Common Inbred Strains of the Laboratory Mouse That Are Susceptible to Infection by Mouse Xenotropic Gammaretroviruses and the Human-Derived Retrovirus XMRV ^▿

Viruses, cells, and virus infectivity.

CAST-X is an X-MLV isolated in our laboratory from the spleen of a CAST/EiJ mouse (46). Cz524 is a novel MLV isolated from the spleen of a CZECHII/EiJ mouse (47). The human xenotropic murine leukemia virus-related virus, XMRV (7), was kindly provided by R. Silverman (Cleveland Clinic, Cleveland, OH). AKR6 X-MLV, CasE#1 (5), Moloney mink cell focus-inducing virus (MoMCF), and Friend MCF (FrMCF) were originally obtained from J. Hartley (NIAID, Bethesda, MD).

Susceptibility to MLVs was tested in NIH 3T3 cells, in Mus dunni cells (20) obtained from M. Lander (NIAID, Bethesda, MD), and in cultured fibroblasts developed from tail biopsy specimens from adult mice (21). Retrovirus stocks carrying the LacZ reporter were generated for the various MLVs by virus infection of GP2-293 cells transfected with pCL-MFG-LacZ (47); the virions produced by these cells represent a mix of infectious virus and particles carrying LacZ. To test for virus susceptibility in a single-round assay, cells were infected 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 the substrate 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal, 2 mg/ml; ICN Biomedicals, Aurora, OH). Infectious titers were expressed as the number of foci of blue cells per milliliter of virus.

Genomic DNA and RNA.

DNAs were extracted from mice obtained from several sources. The following mice were obtained from M. Potter (NCI, Bethesda, MD): M. domesticus mice trapped at various sites (Abu Rawash, Egypt; Haven's Farm, Davidsonville, MD; Lewes, DE; J. J. Downs, Ridgely, MD; Watkin's Star, Centreville, MD), M. molossinus mice (MOL/Li), M. castaneus mice (CAST/Ncr, CAS/Li), M. musculus mice (Vejrumbro, Denmark; Skive, Denmark), and M. spretus mice (Spain). Three wild-trapped mice from Lake Casitas, CA, were obtained from S. Rasheed (University of Southern California, Los Angeles). CAST/Rp mice were obtained from R. Elliott (Roswell Park Cancer Institute, Buffalo, NY), and SAMP8 and SAMR1 samples were obtained from R. Carp (New York State Institute for Basic Research in Developmental Disabilities, Staten Island, NY). F/St mice were obtained from H. Morse (NIAID, Bethesda, MD) in the early 1980s; this strain is now extinct. 129-GIX+ mice were obtained from E. Boyse (Memorial Sloan-Kettering Cancer Center, NY). SIM.R and SIM/Ut mice (44) were obtained from A. Axelrad (University of Toronto, Toronto, Canada). AKR/N and NFS/N mice were obtained from the Small Animal Section, NIH (Bethesda, MD). NFS/N-Xpr1^sxv congenics were developed in our laboratory and carry the Xpr1 allele from M. domesticus (formerly Mus praetextus), provided by M. Potter. DBA/2J, DBA/1J, C58/J, C57BL/10J, and SWR/J mice were obtained from the Jackson Laboratory (Bar Harbor, ME).

DNAs were also isolated from the cell lines NZB-Q and SC-1, obtained from J. Hartley (NIAID, Bethesda, MD), and from cells of M. domesticus (formerly M. praetextus) obtained from J. Rodgers (Baylor College of Medicine, Houston, TX). Mus fragilicauda DNA was provided by J. Hartley. The Chinese hamster/mouse somatic cell hybrid BE7-2 was produced by the fusion of E36 cells and primary fibroblasts of BALB/c cells, and its mouse chromosome content was determined by examining Giemsa-stained metaphase spreads and by testing for marker loci (14).

Additional DNAs provided by H. Morse and S. Chattopadhyay (NIAID, Bethesda, MD) include Mus abbotti and M. musculus mice trapped in Vejrumbro, Denmark; Brno, Czechoslovakia; Belgrade, Yugoslavia; and Viborg, Denmark. DNA from M. macedonicus was provided by R. Elliott. All other DNAs from common inbred and wild-derived strains were purchased from the Jackson Laboratory (Bar Harbor, ME).

Cloning and sequencing of Xpr1, Fv1, and the Bxv1 insertion site.

The full-length Xpr1 genes were amplified by reverse transcription-PCR (RT-PCR) from cells established from SWR tail biopsy specimens and from the cell line SJLSV, a simian virus 40 (SV40)-transformed line of SJL fibroblasts produced by B. Knowles (Institute of Medical Biology, Singapore). Primers used were 5′-ATGAAGTTCGCCGAGCACCTCTC and 5′-AGTGTTAGCTTCGTCATCTGTGTC. A segment of Xpr1 containing the third and fourth putative extracellular loops (ECL3 and ECL4, respectively) was amplified by RT-PCR from NZB mouse spleen RNA using the forward primer 5′-CGAGTATTTACTGCTCCCTTCC and Ex13R, 5′-CGGAAAACCTCAAGGGGG. Xpr1 exon 13 sequences, containing ECL4, were amplified from the DNA of multiple strains and species using primers Ex13F (5′-GCCTATTACTACTGTGCC) and Ex13R. All PCR products were cloned into pCR2.1-TOPO and sequenced.

PCR primers for the Bxv1 provirus-cell junction fragment and the empty locus at the Bxv1 insertion site were designed based on Mus musculus chromosome 1 clone RP24-65D16 (GenBank accession no. AC115959). The Bxv1 provirus is present in the reverse orientation at coordinates 107533 to 116189. A Bxv1 provirus-cell junction fragment was PCR amplified using primers 5′-AGGAGTAGGAACAGGGACTACAGC (Bxv1ENFP; nucleotide coordinates 108635 to 108612) and 5′-GATCTAGAGTCACCTAGGCAACAAGC (Bxv1FLRP; nucleotide coordinates 106765 to 106740). The empty locus at the Bxv1 insertion site was amplified using primers 5′-CCATGCTAAATACCTCAGACTCCAGTG (Bxv1FLFP; nucleotide coordinates 116359 to 116333) and 5′-GATCTAGAGTCACCTAGGCAACAAGC (Bxv1FLRP; nucleotide coordinates 106765 to 106740) (Fig. 1). Representative PCR products were cloned into pCR2.1-TOPO and sequenced.

PCR primers for the 3′ half of the Fv1 gammaretrovirus restriction gene, MHRF and GT17a (45), generate fragments of 510 bp for Fv1ⁿ and Fv1^nr and 1.8 kb for Fv1^b. The 510-bp PCR product from SWR/J mice was cloned into pCR2.1-TOPO and sequenced.

Nucleotide sequence accession numbers.

The sequences of the Xpr1 genes for 19 mouse strains (129, AKR, DBA, F/St, FVB, HRS, LP, LT, MA/My, NOD, PWD, RBA, RIIIS, SOD, SWR, YBR, NZB, SIM.R, SJL) were deposited in GenBank under accession no. HQ323662 to HQ323680. Sequences for the empty Chr 1 Bxv1 site in NFS/N and the SWR Fv1 gene were deposited in GenBank under accession no. HQ323681 and HQ323682.

Inbred strain distribution of the inducible XMV locus Bxv1.

We designed primers to identify cell-virus junction fragments diagnostic of the Bxv1 provirus, as well as primers to identify the empty locus (Fig. 1). Primers were initially tested on DNAs from mouse strains previously shown to lack Bxv1, NFS and NIH 3T3, and on DNAs from C57BL and BALB/c, the strains in which Bxv1 was initially discovered (18). We also tested a hamster/mouse somatic cell hybrid, BE 7-2, carrying fragments of two mouse chromosomes, Chr 1 and Chr 15. This hybrid, like mouse cells carrying Bxv1, produces infectious X-MLV following induction (14), and it contains a single XMV provirus (11). The 1.8-kb cell-virus junction fragment was identified only in DNAs from the hybrid cell line and from mice known to carry Bxv1, whereas the 1.0-kb fragment diagnostic of the empty locus was found only in DNAs from mice or cells lacking Bxv1 (Fig. 1), thus confirming that these primers identify the Bxv1 locus. The sequence of the 1.9-kb cell-virus junction fragment produced by Bxv1ENFP and Bxv1FLRP was identical to the corresponding Chr 1 segment in the sequenced C57BL genome (GenBank accession no. AC115959), and the sequence of the Bxv1FLFP/Bxv1FLRP product from the Bxv1-negative NFS/N mouse represents the provirus-free insertion site (GenBank accession no. HQ323681).

We analyzed 50 laboratory mouse strains, most of which fall into several inbred strain categories defined by their related breeding histories (3) (Fig. 2). The majority of the analyzed strains originated from the fancy mouse colonies of breeder Abbie Lathrop and from the early lines developed by William Castle; one set of Lathrop mice provided to William Castle gave rise to many common strains, such as DBA and BALB/c, and additional Lathrop mice produced the C57/C58-related strains. A third set of strains was derived from Swiss mice. Additional strains were developed from stocks in New Zealand, and we also included strains of other or uncertain ancestry in this analysis. We did not test the various strains derived by intercrossing other inbred strains.

Bxv1 was identified in 17 strains, most of which are Lathrop or Castle strain-derived mice (Fig. 1, 2). Bxv1 was not found in any of the New Zealand or Swiss strain-derived mice. This distribution is consistent with previous analyses that typed 12 strains for the presence or absence of the Chr 1 Bxv1 induction locus (18, 19, 51) and 7 strains for the presence or absence of diagnostic env fragments by Southern blotting (8) (Fig. 2). Previous testing had also identified a Bxv1-like locus in F/St mice, and although the Chr 1 location initially determined for this virus induction locus did not clearly identify it as Bxv1 (31), the present analysis confirms that Bxv1 is present in F/St (Fig. 1).

Wild mouse origin of the Bxv1 provirus.

Laboratory mice are a mosaic of the wild mouse species of house mice, and 3 of the 4 house mouse species (M. molossinus, M. castaneus, M. musculus) carry X-MLV-related sequences and are capable of producing infectious virus (4, 17, 26, 43, 46, 47). It is therefore possible that the Bxv1 provirus originated in one of these wild mouse species. A previous attempt to identify Bxv1 in M. molossinus by screening backcross mice for inducible Chr1-linked proviral loci was unsuccessful because these mice carry too many active proviruses (16). We screened house mouse species and several other Mus species that had not been previously typed for endogenous MLVs for the presence of Bxv1 by PCR (Fig. 1, 3). We identified Bxv1 in 4 wild mouse DNAs, all of which were M. molossinus, although 3 additional breeding lines of M. molossinus mice did not have this provirus.

None of the other house mouse species carries Bxv1. M. molossinus is thought to be a natural hybrid of M. castaneus and M. musculus (52), both of which species carry active XMVs; however, the Bxv1 provirus was not found in any mouse of these 2 species (Fig. 1 and 3). The absence of this provirus in these 2 species and the fact that some, but not all, M. molossinus mice carry Bxv1 indicate that this XMV insertion arose relatively recently in Japanese mice. It is likely that Asian mouse X-MLVs, as well as the ecotropic MLVs carried by these mice (17, 39), were introduced into fancy mouse colonies because of inclusion of Japanese mice in fancy mouse colonies; in fact, Lathrop started her mouse colony with a pair of Japanese waltzing mice (22, 29, 37). It is also probable that the additional active proviruses found in laboratory mouse strains like NZB and MA/My may have been acquired from these Asian mice.

DNAs from M. molossinus and from the common laboratory strains were also typed for alleles of the Fv1 gammaretrovirus restriction gene, a gene that targets the virus capsid and inhibits replication postentry. DNAs were typed using PCR primers that generate products of different sizes for Fv1^b (1.8 kb) and for Fv1ⁿ or Fv1^nr (510 bp) (45). Among the laboratory strains, 28 of the 39 mice typed in this way as Fv1ⁿ were found to lack Bxv1, but 6 of the 10 Fv1^b mice have Bxv1 (Fig. 2). All M. molossinus DNAs tested (MOLC/RkJ, MOLD/RkJ, MSM/Ms, JF1/Ms, MOLF/EiJ) produced the longer Fv1^b-like PCR product, suggesting the possibility that the Bxv1 virus is B tropic. However, M. molossinus cells have not been tested for Fv1-type restriction, and although PCR suggests that this Fv1 resembles the longer Fv1^b at its 3′ end, our previous sequence analysis of the M. molossinus Fv1 gene identified Fv1^nr codons at key sites in the coding region (GenBank accession no. FJ603569). Furthermore, the Fv1 tropism of the virus produced by Bxv1 has not been determined, and the Bxv1 capsid sequence has novel substitutions at key Fv1 target sites, so its sensitivity to the laboratory mouse Fv1 variants cannot be predicted. It is therefore not yet clear in what way the acquisition of the Bxv1 provirus was influenced by Fv1.

Xpr1 receptor alleles in viremic strains of laboratory mice.

Multiple host factors can influence the expression of proviral loci and the spread of virus in vivo, and the detection of infectious X-MLV or viral proteins depends on what proviruses are present and can also be influenced by genes at or near the major histocompatibility complex (MHC) locus and the Fv1 gene (32, 51). For mice carrying active XMVs, another contributing factor is likely to be the XPR1 receptor used by X-MLVs for entry. Because functional variants of the XPR1 receptor for X-MLVs exist in the various mouse species that contributed to the laboratory mouse genome, we screened inbred strains for variants of Xpr1. Two inbred strains, NZB and F/St, are unusual among laboratory strains in that they constitutively express infectious X-MLV throughout life (6, 31). X-MLVs are readily isolated from lymphoid organs of these mice. NZB carries two active XMVs, and the high-virus expression pattern has been attributed to the constitutive expression of one of those XMVs (6), although the presence of a permissive XPR1 variant could also be a contributing factor. Neither of the XMV loci in NZB loci is Bxv1, as indicated by previous genetic analysis of this strain (16) and by our present study (Fig. 1). To determine if a permissive receptor contributes to the observed high-virus phenotype, we sequenced a segment of the NZB Xpr1 gene encoding the receptor determinants in XPR1 ECL3 and ECL4 (28, 46). This sequencing identified the NZB receptor as the restrictive allele Xpr1ⁿ (GenBank accession no. HQ323678), indicating that the viremia observed in these mice has no contribution from receptor-mediated secondary spread.

The second high-virus strain, F/St, produces detectable levels of X-MLV at 1 week of age and high levels from 4 weeks (31). This mouse carries one active XMV that was identified as the Bxv1 provirus (Fig. 1), but this XMV is not associated with viremia in the other strains that carry it. We sequenced exon 13 of F/St Xpr1, which contains ECL4, and found that these mice carry the permissive Xpr1^sxv allele (15, 28) (GenBank accession no. HQ323665). While this might suggest an explanation for the F/St high-virus phenotype, analysis of genetic crosses between F/St and other inbred strains have shown that viremia requires Bxv1 and a recessive gene linked to the major histocompatibility locus (51). Xpr1 is closely linked to Bxv1 on Chr 1; its role in viremia cannot be evaluated from the existing data on F/St mice, and this strain has recently become extinct.

Xpr1 variants in other inbred strains.

Rather than sequence XPR1 from dozens of mouse strains, we first screened DNAs of laboratory mouse strains by PCR for the presence of the permissive Xpr1^sxv allele. There are five functionally distinct Xpr1 alleles in Mus; 4 of these variants are found in the house mouse species and inbred strains, and only 1 of these variants, Xpr1^sxv, encodes a fully permissive receptor (15). The other 3 house mouse receptors (Xpr1^m, Xpr1ⁿ, and Xpr1^c) restrict 2 or more of the viruses that rely on XPR1 for entry, and all 3 of these restrictive receptors have deletions in the ECL4 segment of the receptor (48) (Fig. 4). The deletion found in M. castaneus (Xpr1^c) removes 5 codons, and the different deletions found in NIH 3T3 (Xpr1ⁿ) and M. musculus (Xpr1^m) remove single codons. The m, n, and sxv alleles cannot be reliably distinguished by PCR product size differences, but mixtures of DNAs carrying the m, c, and n alleles with Xpr1^sxv DNA produce PCR product doublets. Because we were specifically interested in identifying the permissive Xpr1^sxv allele in laboratory mouse strains, we did an initial PCR screen by mixing DNAs from each mouse strain with DNA from a mouse carrying wild-mouse Xpr1^sxv (Fig. 4).

Fifty strains were screened by this method, and 6 strains in addition to F/St were implicated as having Xpr1^sxv. Sequencing confirmed the presence of this permissive allele in all 6 strains. Xpr1^sxv was found in 3 of the 10 Swiss mouse DNAs and in the 3 mouse strains SOD1, LP, and LT/SvEi. We also sequenced exon 13 in 9 additional strains of interest. This segment of Xpr1 is hypervariable in rodent species (48), and we wanted to identify the presence of any novel polymorphisms due to substitutions that would be missed by the PCR screen for insertions or deletions. The 9 strains selected for sequencing included AKR and HRS, 2 strains that have been used in studies on viral leukemogenesis, and RIIIS, a strain that has acquired an XMV insertion in its Apobec3 gammaretrovirus resistance gene (36). We also sequenced exon 13 in strains expressing high levels of nonecotropic MLV envelope (Env) glycoprotein. For example, DBA mice carry a provirus associated with the Rmcf locus that blocks the entry of viruses that use the Xpr1 receptor (13) and also express high levels of XenCSA and GIX, cell surface antigens related to nonecotropic MLV Env glycoprotein (30). 129-GIX+ mice also express the Env glycoprotein-related GIX antigen. Strains selected for sequencing also included 129, FVB, and NOD, because they are used as human disease models or for genetic manipulations, MA/My, a strain that carries two active XMV proviruses (16), and YBR, a strain produced from a different fancy mouse stock (Fig. 2). No novel sequence variants were identified in these additional mice; all of these strains were found to carry the exon 13 sequence of Xpr1ⁿ.

In the last few decades, a number of new inbred strains were derived from wild-trapped mice. These strains are unrelated to the fancy mouse-derived older common inbred strains, and they carry the Xpr1 alleles found in their species of origin. Thus, strains derived from M. molossinus and M. musculus carry Xpr1^m and strains derived from M. castaneus carry Xpr1^c (48). Two inbred strains derived from wild-trapped M. domesticus mice, PWD and RBA, were found by sequencing to carry Xpr1^sxv, as expected (GenBank accession no. HQ323672 and HQ323673).

X-MLV infectivity of cells derived from inbred strains SWR, SJL, and SIM.R.

Three of the strains determined to carry the permissive Xpr1^sxv receptor based on initial screening and sequencing of exon 13 were further characterized. We sequenced the entire Xpr1 gene from SJL and SIM.R. The gene in both strains is identical to that originally cloned from M. dunni cells (28). This allele differs from Xpr1ⁿ by one insertion, Δ582T, and two substitutions, E500K and N590D.

The presence of a permissive receptor in these mice does not guarantee infectivity, as failure to infect can result from factors such as receptor interference (reviewed in reference 40). Therefore, we examined cells from all three of these strains for X-MLV infectivity. Cultured tail fibroblasts from SIM.R and SWR and an SV40-transformed line of SJL mouse fibroblasts were tested for susceptibility to LacZ retroviral vectors incorporating the Env glycoproteins of 2 mouse X-MLVs and 5 other virus isolates known to use the Xpr1 receptor. These 5 additional viruses included 2 polytropic MLVs (P-MLVs), 2 wild mouse virus isolates (CasE#1 and Cz524), and XMRV, an X-MLV-related virus originally isolated from human patients with prostate cancer (7). Xpr1ⁿ restricts all these viruses except the P-MLVs, while Xpr1^sxv is permissive for all 7 viruses (47) (Table 1).

The cells from these three inbred mouse strains are susceptible to all viruses (Table 1), confirming that these cells all express a fully permissive XPR1 receptor. Some viruses infected these cells with nearly equal efficiencies, but most of the viruses showed infectivity differences that are consistent with Fv1 restriction. All MLVs are potentially subject to Fv1 restriction, including X-MLVs (35), and mouse-tropic MLVs have been typed as N or B tropic based on their restriction by Fv1^b or Fv1ⁿ, respectively, or as NB tropic if unaffected by either allele. SIM.R was developed as a congenic strain of Swiss noninbred mice and carries the Fv1^b allele from C57BL/6 (44). The presence of Fv1^b in the SIM.R cells and Fv1ⁿ in SJLSV was also determined by PCR. A segment of Fv1 was cloned from SWR cell DNA and identified as Fv1ⁿ by sequencing (GenBank accession no. HQ323682).

Two of the viruses used here are NB tropic, and these two P-MLVs, FrMCF and MoMCF, infected all 5 mouse cells with roughly equal efficiencies (Table 1). The other 5 viruses have unknown Fv1 tropism, but all 5 showed reduced LacZ titers in some or all 3 of the Xpr1^sxv laboratory mouse cells relative to titers in M. dunni, which lacks a functional Fv1 gene Fv1⁰ (45). This was somewhat unexpected, as the LacZ pseudotypes used here were produced by virus infection of GP2-293 cells, which carry ecotropic Moloney MLV gag-pol genes and therefore produce NB-tropic MoMLV capsids as well as capsids of the infecting virus. The fact that none of these 5 pseudotypes are NB tropic indicates that tropism is influenced by the infecting virus. Two of those viruses, CasE#1 and AKR6, more efficiently infected SIM.R than they infected SWR or SJL, suggesting that these viruses are restricted by Fv1ⁿ and may therefore be B tropic. CAST-X, Cz524, and XMRV are restricted in all 3 of these cells relative to levels of infection in M. dunni cells, although one virus, XMRV, appears somewhat more sensitive to Fv1ⁿ than to Fv1^b. The restriction of XMRV by Fv1ⁿ and Fv1^b has been reported previously (10, 47), and these results suggest that CAST-X and Cz524 may have similar Fv1-defined tropisms.