Geographic origin and evolutionary history of China's two predominant HIV-1 circulating recombinant forms, CRF07_BC and CRF08_BC

Abstract

To determine the origin and evolutionary history of two predominant and closely-related circulating recombinant forms (CRF07_BC and CRF08_BC), recombinant structures and phylogenies of 7 unique recombinant forms comprised of subtypes of B' (Thai B linage) and C (designated URFs_BC) from archival specimens of injection drug users (IDUs) collected in 1996 to 1998 from western Yunnan and 4 circulating recombinant forms with B'/C recombinants recently identified (designated nCRFs_BC) in China were compared with those of CRF07_BC and CRF08_BC. The results showed that 5 of 7 URFs_BC and all the nCRFs_BC shared recombination breakpoints with CRF07_BC and/or CRF08_BC. Yunnan URFs_BC consistently occupied the basal branch positions compared with CRF07_BC, CRF08_BC, and nCRFs_BC in phylogenetic trees. The estimated most recent common ancestors (tMRCA) for Yunnan URFs_BC were from ~1987, approximately half a decade earlier than those for CRF07_BC (~1994) and CRF08_BC (~1992). Discrete phylogeographic and spatial diffusion analysis revealed that both CRF07_BC and CRF08 BC came from western Yunnan in the early 1990s. Our results provide compelling evidence for western Yunnan as the geographic origin of CRF07_BC and CRF08_BC, which emerged from a swarm of URFs_BC by a series of recombination events in western Yunnan in the early 1990s.

Figure 1. Precise mapping of recombination breakpoints in the B’/C recombinant forms in western Yunnan.

Detailed subtype structure of the B’/C recombinant forms was estimated as described in Materials and Methods (see also Supplemental Figure S2, S3). Subtype B’ region (blue), subtype C region (magenta). The sites 1 through 8 are the recombination breakpoints shared among CRF07_BC, CRF08_BC, and other B’/C recombinants that are indicated with thick vertical lines. Regions Ia (HXB2: 790–1191), Ib (HXB2: 1642–2010 nt) II (HXB2: 3330–5593 nt). IIIa (HXB2: 6541–7348 nt), and III (HXB2: 6541–8695) are the subtype C segments that were selected for the subregion MCC tree analyses (see Fig. 2) and temporal dynamic analyses of spatial diffusion for B’/C recombinants in order to examine the relationships and the timescale of emergence and diffusion of CRF07_BC/CRF08_BC.

Figure 2. MCC trees based on selected subtype C segments considering B’/C recombinant forms in this study.

Bayesian molecular clock analyses were performed using BEAST v1.74 (see methods for details) for different subtype C regions: (A) the concatenated gag region (Ia + Ib) (HXB2: 790–1191 nt and 1642–2010 nt), (B) the pol region (II) (3330–5593 nt), and (C) the env region (III) (6541–8695 nt).The tree branches are color-coded according to their respective genotypes and geographical locations. The medians of tMRCA are as follows: CRF07_BC ancestor = ~1994, CRF08_BC ancestor = ~1992, CRF57_BC ancestor = ~1993, CRF61_BC ancestor = ~2002, CRF62_BC ancestor = ~2003, CRF64_BC ancestor = ~1995, Yunnan RF ancestor = ~1987. Indian subtype C ancestor = ~1976, and African subtype C ancestor = ~1966 (see Table S1).

Figure 3. Estimated tMRCAs of subtype C and its recombinant lineages in China.

Estimated dates of the MRCA of CRF07_BC and CRF08_BC in comparison with nCRFs_BC and ancestors of URFs_BC were obtained for sub-genomic regions. Estimates were calculated under a GTR substitution model with a relaxed molecular clock and a Bayesian skyline plot coalescent model. Error bars represent the 95% higher probability density credible regions for each estimate. Estimates obtained using region (Ia + Ib:), (II), and (III) are labeled with square, triangle, and diamond respectively.

Figure 4. Temporal dynamics of spatial diffusion for B’/C recombinants in China and related “parental” subtype C lineages.

The temporal dynamics of spatial diffusion for B’/C recombinants and related lineages can be displayed follow the software instructions of SPREAD. The sketch map was drawn using Adobe Illustrator CS6 refer to the Bayesian inference of diffusion process estimated by SPREAD. The dispersal patterns for 1991, 1995, 1998, and 2010 are shown as different colors. Lines between locations represent branches in the MCC tree along which the relevant location transition occurs. Location circle diameters are proportional to the square root of the number of MCC branches maintaining a particular location state at each time-point.