Eastern chimpanzees, but not bonobos, represent a simian immunodeficiency virus reservoir

Abstract

Chimpanzees in west central Africa (Pan troglodytes troglodytes) are endemically infected with simian immunodeficiency viruses (SIVcpzPtt) that have crossed the species barrier to humans and gorillas on at least five occasions, generating pandemic and nonpandemic forms of human immunodeficiency virus type 1 (HIV-1) as well as gorilla SIV (SIVgor). Chimpanzees in east Africa (Pan troglodytes schweinfurthii) are also infected with SIVcpz; however, their viruses (SIVcpzPts) have never been found in humans. To examine whether this is due to a paucity of natural infections, we used noninvasive methods to screen wild-living eastern chimpanzees in the Democratic Republic of the Congo (DRC), Uganda, and Rwanda. We also screened bonobos (Pan paniscus) in the DRC, a species not previously tested for SIV in the wild. Fecal samples (n = 3,108) were collected at 50 field sites, tested for species and subspecies origin, and screened for SIVcpz antibodies and nucleic acids. Of 2,565 samples from eastern chimpanzees, 323 were antibody positive and 92 contained viral RNA. The antibody-positive samples represented 76 individuals from 19 field sites, all sampled north of the Congo River in an area spanning 250,000 km(2). In this region, SIVcpzPts was common and widespread, with seven field sites exhibiting infection rates of 30% or greater. The overall prevalence of SIVcpzPts infection was 13.4% (95% confidence interval, 10.7% to 16.5%). In contrast, none of the 543 bonobo samples from six sites was antibody positive. All newly identified SIVcpzPts strains clustered in strict accordance to their subspecies origin; however, they exhibited considerable genetic diversity, especially in protein domains known to be under strong host selection pressure. Thus, the absence of SIVcpzPts zoonoses cannot be explained by an insufficient primate reservoir. Instead, greater adaptive hurdles may have prevented the successful colonization of humans by P. t. schweinfurthii viruses.

Fig 1

Location of ape study sites. (A) Geographic ranges of chimpanzees (Pan troglodytes) and bonobos (Pan paniscus) in Saharan Africa. The four recognized chimpanzee subspecies are color coded (P. t. verus, gray; P. t. ellioti, purple; P. t. troglodytes, blue; P. t. schweinfurthii, yellow). International borders, major rivers and lakes, and select cities are shown. Asterisks indicate where the closest SIVcpz relatives of HIV-1 groups M (red) and N (green) were identified in wild-living P. t. troglodytes communities (30). A box outlines the study area, which is magnified in panel B. (B) Location of chimpanzee (circles) and bonobo (squares) study sites in the DRC, Uganda, and Rwanda. The ranges of eastern chimpanzees (yellow) and bonobos (orange) are shown as in panel A. Sites where SIVcpz was detected are indicated in red, with white and yellow lettering denoting the recovery of antibody-positive and antibody- and nucleic acid-positive samples, respectively. Previously published SIVcpz-positive and -negative sites in Uganda, Rwanda, and Tanzania are shown in dark red and gray, respectively. Forested areas are shown in green, while arid and semiarid areas are in yellow and brown. Major lakes are shown in black, with major rivers depicted in blue. Dashed white lines indicate national boundaries.

Fig 2

Detection of SIVcpz antibodies in chimpanzee fecal samples. Fecal samples from eastern chimpanzees (middle) and bonobos (right) and human controls (left) were tested by an enhanced chemiluminescence Western blot using HIV-1 antigen-containing strips. Samples are numbered, with letters indicating their collection site as shown in Fig. 1B. Molecular weights of HIV-1 proteins are indicated. The banding patterns of plasma from HIV-1-infected (positive) and uninfected (negative) humans are shown for the control.

Fig 3

SIVcpz strains from the DRC cluster according to their subspecies of origin. A maximum likelihood tree was constructed from partial (232-bp) pol sequences (spanning HXB2 coordinates 4682 to 4913). Newly characterized SIVcpz strains from the DRC are highlighted, with sequences from the same individual color coded (for individual designations and sample numbers, see Table S3 in the supplemental material). Previously characterized SIVcpz, SIVgor, and HIV-1 strains forming the SIVcpzPtt (top cluster) and SIVcpzPts (bottom cluster) lineages are shown in black. The latter includes reference strains from Gombe (TAN1, TAN2, TAN3, TAN5, and TAN13) and Ugalla (UG38) as well as ANT, which is of unknown origin. Asterisks indicate bootstrap support of ≥70%. The scale bar represents 0.05 substitutions per site.

Fig 4

Phylogeny of SIVcpz from the DRC. Maximum likelihood trees were constructed of partial pol (HXB2 coordinates 3887 to 4778) (A), vpu/env (HXB2 coordinates 6062 to 6578) (B), gp41 (HXB2 coordinates 7836 to 8264) (C), and gp41/nef (HXB2 coordinates 8277 to 9047) (D) sequences. Regions of ambiguous alignment were removed from this analysis. New SIVcpzPts strains from the DRC are shown in blue, followed by the sample code in parentheses. Previously characterized SIVcpz, SIVgor, and HIV-1 strains forming the SIVcpzPtt (top cluster) and SIVcpzPts (bottom cluster) lineages are shown in black. Nodes with both bootstrap support of ≥70% and a Bayesian posterior probability of ≥0.95 are indicated by asterisks. The scale bar represents 0.05 substitutions/site.

Fig 5

Generation and biological characterization of a replication-competent SIVcpzPts molecular clone. (A) Individual RT-PCR amplicons (orange boxes) of BF1167 are shown in relation to the SIVcpz genome. Fragments are drawn to scale, with nucleotide sequences numbered starting at the beginning of the R region in the 5′ LTR (see scale bar). Three subgenomic fragments bound by MluI, NcoI, SalI, and ApaI restriction sites were synthesized and then assembled to produce a full-length provirus (blue line). (B) The replication kinetics of BF1167-derived virus in human (top) and chimpanzee (bottom) CD4⁺ T cells are shown in relation to those of HIV-1 (SG3; blue) and SIVcpzPts (TAN2; green) reference strains (x axis, days postinfection; y axis, nanograms of reverse transcriptase [RT] activity per ml of culture supernatant). Average values (and one standard deviation) from different experiments (indicated in parentheses) are shown. (C) TZM-bl cells were pretreated with AMD3100 (inhibitor of CXCR4), TAK779 (inhibitor of CCR5), or both prior to addition of the virus preparations indicated. Virus infectivity is plotted on the vertical axis as a percentage of the untreated control. Virus derived from the reference clones NL4.3 (X4 tropic), YU2 (R5 tropic), and WEAU1.6 (dual tropic) and SIVcpzPtt (MB897) and SIVcpzPts (TAN2) strains were included as controls. BF1167 is an R5-tropic virus.

Fig 6

Evolutionary relationships of BF1167 full-length genome sequences. Maximum likelihood trees were inferred from amino acid (aa) sequence alignments of the major proteins, including Gag (420 aa; HXB2 coordinates 790 to 2280) (A), N-terminal Pol (630 aa; HXB2 coordinates 2295 to 4184) (B), C-terminal Pol/Vif (453 aa; HXB2 coordinates 4185 to 5556) (C), and Env (729 aa; HXB2 coordinates 6324 to 8792) (D); the Pol protein was separated into two fragments at a point where a recombination breakpoint was previously identified in HIV-1 group N. The BF1167 sequence is shown in blue. Previously characterized SIVcpz, SIVgor, and HIV-1 strains forming the SIVcpzPtt (top cluster) and SIVcpzPts (bottom cluster) lineages are shown in black. Nodes with both bootstrap support of ≥70% and a Bayesian posterior probability of ≥0.95 are indicated by asterisks. The scale bar represents 0.05 amino acid replacements/site.

Fig 7

Adaptive requirements of SIVcpz at Gag-30. The codon at position 30 of the Gag matrix protein is shown for 14 SIVcpzPts strains, including seven new viruses from the DRC. Ten of these 14 viruses encoded a Leu using TTA or TTG codons at Gag-30, codons which require at least two nucleotide changes to become Arg (CGN or AGR) or Lys (AAR) codons. In contrast, all sequenced SIVcpzPtt (n = 15) and SIVgor (n = 3) strains contain a Met ATG codon, which requires only a single substitution to change to either human-specific signature.

Fig 8

Adaptive requirements of SIVcpz in transmembrane and cytoplasmic domains of Vpu. Vpu amino acid sequences of SIVcpzPtt, SIVgor, and SIVcpzPts strains are aligned in their transmembrane domain to the corresponding region of the HIV-1 group M consensus as previously described (36). Dashes indicate gaps introduced to optimize the alignment. Gray boxes highlight residues of a conserved helix-helix interaction motif (G/AxxxAxxxAxxxW, where “x” may be any amino acid) that is required to counteract human tetherin (68, 80). The minimum number of mutational steps needed to change the corresponding amino acid to that of the human residue is indicated on the right. Columns on the far right indicate the presence (+) or absence (blank) of previously described transport and/or degradation motifs in the intracellular domain of Vpu. These include a YxxΦ motif scored as Yxx(L/M/V/I/F/W) (8, 21, 56), a β-TrCP ubiquitin-dependent degradation signal scored as D(S/D/E)Gxx(S/D/E) (28, 39), and a putative trafficking signal scored as (D/E)xxxL(L/V/I/M) (33).