Human immunodeficiency virus type 1 (HIV-1) is infectious only for humans and chimpanzees. This is due in part to TRIM5α, which blocks infection early after viral entry, before the establishment of a provirus in Old World monkey (OWM) cells. Rhesus monkey and cynomolgus monkey (CM) TRIM5αs inhibit HIV-1 but not simian immunodeficiency virus isolated from macaque (SIVmac), while African green monkey (AGM) TRIM5α prevents replication of HIV-1 and SIVmac (14
). Human (Hu) TRIM5α shows very weak antiviral activity against those viruses (30
) but strong resistance against N-tropic murine leukemia virus (N-MLV) (14
). Among several splicing variants of TRIM5, an α isoform carries the SPRY or B30.2 domain that determines virus specificity of this intracellular factor (26
HIV-2, simian immunodeficiency virus isolated from sooty mangabey (SIVsmm), and SIVmac have extremely similar genomes (11
). SIVmac was resistant to the restriction by rhesus monkey and CM TRIM5αs (26
), while HIV-2 isolates GH123 and ROD were shown to be sensitive to those TRIM5αs (26
). Since HIV-2 isolates varied considerably in their abilities to grow in OWM cells (3
), we studied the effects of CM and Hu TRIM5αs on eight different HIV-2 isolates (UC1, UC2, UC7, UC12, UC14, 9429, 12741, and GH123) (1
We used Sendai virus (SeV) expressing CM TRIM5α (CM-TRIM5α-SeV) (26
) or Hu TRIM5α (Hu-TRIM5α-SeV). We also generated SeV expressing CM TRIM5α without the SPRY domain [CM-SPRY(−)-SeV] as a negative control. There was no variation in TRIM5α expression levels among T-cell-line Hut78 cells infected with those SeVs (data not shown).
Consistent with previous observations (26
), HIV-2 isolate GH123 grew to lower titers in CM-TRIM5α-SeV-infected cells and to slightly but significantly lower titers in Hu-TRIM5α-SeV-infected cells than they did in CM-SPRY(−)-SeV-infected cells (Fig. ). Of the seven newly tested isolates, only UC12 showed a pattern similar to that of GH123 (Fig. ), while the other six isolates, UC1, UC14, UC2, UC7, 9429, and 12741, grew to almost the same titers in CM-TRIM5α-SeV-, Hu-TRIM5α-SeV-, and CM-SPRY(−)-SeV-infected cells (Fig. ). These results indicated that HIV-2 isolates GH123 and UC12 are sensitive to CM TRIM5α and moderately sensitive to Hu TRIM5α, while the isolates UC1, UC14, UC2, UC7, 9429, and 12741 are resistant to CM and Hu TRIM5αs.
FIG. 1. Hut78 cells (105) were infected with CM-TRIM5α-SeV (open circles), Hu-TRIM5α-SeV (filled triangles), or CM-SPRY(−)-SeV (open squares) at a multiplicity of infection of 10 PFU. The CM-SPRY(−)-SeV domain was generated by (more ...)
It has been suggested that viral capsid protein (CA) was the determinant of TRIM5α restriction (13
). We therefore determined the nucleotide sequences of PCR-amplified DNA fragments corresponding to proviral DNA encoding the CAs of the HIV-2 isolates. Figure shows the deduced amino acid sequences of the HIV-2 CAs. TRIM5α-sensitive GH123 and UC12 had proline at the 120th (GH123) or corresponding 119th (UC12) position, while the other six TRIM5α-resistant isolates showed alanine (UC1 and UC14) or glutamine (UC2, UC7, 9429, and 12741) at the corresponding 119th position. Except for this substitution, there was no substitution observed commonly and specifically for GH123 and UC12.
FIG. 2. (A) Alignments of amino acid sequences of CAs of eight HIV-2 isolates. “/S” after the names of isolates represents sensitivity to CM and Hu TRIM5α, and “/R” indicates resistance to CM and Hu TRIM5α. A dash (more ...)
To determine whether the single amino acid at the 119th or 120th position of the HIV-2 CA affects the virus's susceptibility to CM and Hu TRIM5α, we constructed two mutant GH123 viruses carrying either alanine (GH123/A) or glutamine (GH123/Q) at the 120th position (Fig. ). As shown in Fig. , all three viruses grew substantially and to similar titers in CM-SPRY(−)-SeV-infected MT4 cells. In CM-TRIM5α-SeV-infected cells, both GH123/A and GH123/Q grew to titers more than 250 times higher than those of the wild-type GH123. In the Hu-TRIM5α-SeV-infected cells, both mutant viruses grew to titers approximately fourfold higher than those of the wild type (Fig. ). In contrast, replication of the three viruses could not be detected in the AGM-TRIM5α-SeV-infected cells (Fig. ). Unlike CM TRIM5α, AGM TRIM5α possessed a potent antiviral activity against SIVmac as well as HIV-1 and HIV-2 (26
). These results indicate that the single amino acid at the 120th position of the GH123 CA indeed affects susceptibility to the restriction of virus replication by CM TRIM5α.
FIG. 3. (A) Mutant GH123/A and GH123/Q viruses generated by changing a single amino acid of proline to alanine or glutamine and mutant SIVmac239/P virus generated by changing a single amino acid of glutamine to proline by site-directed mutagenesis. Dashes indicate (more ...)
There is more than 87% amino acid identity in CA between HIV-2 GH123 and SIVmac239. SIVmac239 is resistant to CM TRIM5α (26
) and contains glutamine at the 118th position, which corresponds to the 120th position of the GH123 CA. To determine whether this particular amino acid also affects the resistance of SIVmac239 to CM TRIM5α, we constructed a mutant SIVmav239 carrying a proline at the 118th position (SIVmac239/P) (Fig. ). As shown in Fig. , both wild-type and mutant SIVmac239 grew substantially and to similar titers in the CM-SPRY(−)-SeV-infected MT4 cells. In CM-TRIM5α-SeV-infected cells, the mutant SIVmac239 grew to titers that were approximately 10 times lower than those of the wild type. The glutamine-to-proline mutation also caused an approximately eightfold decrease in the virus titer in Hu-TRIM5α-SeV-infected cells (Fig. ). In contrast, neither virus grew in AGM-TRIM5α-SeV-infected cells (Fig. ). These results indicate that the glutamine at the 118th position of the SIVmac239 CA also affects resistance to the CM and Hu TRIM5αs.
Nineteen HIV-2 and 20 SIVsmm or SIVmac CA sequences were listed in the Los Alamos sequence database (Fig. ). All SIV isolates except for SMM-SL92B carried glutamine at the position corresponding to the 120th position of GH123 (Fig. ). SMM-SL92B, which carries alanine, was most distantly related to all the other HIV-2 and SIV isolates from the phylogenetic analysis (data not shown). These results suggest that all SIV isolates are resistant to CM and Hu TRIM5αs. In contrast, HIV-2 group A isolates showed a mixture of proline, alanine, and glutamine and HIV-2 group B isolates had a mixture of proline and alanine in the CA (Fig. ). Other HIV-2 isolates also carried proline and glutamine, with H2AB-7312A, which had glycine, being the only exception (Fig. ). These results suggest that different HIV-2 isolates have diverse susceptibilities to CM and Hu TRIM5αs. It is likely that glutamine-to-alanine or glutamine-to-proline substitutions occurred after the proposed zoonotic transfer of virus from monkeys to humans (11
To obtain structural insight into the mechanisms by which this amino acid change in the CA alters viral susceptibility to restriction by TRIM5α, three-dimensional (3-D) models of HIV-2 CAs were constructed with the homology-modeling technique based on the crystal structure of the HIV-1 CA N-terminal domain (15
). These models consist of those with proline (GH123, UC12), alanine (UC1, UC14, GH123/A), and glutamine (UC2, UC7, 9429, 12741, GH123/Q) at the 120th position. The thermodynamically optimized 3-D structures showed that the HIV-2 CA N-terminal domains consist of a packed core structure from which seven α-helices and three loops protrude, which is basically the same conformation as that of their HIV-1 counterparts (15
) (Fig. ).
FIG. 4. Structural model of the N-terminal halves of HIV-2 CAs. The 3-D models of 10 HIV-2 CAs were constructed with the homology-modeling technique using “MOE-Align” and “MOE-Homology” in the Molecular Operating Environment (MOE) (more ...)
The 120th amino acid that affects the viral susceptibility to TRIM5α restriction is located in the loop between helices 6 and 7 (L6/7) (Fig. , labeled Q, A, P). Especially worth noting is that the amino acid substitution at the 120th position is predicted to induce marked changes in the configuration of L6/7. The loop with the TRIM5α-sensitive proline (GH123 and UC12, Fig. , labeled P) is positioned most closely to the loop between helices 4 and 5 (L4/5). We obtained results the same as those described above when we constructed HIV-2 CA models based on the HIV-1 CA structure in solution (data not shown) (8
). In HIV-1, L4/5 interacts directly with cyclophilin A (15
). It is possible that TRIM5α recognizes the particular structure formed by two closely aligned L4/5 and L6/7 with proline.
Previously, a single amino acid substitution at the 110th position of N-MLV CA was shown to determine viral susceptibility to Fv1 (20
), another type of restriction factor in mice (2
), as well as to Hu TRIM5α (14
). The fact that the amino acids at analogous positions of both N-MLV and HIV-2 CAs (25
) affected sensitivity to Hu and CM TRIM5αs, respectively, suggested that N-MLV and HIV-2 are recognized by TRIM5α in a similar manner. It would be interesting to investigate whether the 120th position of HIV-2 CA affects its sensitivity to TRIM5α of other OWM cells.
Human TRIM5α has no or very little effect on HIV-1 or SIVmac infection (30
). Unlike HIV-1 (Fig. ), HIV-2 isolates sensitive to CM TRIM5α were slightly more sensitive to Hu TRIM5α than those resistant to CM TRIM5α (Fig. and ). As shown in Fig. , half of the HIV-2 isolates in the database carried proline, which would be sensitive to Hu TRIM5α. This finding may be one of the reasons why HIV-2 is less pathogenic than HIV-1 (9
). Since certain HIV-2 patients with high plasma HIV-2 loads developed AIDS as rapidly as HIV-1 patients (10
), examining the effect of sensitivity to Hu TRIM5α of HIV-2 strains in infected individuals on the rate of disease progression merits attention.