We investigated the genetic diversity of the serogroup 2 simian retroviruses (SRV-2) in four different wild-caught or captive macaque species from six different primate centers within the US over a 23 year time period. We identified at least six different SRV-2 subtypes by molecular comparison of the complete env gene from twenty-two different isolates. Our results indicate that separate introductions of at least six parental virus subtypes have occurred in the captive macaque populations in the U.S. with subsequent horizontal transfer between macaque species and primate centers.
It is most likely that divergent SRV-2 strains were introduced to the United States via importations of different species of infected macaques from different geographical areas. Procurement from common sources, close contact in primate holding facilities, and traffic between primate centers would explain the spread of virus across the captive macaque populations and between macaque species. The introduction of the SRV-2E subtype into the Washington NPRC provides one example for such a virus transfer that was evident from our study. In 1994, a female cynomolgus macaque, A94040, was purchased by the Washington NPRC for breeding purposes. At the time of transfer, this animal was negative for SRV-2 by serology but was later shown to be positive by virus culture. Our analysis of DNA from PBMCs collected in 1997 revealed that A94040 was infected with an SRV-2E subtype, that was not present in other macaques sampled at the Washington NPRC before 1994. The offspring of A94040, M96026, born in 1996, became infected with SRV-2 and analysis of PBMCs collected in 2003 revealed the presence of an SRV-2E isolate identical to that of its mother. Our analysis demonstrated that siblings with the same father as M96026, but a different mother, were infected with SRV-2E isolates that were nearly identical (1–3 aa differences) to that of M96026 and its mother A94040
Our data demonstrates that the env gene of SRV-2 is very stable suggesting a remarkable adaptation of the virus to its host. Within the five isolates of SRV-2E obtained from a cohort of cynomolgus macaques at the Washington NPRC, only 0–3 amino acid differences within the 574 aa envelope protein were detected. In addition, we found no evidence for variation of the viral env gene within a single individual over a 6 year period. Surprisingly, even viral isolates from different primate centers from different macaque species separated in time by as much as 20 years showed a high degree of conservation. The SRV-2B isolates obtained seven years apart from the rhesus macaque, YN91-224, at the Yerkes NPRC and the pig-tailed macaque, T81273, at the Washington NPRC, differed by only one amino acid. Our data confirm earlier studies which showed a remarkable stability of the SRV-2 genome over time by analysing partial env sequences in smaller and more restricted samples [17, 41].
The stability of the viral env gene over time within any given subtype suggests that the different SRV-2 subtypes evolved in the wild over long periods of time in segregrated primate hosts. Such segregation could be dictated by constraints involving different geographical areas, different niches within the same geographical area, and/or different natural host species. In our study, the natural host species for only one of the SRV-2 subtypes was apparent. The SRV-2F subtype was identified in a number of cynomolgus macaques in the wild on the island of Singapore. Interestingly, the SRV-2F subtype was clearly distinct from all the subtypes present in captive populations of cynomolgus, rhesus, Black celebes, or pig-tailed macaques, suggesting that none of these five subtypes originated from a cynomolgus macaque reservoir in Singapore. To date, no SRV-2 reservoir has been identified in wild-living rhesus macaques which are native to India . Thus, the natural host species of SRV are likely to be found in Southeast Asia. However, further analysis of SRV-2 isolates directly from wild-caught animals is needed to understand the natural reservoirs for these viruses in more detail.
Our initial impetus to study the genetic variation within the SRV-2 serotype was to determine whether there was an association between virus subtype and SAIDS-RF. Our data revealed that SAIDS-RF was associated with three SRV-2 subtypes, 2A, 2B and 2C, in multiple species of macaque, including pig-tailed, rhesus, cynomolgus and Black celebes. A total of eight RF cases were examined from five primate centers including the Washington, Oregon, and Yerkes NPRCs, the NIH primate center and the Lovelace Respiratory Research Institute in New Mexico. While SRV-2A was associated with RF in celebes and cynomolgus macaques, the SRV-2B subtype was associated with RF in pig-tailed and rhesus macaques. The SRV-2C subtype was only associated with RF in pig-tailed macaques. No obvious sequence similarities were detected between the SRV-2A, -2B and -2C subtypes which would correlate with the RF association. In two of the RF cases, RF occurred soon after experimental infection with SIV or SHIV. The rhesus macaque YN91-224 which was infected with an SRV-2B subtype was diagnosed with RF after undergoing an experimental infection with SIV at the Yerkes NPRC (personal communication, H. McClure). The SRV-2C infected pig-tailed macaque 442N was diagnosed with RF 24 weeks after infection with a pathogenic strain of SHIV . Thus, our studies revealed only an association between the SRV-2A subtype and SAIDS-RF in Black celebes macaques and between the SRV-2B subtype and SAIDS-RF in pig-tailed macaques, in the absence of other known immunodeficiency agents.
We have recently identified a single case of RF in a rhesus macaque experimentally infected with a pathogenic strain of SIV . This animal was negative for all SRV serotypes using type-specific qPCR assays. Additionally, four cases of SAIDS-RF were reported in 1983 in a colony of Taiwanese rock macaques at the New England NPRC which were endemically infected with SRV-1 . Similarly, a single case of SAIDS-SF, the subcutaneous form of RF, was reported in 1983 in a colony of rhesus macaques endemically infected with the D1/RHE/CA subtype of SRV-1 at the California NPRC . Even though the vast majority of RF cases in the different macaque colonies were associated with SRV-2 serotypes, these findings suggest a broader role for different SRV serotypes and possibly other retroviruses such as lentiviruses as cofactors in the development of RF, albeit with an apparent low efficiency.
Factors which are likely to play a major role in the development of RF, include differences in the severity and type of immunosuppression caused by an SRV-2 or lentivirus infection, and co-infection with RFHV. Current data suggests that the macaque herpesvirus, RFHV, may play a causative role in the etiology of RF [23, 45, 46]. We have developed PCR assays to detect the host-specific variants of RFHV in rhesus (RFHVMm) and pig-tailed (RFHVMn) macaques and have identified RFHV in RF lesions from the macaques infected with both SRV-2B (T82422, 90167, M78114, and YN91-224) and SRV-2C (442N) in this study [, and unpublished results]. It is not known, at this point, whether the macaque cohorts infected with the SRV-2D, -2E and -2F subtypes which were not associated with RF, were also co-infected with RFHV.
Differences in disease pathology and severity have been observed in SRV-2 infections which could impact RF development. Within a cohort infected with a single SRV-2 subtype, different outcomes have been reported, including a viremic state with rapid progression of SAIDS, a low-grade viremia with a chronic milder form of the disease, and a strong antibody response with no overt signs of disease . Similarly, the same SRV-2 subtype can elicit differences in disease severities in different macaque species. The D2/CEL/OR isolate, for example, caused severe immunodeficiency in Celebes black macaques, but when the same isolate was transmitted to rhesus macaques, the animals seroconverted and remained virus- and symptom-free . Conversely, the D2/RHE/OR isolate caused mild disease in rhesus macaques but severe fatal immunodeficiency disease in Japanese macaques (Macaca fuscata). On the other hand, the closely related SRV-2B isolates, D2/RHE/OR and D2/RHE/OR/V1, which differ in only 17 amino acids over their entire genomes, were found to induce vastly different disease outcomes in rhesus macaques and to also display differences in tropism in cell culture assays . The D2/RHE/OR variant was associated with only mild disease while the V1 variant caused severe SAIDS.
In our study, the env sequence from the different SRV-2 isolates was highly conserved overall, with 93–96% amino acid identity between isolates from different subtypes and 97–100% amino acid identity between isolates within a subtype. A hypervariable region was detected between aa 284–321 near the C-terminus of the gp70 protein. Even within this variable region, most amino acid changes were conservative or conserved with the SRV-1 or SRV-3 sequences or both, underlining the stability of the env protein. Interestingly, all potential N-linked glycosylation sites were conserved between the SRV-2 isolates, and even with the related SRV-1 and SRV-3 serotypes. This is in stark contrast to HIV and SIV which exhibit extreme variation in number and precise location of N-linked glycosylation sites . Such variation is believed to be a major pathway for immune evasion for lentiviruses. Thus, the strict conservation of glycosylation sites between the various SRV-2 isolates may help explain the high efficiency of neutralizing antibodies against SRV-2 infections .
Our analysis showed that 15.5% of the amino acid positions within the receptor-binding surface-exposed (SU) subunit gp70 were variable while only 8.7% of the amino acid positions within the transmembrane (TM) subunit gp22 differed between isolates. Differences were more concentrated in the C-terminal portion of gp70 (last 100 aa) but also affected the N-terminal signal peptide domain, as well as the known B- and T-cell epitopes (aa96-102, aa127-153). The B- and T-cell epitope at aa96-102 is of particular importance since naturally occurring neutralizing antibodies to SRV-2 are directed against this area  and it confers binding to the RD114/simian type D retrovirus receptor, a neutral amino acid transporter [51, 52]. Three amino acid positions in this seven amino acid domain displayed variations between isolates. We identified three sequence variants within the T-cell epitope at aa127-152, but only one difference in the 2nd T-cell epitope located at aa233-249 . Although some of these differences are of conservative nature, the remaining changes could affect the ability of the virus subtypes to elicit an immune response and lead to differences in disease outcome.
The gp22 subunit was far less variable between isolates than the gp70 subunit which likely is the result of sterical restrictions imposed by a string of conserved functional domains. The TM subunits of most retroviruses, including SRV-3, contain an N-terminal hydrophobic fusion peptide followed by a putative coiled coil-forming sequence, a disulfide-bonded loop connected to a shorter C-terminal alpha helix, a region with one or more N-linked glycosylation sites, a hydrophobic membrane-spanning sequence, and, as in SRVs, a cytoplasmic tail. Absolute sequence conservation between different SRV-2 subtypes and the SRV-1 and SRV-3 serotypes was seen within a region of the C-terminal gp20 domain which has been shown to be crucially important for the interaction between the SU and TM domains and for the fusion of viral and cellular membranes in MPMV. Using a threading algorithm, we showed that the homologous region within SRV-2, between aa426 and aa471, was structurally similar to a region in other retroviruses including MMLV and HTLV-1 [31, 32, 54] and the filovirus Ebola . In these and other distantly related retroviruses [34, 35] as well as in the orthomyxovirus influenza, the conserved disulfide-bonded loop plays a highly pivotal role in stabilizing a chain reversal which provides a hinge-like function that brings the fusion peptide into proximity to the target cell membrane during the fusion process . We propose that SRV-2, in analogy, uses a similar mechanism for host cell membrane fusion.