Loss of CXCR6 coreceptor usage characterizes pathogenic lentiviruses

Abstract

Pandemic HIV-1 originated from the cross-species transmission of SIVcpz, which infects chimpanzees, while SIVcpz itself emerged following the cross-species transmission and recombination of monkey SIVs, with env contributed by the SIVgsn/mus/mon lineage that infects greater spot-nosed, mustached and mona monkeys. SIVcpz and HIV-1 are pathogenic in their respective hosts, while the phenotype of their SIVgsn/mus/mon ancestors is unknown. However, two well-studied SIV infected natural hosts, sooty mangabeys (SMs) and African green monkeys (AGMs), typically remain healthy despite high viral loads; these species express low levels of the canonical coreceptor CCR5, and recent work shows that CXCR6 is a major coreceptor for SIV in these hosts. It is not known what coreceptors were used by the precursors of SIVcpz, whether coreceptor use changed during emergence of the SIVcpz/HIV-1 lineage, and what T cell subsets express CXCR6 in natural hosts. Using species-matched coreceptors and CD4, we show here that SIVcpz uses only CCR5 for entry and, like HIV-1, cannot use CXCR6. In contrast, SIVmus efficiently uses both CXCR6 and CCR5. Coreceptor selectivity was determined by Env, with CXCR6 use abrogated by Pro326 in the V3 crown, which is absent in monkey SIVs but highly conserved in SIVcpz/HIV-1. To characterize which cells express CXCR6, we generated a novel antibody that recognizes CXCR6 of multiple primate species. Testing lymphocytes from SM, the best-studied natural host, we found that CXCR6 is restricted to CD4+ effector memory cells, and is expressed by a sub-population distinct from those expressing CCR5. Thus, efficient CXCR6 use, previously identified in SM and AGM infection, also characterizes a member of the SIV lineage that gave rise to SIVcpz/HIV-1. Loss of CXCR6 usage by SIVcpz may have altered its cell tropism, shifting virus from CXCR6-expressing cells that may support replication without disrupting immune function or homeostasis, towards CCR5-expressing cells with pathogenic consequences.

Fig 1. Coreceptor and CD4 sequences vary in sequence between human, CPZ, GSN and MUS.

CXCR6 and CCR5 were cloned from CPZ, GSN and MUS genomic DNA. Sequences were aligned to the human sequence, and shown are the N-terminus (N term) and the second extracellular loop (ECL2), the two domains thought to interact with the virus glycoprotein, based on studies of human CCR5 and HIV-1 gp120. GSN and MUS CD4 exons were sequenced from genomic DNA and aligned to known sequences of human CD4 and chimpanzee CD4; shown is domain 1 that interacts with HIV/SIV gp120. Residues identical to the human sequence are represented as dots.

Fig 2. SIVcpz is restricted to CPZ CCR5, while SIVmus1085 can use MUS and GSN CXCR6 and CCR5 for entry.

A) Left panel: CF2thLuc cells that contain a Tat-driven luciferase reporter were transfected with expression plasmids containing chimpanzee coreceptor and CD4 or empty vector (No CoR). 48 hours later cells were infected with one of four diverse SIVcpz isolates derived from infectious molecular clones. Entry was quantified 48 hours later by lysing cells and measuring luciferase content by relative light units (RLU). Right panel: 293T cells were transfected with cpzCD4 and coreceptor and infected with luciferase reporter pseudotypes expressing a promiscuous SIVsmm Env that can use a wide repertoire of coreceptors for entry. B-C) 293T cells were transfected with expression plasmids containing coreceptor or empty vector (No CoR), in conjunction with CD4, of MUS (B) or GSN (C) origin. 48 hours later cells were infected with luciferase reporter viruses carrying SIVmus1085 Envs (left panel) or a promiscuous SIVsmm (right panel). Entry was quantified 72 hours later by lysing cells and measuring luciferase content by relative light units (RLU). Infections were carried out in triplicate and data represent means ± one standard deviation.

Fig 3. The inability of SIVcpz to use CXCR6 for entry is determined by Env, not cpzCXCR6.

293T cells were transfected with expression plasmids containing CD4 and coreceptor. The species origin of the CD4 and coreceptor are indicated below the graph (C, chimpanzee; M, mustached monkey;–, empty vector control). 48 hours post transfection, cells were infected with luciferase reporter pseudotypes carrying the SIVcpz MT145 Env (A) or the SIVmus1085 1–54 Env (B). Entry was quantified 72 hours later by lysing cells and measuring luciferase content by relative light units (RLU). Infections were carried out in triplicate and data represent means ± one standard deviation.

Fig 4. Amino acid sequences of SIV and HIV V3 loops.

A) The V3 loop amino acid sequences of SIVcpz and SIVmus sequences used in this study (bold text) were aligned with other members of the SIVgsn/mus/mon lineage (plain text) using the ClustalW algorithm. The V3 loop of the previously published SIVmus-01CM1085 that was isolated from the same animal as the SIVmus1085 envs cloned here is identical to that of SIVmus1085 4–1 and 4–12. Residue 326 (SIVmac239 numbering) of the V3 crown is boxed. Conserved residues are shaded in grey. B) The predominant amino acid residues that occur at V3 crown residue 326 (SIVmac239 numbering) or 313 (HIV HXB2 numbering) for SIVs and HIVs where use of CXCR6 has been tested. *SIVmac can use other species’ CXCR6 for entry, but does not efficiently enter through rmCXCR6 due to a S31R polymorphism in the N terminus of rmCXCR6.

Fig 5. SIVmus1085 Env determinants of CXCR6 usage.

A) SIVmus1085 1–54 and SIVcpz MT145 env constructs were mutated to generate V3 loop mutants SIVmusA326P and SIVcpzP326A. Luciferase reporter pseudotyped viruses carrying these Envs, or SIVmus 1085 1–54 or SIVcpz MT145, were used to infect 293T cells expressing musCD4 and coreceptor (M) or cpzCD4 and coreceptor (C). Entry was quantified 72 hours post-infection by relative light units (RLU). B) The V3 loop of SIVcpz MT145 was replaced with the V3 loop of SIVmus1085 1–54 to generate SIVcpz-musV3. Pseudotypes carrying the SIVcpz-musV3 Env or SIVcpz MT145, were used to infect 293T cells expressing musCD4 and coreceptor (M) or cpzCD4 and coreceptor (C). Due to low infectivity of SIVcpz-musV3, entry was normalized to % entry through musCD4 and musCCR5 to enable comparison. Infections were carried out in triplicate and data represent means ± one standard deviation.

Fig 6. Monoclonal antibody 20D8 recognizes multiple primate species’ CXCR6 molecules.

293T cells were transfected with expression plasmids containing one of eight primate CXCR6 genes, SM CCR5 or empty vector. 48 hours later, cells were stained with mouse anti-primate CXCR6 antibody 20D8 followed by a goat anti-mouse secondary (A), secondary only (B) or a commercially available anti-human CXCR6 antibody (C).

Fig 7. CXCR6 is expressed on natural host CD4+ T cells that are distinct from CCR5-expressing cells.

A) CCR5 (y-axis) and CXCR6 (x-axis) expression on resting SM CD4+ T cells (left) and CD8+ T cells (right). Numbers in the quadrant are the percent of CD4+ or CD8+ T cells expressing the respective combination of coreceptors. One representative animal is shown. B) Summary of CCR5 and CXCR6 single- and co-expressing cells as the percent of total CD4+ and CD8+ T cells from six SMs. Each symbol represents cells from a different individual SM. C) Resting SM PBMC were stained using antibodies to define CXCR6 expression of CD4+ memory subsets: naive (Tn: CD45RA+/ CCR7+/ CD28+/ CD95-), central memory (Tcm: CD45RA-/ CCR7+) and effector memory (Tem: CD45RA-/ CCR7-). D) Summary of CXCR6 expression of CD4+ Tn, Tcm and Tem from six SMs. Each symbol represents cells from a different individual SM. Data show individual percentages, along with mean and standard deviation.