The Contribution of RING and B-box 2 Domains to Retroviral Restriction Mediated by Monkey TRIM5 (cid:1) *

TRIM5 (cid:1) is a cytoplasmic protein that mediates a post-entry block to infection by some retroviruses. TRIM5 (cid:1) contains a tripartite motif (TRIM), which includes RING, B-box 2, and coiled-coil domains, and a C-termi-nal B30.2 (SPRY) domain. We investigated the contribution of the RING and B-box 2 domains to the antiretroviral activity of rhesus monkey TRIM5 (cid:1) (TRIM5 (cid:1) rh ), which potently restricts infection by human immunodeficiency virus, type 1 (HIV-1) and simian immunodeficiency virus of African green monkeys (SIV agm ). Disrup- tion of the RING domain caused mislocalization of TRIM5 (cid:1) rh so that the cytoplasmic level of the protein was decreased compared with that of the wild-type protein. Nonetheless, partial ability to restrict HIV-1 and SIV agm was retained by the RING domain mutants. By contrast, although TRIM5 (cid:1) rh mutants with disrupted B-box 2 domains were efficiently expressed and cor-rectly localized to the cytoplasm, antiretroviral activity was absent. The B-box 2 mutants colocalized and associated with wild-type

Host cell factors influence the infectivity of retroviruses. Following entry into the host cell, uncoating of the viral core, reverse transcription, nuclear access, and integration of the viral DNA into the host genome must occur to establish a permanent infection (1,2). Early, post-entry restrictions to retrovirus infection can determine tropism at the species level. HIV-1 1 and SIV agm encounter a post-entry block in Old World monkeys, whereas SIV mac is blocked in most New World monkey cells (3)(4)(5)(6). These species-specific restrictions occur prior to or concurrent with reverse transcription; at most, low levels of early reverse transcripts are made in restricted cells (3,5,7,8). The viral determinant of susceptibility to these blocks is the capsid protein (7, 9 -13). The early restriction to HIV-1 and SIV is mediated by dominant host factors, the activity of which can be titrated by the introduction of virus-like particles containing proteolytically processed capsid proteins of the restricted viruses (7, 12, 14 -18). Thus, in the cells of specific monkey species, host restriction factors apparently interact, directly or indirectly, with the HIV-1 or SIV capsid and prevent its progression along the infectious pathway.
A genetic screen identified TRIM5␣ as a major restriction factor in monkey cells that acts on HIV-1 and, to a lesser extent, on SIV mac (19). The expression of rhesus monkey TRIM5␣ was shown to be sufficient to confer potent resistance to HIV-1 infection in otherwise susceptible cells. Moreover, TRIM5␣ was necessary for the maintenance of the block to the early phase of HIV-1 infection in Old World monkey cells, as demonstrated by interference with TRIM5␣ expression in these cells. Differences among the sequences of TRIM5␣ proteins of primate species account for the different abilities to restrict infection by particular retroviruses (6, 20 -23).
TRIM5 is a member of a family of proteins that contain a tripartite motif, hence the designation TRIM (24). The tripartite motif includes a RING domain, B-box 2 domain and coiledcoil (cc) domain; TRIM proteins have also been called RBCC proteins. TRIM proteins exhibit the propensity to assemble into cytoplasmic or nuclear bodies (24). Many cytoplasmic TRIM proteins contain a C-terminal B30.2 or SPRY domain. Differential splicing of the TRIM5 primary transcript gives rise to the expression of several isoforms of the protein product. The TRIM5␣ isoform is the largest product (ϳ493 amino acid residues) and contains the B30.2(SPRY) domain. The B30.2(SPRY) domain of rhesus monkey TRIM5␣ is essential for anti-HIV-1 activity (19). Moreover, the difference in the anti-HIV-1 potency of rhesus and human TRIM5␣ proteins is determined by B30.2(SPRY) sequences (25)(26)(27). An intact RING domain also contributes, either directly or indirectly, to the anti-HIV-1 activity of rhesus monkey TRIM5␣ (19). Here we investigate the contribution of the RING, B-box 2, and coiled-coil domains to the antiviral activity of rhesus monkey TRIM5␣.

EXPERIMENTAL PROCEDURES
Plasmid Construction-The plasmids expressing the wild-type and mutant TRIM5␣ rh proteins were constructed using PCR-directed mutagenesis. The TRIM5␣ rh cDNA was PCR-amplified (19) and digested with EcoRI and ClaI, whose sites were introduced in each of the PCR primers. These fragments were cloned into the EcoRI and ClaI site of pLPCX (Stratagene). The plasmid expressing TRIM5␣ rh -V5 was constructed by inserting a PCR-amplified TRIM5␣ rh cDNA into the Viral Power plasmid using directional TOPO® cloning (Invitrogen).
Creation of Cells Stably Expressing TRIM5␣ Variants-Retroviral vectors encoding wild-type or mutant TRIM5␣ rh proteins were created using the pLPCX vector (19). The pLPCX vectors contain only the amino acid-coding sequence of the TRIM5␣ cDNA. The TRIM5␣ rh proteins encoded by the pLPCX vectors possess C-terminal epitope tags derived from influenza hemagglutinin (HA). Recombinant viruses were produced in 293T cells by cotransfecting the pLPCX plasmids with the pVPack-GP and pVPack-VSV-G packaging plasmids (Stratagene). The pVPack-VSV-G plasmid encodes the vesicular stomatitis virus (VSV) G envelope glycoprotein, which allows efficient entry into a wide range of vertebrate cells.
TRIM5␣ rh -V5 was made using the Viral Power system (Invitrogen). The TRIM5␣ rh -V5 protein possesses a V5 C-terminal epitope tag. Recombinant lentiviruses were produced according to the manufacturer's protocol (Invitrogen). The resulting virus particles were used to transduce ϳ1 ϫ 10 6 HeLa cells in the presence of 5 g/ml Polybrene. Cells were selected in either 1 g/ml puromycin (Sigma) or 1 g/ml puromycin and 2 g/ml blasticidin (Invitrogen).
Infection with Viruses Expressing GFP-Recombinant HIV-1, SIV mac , and SIV agm expressing GFP were prepared as described previously (6,19). HIV-1 viral stocks were quantified by measuring reverse transcriptase activity. For infections, 3 ϫ 10 4 HeLa cells seeded in 24-well plates were incubated in the presence of virus for 24 h. Cells were washed and returned to culture for 48 h and then subjected to FACS analysis with a FACScan (BD Biosciences). Protein Analysis-Cellular proteins were extracted with radioimmune precipitation assay (RIPA) buffer (10 mM Tris, pH 7.4, 100 mM NaCl, 1% sodium deoxycholate, 0.1% SDS, 1% Nonidet P-40, 2 mg/ml aprotinin, 2 mg/ml leupeptin, 1 mg/ml pepstatin A, 100 mg/ml phenylmethylsulfonyl fluoride). The cell lysates were analyzed by SDS-PAGE (10% acrylamide), followed by blotting onto nitrocellulose membranes (Amersham Biosciences). Detection of protein by Western blotting utilized monoclonal antibodies that are specifically reactive with the HA (Roche Applied Science) or V5 (Invitrogen) epitope tags, and monoclonal antibodies to ␤-actin (Sigma). Detection of proteins was performed by enhanced chemiluminescence (PerkinElmer Life Sciences), using the following secondary antibodies obtained from Amersham Biosciences: anti-mouse (for V5 and ␤-actin) and anti-rat (for HA).
Colocalization Experiments-Colocalization was studied as previously described (28). Briefly, cells were grown overnight on 12-mmdiameter coverslips and fixed in 3.9% paraformaldehyde (Sigma) in phosphate-buffered saline (PBS, Cellgro) for 30 min. Cells were washed in PBS, incubated in 0.1 M glycine (Sigma) for 10 min, washed in PBS, and permeabilized with 0.05% saponin (Sigma) for 30 min. Samples were blocked with 10% donkey serum (Dako, Carpentaria, CA) for 30 min, and incubated for 1 h with antibodies. The anti-HA fluorescein isothiocyanate-conjugated 3F10 antibody (Roche Applied Science) and anti-V5 Cy3-conjugated antibody (Sigma) were used to stain HA-or V5-tagged TRIM5␣ proteins, respectively. Subsequently, samples were mounted for fluorescence microscopy by using the ProLong Antifade Kit (Molecular Probes, Eugene, OR). Images were obtained with a Bio-Rad Radiance 2000 laser scanning confocal microscope with Nikon 60ϫ numerical aperture 1.4 optics.
Immunoprecipitations-At 48 h after transfection, 293T cells were removed from the plate and washed with PBS. 293T cells from nearly confluent 100-mm plates were lysed in 500 l of RIPA buffer. Insoluble material was pelleted at 22,000 ϫ g for 2 h, and the supernatant was used for immunoprecipitation. Equal amounts of protein (ϳ200 -500 g, as determined by the Bio-Rad assay) were incubated with 30 l of protein A-Sepharose (50%) and 5 l of anti-HA antibody (Roche Applied Science) for 2 h at 4°C. The immunoprecipitates were then washed three times with RIPA buffer and twice with PBS. After the final supernatant was removed, 30 l of 2ϫ sample buffer (120 mM Tris-HCl, pH 6.8, 20% glycerol, 4% SDS, and 0.02% bromphenol blue) was added, and the precipitate was boiled for 5 min to release the precipitated proteins. After centrifugation, the resulting supernatant was analyzed by SDS-PAGE and Western blotting.

Role of the RING and B-box 2 Domains in Retrovirus
Restriction-TRIM5␣ rh contains a tripartite motif that includes a RING (residues 15-59) and B-box 2 (residues 97-129) domains (19, 24, 29 -31). To investigate the role of these domains in retrovirus restriction, we generated rhesus monkey TRIM5␣ mutants with alterations affecting either one or both of these domains (Fig. 1A). The TRIM5␣ rh -HA ⌬93 protein lacks the RING domain, and the TRIM5␣ rh -HA ⌬132 protein lacks the RING and B-box 2 domains. We also altered specific cysteines and histidines within these domains; these residues are critical for zinc binding and are thought to contribute to the proper folding of these domains (29). In the C15A/C18A mutant, two key cysteines within the RING domain of TRIM5␣ rh -HA were altered to alanines. In the C97A/H100A mutant, cysteine and histidine residues conserved among B-box 2 domains were changed to alanines. The wild-type and mutant TRIM5␣ rh -HA proteins were expressed stably in HeLa cells (Fig. 1B). The TRIM5␣ rh -HA ⌬93, C15A/C18A, and C97A/H100A mutants were expressed at levels comparable to that of the wild-type TRIM5␣ rh -HA protein. The TRIM5␣ rh -HA ⌬132 protein was expressed at higher levels than that of the wild-type protein.
The HeLa cells expressing the wild-type and mutant TRIM5␣ rh -HA proteins were incubated with recombinant retroviruses (HIV-1-GFP, SIV mac -GFP, and SIV agm -GFP) pseudotyped with the vesicular stomatitis virus G glycoprotein and expressing GFP. The wild-type TRIM5␣ rh -HA protein potently restricted HIV-1-GFP infection ( Fig. 2A). Cells expressing the C15A/C18A or ⌬93 mutants were less susceptible to HIV-1-GFP infection than the control cells transduced the empty pLPCX vector, although the level of restriction observed was significantly less potent than that observed in the cells expressing wild-type TRIM5␣ rh -HA. HIV-1-GFP infection of cells expressing the TRIM5␣ rh -HA ⌬132 and C97A/H100A proteins was slightly more efficient than that of cells transduced with the empty pLPCX vector. We conclude that disruption of the RING domain partially attenuates the HIV-1-restricting activity of TRIM5␣ rh , whereas disruption of the B-box 2 domain completely eliminates HIV-1 restriction.
One explanation for the partial anti-HIV-1 activity of the TRIM5␣ rh -HA RING domain mutants is functional complementation by the endogenous human TRIM5␣ in the HeLa cells. To investigate this possibility, the TRIM5␣ rh -HA C15A/ C18A mutant and wild-type TRIM5␣ rh -HA were expressed in Cf2Th canine thymocytes. Dogs do not have a TRIM5 gene (32). The Cf2Th cells were incubated with HIV-1-GFP. The wildtype TRIM5␣ rh -HA protein potently blocked HIV-1 infection, whereas the C15A/C18A mutant partially restricted HIV-1 (Fig. 3). The degree and pattern of HIV-1 restriction observed for the wild-type and RING domain mutant TRIM5␣ rh proteins in canine cells were comparable to those observed in human cells. Therefore, the partial HIV-1-restricting activity seen for RING domain mutants is not dependent upon complementation by endogenous TRIM5␣. It is possible, however, that another TRIM protein in canine cells complemented the activity of the RING domain-deleted TRIM5␣ rh mutant.
The rhesus monkey TRIM5␣ protein potently blocks infection by SIV agm (13,15). The activities of the mutant TRIM5␣ rh -HA proteins in restricting SIV agm -GFP infection were similar to those observed for HIV-1-GFP infection (Fig.  2B). The TRIM5␣ rh -HA ⌬93 and C15A/C18A mutants slightly inhibited SIV agm -GFP infection, compared with the level of infection observed in control cells transduced with the pLPCX vector. By contrast, cells expressing the ⌬132 and C97A/H100A mutants were infected more efficiently than the control cells transduced with the empty vector. As was observed for restriction of HIV-1, restriction of SIV agm by rhesus monkey TRIM5␣ apparently requires the B-box 2 domain and is potentiated by the RING domain. The moderate inhibition of SIV mac infection previously observed for the rhesus monkey TRIM5␣ protein  (19) was abrogated by disruption of either the RING or B-box 2 domain (Fig. 2C).
Subcellular Localization of TRIM5␣ rh Mutants-The subcellular localization of the wild-type and mutant TRIM5␣ rh -HA proteins was examined by staining cells with an antibody directed against the HA epitope tag. As has been described for several TRIM proteins (24), TRIM5␣ rh -HA formed cytoplasmic bodies and also exhibited diffuse cytoplasmic staining (Fig. 4B). Compared with the wild-type TRIM5␣ rh -HA protein, the TRIM5␣ rh -HA ⌬93 and C15A/C18A mutants exhibited less intense cytoplasmic staining. The bodies formed by TRIM5␣ rh -HA ⌬93 were larger than those of the wild-type TRIM5␣ rh protein, and were found in the nucleus as well as the cytoplasm (Fig. 4C). Cells expressing TRIM5␣ rh -HA C15A/ C18A exhibited bodies that were larger and more numerous than those formed by the wild-type protein; the bodies formed by the C15A/C18A mutant were located in both the nucleus and cytoplasm (Fig. 4D). Thus, disruption of the RING domain causes significant changes in the subcellular localization of TRIM5␣ rh ; notably, decreases in the levels of diffuse cytoplasmic staining were observed for these mutants compared with that of the wild-type TRIM5␣ rh protein.
The TRIM5␣ rh -HA ⌬132 mutant lacking the RING and Bbox 2 domains exhibited diffuse nuclear and cytoplasmic staining; the intensity of staining was consistent with the high steady-state level of expression observed for this protein (Fig.  4E). The TRIM5␣ rh -HA C97A/H100A mutant exhibited a diffuse pattern of cytoplasmic staining with occasional small speckles evident (Fig. 4F). Thus, disruption of the B-box 2 domain, either alone or in combination with the RING domain, alters the pattern of TRIM5␣ rh localization in cells; nonetheless, the cytoplasmic levels of the B-box 2 mutants were equal to or greater than that of the wild-type TRIM5␣ rh protein.
Dominant-negative Activity of TRIM5␣ Mutants-We wished to investigate whether the RING or B-box 2 mutants might exhibit a dominant-negative phenotype with respect to the antiviral function of wild-type TRIM5␣ rh . To this end, we transduced the HeLa cell lines expressing the wild-type or mutant TRIM5␣ rh -HA proteins with a lentivirus vector expressing a wild-type TRIM5␣ rh protein with a V5 epitope tag at the C terminus (TRIM5␣ rh -V5). The expression level of TRIM5␣ rh -V5 varied in the transduced cell lines (Fig. 5), possibly reflecting the effect of the initially expressed TRIM5␣ rh -HA proteins on transduction efficiency or on the ability of the cells to tolerate TRIM5␣ rh -V5 expression.
To determine whether the expression of the mutant TRIM5␣ rh -HA proteins might affect TRIM5␣ rh -V5 function, the cell lines were incubated with different amounts of HIV-1-GFP, SIV agm -GFP and SIV mac -GFP. For all three virus infections, the percentage of GFP-positive cells in the cell lines initially containing the pLPCX vector and transduced with the vector expressing wild-type TRIM5␣ rh -V5 was lower than that in control cells expressing only the vector (Fig. 6, A-C). For all three viruses, the percentages of GFP-positive cells were comparable for the cells initially containing the LPCX vector and the cells initially containing the RING domain TRIM5␣ rh -HA mutants, both of which were subsequently transduced with the vector expressing TRIM5␣ rh -V5. These results indicate that the RING domain mutants exert minimal dominant-negative effects on the ability of TRIM5␣ rh -V5 to restrict HIV-1, SIV agm , or SIV mac infection. By contrast, infection by all three viruses was more efficient in the cells initially expressing the TRIM5␣ rh -HA ⌬132 and C97A/H100A mutants and subsequently transduced with the vector expressing TRIM5␣ rh -V5, compared with the infection of cells initially containing the empty pLPCX vector and transduced with the vector expressing TRIM5␣ rh -V5. In the case of SIV agm and SIV mac infections, the percentages of GFP-positive cells in the ⌬132and C97A/H100A-expressing cells transduced with TRIM5␣ rh -V5 were similar to those seen in the control cells containing only the empty pLPCX vector. We conclude that the TRIM5␣ rh -HA ⌬132 and C97A/H100A mutants can inhibit the antiretroviral activity of the wild-type TRIM5␣ rh -V5 protein in a dominant-negative manner.
We examined the subcellular localization of the wild-type and mutant TRIM5␣ rh -HA proteins in the cells expressing the wild-type TRIM5␣ rh -V5 protein by confocal microscopy (Fig. 7). The HA and V5 epitope-tagged proteins were detected by direct immunofluorescence, using anti-HA-fluorescein isothiocyanate and anti-V5-Cy3, respectively. Thus, the TRIM5␣ rh -HA proteins exhibit green fluorescence (left panels, Fig. 7) and the TRIM5␣ rh -V5 proteins exhibit red fluorescence (middle panels, Fig. 7). The right panels represent overlays of the green and red fluorescence from the HA-tagged and V5-tagged TRIM5␣ rh proteins, respectively. The TRIM5␣ rh -V5 expressed in the cells with the empty pLPCX vector exhibited diffuse cytoplasmic staining as well as small, scattered cytoplasmic bodies (Fig.  7A), as seen previously for the TRIM5␣ rh -HA protein (Fig. 4B). As expected, the wild-type TRIM5␣ rh -HA and TRIM5␣ rh -V5 proteins colocalized in the cytoplasm (Fig. 7B). In cells coexpressing TRIM5␣ rh -HA ⌬93 and TRIM5␣ rh -V5, both proteins exhibited diffuse cytoplasmic staining (Fig. 7C). Both proteins also colocalized in nuclear bodies, indicating that the coexpression of the TRIM5␣ rh -HA ⌬93 mutant causes relocalization of some of the wild-type TRIM5␣ rh V5 protein to these nuclear structures. Minimal colocalization was observed for the TRIM5␣ rh -HA C15A/C18A mutant and the TRIM5␣ rh -V5 pro-tein (Fig. 7D). In cells coexpressing TRIM5␣ rh -HA ⌬132 and TRIM5␣ rh -V5, the former protein exhibited diffuse nuclear and cytoplasmic staining (Fig. 7E). The wild-type TRIM5␣ rh -V5 protein in these cells did not form cytoplasmic bodies, but maintained a diffuse cytoplasmic pattern of staining that colocalized with the cytoplasmic portion of the TRIM5␣ rh -HA ⌬132 protein. The TRIM5␣ rh -HA C97A/H100A and wild-type TRIM5␣ rh -V5 proteins colocalized in a diffuse cytoplasmic pattern of staining (Fig. 7F). As was seen in cells expressing the TRIM5␣ rh -HA ⌬132 protein, the wild-type TRIM5␣ rh -V5 protein did not form cytoplasmic bodies in the cells coexpressing TRIM5␣ rh -HA C97A/H100A. FIG. 7. Changes in wild-type TRIM5␣ rh localization in the presence of TRIM5␣ rh mutants. HeLa cells stably expressing the wildtype TRIM5␣ rh -V5 protein and wild-type or mutant TRIM5␣ rh -HA proteins were fixed and stained using a fluorescein isothiocyanateconjugated anti-HA antibody (green) and a Cy3-conjugated anti-V5 antibody (red), respectively. Confocal microscopic images are shown of HeLa cells expressing the wild-type TRIM5␣ rh -V5 protein and one of the following: the empty pLPCX vector (A), wild-type TRIM5␣ rh -HA (B), TRIM5␣ rh -HA ⌬93 (C), TRIM5␣ rh -HA C15A/C18A (D), TRIM5␣ rh -HA ⌬132 (E), or TRIM5␣ rh -HA C97A/H100A (F). The overlay of green and red fluorescence is shown in the right panels.
The above studies suggest that the TRIM5␣ rh -HA mutants with disrupted B-box 2 domains exhibit dominant-negative phenotypes. To examine this dominant-negative activity in a primary monkey cell that naturally expresses TRIM5␣ rh , primary rhesus lung (PRL) fibroblasts were transduced with the empty pLPCX vector or with vectors expressing the TRIM5␣ rh -HA ⌬93 or TRIM5␣ rh -HA ⌬132 proteins. PRL cells have been previously shown to restrict HIV-1 infection potently; furthermore, this restriction can be relieved by TRIM5directed siRNA or by the expression of the TRIM5␥ isoform (19). The expression of the TRIM5␣-HA ⌬132 protein in the PRL cells significantly increased the susceptibility of these cells to HIV-1 infection (Fig. 8). By contrast, expression of the TRIM5␣ rh -HA ⌬93 protein resulted in only a minor increase in the level of HIV-1 infection, compared with control PRL cells transduced with the empty pLPCX vector (Fig. 8). Thus, the TRIM5␣ rh -HA ⌬132 protein exhibits dominant-negative activity in interfering with the HIV-1-restricting function of the wild-type rhesus monkey TRIM5␣ protein.
Hetero-oligomerization of Mutant and Wild-type TRIM5␣ Proteins-TRIM5 has been reported to form homo-oligomers (24). To determine whether the TRIM5␣ rh -HA mutants studied above can associate with the wild-type TRIM5␣ rh -V5 protein, the TRIM5␣ rh -HA protein variants were precipitated and coimmunoprecipitation of the wild-type TRIM5␣ rh -V5 protein was examined. In addition to the TRIM5␣ rh -HA mutants discussed above, we studied TRIM5␣ rh -HA ⌬297, which consists of the TRIM5␣ rh B30.2(SPRY) domain alone. Of note, the TRIM5␣ rh -HA ⌬297 mutant lacks the coiled-coil domain, which has been implicated in the self-association of some TRIM proteins (24,(33)(34)(35). To obtain comparable levels of expression of the wild-type and mutant TRIM5␣ rh proteins, 293T cells were cotransfected with plasmids encoding TRIM5␣ rh -V5 and either wild-type or mutant TRIM5␣ rh -HA proteins. Forty-eight hours afterward, the cells were lysed. One portion of the cell lysate was Western blotted with antibodies direct against the HA or V5 epitope tags or against ␤-actin. Fig. 9A shows that all of the TRIM5␣ rh -HA variants and the wild-type TRIM5␣ rh -V5 protein were efficiently expressed in the 293T cells. The remaining portion of the cell lysates was incubated with an anti-HA antibody and Protein A-Sepharose beads. The precipitated proteins were analyzed by SDS-PAGE followed by Western blotting with an anti-V5 antibody. Fig. 9B shows that the wild-type TRIM5␣ rh -HA protein as well as the RING and B-box 2 mutant TRIM5␣ rh -HA proteins associate efficiently with the wildtype TRIM5␣ rh -V5 protein. By contrast, the TRIM5␣ rh -HA ⌬297 mutant containing only the B30.2(SPRY) domain did not associate detectably with the wild-type TRIM5␣ rh -V5 protein.
These results are consistent with a role for the coiled-coil domain in self-association of TRIM5␣ rh .

DISCUSSION
The tripartite motif components of the RING and B-box 2 domains were investigated for their contribution to the antiretroviral function of TRIM5␣ rh . Disruption or deletion of the RING domain resulted in partial activity against HIV-1 and SIV agm , even in cell types that do not express an endogenous TRIM5␣ protein. In other proteins, RING domains contribute to a variety of functions, including protein-protein interactions, translational repression, and ubiquitin ligation (24,29,30,35). The basis for the partial loss of antiretroviral activity of TRIM5␣ rh associated with RING domain disruption is unknown. Because TRIM5␣ is known to act between the introduction of the viral capsid into the cytoplasm and reverse transcription (19), cytoplasmic levels of TRIM5␣ are likely important for antiretroviral function. It is probable that at least part of the decrease in HIV-1-suppressive activity of the TRIM5␣ rh RING domain mutants is due to mislocalization of the mutant proteins in the cell. Levels of diffusely staining, FIG. 9. Association of coexpressed wild-type and mutant TRIM5␣ rh proteins. 293T cells were cotransfected with a plasmid encoding wild-type TRIM5␣ rh -V5 protein and a plasmid encoding either wild-type or mutant TRIM5␣ rh -HA proteins. The 293T cells were lysed at 48 h after transfection. A, cell lysates were Western blotted with antibodies against the HA tag (top), the V5 tag (middle), or ␤-actin (bottom). B, cell lysates were used for immunoprecipitation by the anti-HA antibody. The precipitates were Western blotted with an anti-V5 antibody. cytoplasmic TRIM5␣ rh proteins were decreased for the RINGdisrupted or -deleted mutants, compared with those observed for the wild-type TRIM5␣ rh protein. The RING mutants instead accumulate in large nuclear or perinuclear bodies. Thus, deletion or disruption of the RING domain decreases the amount of TRIM5␣ rh available in the cytoplasm to mediate restriction of the incoming retroviral capsid. Whether the RING domain contributes in ways besides subcellular localization to TRIM5␣ rh antiviral function remains to be investigated. The RING domain may also be important for other, as yet undetermined, functions of TRIM5␣ or the other TRIM5 isoforms that lack antiretroviral activity.
The phenotypes of the TRIM5␣ rh mutants in which the B-box 2 domain is disrupted distinctly differ from those of the RING domain mutants. Deletion of the RING and B-box 2 domains, or disruption of only the B-box 2 domain, completely eliminated the antiretroviral activity of TRIM5␣ rh . These mutants were expressed efficiently, at least as well as wild-type TRIM5␣ rh , and localized in the cytoplasm. Apparently, the B-box 2 plays an important role in retroviral restriction mediated by TRIM5␣ rh , although further studies are required to elucidate this role. In promyelocytic protein (TRIM19), disruption of either of the two B-boxes interferes with nuclear body formation without affecting homo-oligomerization. In contrast, the structural integrity of the B-box contributes to the self-association of the ret finger protein TRIM27. The B-box 2 mutants associate with the wild-type TRIM5␣ rh protein, suggesting that these may retain the ability to oligomerize. The B-box 2 potentially contributes to the TRIM5␣ rh interaction with the retroviral capsid or host cofactors for restriction. B-box domains have been reported to be important for the association of TRIM1, TRIM18, and TRIM20 with elements of cell signaling pathways (33,34).
Our observation that the TRIM5␣ rh ⌬132 but not the ⌬297 mutant associates with wild-type TRIM5␣ rh implies that the coiled-coil domain makes a major contribution to TRIM5␣ rh self-association. Coiled-coil domains have been implicated in the ability of other TRIM proteins to form homo-oligomers (24,(33)(34)(35).
Our results suggest that B-box 2 mutants can interfere with the antiretroviral activity of the wild-type TRIM5␣ rh protein in a dominant-negative manner. The mild increase in the susceptibility of HeLa cells expressing the B-box 2 mutants to HIV-1 and SIV agm infection may reflect dominant-negative inhibition of the modest antiviral activity of the endogenous human TRIM5␣ in these cells (19). The TRIM5␥ rh isoform, which lacks the B30.2(SPRY) domain and exhibits no antiretroviral activity, has been shown to exhibit dominant-negative activity (19). TRIM5␥ rh can associate with TRIM5␣ rh when the two proteins are coexpressed in cells (data not shown). The ability of TRIM5 variants to associate with and to interfere with the antiretroviral activity of wild-type TRIM5␣ may be related. Of interest, all of the TRIM5 variants that retain these two properties contain a coiled-coil domain.
The levels of expression of TRIM5␣ rh ⌬132 and C97A/H100A mutants were greater than or equal to those of the wild-type TRIM5␣ rh in the cells where restriction by the wild-type protein was relieved. These levels of expression likely contribute to the ability of the dominant-negative mutants to compete with the wild-type TRIM5␣ rh protein for factors, binding to which is necessary for antiviral function. Candidates for such factors include the viral capsid, TRIM5␣ itself, and cellular cofactors. Future studies will be directed toward a more complete understanding of the importance of specific elements of the tripartite motif of TRIM5␣ in retrovirus restriction.