Feline leukemia virus infection requires a post-receptor binding envelope-dependent cellular component

Gammaretrovirus receptors have been suggested to contain the necessary determinants to mediate virus binding and entry. Here, we show that murine NIH 3T3 and baby hamster kidney (BHK) cells overexpressing receptors for subgroup A, B, and C feline leukemia viruses (FeLVs) are weakly susceptible (10(1) to 10(2) CFU/ml) to FeLV pseudotype viruses containing murine leukemia virus (MLV) core (Gag-Pol) proteins, whereas FeLV receptor-expressing murine Mus dunni tail fibroblast (MDTF) cells are highly susceptible (10(4) to 10(6) CFU/ml). However, NIH 3T3 cells expressing the FeLV subgroup B receptor PiT1 are highly susceptible to gibbon ape leukemia virus pseudotype virus, which differs from the FeLV pseudotype viruses only in the envelope protein. FeLV resistance is not caused by a defect in envelope binding, low receptor expression levels, or N-linked glycosylation. Resistance is not alleviated by substitution of the MLV core in the FeLV pseudotype virus with FeLV core proteins. Interestingly, FeLV resistance is alleviated by fusion of receptor-expressing NIH 3T3 and BHK cells with MDTF or human TE671 cells, suggesting the absence of an additional cellular component in NIH 3T3 and BHK cells that is required for FeLV infection. The putative FeLV-specific cellular component is not a secreted factor, as MDTF conditioned medium does not alleviate the block to FeLV infection. Together, our findings suggest that FeLV infection requires an additional envelope-dependent cellular component that is absent in NIH 3T3 and BHK cells but that is present in MDTF and TE671 cells.     

Subtle mutational changes in the SU protein of a natural feline leukemia virus subgroup A isolate alter disease spectrum

FeLV-945 is a representative isolate of the natural feline leukemia virus (FeLV) variant predominant in non-T-cell malignant, proliferative, and degenerative diseases in a geographic cohort. The FeLV-945 surface glycoprotein (SU) is closely related to natural horizontally transmissible FeLV subgroup A (FeLV-A) but was found to differ from a prototype to a larger extent than the members of FeLV-A differ among themselves. The sequence differences included point mutations restricted largely to the functional domains of SU, i.e., VRA, VRB, and PRR. Despite the sequence differences in these critical domains, measurements of receptor utilization, including host range and superinfection interference, confirmed the assignment of FeLV-945 to subgroup A. Other proviruses isolated from the cohort contained similar sequence hallmarks and were assigned to FeLV subgroup A. A provirus from cat 1046 contained a histidine-to-proline change at SU residue 6 within an SPHQ motif that was previously identified as a critical mediator of fusion events during virus entry. The 1046 pseudotype virus entered cells only in the presence of the soluble cofactor FeLIX provided in trans, but it retained an ecotropic host range even in the presence of FeLIX. The mutational changes in FeLV-945 were shown to confer significant functional differences compared to prototype FeLV-A viruses. The substitution of FeLV-945 envelope gene sequences for FeLV-A/61E sequences conferred a small but statistically significant replicative advantage in some feline cells. Moreover, substitution of the unique FeLV-945 long terminal repeat and envelope gene for those of FeLV-A/61E altered the disease spectrum entirely, from a thymic lymphoma of a T-cell origin to an as yet uncharacterized multicentric lymphoma that did not contain T cells.     

Reevaluation of host ranges of feline leukemia virus subgroups

We reevaluated the host ranges of feline leukemia virus (FeLV) subgroups A, B and C using pseudotype assays based on recombinant NB-tropic murine leukemia virus, which is not usually blocked after viral entry in mammalian cells. Pseudotype viruses of FeLV-B and -C infected a variety of cell lines from many mammalian species. Unexpectedly, FeLV-A pseudotype viruses of two independent isolates from the UK and US also infected a variety of non-feline cell lines including cells from humans, rabbits, pigs and minks. Moreover, both isolates of FeLV-A productively infected human embryonic kidney 293 and mink Mv-1-Lu cells. We conclude that FeLV-A is not strictly ecotropic.     

A putative thiamine transport protein is a receptor for feline leukemia virus subgroup A

Feline leukemia virus (FeLV) is a horizontally transmitted virus that causes a variety of proliferative and immunosuppressive diseases in cats. There are four subgroups of FeLV, A, B, C, and T, each of which has a distinct receptor requirement. The receptors for all but the FeLV-A subgroup have been defined previously. Here, we report the identification of the cellular receptor for FeLV-A, which is the most transmissible form of FeLV. The receptor cDNA was isolated using a gene transfer approach, which involved introducing sequences from a feline cell line permissive to FeLV-A into a murine cell line that was not permissive. The feline cDNA identified by this method was approximately 3.5 kb, and included an open reading frame predicted to encode a protein of 490 amino acids. This feline cDNA conferred susceptibility to FeLV-A when reintroduced into nonpermissive cells, but it did not render these cells permissive to any other FeLV subgroup. Moreover, these cells specifically bound FeLV-A-pseudotyped virus particles, indicating that the cDNA encodes a binding receptor for FeLV-A. The feline cDNA shares approximately 93% amino acid sequence identity with the human thiamine transport protein 1 (THTR1). The human THTR1 receptor was also functional as a receptor for FeLV-A, albeit with reduced efficiency compared to the feline orthologue. On the basis of these data, which strongly suggest the feline protein is the orthologue of human THTR1, we have named the feline receptor feTHTR1. Identification of this receptor will allow more detailed studies of the early events in FeLV transmission and may provide insights into FeLV pathogenesis.     

Pathogenicity Induced by Feline Leukemia Virus, Rickard Strain, Subgroup A Plasmid DNA (pFRA)

A new provirus clone of feline leukemia virus (FeLV), which we named FeLV-A (Rickard) or FRA, was characterized with respect to viral interference group, host range, complete genome sequence, and in vivo pathogenicity in specific-pathogen-free newborn cats. The in vitro studies indicated the virus to be an ecotropic subgroup A FeLV with 98% nucleotide sequence homology to another FeLV-A clone (F6A/61E), which had also been fully sequenced previously. Since subgroup B polytropic FeLVs (FeLV-B) are known to arise via recombination between ecotropic FeLV-A and endogenous FeLV (enFeLV) env elements, the in vivo studies were conducted by direct intradermal inoculation of the FRA plasmid DNA so as to eliminate the possibility of coinoculation of any FeLV-B which may be present in the inoculum prepared by propagating FeLV-A in feline cell cultures. The following observations were made from the in vivo experiments: (i) subgroup conversion from FeLV-A to FeLV-A and FeLV-B, as determined by the interference assay, appeared to occur in plasma between 10 and 16 weeks postinoculation (p.i.); (ii) FeLV-B-like recombinants (rFeLVs), however, could be detected in DNA isolated from buffy coats and bone marrow by PCR as early as 1 to 2 weeks p.i.; (iii) while a mixture of rFeLV species containing various amounts of N-terminal substitution of the endogenous FeLV-derived env sequences were detected at 8 weeks p.i., rFeLV species harboring relatively greater amounts of such substitution appeared to predominate at later infection time points; (iv) the deduced amino acid sequence of rFeLV clones manifested striking similarity to natural FeLV-B isolates, within the mid-SU region of the env sequenced in this work; and (v) four of the five cats, which were kept for determination of tumor incidence, developed thymic lymphosarcomas within 28 to 55 weeks p.i., with all tumor DNAs harboring both FeLV-A and rFeLV proviruses. These results provide direct evidence for how FeLV-B species evolve in vivo from FeLV-A and present a new experimental approach for efficient induction of thymic tumors in cats, which should be useful for the study of retroviral lymphomagenesis in this outbred species.     

Protection of cats against feline leukemia virus by vaccination with a canarypox virus recombinant, ALVAC-FL.

Two ALVAC (canarypox virus)-based recombinant viruses expressing the feline leukemia virus (FeLV) subgroup A env and gag genes were assessed for their protective efficacy in cats. Both recombinant viruses contained the entire gag gene. ALVAC-FL also expressed the entire envelope glycoprotein, while ALVAC-FL(dl IS) expressed an env-specific gene product deleted of the putative immunosuppressive region. Although only 50% of the cats vaccinated with ALVAC-FL(dl IS) were protected against persistent viremia after oronasal exposure to a homologous FeLV isolate, all cats administered ALVAC-FL resisted the challenge exposure. Significantly, protection was afforded in the absence of detectable FeLV-neutralizing antibodies. These results represent the first effective vaccination of cats against FeLV with a poxvirus-based recombinant vector and have implications that are relevant not only to FeLV vaccine development but also to developing vaccines against other retroviruses, including human immunodeficiency virus.     

Comparative analysis of the genomes of feline leukemia viruses.

The genomes of several strains of feline leukemia virus (FeLV) were compared by two-dimensional polyacrylamide gel electrophoresis of the large RNase T1-resistant oligonucleotides of the 70S RNA. Differences between each strain of FeLV tested were detected by this method. We estimate that the degree of sequence identity between the viruses is: FeLV A (Glasgow-1) to FeLV B (Snyder-Theilen), 52%; FeLV A (Glasgow-1) to FeLV C(Sarma), 66%; FeLV B(Snyder-Theilen) to FeLV C (Sarma), 37%. The fingerprints of two independent isolates of FeLV strains of subgroup A (Glasgow-1 and Rickard) were detectably different. We conclude that the RNase T1 oligonucleotide fingerprint pattern provides a useful tool for identification of FeLV strains.     

Specificity in receptor usage by T-cell-tropic feline leukemia viruses: implications for the in vivo tropism of immunodeficiency-inducing variants

Cytopathic, T-cell-tropic feline leukemia viruses (FeLV-T) evolve from FeLV-A in infected animals and demonstrate host cell specificities that are distinct from those of their parent viruses. We recently identified two cellular proteins, FeLIX and Pit1, required for productive infection by these immunodeficiency-inducing FeLV-T variants (M. M. Anderson, A. S. Lauring, C. C. Burns, and J. Overbaugh, Science 287:1828-1830, 2000). FeLV-T is the first example of a naturally occurring type C retrovirus that requires two proteins to gain entry into target cells. FeLIX is an endogenous protein that is highly related to the N-terminal portion of the FeLV envelope protein, which includes the receptor-binding domain. Pit1 is a multiple-transmembrane phosphate transport protein that also functions as a receptor for FeLV-B. The FeLV-B envelope gene is derived by recombination with endogenous FeLV-like sequences, and its product can functionally substitute for FeLIX in facilitating entry through the Pit1 receptor. In the present study, we tested other retrovirus envelope surface units (SUs) with their cognate receptors to determine whether they also could mediate infection by FeLV-T. Cells were engineered to coexpress the transmembrane form of the envelope proteins and their cognate receptors, or SU protein was added as a soluble protein to cells expressing the receptor. Of the FeLV, murine leukemia virus, and gibbon ape leukemia virus envelopes tested, we found that only those with receptor-binding domains derived from endogenous FeLV could render cells permissive for FeLV-T. We also found that there is a strong preference for Pit1 as the transmembrane receptor. Specifically, FeLV-B SUs could efficiently mediate infection of cells expressing the Pit1 receptor but could only inefficiently mediate infection of cells expressing the Pit2 receptor, even though these SUs are able to bind to Pit2. Expression analysis of feline Pit1 and FeLIX suggests that FeLIX is likely the primary determinant of FeLV-T tropism. These results are discussed in terms of current models for retrovirus entry and the interrelationship among FeLV variants that evolve in vivo.     

Molecular cloning and characterization of endogenous feline leukemia virus sequences from a cat genomic library.

Recombinant bacteriophage lambda clones from a cat genomic library derived from placental DNA of a specific pathogen-free cat were screened to identify endogenous feline leukemia virus (FeLV) sequences. Restriction endonuclease mapping of four different clones indicates that there are a number of similarities among them, notably the presence of a 6.0- to 6.4-kilobase pair (kbp) EcoRI hybridizing fragment containing portions of sequences homologous to the gag, pol, env, and long terminal repeat-like elements of the infectious FeLV. The endogenous FeLV sequences isolated are approximately 4 kbp in length and are significantly shorter than the cloned infectious FeLV isolates, which are 8.5 to 8.7 kbp in length. The endogenous elements have 3.3- to 3.6-kbp deletions in the gag-pol region and approximately 0.7- to 1.0-kbp deletions in the env region. These deletions would render them incapable of encoding an infectious virus and may therefore be related to the non-inducibility of FeLV from uninfected cat cells and the subgenomic expression of these endogenous sequences in placental tissue. It appears that there is conservation in the ordering of restriction sites previously reported in the proviruses of the infectious FeLVs in sequences corresponding to the pol and env boundary as well as the region spanning the env gene of the endogenous clones, whereas a greater divergence occurs among restriction sites mapped to the gag and part of the pol regions of the infectious FeLV. Such deleted, FeLV-related subsets of DNA sequences could have originated either by germ-line integration of a complete ecotropic virus followed by deletion, or by integration of a preexisting, defective, deleted variant of the infectious virus.     

Pathogenesis of feline leukemia virus T17: contrasting fates of helper, v-myc, and v-tcr proviruses in secondary tumors.

A naturally occurring feline thymic lymphosarcoma (T17) provided the unique observation of a T-cell antigen receptor beta-chain gene (v-tcr) transduced by a retrovirus. The primary tumor contained three classes of feline leukemia virus (FeLV) provirus, which have now been characterized in more detail as (i) v-tcr-containing recombinant proviruses, (ii) v-myc-containing recombinant proviruses, and (iii) apparently full-length helper FeLV proviruses. The two transductions appear to have been independent events, with distinct recombinational junctions and no sequence overlap in the host-derived inserts. The T17 tumor cell line releases large numbers of FeLV particles of low infectivity; all three genomes are encapsidated, but passage of FeLV-T17 on feline fibroblast and lymphoma cells led to selective loss of the recombinant viruses. The oncogenic potential of the T17 virus complex was, therefore, tested by infection of neonatal cats with virus harvested directly from the primary T17 tumor cell line. A single inoculation of FeLV-T17 caused persistent low-grade infection culminating in thymic lymphosarcoma and acute thymic atrophy, which was accelerated by coinfection with the weakly pathogenic FeLV subgroup A (FeLV-A)/Glasgow-1 helper. Molecularly cloned FeLV-tcr virus (T-31) rescued for replication by a weakly pathogenic FeLV-A/Glasgow-1 helper virus was similarly tested in vivo and induced thymic atrophy and thymic lymphosarcomas. Most FeLV-T17-induced tumors manifested either v-myc or an activated c-myc allele and had undergone rearrangement of endogenous T-cell antigen receptor beta-chain genes, supporting the proposition that the oncogenic effects of c-myc linked to the FeLV long terminal repeat are targeted to a specific window in T-cell differentiation. However, neither the FeLV-T17-induced tumors nor the T-31 + FeLV-A-induced tumors contained clonally represented v-tcr sequences. Only one of the FeLV-T17-induced tumors contained detectable v-tcr proviruses, at a low copy number. While v-tcr does not have a readily transmissible oncogenic function, a more restricted role is not excluded, perhaps involving antigenic peptide-major histocompatibility complex recognition by the T-cell receptor complex. Such a function could be obscured by the genetic diversity of the outbred domestic cat host.     

A serosurvey of viral infections in lions (Panthera leo), from Queen Elizabeth National Park, Uganda

Serum samples from 14 lions (Panthera leo) from Queen Elizabeth National Park, Uganda, were collected during 1998 and 1999 to determine infectious disease exposure in this threatened population. Sera were analyzed for antibodies against feline immunodeficiency virus (FIV), feline calicivirus (FCV), feline herpesvirus 1 (feline rhinotracheitis: FHV1), feline/canine parvovirus (FPV/CPV), feline infectious peritonitis virus (feline coronavirus: FIPV), and canine distemper virus (CDV) or for the presence of feline leukemia virus (FeLV) antigens. Ten lions (71%) had antibodies against FIV, 11 (79%) had antibodies against CDV, 11 (79%) had antibodies against FCV, nine (64%) had antibodies against FHV1, and five (36%) had antibodies against FPV. Two of the 11 CDV-seropositive lions were subadults, indicating recent exposure of this population to CDV or a CDV-like virus. No lions had evidence of exposure to FeLV or FIPV. These results indicate that this endangered population has extensive exposure to common feline and canine viruses.     

A novel truncated env gene isolated from a feline leukemia virus-induced thymic lymphosarcoma

We PCR amplified the exogenous feline leukemia virus (FeLV)-related env gene species from lymphosarcomas induced by intradermally administered plasmid DNA of either the prototype FeLV, subgroup A molecular clone, F6A, or a new molecular clone, FeLV-A, Rickard strain (FRA). Of the nine tumors examined, six showed the presence of deleted env species of variable sizes in the tumor DNA. One env mutant, which was detected in a FRA-induced thymic lymphosarcoma, had a large internal deletion beginning from almost the N-terminal surface glycoprotein (SU) up to the middle region of the transmembrane (TM) protein of the env gene. The deduced polypeptide of this truncated env (tenv) retained the complete signal peptide and seven amino acids of the N-terminal mature SU of FRA env gene, followed by eight amino acids from the frameshift in the TM region. To study the biological function of tenv, we used a murine retrovirus vector to produce amphotropic virions. Infection of feline fibroblasts (H927), human fibrosarcoma cells (HT1080), or human B-lymphoma cells (Raji) led to pronounced cytotoxicity, while the tenv virus did not induce significant cytotoxicity to feline T-lymphoma cells (3201B) or human T-lymphoma cells (CEM). Together, these results convincingly demonstrated that the genetic events that led to truncation in the env gene occurred de novo in FeLV lymphomagenesis and that such a product, tenv could induce cytotoxicity to fibroblastic and B-lymphoid cells but not to T-lymphoid tumor cells. This type of selective toxicity might be potentially important in the development of the neoplastic disease.     

Identification of a putative receptor for subgroup A feline leukemia virus on feline T cells.

Retrovirus infection is initiated by the binding of virus envelope glycoprotein to a receptor molecule present on cell membranes. To characterize a receptor for feline leukemia virus (FeLV), we extensively purified the viral envelope glycoprotein, gp70, from culture supernatants of FeLV-61E (subgroup A)-infected cells by immunoaffinity chromatography. Binding of purified 125I-labeled gp70 to the feline T-cell line 3201 was specific and saturable, and Scatchard analysis revealed a single class of receptor binding sites with an average number of 1.6 x 10(5) receptors per cell and an apparent affinity constant (Ka) of 1.15 x 10(9) M-1. Cross-linking experiments identified a putative gp70-receptor complex of 135 to 140 kDa. Similarly, coprecipitation of 125I-labeled cell surface proteins with purified gp70 and a neutralizing but noninterfering anti-gp70 monoclonal antibody revealed a single cell surface protein of approximately 70 kDa. These results indicate that FeLV-A binds to feline T cells via a 70-kDa cell surface protein, its presumptive receptor.     

Molecular analysis and pathogenesis of the feline aplastic anemia retrovirus, feline leukemia virus C-Sarma.

We describe the molecular cloning of an anemogenic feline leukemia virus (FeLV), FeLV-C-Sarma, from the productively infected human rhabdomyosarcoma cell line RD(FeLV-C-S). Molecularly cloned FeLV-C-S proviral DNA yielded infectious virus (mcFeLV-C-S) after transfection of mammalian cells, and virus interference studies using transfection-derived virus demonstrated that our clone encodes FeLV belonging to the C subgroup. mcFeLV-C-S did not induce viremia in eight 8-week-old outbred specific-pathogen-free (SPF) cats. It did, however, induce viremia and a rapid, fatal aplastic anemia due to profound suppression of erythroid stem cell growth in 9 of 10 inoculated newborn, SPF cats within 3 to 8 weeks (21 to 58 days) postinoculation. Thus, the genome of mcFeLV-C-S encodes the determinants responsible for the genetically dominant induction of irreversible erythroid aplasia in outbred cats. A potential clue to the pathogenic determinants of this virus comes from previous work indicating that all FeLV isolates belonging to the C subgroup, an envelop-gene-determined property, and only those belonging to the C subgroup, are potent, consistent inducers of aplastic anemia in cats. To approach the molecular mechanism underlying the induction of this disease, we first determined the nucleotide sequence of the envelope genes and 3' long terminal repeat of FeLV-C-S and compared it with that of FeLV-B-Gardner-Arnstein (mcFeLV-B-GA), a subgroup-B feline leukemia virus that consistently induces a different disease, myelodysplastic anemia, in neonatal SPF cats. Our analysis revealed that the p15E genes and long terminal repeats of the two FeLV strains are highly homologous, whereas there are major differences in the gp70 proteins, including five regions of significant amino acid differences and apparent sequence substitution. Some of these changes are also reflected in predicted glycosylation sites; the gp70 protein of FeLV-B-GA has 11 potential glycosylation sites, only 8 of which are present in FeLV-C-S.     

Feline Leukemia Virus-B Envelope Together With its GlycoGag and Human Immunodeficiency Virus-1 Nef Mediate Resistance to Feline SERINC5

Human SERINC5 (SER5) protein is a recently described restriction factor against human immunodeficiency virus-1 (HIV-1), which is antagonized by HIV-1 Nef protein. Other retroviral accessory proteins such as the glycosylated Gag (glycoGag) from the murine leukemia virus (MLV) can also antagonize SER5. In addition, some viruses escape SER5 restriction by expressing a SER5-insensitive envelope (Env) glycoprotein. Here, we studied the activity of human and feline SER5 on HIV-1 and on the two pathogenic retroviruses in cats, feline immunodeficiency virus (FIV) and feline leukemia virus (FeLV). HIV-1 in absence of Nef is restricted by SER5 from domestic cats and protected by its Nef protein. The sensitivity of feline retroviruses FIV and FeLV to human and feline SER5 is considerably different: FIV is sensitive to feline and human SER5 and lacks an obvious mechanism to counteract SER5 activity, while FeLV is relatively resistant to SER5 inhibition. We speculated that similar to MLV, FeLV-A or FeLV-B express glycoGag proteins and investigated their function against human and feline SER5 in wild type and envelope deficient virus variants. We found that the endogenous FeLV recombinant virus, FeLV-B but not wild type exogenous FeLV-A envelope mediates a strong resistance against human and feline SER5. GlycoGag has an additional but moderate role to enhance viral infectivity in the presence of SER5 that seems to be dependent on the FeLV envelope. These findings may explain, why in vivo FeLV-B has a selective advantage and causes higher FeLV levels in infected cats compared to infections of FeLV-A only.     

Isolation of an RD-114-Like Oncornavirus from a Cat Cell Line

A clone of cells derived from a continuous line of cat cells (CCC) spontaneously produced an RNA C-type virus (CCC virus) which did not have the group-specific antigen of the standard strains of feline leukemia viruses but did have that of the RD-114 virus. Single-hit infection of a virus yielding CCC cell with only the feline leukemia virus pseudotype of murine sarcoma virus [MSV(FeLV)] resulted in the release of a pseudotype of MSV coated with the CCC virus envelope. Host range, transmission of virus, helper functions, interference properties, and specific neutralization showed that the CCC and the RD-114 isolates as well as their respective MSV pseudotypes are closely similar if not identical. Parental, virus-negative cells frozen before the existence of RD-114 were chemically induced to yield CCC-like virus de novo. Infection of susceptible human cells with the chemically induced virus resulted in interference with the CCC virus pseudotype of MSV but not with the FeLV pseudotype of MSV.     

Highly efficient transduction of repopulating bone marrow cells using rapidly concentrated polymer-complexed retrovirus

Using the cationic polymer, Polybrene, and the anionic polymer, chondroitin sulfate C, we concentrated recombinant retrovirus pseudotyped with an ecotropic envelope, which is susceptible to inactivation by high-speed concentration methods. To evaluate gene marking, murine bone marrow was harvested from C3H mice, transduced with polymer-concentrated GFP virus, and transplanted into lethally irradiated recipients. Total gene marking in mice averaged 30-35% at 8 weeks post-transplant and transgene expression remained stable for over 16 weeks. Using the polymer concentration method, a second retroviral vector encoding the drug resistant variant of dihydrofolate reductase (L22Y-DHFR) was concentrated and tested. Approximately 40% of transduced murine bone marrow progenitor cells were protected against trimetrexate concentrations that completely eliminated the growth of non-modified cells. These results show that anionic and cationic polymers can be combined to rapidly concentrate viruses that are normally difficult to concentrate, and the concentrated virus efficiently transduces hematopoietic stem cells.     

Sindbis virus vectors designed to express a foreign protein as a cleavable component of the viral structural polyprotein

Alphavirus-based expression vectors commonly use a duplicated 26S promoter to drive expression of a foreign gene. Here we describe an expression strategy in which the foreign sequences are linked to the gene encoding the 2A protease of foot-and-mouth disease virus and then inserted in frame between the capsid and E3 genes of Sindbis virus. During replication, the 2A fusion protein is synthesized as a component of the viral structural polyprotein that is then released by intramolecular cleavages mediated by the capsid and 2A proteases. Recombinant Sindbis viruses that expressed fusion proteins composed of 2A linked to the green fluorescent protein (GFP) and to the VP7 protein of bluetongue virus were constructed. Viruses engineered to express GFP and VP7 from a duplicate 26S promoter were also constructed. All four viruses expressed the transgene and grew to similar titers in cultured cells. However, the GFP/2A- and VP7/2A-expressing viruses displayed greater expression stability and were less attenuated in newborn mice than the cognate double-subgenomic promoter-based viruses. By combining the two expression strategies, we constructed bivalent viruses that incorporated and expressed both transgenes. The bivalent viruses grew to lower titers in cultured cells and were essentially avirulent in newborn mice. Groups of mice were vaccinated with each VP7- and VP7/2A-expressing virus, and antibody responses to native VP7 were measured in an indirect enzyme-linked immunosorbent assay. Despite their genetic and phenotypic differences, all viruses induced similarly high titers of VP7-specific antibodies. These results demonstrate that 2A fusion protein-expressing alphaviruses may be particularly well suited for applications that require enduring expression of a single protein or coexpression of two alternative proteins.