Distal Leucines Are Key Functional Determinants of Alix-Binding Simian Immunodeficiency Virus SIV_smE543 and SIV_mac239 Type 3 L Domains

Abstract

In addition to PTAP L domains, primate lentiviruses carry Alix-binding motifs that include the recently described type 3 SREKPYKEVTEDLLHLNSLF sequence. We examined the requirements for the type 3 sequence motif in simian immunodeficiency virus SIV_smE543 and identified the ₄₉₉LNSLF₅₀₃ sequence as a key functional determinant. Mutation of distal leucines ₄₉₉L and ₅₀₂L (LL mutant) caused an inhibitory effect on Alix-dependent SIV_smE543 release that was quantitatively similar to that observed following disruption of the type 3 L domain or RNA interference (RNAi) depletion of Alix. Similar results were obtained with the SIV_mac239 LL mutant. Thus, distal leucines are key determinants of SIV_smE543 and SIV_mac239 type 3 L domains.

TEXT

Retroviruses acquire their envelopes at the host cell membrane of infected cells. To this end, they utilize short sequences designated L (late) domains to recruit members of the endosomal sorting complex required for transport (ESCRT) to catalyze membrane modeling events that lead to virus release (1, 2, 5, 11, 16). Three types of L domains have so far been identified: the PT/SAP, PPXY, and LYPX_nL motifs. They bind the host proteins Tsg101, Nedd4-like ubiquitin ligase family members, and Alix, respectively (6, 13, 15, 23, 25, 26). These interactions lead to the recruitment of members of the ESCRT pathway and require the activity of the AAA ATPase VPS4 (14, 23, 24, 26).

All retroviral Gag proteins contain at least one L domain, though most carry multiple L domain motifs, believed to function synergistically and ensure efficient viral release (4, 7, 8, 13, 21, 27). Structural studies have been used to determine L domain sequence and functional requirements. Only limited sequence variability was noted in motifs that bind Tsg101, including PTAP, PSAP, and PXXP motifs found in Hrs, the natural partner for Tsg101 (3, 10, 18–20, 28). However, more sequence divergence was observed in L domains binding Alix, with the identification of three sequences: YPDL in equine infectious anemia virus (EIAV) (23, 26, 29), LYPLASLRSLF in human immunodeficiency virus type 1 (HIV-1), and the recently described SREKPYKEVTEDLLHLNSLF sequence in the simian immunodeficiency virus SIV_mac239 (30). They have been dubbed types 1, 2, and 3, respectively. The last type contains an anchoring tyrosine, followed by one or two hydrophobic residues (valine and/or leucine) (proximal leucines), and both residues are key functional determinants for Alix (30). We observed that the Alix-binding type 3 L domain also contains a downstream LXXLF sequence carrying additional leucines (distal leucines) that are present in SIV_mac239, SIV_agmTan-1, and SIV_smE543. Similarly, the HIV-1 p6 domain also harbors an LXXLF motif in its LYPLASLRSLF L domain (underlined), which is critical for Alix binding and function (23, 26). In this study, we examined the role of distal leucine residues in the LNSLF sequence within the Alix-binding type 3 L domains of both SIV_smE543 (9) and SIV_mac239 (12).

The ₄₉₉LNSLF₅₀₃ sequence is required for PTAP-independent SIV_smE543 release.

L domain function has recently been assessed in simian retroviruses and has led to the discovery of a new Alix-binding L domain in SIV_mac239 (30). We examined the L domain function in SIV_smE543 (9). The Gag protein of this virus contains a total of two PTAP L domains located in the SIV_smE543 MA and p6 domains, respectively. Substitution of LTAA for the PTAP motif in p6 (_smL2 mutant) eliminated ∼50% of virus release from 293T (Fig. 1A, lane 3) and CEM cells (data not shown), whereas substitutions in MA had little effect on release (_smL1 mutant), indicating that the p6 motif plays a dominant role in virus release. SIV_smE543 devoid of both motifs (_smL1L2 mutant) retained ∼50% of wild-type (WT) virus release (Fig. 1A, lane 4), suggesting that it harbors an additional L domain sequence(s). We searched for an Alix-binding L domain in the C-terminal region of the SIV_smE543 p6 domain and found a ₄₈₈PYKEVTEDLLHLNSLF₅₀₃L sequence, which is similar to the recently described Alix-binding type 3 L domain in SIV_mac239 (Fig. 1B). To investigate the function of this motif, we replaced distal leucines ₄₉₉L, ₅₀₂L, and ₅₀₃F in the ₄₉₉LNSLF₅₀₃ motif with prolines and a serine (_smLL mutant) (Fig. 1B) and examined the effect of such changes. As expected, an _smLL mutation in the context of WT SIV_smE543 had only a minimal affect on virus release (Fig. 1C, lane 2). Conversely, when the LL mutation was combined with L1 and L2 mutations (_smL1L2/LL), virus release was obliterated (Fig. 1C, lanes 6 and 8). The effect of these mutations was comparable to that seen following the substitution of serines for the ₄₈₉Y and ₄₉₆L residues in the type 3 L domain (Fig. 1D, lanes 6 and 8), thus underscoring the importance of distal leucines in virus release. Similar results were obtained when distal leucines were mutated in the SIV_mac239 LNSLF motif (_macL2/LL and _macL1L2/LL mutants) (Fig. 1E). These results, which were observed in three independent experiments (see quantification panels under the blots), indicate that distal leucines are an active part of the type 3 L domain and critical for Alix-dependent SIV_smE543 and SIV_mac239 release.

Fig. 1.

Distal leucines in the type 3 L domain are key participants in the release of SIV_smE543 and SIV_mac239. (A) PTAP L domain motifs drive SIV_smE543 release from 293T. (Left) Schematic representation of SIV_smE543 Gag showing the positions of PTAP motifs. (Middle) 293T cells were transfected with WT SIV_smE543 proviral DNA (lane 1), the MA-PTAP motif mutant (_smL1) (lane 2), the p6-PTAP mutant (_smL2) (lane 3), or the double mutant (_smL1L2) (lane 4). (Right) Quantification of the release of L domain mutants. (B) Sequence comparison of Alix-binding type 3 L domains in SIV_smE543, SIV_mac239, and SIV_agmTAN-1. The three lentiviruses carry LXXLF motifs (boldface) downstream of their defined type 3 L domains (anchorage tyrosine in boldface). Dots indicate absent residues. Positions of YL and LL mutants are indicated. (C) The distal leucine _smLL mutant is as defective as the _smYL mutant in Alix-mediated release. 293T cells were transfected with WT SIV_smE543 proviral DNA (lane 1), the _smLL mutant (lane 2), the MA-PTAP _smL1 mutant (lane 3), the p6-PTAP _smL2 mutant (lane 5), or the indicated double mutant. (D) Cells were transfected with the YL mutant (lane 2) or the indicated single or double mutants. (E) Analysis of the SIV_{mac239 mac}L1, _macL2, and _macLL mutants. Cells were transfected with the SIV_mac239 PTAP motif single mutants _macL1, _macL2, the LXXLF motif single mutant _macLL (lanes 2, 3, and 5), or the indicated double mutants. Cells and viruses were collected 24 h posttransfection, and their protein content was analyzed by SDS-PAGE and Western blotting using an anti-SIV serum. Virus release efficiencies (values in percentages) were quantified as the ratio of virion-associated Gag and cellular Gag from three independent experiments and are shown in the panels under the blots.

Leucines in the LNSLF motif are critical for SIV_smE543 and SIV_mac239 Gag interactions with Alix.

Next, we assessed whether the ₄₉₉L and LF₅₀₃ residues are involved in Gag interactions with Alix. Cell lysates from 293T cells expressing SIV_{smE543 sm}L1, _smL2, and _smLL mutants were tested for their ability to interact in immunoprecipitation assays with Flag-tagged Alix in the presence of 1% NP-40 detergent. Similar to observations with WT SIV_smE543 Gag, the _smL1 and _smL2 mutants were captured by Flag-Alix proteins (Fig. 2A, lanes 2, 4, and 6). In contrast, SIV_smE543 and SIV_mac239 LL mutants (_smLL and _macLL) were not captured (Fig. 2A, lane 8, and C, lane 4). Likewise, substitutions of alanines for ₄₈₉Y and ₄₉₆L residues within SIV_smE543 and SIV_mac239 (_smYL and _macYL mutants) eliminated binding to Alix (Fig. 2B, lane 4) (30). These results indicate that the Y and L residues are as critical as the distal leucines for the SIV_smE543 and SIV_mac239 type 3 L domain binding to Alix.

Fig. 2.

Distal leucine residues in the Alix-binding type 3 L domain are required for SIV_smE543 and SIV_mac239 Gag-Alix interactions. 293T cells were transfected with either WT SIV_smE543 (A and B) or WT SIV_mac (C) proviral DNA, or with the _smL1, _smL2, or _smLL mutants (panel A, lanes 3 to 8), the _smYL mutant (panel B, lanes 3 and 4), or the _macLL mutant (panel C, lanes 3 and 4) in the presence or absence of Flag-tagged Alix. Cells were lysed in radioimmunoprecipitation assay (RIPA) buffer, and lysates were incubated with anti-Flag antibody-conjugated beads. Both input and immunoprecipitated complexes were analyzed by SDS-PAGE and Western blotting using the indicated antibodies.

Alix enhances the release of SIV_smE543 and SIV_mac239 in an LXXLF-dependent manner.

To examine whether distal leucines ₄₉₉L and ₅₀₂LF₅₀₃ are involved in Alix-mediated virus release, we used a virus rescue assay to test the effect of Flag-tagged Alix overexpression (22) on the release of a SIV_smE543 mutant that lacked the PTAP motif activity. Alix overexpression enhanced the release of SIV_smE543 lacking either the p6-located dominant PTAP motif (_smL2 mutant) or both PTAP motifs in Gag (_smL1L2 mutant) (Fig. 3A, lanes 2 and 6). Virus release augmentation was inhibited by the mutation of ₄₉₉L and ₅₀₂LF₅₀₃ residues in the ₄₉₉LNSLF₅₀₃ motif (_smL2/LL and _smL1L2/LL mutants) (Fig. 3A, lanes 4 and 8) in a manner similar to that seen following the substitution of serines for ₄₈₉Y and ₄₉₆L (Fig. 3B). A similar result was obtained when the LNSLF motif was disrupted in SIV_mac239 (Fig. 3C). These results were confirmed in three independent experiments, as shown in the release quantification panel in Fig. 3A. Together, they indicate that SIV_smE543 and SIV_mac239 are less responsive to Alix overexpression when distal leucines ₄₉₉L and ₅₀₂LF₅₀₃ are mutated.

Fig. 3.

Alix enhances the release of SIV_smE543 and SIV_mac239 in an LNSLF-dependent manner. 293T cells were transfected with either SIV_smE543 carrying disrupted PTAP motifs (_smL2 or _smL1L2 mutants) or SIV_smE543 carrying mutations in the PTAP and LXXLF motifs (_smL2/LL or _smL1L2/LL mutant) (A), SIV_smE543 carrying mutations in both the p6-bound PTAP and the Alix-binding YL motifs (_smL2YL mutant) (B), or SIV_mac239 carrying mutations in the PTAP and LXXLF motifs (_macL2 and _macL2/LL mutants) (C) in the presence or absence of Flag-Alix. Cells and viruses were collected 24 h posttransfection, and their protein content was analyzed by SDS-PAGE and Western blotting using an anti-SIV serum. Expression of Alix was detected using an anti-Flag antibody, and cellular tubulin levels were analyzed using a mouse monoclonal antitubulin antibody. Alix-mediated virus enhancement was quantified from three independent experiments, and the results are summarized in panel A (under the Western blot in panel A).

LNSLF-dependent SIV_smE543 release requires cellular Alix.

When both PTAP L domains were disrupted in SIV_smE543 (_smL1L2 mutant), ∼50% of virus release was retained, as long as the Gag possessed an intact ₄₈₈PYKEVTEDLLHLNSLF motif (Fig. 1 and 3). These findings suggest that the SIV_smE543 Gag utilizes the type 3 L domain sequence to recruit the host cell Alix and promote virus release. To examine whether distal leucines in the ₄₉₉LNSLF motif are involved in the utilization of the host Alix, we knocked down Alix using RNAi and examined its effect on the WT or mutant viruses carrying either one (_smL1 and _smL2 mutants) or no (_smL1L2 mutant) PTAP motif. As expected, RNAi depletion of cellular Alix had only a modest effect on the release of WT or L1 mutant viruses because of the second PTAP L domain dominance in the p6 region of Gag (Fig. 4A, lanes 3 and 6, and the release quantification below the blots). Conversely, depletion of cellular Alix decreased the release of SIV_smE543 L2 and L1L2 mutants to nearly undetectable levels (Fig. 4B, lanes 2 and 5), demonstrating that SIV_smE543 LNSLF is critical for the utilization of the host cell Alix during virus release.

Fig. 4.

(A and B) LNSLF-dependent SIV_smE543 release requires cellular Alix. 293T cells were transfected with control RNAi (lanes 1 and 4), RNAi directed against Tsg101 (lanes 3 and 6), or Alix (lanes 2 and 5) at 24-h intervals. At the second transfection, cells were cotransfected with _smWT SIV_smE543 or the _smL1, _smL2, or _smL1L2 L domain mutants. Cells and viruses were collected 24 h posttransfection, and their protein content was analyzed by SDS-PAGE and Western blotting using an anti-SIV serum. Cellular Alix and Tsg101 were detected using a rabbit anti-Alix and mouse anti-Tsg101 antibody, respectively. Alix-mediated virus release rescue was quantified from three independent experiments, and the results are summarized in the panels below the blots.

Conclusions.

Three types of Alix-binding L domains have been identified: YPDL, LYPLASLRSLF and SREKPYKEVTEDLLHLLNSLF motifs (key residues underlined) found within EIAV, HIV-1, and SIV_mac239, respectively (23, 26, 30). These motifs share key binding and functional determinants. A tyrosine considered an “anchorage” of Alix (29, 30) is followed by hydrophobic residues, often leucines, which make contact with the Alix V domain (30). Here we report that two leucines and a phenylalanine residue in the SIV_{smE543 488}PYKEVTEDLLHLNSLF₅₀₃ L domain (italics) play a key role in Alix function (Fig. 1 to 3), since substitutions at these residues resulted in loss of both binding and function. We conclude that in addition to the tyrosine and proximal hydrophobic residues in the type 3 L domain, additional relatively distant residues (distal leucines) are critical functional determinants of Alix. Interestingly, both leucines are found within an LXXLF sequence that is part of the type 3 L domains of SIV_smE543, SIV_mac239, and SIV_agmTAN-1 (30). An LXXLF motif is also found in HIV-1 and is critical for Alix recruitment and Alix-dependent release (23, 26). These observations confirm the conservation of the LXXLF sequence across primate lentiviruses, underscoring its importance in Alix-mediated virus release.

The Alix-binding type 3 L domain shares key structural and functional determinants with types 1 and 2, including the tyrosine (Y) and the hydrophobic proximal leucines that make contact within the V domain in order to stabilize binding with Alix (29, 30). Our data demonstrate that both functional determinants are required, but not sufficient, for the type 3 L domain binding to Alix, suggesting that further stabilization is necessary for recruitment and function. Consistent with this hypothesis, we found that leucines located 10 and 13 residues downstream of the tyrosine (distal leucines) in the PYKEVTEDLLHLNSLF sequence, are required for Alix binding to SIV_smE543 and SIV_mac239 Gag proteins (Fig. 2). Similarly, leucines in the HIV-1 LYPLASLRSLF L domain (underlined) are also critical for Alix binding (17, 23, 26, 29). In contrast, distant leucines appear to be dispensable for the EIAV-borne type 1 L domain. One possible explanation for these discrepant requirements is the ability of the latter L domain to bind Alix with a high affinity (29). Together, these observations suggest that distal leucines in type 2 and 3 L domains are involved in further stabilization of Alix-p6 interactions, possibly to compensate for their low-affinity binding to p6.

Virus release was completely eliminated when mutation of distal leucines was combined with the disruption of both PTAP L domains in SIV_smE543. This suggested that the entire PYKEVTEDLLHLNSLF sequence functions as an additional L domain in p6 (Fig. 1). Such a notion is further supported by the reliance of the type 3 L domain on cellular Alix to drive virus release in the absence of PTAP motifs (Fig. 4). In summary, our data identify distal leucine residues in the SIV_smE543 and SIV_mac239 type 3 Alix-binding L domains as key functional determinants.