Generation of Simian-Tropic HIV-1 by Restriction Factor Evasion

See allHide authors and affiliations

Science  06 Oct 2006:
Vol. 314, Issue 5796, pp. 95
DOI: 10.1126/science.1130994


Because HIV-1 does not infect most nonhuman primates, animal modeling of human HIV infection and AIDS has primarily consisted of experimentally infecting macaques with related simian immunodeficiency viruses (SIVMAC). However, the usefulness of such models is limited by the substantial divergence between SIVMAC and HIV-1. We derived an HIV-1‐based virus that includes only small portions of SIVMAC yet replicates robustly in both transformed and primary rhesus macaque T cells. Derivation of simian-tropic HIV-1 (stHIV-1) has important implications for understanding primate lentivirus zoonosis and should allow the development of improved animal models for studies of AIDS and the evaluation of vaccines and treatments.

Although the evolution of intrinsic cellular defenses has likely prevented colonization of humans by many viruses (1), it also poses problems for AIDS research, insofar as the same intrinsic factors may inhibit HIV-1 replication in animal models. Indeed, HIV-1 does not replicate in most nonhuman primates, so research has relied on animal models of AIDS that are based on related simian viruses, which have inherent limitations. HIV-1 infection of rhesus macaque (rh) cells in vitro fails in part because a block is imposed by the cellular factor rhTRIM5α, which targets incoming HIV-1 capsids (CA) (2). A second block may be imposed by cellular APOBEC3 cytidine deaminase enzymes (3) because of a failure of HIV-1 Vif to bind rhAPOBEC3G and induce its degradation as it does in human cells (4). However, it is not known whether overcoming these blocks would allow HIV-1 replication in rhesus macaque cells.

In an attempt to generate a simiantropic HIV-1, we incorporated sequences from a simian lentivirus (SIVMAC239) into HIV-1 molecular clones. In these chimeric HIV-1–based genomes, “SCA” and “SVif” indicate that the CA and Vif sequences originated from SIVMAC239 (Fig. 1A). Replacement of HIV-1 CA, generating HIV(SCA) (5), attenuated replication, but after serial in vitro passage in human CEMx174 cells (19 times over 16 weeks) HIV(SCA) replicated robustly and could rapidly induce abundant cytopathic effects, unlike the starting virus. A fragment of this adapted HIV(SCA) genome (BssHII-ApaI, Fig. 1A) was reintroduced into an HIV-1 genomic clone, encoding green fluorescent protein (GFP) in place of the nonessential Nef protein, to generate an improved and easily monitored HIV(SCA) virus (5).

Fig. 1.

(A) Schematic of the HIV(SCA, SVif) or stHIV-1 genome. Shaded regions are from SIVMAC239. (B) Single-cycle infectivity of GFP-expressing HIV-1 and derivatives measured by using target CEMx174 cells expressing rhTRIM5α (black). Alternatively, viruses were generated in 293T cells in the presence of rhAPOBEC3G (gray). Values are plotted as a proportion of those obtained by using producer and target cells that did not express rhTRIM5α or rhAPOBEC3G. Error bars indicate standard deviations. (C) Reverse transcriptase (RT) accumulation in culture supernatants of human cells (CEMx174) or rhesus cells (221 and PBL) after challenge with HIV-1, SIVMAC239, or stHIV-1.

Subsequently, HIV-1 vif was replaced to generate viral clones, namely HIV, HIV(SCA), HIV(SVif), and HIV(SCA, SVif), encoding the four possible combinations of HIV-1 and SIVMAC239 CA and Vif proteins. In single-cycle infection assays (Fig. 1B), expression of rhAPOBEC3G during virus production reduced HIV-1 Vif–encoding virus infectivity by >30-fold but only marginally affected SIVMAC239 Vif–encoding counterparts. Conversely, rhTRIM5α expression in target cells inhibited HIV-1 CA–encoding virus infection by about 30-fold, but SIVMAC239 CA–encoding viruses were unaffected (Fig. 1B).

HIV, HIV(SCA), and HIV(SVif) replicated rapidly in CEMx174 cells, whereas HIV(SCA, SVif) replicated more slowly. Conversely, HIV and HIV(SVif) did not replicate in a rhesus macaque T cell line, 221, whereas HIV(SCA) replication was low and transient (fig. S1). Only HIV(SCA, SVif) initiated a slowly spreading infection in 221 cells, with cytopathic effects becoming evident at 16 to 20 days post-infection (fig. S1). HIV(SCA, SVif) replication became robust after adaptation (two passages in CEMx174 cells and three passages in 221 cells) and, thereafter, a BssHII-SalI fragment of this adapted HIV(SCA, SVif) genome was reintroduced into an HIV-1 genome encoding Nef, generating the simian tropic HIV-1 (stHIV-1) clone. Importantly, stHIV-1 differed from the parental HIV-1 equivalent in only minor ways, other than the CA and Vif substitutions. Specifically, stHIV-1 had few coding (amino acids Lys110→Ile110, Ala208→Val208, and Pro371→Leu371) and silent mutations in gag (nucleotides 291, 321, and 477) and in pol (nucleotides 1248, 2157, and 3411). No sequence changes in or around vif were evident.

The stHIV-1 clone replicated robustly in both human CEMx174 cells and macaque 221 cells (Fig. 1C). Moreover, stHIV-1 and SIVMAC239 replicated with similar rapid kinetics in peripheral blood lymphocytes (PBL) from all five macaque donors tested (Fig. 1C). In contrast, HIV-1 replication was either extremely low (two donors) or undetectable (three donors).

Overall, these findings suggest that avoidance of CA- and Vif-based restriction, by chance or adaptation, may be sufficient to allow cross-species transmission of primate lentiviruses. Additionally, the generation of a macaque cell-tropic virus whose genome is 88% HIV-1 derived should allow the development of more authentic animal models of human AIDS and facilitate the preclinical development of new therapies and vaccines.

Supporting Online Material

Materials and Methods

Fig. S1



View Abstract

Navigate This Article