pets or farm animals. Cells from multiple species were tested for virus susceptibility (Table (Table1),1), and Xpr1 sequences were determined or acquired from sequence databases (Fig. (Fig.11).
Alignment of these Xpr1 genes confirms that there is striking hypervariability in ECL4, as seen in the murid genes (Fig. (Fig.1);1); however, this 13-residue ECL contains 3 nonvariant residues, S578, T580, and G589. Such nonvariant residues are often found near hypervariable regions in different virus receptors (40), and while mutational analysis has generally shown that such hypervariable residues are important for entry, it is also possible that these residues are not the actual receptor determinants. Instead, these variable residues may be negative control regions that govern accessibility to nearby conserved residues that serve as the actual viral attachment sites (40). Reliance on conserved rather than variable residues could provide the virus with a more reliable port of entry. To determine if these three conserved ECL4 residues are needed for receptor function, we mutated each of these sites in Xpr1sxv (Fig. (Fig.7).7). When these Xpr1 genes were expressed in hamster cells, however, all three were highly permissive receptors, suggesting these residues do not contribute significantly to the receptor interface.
Using the same seven X/P-MLV isolates that were used to test rodent cells for susceptibility, we identified distinctive patterns of virus susceptibility among the nonrodent mammals, including several patterns not seen in mice: for example, the restriction of P-MLVs and both wild mouse isolates by dog and buffalo cells and the restriction of P-MLVs by bat cells (Table (Table1).1). All of the cells were susceptible to X-MLV, as has been noted previously for some of them (6, 23, 35). The cells with restrictive receptors all restricted P-MLVs, and some of these cells also restricted CasE#1 or restricted both wild mouse MLVs. This indicates that X-MLVs are the most accommodating of receptor polymorphisms, and P-MLVs are the least accommodating. Comparison of the permissive and restrictive amino acid sequences identified 15 different replacements unique to restrictive genes in the regions involved in receptor function, namely, ECL4 and the 9-residue segment at the C-terminal end of ECL3 (data not shown). None of these replacements showed obvious correlations with specific patterns of restriction of P-MLVs, CasE#1 and Cz524.
Cells from two species were uniquely resistant to XMRV. As reported previously (34, 49), Chinese hamster cells are generally resistant to infection by X/P-MLVs, but interference with glycosylation by mutation or glycosylation inhibitors results in susceptibility to X-MLVs, but not to XMRV, AKR6 X-MLV, or P-MLVs. The current screening determined that gerbil GeLu cells are susceptible to all mouse X-MLVs, including AKR6, but are resistant to XMRV (Table (Table11 and Fig. Fig.8).8). Expression of the gerbil XPR1 receptor in E36 hamster cells reproduced the gerbil susceptibility pattern (Fig. (Fig.8).8). To determine if there are additional factors in gerbil cells that restrict XMRV, we made a stable gerbil cell transfectant expressing exogenously introduced Xpr1. This transfectant expressed AGMK COS-1 Xpr1, which supports infection by both X-MLV and XMRV. These transfectants showed a small increase in susceptibility to X-MLV and were efficiently infected by XMRV, supporting the conclusion that gerbil cell resistance to XMRV is receptor mediated (Fig. (Fig.88).
We compared the sequences of the full-length XMRV-restricting Xpr1 genes of Chinese hamsters and gerbils. Both genes carry K500 and T582, residues that function as independent X-MLV receptor determinants in the mouse and that also function as XMRV receptor determinants in the mouse (Fig. (Fig.5).5). The presence of both of these residues in hamster and gerbil cells suggests that infectivity by XMRV is restricted in these two cells by other residues. The XPR1 genes of these species differ from XMRV-permissive XPR1 genes at multiple sites, but these two restrictive genes have two shared substitutions not found in any XMRV-susceptible species (Fig. (Fig.1).1). Both of these residues, A583 and Q585, are in the ECL4 receptor-determining region, suggesting that one or both of these substitutions are responsible for this restriction. These results indicate that XMRV has unique receptor requirements that distinguish it from its X-MLV relatives.
Although no cellular function has been identified for mammalian XPR1, retrovirus receptors control the spread of elements responsible for genetic variation and disease and thus have important roles in the evolution of host species subject to infection (30). In this study, we describe XPR1 sequence and functional variation in the natural hosts of these gammaretroviruses and also in other mammalian species that could serve as virus reservoirs or targets for trans-species transmission. This analysis has allowed us to address two questions: (i) how receptor polymorphism helps natural populations adapt to exposure to infectious disease-inducing retroviruses and (ii) whether the potential for transspecies transmission in different mammalian species is restricted or facilitated by functional XPR1 polymorphisms. In the course of these studies, we characterized a novel restrictive Xpr1 allele in Mus and identified a correlation between exposure to MLV infection and appearance of restrictive XPR1s caused by deletion. We used mutagenesis and phylogenetics to evaluate the functional contributions made by constrained, variable, and deleted residues in XPR1.
There are numerous examples of trans-species transmissions of retroviruses. The most well-known example is, of course, the derivation of HIV-1 from simian lentiviral precursors, but there are other examples, and retroviruses that cluster with mouse gammaretroviruses are found in multiple vertebrates. Martin and colleagues (33) found MLV-related ERVs in approximately one-fourth of vertebrate taxa and identified recent zoonotic transmissions from mammals to birds and from eutherians to metatherians. It would not be surprising to find more examples of interspecies transmissions involving MLVs, since MLV-infected house mouse species have a worldwide geographic distribution (32), and all mammalian species tested are permissive to infection by some or all X/P-MLVs.
The XMRV virus found in some human patients may have been acquired directly from mice or after transmission from mice to another species in contact with humans. We therefore examined the reported worldwide incidence of prostate cancer (7) relative to the geographic distribution of the various Xpr1 alleles of house mice. The highest rates of this disease are found in the United States, and the lowest rates are found in Asian countries like Japan, India, and China. In Europe, rates are highest in Austria, France, and Scandinavia and lowest in Eastern Europe. This distribution roughly corresponds to the distribution of Xpr1 receptor variants in mouse populations, with the most permissive allele, sxv, found in high-tumor-incidence areas like America and Western Europe, and with the allele most restrictive of X-MLVs, Xpr1m, found in low-tumor areas like Japan and Eastern Europe. This Eurasian group of mice also carries receptor-blocking genes (47) that further suggest that these mice might provide a poor reservoir for zoonotic transmission to humans. While the American and Western European mice with Xpr1sxv carry only poorly expressed endogenous PMVs and few or no endogenous XMVs, wild mice are also known to carry viruses that have not become endogenized (12, 36, 46), and at least one virus in the X/P-MLV family, CasE#1, was isolated from a mouse trapped in California. Clearly, more studies are needed to examine wild mouse populations for infectious virus.
The isolation of XMRV from human patients raises the question of whether adaptation to this species has altered its tropism. XMRV shows 95% overall identity to X-MLVs and 93% identity in the 5′ env receptor binding domain. XMRV, as shown here, can utilize the receptor determinants for X-MLVs defined by residues K500 and T582 to enter mouse cells, although K500 is a more effective receptor for XMRV. Unlike the mouse X-MLVs that are able to infect all mammals, however, XMRV is restricted by hamster and gerbil cells and these XPR1 genes share two substitutions, T583A and L/K585Q. Residues A583 and Q585 are not found in other mammals, except for a few rodents that have not been tested for XMRV susceptibility. Other evidence implicates these two codons in virus entry: both residues are deleted in Xpr1c, 583 is under positive selection, and the T583K substitution found in M. molossinus was associated with P-MLV restriction. This XMRV tropism difference demonstrates that the XMRV-receptor interaction differs from that of the mouse viruses and raises the possibility that this may represent an XMRV adaptation acquired through contact with humans or with an as-yet-undiscovered species before transmission to humans.
There are six XPR1 codons important for virus entry identified by mutagenesis of Mus alleles (500, 507, 508, 579, 582, and 583) (31, 48, 49). Expanding this receptor analysis to other mammals should be fruitful; analysis of the gerbil XPR1 has implicated A583 and Q585 in XMRV entry. That further studies on the human and other XPR1 genes could provide additional information on the receptor virus interface is suggested by several observations: the extensive sequence variation in permissive mammalian XPR1s in the regions involved in entry; the fact that various mammalian cells, like bat and dog cells, show resistance patterns not found in mice; the fact that studies on human/hamster Xpr1 chimeras implicate as yet unidentified ECL4 residues in P-MLV entry (44); and the fact that X/P-MLV interference patterns differ in cells with different Xpr1 genes (5, 6; C. A. Kozak, unpublished data).
Phylogenetic analysis of Xpr1 indicates that it is under diversifying selection in rodents, suggesting that this gene has had a defensive role in rodent evolution. There is ample evidence that some wild mouse populations have been exposed to viruses that use the XPR1 receptor: Some species carry MLV ERVs acquired from past infections (20, 42), some of which can produce infectious virus (4, 19, 48, 49), and some wild mice also carry infectious MLVs that have not become endogenized (12, 46, 36). For such populations, survival is enhanced by host factors that restrict virus, and the XPR1 receptor is clearly one of those factors. Among the five Mus XPR1 variants, the one with the broadest susceptibility phenotype, Xpr1sxv, is widely distributed among the Eurasian Mus species, and the sxv ECL4 sequence is also found in multiple species that predate Mus. The species with this ECL4 sequence either lack X/P-MLV ERVs or carry only PMVs, ERVs not known to produce infectious virus (Fig. (Fig.3)3) (17, 20).
The four restrictive Mus XPR1 polymorphisms appeared at two distinct points in Mus evolution. First, Xpr1p appeared about 7.5 million years ago (MYA), shortly after the divergence of Mus from other Murinae (28, 45), but this allele is confined to two species of Southeast Asian mice. The presence of this restrictive receptor in mice that do not carry XMV or PMV ERVs (20) suggests that either the fixation of this variant is unrelated to receptor function, or these mice were exposed to an as-yet-undescribed retrovirus infection. The other three restrictive XPR1s arose later in Mus evolution, in the house mouse complex, which appeared about 0.5 MYA (13). House mice are distinguishable from other Mus species by two notable features. First, these mice are behaviorally different from other Mus species in their dependence on humans; these mice live in our houses, barns, warehouses, and ships and travel wherever we go. Second, these house mice have all been exposed to MLV infection and carry numerous endogenous copies of X/P-MLVs, some of which have remained active. Acquisition of these germ line viral sequences is roughly coincident with the appearance of the restrictive house mouse variants Xpr1m and Xpr1c, both of which, like the ERV sequences in these species, show an apparent species-wide distribution. Thus, the mutations in these two variants restricting Xpr1 receptor function likely provided a survival advantage in the face of endemic infection by potentially mutagenic and pathogenic gammaretroviruses and may have contributed to the “arms race” between virus and host by providing the selective pressure that produced viral variants with altered receptor specificities.
We had expected Xpr1n to be widespread in wild M. domesticus for several reasons. First, Xpr1n was initially identified in laboratory mice, and although M. musculus, M. castaneus, and M. domesticus all contributed to the fancy mouse colonies used to generate the common inbred strains, M. domesticus is, by far, the most significant contributor to the laboratory mouse genome (50). Second, the complete absence of XMV env genes in M. domesticus is consistent with Xpr1n restriction of X-MLV infection. The failure to identify Xpr1n in any M. domesticus mouse trapped in geographically disparate locations in Western Europe and the Americas indicates, however, that this allele is not responsible for the absence of XMVs in these mice, an absence that also marks all other species with Xpr1sxv. The failure to identify Xpr1n in the wild may be a consequence of limited sampling, but the fact that Xpr1n does not show the apparent species-wide distribution of Xpr1m and Xpr1c suggests that this allele evolved only recently in Mus, perhaps in the fancy mouse colonies that provided progenitors of the common laboratory strains (50). Testing of additional Eurasian wild mouse populations and laboratory breeding stocks may identify the origin of this restrictive allele.
The presence of different deletion mutations in the 13-residue ECL4 of the three restrictive alleles found in house mice is unusual in the natural history of this gene. No ECL4 deletion mutations were found in XPR1 genes of nonrodent mammals, and none appeared during the 7 million years of Mus evolution prior to the house mouse radiation. These independent deletion mutations are thus unlikely to have resulted from inherent structural features in this region of the gene, because such deletions would not then be restricted to house mice. The appearance of these variants, each of which blocks two or more viruses in the XPR1 family, coincides with the acquisition of germ line MLV env genes, suggesting that, in the face of infection, replacement mutations may not be an effective way to disable or alter receptor function. After all, permissive receptors have sustained multiple ECL4 substitutions. That all three of the restrictive alleles carry deletion mutations suggests either that these six residues are critical for entry or that the size of the ECL4 loop may affect receptor function.
Finally, these studies do not provide any special insight into the origins or evolutionary relationships of the two major host range variants of MLVs that use XPR1, X-MLV and P-MLV. The fact that P-MLVs are found in M. domesticus that were thought to carry Xpr1n prompted the reasonable suggestion that these viruses may have evolved in response to XPR1 mutations that enabled European mice to evade X-MLV infections (31, 49). Our present studies, however, indicate that the acquisition of P-MLV ERVs predates Xpr1n. The observed distribution of ERVs and XPR1 variants also does not explain why P-MLVs integrated selectively into European mice carrying Xpr1sxv but are not found in other Eurasian species carrying the same receptor variant.
This research was supported by the Intramural Research Program of the NIH, NIAID.
Published ahead of print on 15 September 2010.
†Supplemental material for this article may be found at http://jvi.asm.org/.