Patterns of genomic sequence diversity among their simian immunodeficiency viruses suggest that L'Hoest monkeys (Cercopithecus lhoesti) are a natural lentivirus reservoir

Abstract

Recently, we described a novel simian immunodeficiency virus (SIVlhoest) from a wild-caught L'Hoest monkey (Cercopithecus lhoesti) from a North American zoo. To investigate whether L'Hoest monkeys are the natural host for these viruses, we have screened blood samples from 14 wild animals from the Democratic Republic of Congo. Eight (57%) were found to be seropositive for SIV. Nearly full-length genome sequences were obtained for SIV isolates from three of these monkeys and compared to the original isolate and to other SIVs. The four samples of SIVlhoest formed a distinct cluster in phylogenetic trees. Two of these isolates differed on average at only about 5% of nucleotides, suggesting that they were epidemiologically linked; otherwise, the SIVlhoest isolates differed on average by 18%. Both the level of diversity and the pattern of its variation along the genome were very similar to those seen among isolates of SIVagm from vervet monkeys, pointing to similarities in the nature of, and constraints on, SIV evolution in these two species. Discordant phylogenetic relationships among the SIVlhoest isolates for different genomic regions indicated that mosaic viruses have been generated by recombination, implying that individual monkeys have been coinfected by more than one strain of SIV. Taken together, these observations provide strong evidence that L'Hoest monkeys constitute a natural reservoir for SIV.

FIG. 1

Radioimmunoprecipitation using SIVmac whole-cell lysate. 443, 444, 447, 448, 456, 485, 489, 524, 539, 545, 546, 548, 555, and 554 are plasma samples from L'Hoest monkeys which were wild caught in the northeastern part of the Democratic Republic of Congo.

FIG. 2

Comparison of predicted protein sequences of the surface subunit of envelope proteins of SIVlhoest7, SIVlhoest447, SIVlhoest485, and SIVlhoest524. Conserved cysteines are boxed, and conserved N-linked glycosylation sites are underlined. The sequence of the original isolate, SIVlhoest7 (21), is shown on top. Differences, relative to this sequence, in SIVlhoest isolates 447, 485, and 524 are shown aligned below. Dots indicate amino acid identity at a residue, and dashes indicate gaps introduced to optimize alignment. Variable regions analogous to those observed in HIV-1 and other SIVs are indicated, and the cleavage site for the transmembrane glycoprotein (TM) is shown.

FIG. 3

Diversity plots, showing extent of nucleotide sequence difference along the SIV genome, for windows of 300 nucleotides moved in steps of 50. gag, pol, vif, vpr, env, and nef gene sequences were aligned (based on protein alignments) and concatenated. Regions of gene overlap, as well as regions of uncertain alignment, and sites with a gap in any sequence, were excluded. Comparisons involved SIVlhoest isolate 447 versus isolates 7 (solid black), 485 (dashed black), and 524 (grey) (A) and SIVlhoest isolate 7 versus isolate 524 (solid black), SIVagmVer isolate 3 versus isolate 155 (grey), and SIVcpz isolate Gab-1 versus isolate US (dashed black) (B).

FIG. 4

Phylogenetic relationship of the four SIVlhoest isolates (SIVlhoest7, SIVlhoest447, SIVlhoest485, and SIVlhoest524) to other representatives of the major lentivirus lineages: SIVcpzPtt/Gab-1 (GenBank accession no. X52154), SIVcpzPts/Ant (accession no. U42720), SIVcpzPtt/US (accession no. AF103818), SIVsmPBj (accession no. M31325), HIV-2 subtype A (ROD; accession no. M15390), HIV-2 subtype B (EHOA; accession no. U27200), SIVagmVer3 (accession no. M30931), SIVagmVerTyo-1 (accession no. X07805), SIVagmVer9063 (accession no. L40990), SIVagmVer155 (accession no. M29975), SIVagmGri (gri-1; accession no. M58410), SIVagmTan (tan-1; accession no. U58991), SIVsyk (173; accession no. L06042), SIVmnd (GB1; accession no. M27470), SIVlhoest7 (accession no. AF075269), and SIVsun (accession no. AF131870). The trees were derived by the neighbor-joining method (44) applied to protein distances estimated using Kimura's correction (33) with 1,000 bootstrap replicates, implemented using ClustalX (47). The trees were midpoint rooted. Bootstrap values greater than 80% are shown. Trees derived by maximum-likelihood analysis implemented with PROTML (MOLPHY, version 2.2, J. Adachi and M. Hasegawa, Institute of Statistical Mathematics, Tokyo, Japan, 1994) using the Jones-Taylor-Thornton model (30) with data frequencies differed in no significant way. Horizontal branch lengths are drawn to scale, with the bar indicating 0.1 amino acid replacements per site.