Topographical rearrangements of visna virus envelope glycoprotein during antigenic drift

Maedi-visna virus persistence: Antigenic variation and latency

Maedi-visna virus (MVV), a lentivirus of sheep, shares with other lentiviruses the ability to establish a lifelong infection. In this study five sheep were infected intravenously with MVV and housed together with a number of uninfected sheep for natural transmission. All virus isolates from ten sheep that had been infected naturally had multiple mutations in the principal neutralization domain in Env and were antigenic variants, while three of four isolates from the carrier sheep had identical sequences to the infecting strain and were not antigenic variants. There was evidence of positive selection in the gene, particularly in amino acids comprising the neutralization epitope and some adjacent glycosylation sites. Together these results suggest that virus persistence is acquired by a reservoir of latent viruses, and that there is selection for antigenic variants of virus that is transmitted naturally.

Comparison of a maedi-visna virus CA-TM fusion protein ELISA with other assays for detecting sheep infected with North American ovine lentivirus strains

A maedi-visna virus CA-TM fusion protein ELISA (MVV ELISA) was evaluated for the detection of antibody in sheep infected with North American ovine lentivirus (OvLV). The results of the MVV ELISA were compared with other assays for OvLV antibody and with viral infection in an intensively studied group of 38 sheep with a high prevalence of OvLV infection and disease. The sensitivity, specificity, and concordance of assays for OvLV antibody (MVV ELISA, indirect ELISA, Western blot, and AGID), virus (virus isolation, PCR, antigen ELISA), and OvLV-induced disease in each animal were compared with OvLV infection status as defined by a positive result in two or more of the assays. Five sheep met the criteria for absence of OvLV infection. The sensitivity of the MVV ELISA in detecting OvLV infected sheep was 64%, whereas the sensitivity of the other three tests for antibody ranged from 85 to 94%. All the antibody assays were 100% specific in this group of animals. Of the assays for virus, the PCR test had the highest sensitivity and the best concordance with OvLV infection, but it also had the lowest specificity of any of the virus or antibody assays. Among the antibody tests, the concordance of the MVV ELISA compared most favorably with the AGID test for detecting OvLV-infected sheep. Analysis of serum samples from 28 lambs experimentally-infected with one of three North American strains of OvLV suggested that there were no significant strain differences detectable by antibody assay. Twenty virus-inoculated lambs were positive by both the MVV ELISA and the AGID test, five lambs were MVV ELISA negative and AGID test positive, and three lambs were MVV ELISA positive and AGID test negative. No pre-inoculation samples were positive by either assay. In a longitudinal study involving seven lambs, antibodies to OvLV were detected by AGID 3-5 weeks post-inoculation, but were not detected by MVV ELISA until 5-10 weeks post-inoculation. Among 128 naturally and experimentally-infected sheep that were seropositive in the AGID test, the overall sensitivity of the MVV ELISA was higher in the naturally infected sheep (84%) than in the experimentally infected sheep (69%). The data indicated that the MVV ELISA represents a less sensitive, but specific alternative for the detection of OvLV antibodies.

Immune response to recombinant visna virus Gag and Env precursor proteins synthesized in insect cells

Two different recombinant visna virus (VV) gag-baculoviruses were constructed for the expression of precursor VV Gag in insect cells. Both recombinant Gag viruses expressed proteins migrating on SDS PAGE at the predicted rate for VV Gag precursor, Pr50gag. However, differences were seen in the morphology of the virus-like particles produced. Monoclonal antibody directed against the VV Gag capsid protein (p25) and sera from sheep infected with ovine lentiviruses reacted to both 50-kDa proteins. A recombinant VV env-baculovirus was constructed, substituting sequences encoding the signal peptide of VV Env with the murine IFN-gamma analogue. Sera from ovine lentivirus infected sheep reacted in immunoblots with two proteins of approximately 100 and 200 kDa found in the plasma membrane of insect cells infected with env-recombinant virus. Sheep immunized with either the recombinant Gag or the Env proteins developed high antibody titers to VV in ELISA. The serum of sheep and ascitic fluid of mice immunized with the recombinant Gag reacted with native Pr50gag and the processed Gag proteins in immunoblots, whereas serum of the recombinant Env immunized sheep reacted with VV gp135 and a putative oligomer of gp135. The immunized sheep responded specifically to visna virus by lymphocyte proliferation in vitro.

Envelope sequence variation, neutralizing antibodies, and primate lentivirus persistence

Studies in ungulate lentivirus systems clearly indicate that neutralization escape variants emerge over time in chronically infected animals. Studies in the EIAV system, in particular, have provided strong evidence that the humoral branch of the immune system is at least one selective force acting on an array of viral variants. In previous studies with the ungulate lentiviruses, molecularly cloned virus was never used, and plaque-purified virus was only sometimes used; the genetic determinants responsible for antigenic variation and immune selection were not determined. While molecular clones are available for HIV-1, immune selection studies have been hampered in this system by the fact that HIV-1 is infectious only for chimpanzees, which do not develop disease and are available in only limited numbers. Experiments on immune selection in humans are generally complicated by lack of knowledge on the time of infection and the genetic make-up of the infecting virus. Our studies on SIV immune selection summarized in this review provide definitive evidence that neutralization-resistant variants emerge in an individual during persistent infection by primate lentiviruses. By cloning viral envelope genes from rhesus monkeys over time and obtaining sequential serum samples from them, we have been able to study not only the evolution of envelope sequences but also the emergence of neutralization-resistant variants. Reciprocal neutralization studies were performed using parental and variant specific sera, and immune selection was demonstrated using molecularly cloned virus of defined sequence. During the course of persistent infection with SIV and HIV, there is clear selective pressure for change in discrete variable regions of envelope. The host neutralizing antibody response appears to be at least one of the selective forces driving sequence change in envelope since one result of the sequence variation is the emergence of neutralization escape mutants. This indicates that neutralizing antibodies do serve to limit HIV and SIV replication during the lengthy asymptomatic stage of infection. The coincidence of neutralization domains of HIV and/or SIV with variable regions V1, V2, V3, V4, V5, and V6 suggests a direct relationship between neutralization domains and the emergence of sequence variants. However, different selective forces may be responsible all or in part for driving sequence changes in some variable domains (summarized in Table 2). For example, alterations in cell and/or tissue tropism may be responsible at least in part for driving change in V3 and the cytotoxic T-lymphocyte response may be responsible for driving change in the signal peptide (V0; Henderson et al. 1992; Wei and Cresswell 1992).(ABSTRACT TRUNCATED AT 400 WORDS)

Pathogen evolution within host individuals as a primary cause of senescence

This paper discusses a novel theory of senescence: the community of pathogens within each host individual evolves during the life-time of the host, and in doing so progressively reduces host vigour. I marshal evidence that asymptomatic host individuals maintain persistent populations of viral pathogens; that these pathogens replicate; that they are often extremely variable; that selection within hosts causes the evolution of pathogens better able to exploit the host; that selection is host-specific; and that such evolving infections cause appreciable and progressive deterioration. Experimental approaches to testing the theory are discussed.

Detection of ovine lentivirus in seronegative sheep by in situ hybridization, PCR, and cocultivation with susceptible cells

Serological surveys for ovine lentivirus (OvLV), a worldwide cause of pneumonia and chronic debilitation in sheep, have demonstrated a wide range of seroprevalence rates. This study analyzed OvLV infection in a purebred sheep flock with a history of OvLV disease (flock 1), and compared the prevalence with that of a flock lacking previous OvLV-associated disease (flock 2). Serological tests (ELISA and Western blot assay) indicated that 25% of sheep of all ages in flock 1 (Group A) and 33% of animals of all ages in flock 2 (Group B) had antibodies to OvLV. In situ hybridization, however, detected viral RNA in a much larger proportion of sheep (72 and 67%, respectively). Animals less than 1 year of age rarely had antibodies to OvLV, although most harbored viral RNA. Twenty animals in this age group from flock 1 (Group C) were therefore studied more closely for infection. These yearling animals were tested serologically by ELISA and their peripheral blood-derived macrophages were cultured for 14 days to amplify any infection in these target cells. The macrophages were then tested by in situ hybridization, PCR, and cocultivation with susceptible target cells. The results of these tests showed that while only 10% of animals in Group C were seropositive, 70% were positive by in situ hybridization, PCR, and cocultivation. These data suggest that latent OvLV infection is common in sheep and that infection is frequently undetected by serological tests.

Detection of antibodies to ovine lentivirus using a recombinant antigen derived from the env gene

The Western blot assay was performed to characterize antibodies to the transmembrane glycoprotein (TGP) of ovine progressive pneumonia virus (OPPV) by using glutathione-S-transferase-TGP (GST-TGP) protein. The GST-TGP protein was efficiently expressed in Escherichia coli (E. coli) and was highly immunoreactive in the Western blot assay. This assay detected antibodies in 97% (103/106) of the sera from agarose gel immunodiffusion (AGID) positive OPP animals. Like human AIDS patients, antibodies to TGP appear to be one of the major serological markers in OPP infected animals.

Comparison of immunoblots with neutralizing and complement fixing antibodies in experimental and natural cases of visna-maedi

In this study the humoral antibody response in visna-maedi virus disease in sheep during long-term infection was analyzed utilizing immunoblot assays, neutralization tests and complement fixation tests. In immunoblot assays antibodies to several virus specific protein bands were detected, both against the viral envelope glycoproteins and internal proteins of the virus. The immunoblot reaction pattern resembled that found in HIV-1 infection in humans, consistent with reported similar molecular weight of the major proteins of these two viruses. The immunoblot band pattern was compared with the pattern of complement fixing and neutralizing antibodies through the preclinical and clinical course in natural and experimental cases of visna-maedi. Of six immunoblot bands identified as virus specific, the antibody response against three gag products and the major env glycoprotein appeared early in infection, at a similar time as the complement fixing antibodies. The response against two proteins, one presumably the transmembrane protein and the other possibly a gag precursor, was delayed.

Identification of cell membrane proteins that bind visna virus

Visna virus infects cells of ovine origin by attaching to a cell surface receptor via its envelope glycoprotein. The identity of the visna virus receptor is not known. To identify the molecule responsible for binding the virus to target cells, virus overlay protein blot assays were used to examine the molecular weights of cell surface molecules which bind purified virus. Molecules on the surface of goat synovial membrane (GSM) cells and sheep choroid plexus (SCP) cells of approximately 15, 30, and 50 kDa bound to visna virus. The binding of visna virus to these proteins was reduced by preincubating virus with neutralizing antibodies. 125I-labeled cell membrane preparations of GSM and SCP cells were used to affinity purify these virus-binding proteins. These proteins were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and had molecular masses of 15, 30, and 50 kDa. Antibodies to the 50-kDa protein bound to the surface of both live SCP and GSM cells in immunofluorescence assays. In addition, antibodies to the 50-kDa protein blocked the binding of [35S]methionine-labeled visna virus to SCP cells in culture. Antibodies raised against the 15- and 30-kDa proteins did not block virus binding to cells. The blocking activity of antibody of the 50-kDa protein provided data that this protein is the molecule which visna virus recognizes and binds to on the surface of target cells.

Identification of the fusion domain in the visna virus transmembrane protein

Visna virus, a lentivirus of sheep, causes fusion of susceptible cells. Fusion has previously been shown to be mediated by the viral envelope glycoprotein. The transmembrane protein of visna virus contains a hydrophobic region at its amino terminus. This region is similar to the fusion epitopes of the orthomyxoviruses and paramyxoviruses. This region is located in a position similar to that of the fusion epitopes in the transmembrane proteins of HIV-1 and SIV. To determine the role of this hydrophobic region in visna virus-induced cell fusion, a peptide of 24 amino acids corresponding to this region was synthesized. The peptide alone induces fusion of goat cells. Antibodies to this peptide inhibit both viral-induced cell fusion and peptide fusion in goat cells. Further, the direct fusion of cells by this peptide is a unique observation and may be useful for studying the fusion epitopes of other lentiviruses. Thus, this hydrophobic region appears to be one epitope responsible for visna virus-induced cell fusion.

Epitope analysis of capsid and matrix proteins of North American ovine lentivirus field isolates

Monoclonal antibodies (MAbs) directed against two phenotypically distinct ovine lentivirus (OvLV) strains were generated by fusion of BALB/c SP2/0-Ag 14 myeloma cells with spleen cells from mice immunized with purified OvLV. Hybridomas were selected by indirect enzyme-linked immunosorbent assay (ELISA) and analysis of reactivity on immunoblots. The majority (17 of 21) of the MAbs recognized the gag-encoded capsid protein, CA p27, of both strains. Four other MAbs recognized a smaller structural protein, presumably a matrix protein, MA p17. Three distinct epitopes on CA p27 and one on MA p17 were distinguished by the MAbs with competition ELISA. MAbs from each epitope group were able to recognize 17 North American field isolates of OvLV and the closely related caprine arthritis-encephalitis virus (CAEV). Analysis of the data indicated that these epitopes were highly conserved among naturally occurring isolates. A representative MAb from each epitope group of anti-CA p27 MAbs reacted with four field strains of OvLV and CAEV on immunoblots. An anti-MA p17 MAb recognized the same OvLV strains on immunoblots but failed to recognize CAEV. MAbs which recognize conserved epitopes of gag-encoded lentivirus proteins (CA p27 and MA p17) are valuable tools. These MAbs can be used to develop sensitive diagnostic assays and to study the pathogenesis of lentivirus infections in sheep and goats.

Naturally occurring HIV-1 isolates with differences in replicative capacity are distinguished by in situ hybridization of infected cells

Replication of human immunodeficiency virus type 1 (HIV-1) isolates in peripheral blood mononuclear cells (PBMC) has been studied by in situ hybridization using the riboprobe BH10-R3 from HTLV-IIIB. Two series of isolates were tested: (a) 20 isolates from individuals with varying severity of HIV-1 infection and (b) sequential isolates from 5 subjects showing signs of clinical progression over a 45 month observation period. The results show that HIV-1 isolates with distinct replicative capacity can be distinguished by the intensity of radioactive labeling over single infected cells after in situ hybridization. Sequential isolates from patients with clinically progressive HIV-1 infection show a gradual increase in replicative capacity over time. In PBMC cultures infected with such sequential isolates, intensity of radioactive label over single infected cells increases and is strongest with isolates obtained at the time of low CD4 counts in blood. The results suggest that the restriction of virus replication that operates in the early stages of HIV-1 infection is gradually lost with progression of the disease.

Subcellular localization of rev-gene product in visna virus-infected cells

The 1.4-kb mRNA of visna lentivirus is expressed early during the lytic infection of sheep choroid plexus cell cultures. It encodes for visna early gene 1 (VEG1) product, since renamed rev gene product (or Rev), based on significant amino acid sequence homologies between this protein and the proteins of simian immunodeficiency virus of macaque and human immunodeficiency virus type 2. In this report, we examined the subcellular localization and time course appearance of the Rev protein in visna virus-infected cells. Immunoprecipitation assays of [35S]methionine-labeled cell lysates with antisera raised against the Rev protein revealed a polypeptide of 19 kDa (p19rev). This protein was predominant early in the viral replication cycle and accumulated preferentially in the cytoplasmic/membrane fraction of infected cells. Indirect immunofluorescence staining of infected cells confirmed the cytoplasmic location of visna Rev protein and could reveal in some stained cells a higher concentration of Rev at the cellular plasma membrane. The regulating protein, still present late in the viral lytic cycle, is packaged into mature viral particles along with the structural gag and env gene products.

Emergence of viruses resistant to neutralization by V3-specific antibodies in experimental human immunodeficiency virus type 1 IIIB infection of chimpanzees

Emergence in two chimpanzees of human immunodeficiency virus type 1 (HIV-1) IIIB variants resistant to neutralization by the preexisting antibody is described. Viruses isolated from the HIV-1 IIIB gp120-vaccinated and -challenged animal were more resistant to neutralization by the chimpanzee's own serum than viruses isolated from the naive infected animal, indicating immune pressure as the selective mechanism. However, all reisolated viruses were 16- to 256-fold more neutralization resistant than the inoculum virus to antibodies binding to the third variable domain (V3) of the HIV-1 external envelope. Early chimpanzee serum samples that neutralized the inoculum strain but not the reisolated viruses were found to bind an HIV-1 IIIB common nonapeptide (IQRGPGRAF) derived from the gp120 isolate-specific V3 domain shown to induce isolate-specific neutralization in other animals. Amplification of the V3 coding sequence by polymerase chain reaction and subsequent sequence analysis of the neutralization-resistant variants obtained from in vivo-infected animals indicated that early resistance to neutralization by an HIV-1 IIIB monoclonal antibody (0.5 beta) was conferred by changes outside the direct binding site for the selective neutralizing antibody. The reisolated neutralization-resistant isolates consisted of the lower-replication-competent virus subpopulations of the HIV-1 IIIB stock, as confirmed by biological and sequence analyses. In vitro passage of the HIV-1 IIIB stock through chimpanzee and human peripheral blood mononuclear cell cultures void of HIV-specific antibody resulted in homogenic amplification of the more-replication-competent subpopulation preexisting in the original viral stock, suggesting a role for the immune system in suppressing the more-replication-competent viruses.

Analysis of antibody responses to phenotypically distinct lentiviruses

To define the immune responses against phenotypically and pathogenically distinct lentiviruses, we used an immunoblotting assay to study antibodies to viral proteins of ovine lentivirus (OvLV) in 16 experimentally and 12 naturally infected sheep. Two distinct phenotypes of OvLV were used to experimentally infect lambs: strain 85/34, a "rapid/high" isolate which rapidly induced lysis in infected primary macrophage cultures and replicated to relatively high titers, and strains 84/28 and 85/14, "slow/low" isolates which induced slowly progressive syncytia with minimal lysis in vitro and replicated only to low titers in the same cell type. Serum antibodies against four major viral structural proteins, gp105, p25, p16, and p14, were detected. In a longitudinal study of experimentally infected lambs, the antibody to p25 (major gag protein) usually appeared first (average, about 3 weeks postinoculation [p.i.]) and was followed in about 2 weeks by p16, p14, and gp105 almost simultaneously. Six of 16 animals did not develop anti-p14 antibody by the time of necropsy at 9 to 29 weeks p.i. Two of 10 lambs which developed antibody to p14 had the antibody only transiently from 3 to 8 or 13 weeks p.i. and lost it by the time of necropsy at 21 or 22 weeks p.i. In contrast, antibodies to the other three structural proteins remained fairly constant until the time of necropsy. There were differences in the antibody responses of the experimentally infected lambs to the two phenotypes of OvLV. Seven of 10 (70%) lambs which were inoculated with the rapid/high strain developed antibody to p14, whereas only 17% of the lambs inoculated with the slow/low strains had antibody to this protein. In the longitudinal study, no decline was observed in the activity of any specific antibody such as that which occurs with anti-p24 antibody in human immunodeficiency virus infection, except in the case of anti-p14 antibody in two lambs. There were no significant differences in antibody titers against p25, p16, and p14 in final blood samples between rapid/high virus- and slow/low virus-infected groups. However, the rapid/high virus-infected group developed a fivefold-higher geometric mean titer of anti-env product (gp 105) antibody than did the slow/low virus-infected group (P </= 0.1). Antibody titers to all major structural proteins, except p14, in the naturally infected sheep were markedly lower than those in experimentally induced OvLV infections (P </= 0.01). The failure of the slow/low virus-infected group to develop anti-p14 antibody may suggest diminished viral replication in vivo or a failure of the host to recognize p14 in the slow/low virus-infected group. Since the geometric mean antibody titer to gp105 was threefold higher in lambs with lymphoid interstitial pneumonia than in those without lesions and since no differences were observed in the titers of other antiviral antibodies between these groups, we found no evidence to suggest that levels of such antibodies correlated with protection from OvLV-induced disease.

Pathogenesis of lentivirus-induced arthritis. A review

While it has been known for some years that there is an association between lentiviruses and slowly progressive joint diseases in ruminants, the realization that the human immunodeficiency virus, the cause of AIDS, is a lentivirus has made this group of virus the focus of a considerable research effort. The manifestations of lentivirus infection in animals are discussed and reference is made to the possibility of using them as models for human rheumatoid arthritis and for AIDS.

Expression of viral antigens in the central nervous system of visna-infected sheep: an immunohistochemical study on experimental visna induced by virus strains of increased neurovirulence

Icelandic sheep were infected by intracerebral inoculation with visna virus strains of increased neurovirulence. The character and severity of pathological lesions were studied in brains from four sheep that developed clinical signs 5 to 12 weeks after infection. Viral antigens were identified by immunostaining using mouse monoclonal antibodies against two core proteins and the Avidin-Biotin method of detection. The pathological lesions were in general more severe than observed following infection with the parent strain K1514. Primary demyelination, a late manifestation of infection with K1514, was detected. Thus, in addition to causing more severe pathological lesions, these neurovirulent strains apparently have an increased potential to induce primary demyelination. Viral antigens were detected in lymphocytes, plasma cells, macrophages, endothelial cells, pericytes, fibroblasts and choroidal epithelial cells. Neurons and glial cells were antigen negative. The spectrum of infected cells in the brain was similar to that observed in infections with human immunodeficiency virus. These results do not support the view that the demyelination is caused by immunological damage to infected oligodendrocytes. A perturbation of the function of oligodendrocytes through a non-productive infection could be the underlying pathogenetic mechanism and/or a non-specific demyelination due to the intense inflammatory reaction.

Antigenic variation of neutralization-sensitive epitopes of caprine arthritis-encephalitis lentivirus during persistent arthritis

Caprine arthritis-encephalitis virus (CAEV), a naturally occurring lentivirus of goats, causes disease characterized by virus persistence and recurrent arthritis. These studies demonstrate in vitro neutralization of CAEV infectivity by serum from goats infected with CAEV. Serum neutralizing activity was not detectable until 10 to 36 months postinfection, and titers were relatively low (less than or equal to 1:8). Serum neutralization was caused by antibody and was virus specific. Antigenic variants of CAEV were isolated from cell-free joint fluid of arthritic goats 9 to 18 months postinfection. The delayed appearance of neutralizing antibody and the subsequent development of antigenic variants may promote CAEV persistence in vivo and provide a stimulus for recurrent arthritis.

Development and testing of AIDS vaccines

Recent advances in delineating the molecular biology of human immunodeficiency virus type 1 (HIV-1) have led to innovative approaches to development of a vaccine for acquired immunodeficiency syndrome (AIDS). However, the lack of understanding of mechanisms of protective immunity against HIV-1, the magnitude of genetic variation of the virus, and the lack of effective animal models for HIV-1 infection and AIDS have impeded progress. The testing of AIDS vaccines also presents challenges. These include liability concerns over vaccine-related injuries; identification of suitable populations for phase 3 efficacy studies; balancing the ethical obligation to counsel research subjects to avoid high-risk behavior with the necessity to obtain vaccine efficacy data; and the effect of vaccine-induced seroconversion on the recruiting and welfare of trial volunteers. Several candidate AIDS vaccines are nevertheless currently under development, and some are undergoing phase 1 clinical trials. Rapid progress will depend on continued scientific advancement in conjunction with maximum use of resources, open information and reagent exchange, and a spirit of international collaboration.

Generation of a neutralization-resistant variant of HIV-1 is due to selection for a point mutation in the envelope gene

Transmission and growth of HIV-1 produced from the biologically active clone HTLV-III/HXB2D in the constant presence of a neutralizing antiserum yielded a viral population specifically resistant to neutralization by the same antiserum. Molecular clones MX-1 and -2, containing the entire envelope gene, were obtained from cultures of the resistant variant. The coding regions for the large envelope protein and most of the transmembrane envelope protein of two such clones were substituted for the homologous segment of HXB2D. Infectious viruses from these constructs were also specifically resistant to neutralization by the selecting antiserum. The exchanged fragment contained only one base change, resulting in an Ala----Thr replacement at position 582. When this substitution was introduced into HXB2D it conferred the resistant phenotype. Thus, small differences may be selected for in vivo by the host immune response and result in relatively large differences in susceptibility of the virus to such a response.

Antigenic variation and lentivirus persistence: variations in envelope gene sequences during EIAV infection resemble changes reported for sequential isolates of HIV

The extent and nature of genomic variation among nine antigenically distinct EIAV isolates recovered during sequential clinical episodes from two experimentally infected ponies were examined by restriction fragment analysis and nucleotide sequencing. Only minor variations in restriction enzyme patterns were observed among the viral genomes. In contrast, env gene sequences of four isolates from one pony revealed numerous clustered base substitutions. Divergence in env gene nucleotide and deduced amino acid sequences between pairs of virus isolates ranged from 0.62 to 3.4% env gene mutation rates for isolates recovered during sequential febrile episodes were calculated to be greater than 10(-2) base substitutions per site per year. The degree and nature of env gene variation in EIAV is remarkably similar to the human immunodeficiency virus, suggesting common mechanisms for env gene variation among lentiviruses.