Two Surface-Exposed Elements of the B30.2/SPRY Domain as Potency Determinants of N-Tropic Murine Leukemia Virus Restriction by Human TRIM5α

Primates express dominant restriction factors that block retrovirus infection soon after entry but prior to reverse transcription (1, 2, 4). Most early restriction in primates is mediated by TRIM5α (5, 6, 8, 10, 12). TRIM5α is a member of the tripartite-motif family of proteins and contains RING, B-box 2, and coiled-coil (RBCC) domains (11). TRIM5α also contains a C-terminal B30.2/SPRY domain, which is essential for antiviral activity and has been implicated in interaction with the targeted viral capsid (9, 13, 16). Species-specific differences in TRIM5α account for the patterns of restriction of retroviruses in Old World and New World primates. For example, TRIM5α_hu from humans potently restricts N-tropic murine leukemia virus (N-MLV), whereas TRIM5α_rh from rhesus monkeys exhibits much weaker activity against this virus (6, 7, 10, 18). On the other hand, TRIM5α_rh potently blocks human immunodeficiency virus type 1 (HIV-1), which is only weakly inhibited by TRIM5α_hu (15, 16). Four variable regions (v1 to v4) are found in the B30.2 domains of TRIM5α proteins from different primates (12, 14). Differences in the v1 regions of TRIM5α_hu and TRIM5α_rh account for the differences in anti-HIV-1 potency exhibited by these TRIM5α variants (9, 17, 19).

The TRIM5α region that dictates the potency of TRIM5α_hu for N-MLV restriction has not been defined. To this end, we generated several chimerae between the human and rhesus TRIM5α proteins (Fig. 1A) and examined the abilities of these chimerae to restrict N-MLV and B-tropic murine leukemia virus (B-MLV) infection in transduced MDTF cells. The expression of each construct was confirmed by Western blot analysis (Fig. 1B). TRIM5α_hu restricted N-MLV significantly more potently than TRIM5α_rh, even though the steady-state levels of TRIM5α_hu expression were lower than those of TRIM5α_rh (Fig. 1C). The R(H286-493) mutant, in which the TRIM5α_rh B30.2 domain is replaced with that of TRIM5α_hu, restricted N-MLV infection significantly more efficiently than TRIM5α_rh. Thus, the B30.2 domain of TRIM5α_hu contains the determinant(s) for potent N-MLV restriction.

To define further the region(s) of the B30.2 domain involved in N-MLV restriction, two reciprocal chimerae, R(H286-371) and H(R286-371), were expressed in MDTF cells. The cells were challenged with N- and B-MLV (Fig. 1C). N-MLV infection of the R(H286-371)-expressing cell line was reduced approximately two- to threefold compared with that of the TRIM5α_rh-expressing cell line. The block to N-MLV infection in MDTF cells expressing the H(R286-371) chimera was almost as great as that in the TRIM5α_hu-expressing cells. No difference in susceptibility to B-MLV infection was observed in any of the MDTF cell lines expressing TRIM5α variants (Fig. 1D). These findings indicate that potency determinants for N-MLV restriction reside within multiple regions of the TRIM5α B30.2 domain, consistent with previous studies (19).

The v1 region in the TRIM5α_rh B30.2 domain is a major determinant of anti-HIV-1 potency (9, 17, 19). To test whether any of the TRIM5α_hu B30.2 variable regions are involved in N-MLV restriction, we generated several chimerae in which a small segment of one TRIM5α_rh variable region was replaced by the corresponding sequence from TRIM5α_hu (Fig. 2A). We assayed for a gain in N-MLV restriction in transduced MDTF cells compared to wild-type TRIM5α_rh. The steady-state levels of expression of the TRIM5α variants are shown in Fig. 2B. The Rh(SYQ/PCK) and Rh(GSFA/SFSV) mutants exhibited only modest improvements in blocking N-MLV infection compared to TRIM5α_rh (Fig. 2C); apparently, these differences do not account for the potency of TRIM5α_hu. By contrast, the Rh(LFTFPSLT/RYQT-FV) and Rh(QYV/ECA) mutants restricted N-MLV infection as potently as TRIM5α_hu. B-MLV infection was unaffected by expression of the TRIM5α variants (Fig. 2D). We conclude that at least two potency determinants for N-MLV restriction reside within the TRIM5α_hu B30.2 domain, one within the v1 region between residues 335 and 340 and another in the N-terminal portion of v3.

The presence of a positively charged arginine residue at position 110 of the viral capsid renders N-MLV susceptible to TRIM5α_hu restriction (10). Alteration of this N-MLV capsid residue from an arginine to the corresponding residue in B-MLV, glutamic acid, creates NBNN-MLV, which partially escapes from TRIM5α_hu restriction (10). Conversely, substitution of arginine for the glutamic acid at residue 110 of the B-MLV capsid creates BNBB-MLV, which is susceptible to TRIM5α_hu restriction (10). The correlation between a positive charge at position 110 of the MLV capsid and restriction suggested the hypothesis that the negative charges within the TRIM5α v3 region (Fig. 2A) are involved in an electrostatic interaction with the viral capsid. To test this hypothesis, we generated several mutants with changes in residues 409 and 410 in the TRIM5α_rh v3 region and examined their abilities to restrict N-MLV, B-MLV, NBNN-MLV,and BNBB-MLV. Mutant TRIM5α constructs were expressed stably in MDTF cells (Fig. 3A). Introduction of a negatively charged glutamic acid at residue 409 of TRIM5α_rh created Rh(Q409E), which restricted N-MLV infection to a level comparable to that of TRIM5α_hu. However, alteration of residue 409 to a positively charged arginine residue, Rh(Q409R), completely abrogated restriction of N-MLV. Likewise, alteration of the adjacent TRIM5α_rh residue 410 from a glutamic acid to either a positively charged residue, Rh(E410R) or Rh(E410K), or an uncharged residue, Rh(E410A), also abrogated restriction of N-MLV. On the other hand, the Rh(E410D) mutant, which retains the negative charge at residue 410, restricted N-MLV as efficiently as TRIM5α_rh. None of the TRIM5α_rh mutants restricted B-MLV infection (Fig. 3C). These data suggest that acidic residues within the TRIM5α v3 region enable potent N-MLV restriction.

To investigate whether the N-MLV restriction observed in the mutant TRIM5α_rh-expressing cell lines is dependent upon the charge at position 110 of the viral capsid, we tested whether NBNN-MLV and BNBB-MLV are sensitive to restriction in the mutant TRIM5α_rh-expressing cell lines. NBNN-MLV was able to completely escape restriction by TRIM5α_rh and all TRIM5α_rh mutants; however, only a partial escape was observed in the TRIM5α_hu-expressing cell line (Fig. 3D). Conversely, potent restriction of BNBB-MLV was observed in the TRIM5α_hu- and Rh(Q409E)-expressing cell lines, whereas only a partial restriction was observed in the TRIM5α_rh- and Rh(E410D)-expressing cells (Fig. 3E). There was no restriction of BNBB-MLV in the cells expressing Rh(Q409R), Rh(E410R), Rh(E410K), Rh(E410A), or Rh(Q409R/E410R). Thus, there is a correlation between the presence of negative charges at residues 409 and 410 of TRIM5α_rh and restriction of an MLV containing a positively charged residue at position 110 of the viral capsid.

Our results indicate that two variable regions, v1 and v3, in the B30.2 domain of TRIM5α_hu contribute to the potent restriction of N-MLV infection. Recently, the B30.2/SPRY domain has been shown to assume a β-sandwich, lectin-like fold (5). By analogy with this structure, the major variable regions (v1 to v4) of the TRIM protein B30.2 domains (14) are predicted to be surface-exposed elements on the same face of the domain (Fig. 4A); the variable loops surround a putative ligand-binding site (5). Thus, the v1 and v3 regions of TRIM5α are well positioned to serve as determinants of interaction with the retroviral capsid.

The presence of negatively charged residues at positions 409 and 410 of the TRIM5α_rh B30.2 domain potentiates restriction of MLVs that have an arginine at position 110 in the capsid. Alteration of either of these acidic TRIM5α_rh residues to a positively charged arginine completely abrogated TRIM5α-mediated restriction of N-MLV. It is tempting to speculate that the N terminus of the TRIM5α v3 region interacts electrostatically with the N-MLV capsid, perhaps involving capsid residue 110 (Fig. 4B). Electrostatics may also contribute to Fv-1ⁿ restriction in mice; for example, MLVs containing arginine 110 in the capsid are insensitive to Fv-1ⁿ restriction unless the positively charged lysine at position 358 of Fv-1 is removed (3). Although alteration of both TRIM5α_rh residues 409 and 410 to basic residues abrogated N-MLV restriction, it did not confer the ability to restrict B-MLV. Additional studies should clarify how shape and bond formation govern the recognition of targeted retroviral capsids by TRIM5α proteins.