TRIM5alpha Modulates Immunodeficiency Virus Control in Rhesus Monkeys

Abstract

The cytoplasmic TRIM5alpha proteins of certain mammalian lineages efficiently recognize the incoming capsids of particular retroviruses and potently restrict infection in a species-specific manner. Successful retroviruses have evolved capsids that are less efficiently recognized by the TRIM5alpha proteins of the natural hosts. To address whether TRIM5alpha contributes to the outcome of retroviral infection in a susceptible host species, we investigated the impact of TRIM5 polymorphisms in rhesus monkeys on the course of a simian immunodeficiency virus (SIV) infection. Full-length TRIM5alpha cDNAs were derived from each of 79 outbred monkeys and sequenced. Associations were explored between the expression of particular TRIM5 alleles and both the permissiveness of cells to SIV infection in vitro and clinical sequelae of SIV infection in vivo. Natural variation in the TRIM5alpha B30.2(SPRY) domain influenced the efficiency of SIVmac capsid binding and the in vitro susceptibility of cells from the monkeys to SIVmac infection. We also show the importance in vivo of the interaction of SIVmac with different allelic forms of TRIM5, demonstrating that particular alleles are associated with as much as 1.3 median log difference in set-point viral loads in SIVmac-infected rhesus monkeys. Moreover, these allelic forms of TRIM5 were associated with the extent of loss of central memory (CM) CD4+ T cells and the rate of progression to AIDS in the infected monkeys. These findings demonstrate a central role for TRIM5alpha in limiting the replication of an immunodeficiency virus infection in a primate host.

Figure 1. A schematic illustrating TRIM5α and its functional domains.

The major domains of TRIM5α are shaded in light grey, and the linker 1 (L1) and linker 2 (L2) regions are designated. The dark grey boxes correspond to variable regions (V1–V4) in the B30.2 (SPRY) domain of the molecule. Locations of coding SNPs are indicated by vertical lines. Nonsynonymous (ns) SNPs with corresponding amino acid changes relative to the major allele are shown above and synonymous SNPs are indicated below by asterisks. These alleles were defined by sequencing full-length cDNA and genomic DNA generated from either B-lymphoblastoid cell lines (B-LCL) or PBMCs of these monkeys.

Figure 2. Effect of TRIM5α variants on susceptibility of B-LCLs to SIVmac239 infection.

The TRIM5 alleles expressed by selected rhesus monkey B-LCLs were first characterized by cDNA sequencing. These B-LCLs were then evaluated for their relative susceptibility to SIVmac239 replication by incubation for 48 h with a VSV-G-pseudotyped SIVmac239-GFP construct. The percentage of cells expressing GFP was assessed by flow cytometric analysis in B-LCLs homozygous (A) or heterozygous (B) for TRIM5 alleles 1–11. The B-LCLs were then divided into three groups: “1–5/1–5” expressing only TRIM5 alleles 1–5; “1–5/6–11” expressing a TRIM5 allele from the 1–5 groups and another from the 6–11 groups, “6–11/6–11” expressing only TRIM5 alleles 6–11 (C). The sequences of amino acid residues 328–479 of the TRIM5α B30.2 (SPRY) domain encoded by alleles 1–5 are identical to each other and to those previously described as rhesus monkey TRIM5α (AY625001). In contrast, a 2-amino acid deletion (339–340) and 3 nsSNPs (A333S, P341Q and S422L) are present in the B30.2 (SPRY) domain of TRIM5α variants encoded by alleles 6–11. The comparison of the values from the 3 groups of B-LCL was analyzed using a non-parametric one-way ANOVA Kruskal-Wallis test with a Dunn's multiple comparison test.

Figure 3. Effects of TRIM5α variants on susceptibility of rhesus monkey B-LCLs to VSV-G-pseudotyped retrovirus constructs.

Rhesus monkey B-LCLs expressing selected TRIM5 alleles were divided into two groups, one expressing only TRIM5 alleles 1–5, and the other expressing only TRIM5 alleles 6–11. They were then evaluated for their relative susceptibility to retrovirus infection by incubation for 48 h with VSV-G-pseudotyped recombinant SIVsmE543-GFP (A), HIV-1-GFP (B), EIAV-GFP (C), FIV-GFP (D), N-MLV-GFP (E) or B-MLV-GFP (F) and assessed for % of cells expressing GFP by flow cytometric analysis. The comparison of the values from the 2 groups of B-LCL was analyzed using the Mann-Whitney test.

Figure 4. Antiretroviral activity of TRIM5α variants expressed in feline renal fibroblast (CRFK) cells.

Feline renal fibroblast (CRFK) cells stably expressing TRIM5α variants encoded by allele 1, 5, 7 or 11, or control cells transduced with the empty pLPCX vector were incubated with a VSV-G-pseudotyped recombinant virus construct for 48 h and then subjected to flow cytometric analysis. (A) To assess the effects of TRIM5α variants on the infection of primate immunodeficiency viruses (PIVs), SIVmac239-GFP, SIVsmE543-GFP, and HIV-1-GFP were tested. The results of a repeat experiment are shown. Six dilutions of the stock of SIVmac239-GFP, SIVsmE543-GFP, and HIV-1-GFP were used to infect the transduced cell lines. (B) CRFK cells expressing different TRIM5α variants or control cells transduced with the empty pLPCX vector were incubated with EIAV-GFP, FIV-GFP, N-MLV-GFP or B-MLV-GFP. Infected GFP-positive cells were assessed by flow cytometric analysis. Represent data from repeated experiments are shown. Six dilutions of EIAV-GFP, FIV-GFP, N-MLV-GFP, and B-MLV-GFP were used to achieve 100% transduction of the control cells with the empty pLPCX vector. (C) The expression of TRIM5α protein by the transduced cells was determined by Western blotting using an anti-HA antibody. Levels of β-actin were also assessed as a control for the loading of total protein.

Figure 5. Effect of TRIM5α variation on SIVmac239 capsid binding.

Serial dilutions of 293T cell lysates (input) containing TRIM5α-HA proteins encoded by allele 1 or allele 7 were incubated with equal amounts of SIVmac239 CA-NC complexes assembled in vitro. After incubation, the mixtures were pelleted through a sucrose cushion. The amounts of TRIM5α in both input and pellet were subjected to Western blotting and densitometry. The results of two separate experiments are shown. The Western blot (A) of experiment 1 was quantitatively analyzed by a densitometer with the results shown in (B). The last two samples of the allele 7 variant were beyond the linear range of the densitometry and, therefore, were not included in the final analysis in (B). Both pellet and input amounts are shown in arbitrary densitometric units.

Figure 6. Association of the expression of groups of TRIM5 alleles by rhesus monkeys with viral replication following SIVmac251 infection.

(A) Plasma virus RNA levels were assessed in a cohort of Indian-origin rhesus monkeys between days 1 and 178 after challenge. Area-under-the curve calculations for the plasma SIV RNA levels were assessed in these monkeys. The plasma SIV RNA levels were also assessed on days 14 and 70 following challenge, representing peak and set-point viral load, respectively. (B) The loss of total CD4+ T cells and central memory CD4+ T cells was monitored in the same cohorts of rhesus monkeys following SIVmac251 infection. Data are displayed two ways: dividing the monkeys into 3 cohorts as described in the legend to Fig. 2 or dividing the monkeys into 2 cohorts, one of animlas expressing only alleles 1–5 and the other of animals expressing at least 1 allele of the groups 6–11. (C) The survival of monkeys following SIVmac251 infection is shown in Kaplan-Meier (KM) curves. The P-value corresponds to the log-rank test of equality of the survival curves. The monkeys are divided into either 3 or 2 groups as described above. The comparison of the values from the groups of monkeys was analyzed using a non-parametric one-way ANOVA Kruskal-Wallis test with Dunn's multiple comparison test or the Mann-Whitney test.