Regulation of APOBEC3 Proteins by a Novel YXXL Motif in Human Immunodeficiency Virus Type 1 Vif and Simian Immunodeficiency Virus SIVagm Vif

Abstract

The APOBEC3 cytidine deaminases are potent antiviral factors that restrict the replication of human immunodeficiency virus type 1 (HIV-1). In HIV-1-infected CD4⁺ T cells, the viral accessory protein Vif binds to APOBEC3G (A3G), APOBEC3F (A3F), and APOBEC3C (A3C) and targets these proteins for polyubiquitination by forming an E3 ubiquitin ligase with cullin 5. Previous studies identified regions of HIV-1 Vif, ₄₀YRHHY₄₄ and ₁₂QVDRMR₁₇, which are important for interaction with A3G and A3F, respectively, and showed that Vif residues 54 to 71 are sufficient for A3G binding. Here, we identify ₆₉YXXL₇₂ as a novel conserved motif in HIV-1 Vif that mediates binding to human A3G and its subsequent degradation. Studies on other APOBEC3 proteins revealed that Tyr69 and Leu72 are important for the degradation of A3F and A3C as well. Similar to A3F, A3C regulation is also mediated by Vif residues ₁₂QVDRMR₁₇. Simian immunodeficiency virus (SIV) Vif was shown to bind and degrade African green monkey A3G (agmA3G) and, unexpectedly, human A3C. The YXXL motif of SIVagm Vif was important for the inactivation of agmA3G and human A3C. Unlike HIV-1 Vif, however, SIVagm Vif does not require Tyr40 and His43 for agmA3G degradation. Tyr69 in the YXXL motif was critical for binding of recombinant glutathione S-transferase-Vif(1-94) to A3G in vitro. These results suggest that the YXXL motif in Vif is a potential target for small-molecule inhibitors to block Vif interaction with A3G, A3F, and A3C, and thereby protect cells against HIV-1 infection.

Members of the human cytidine deaminase apolipoprotein B mRNA-editing catalytic polypeptide-like 3 (APOBEC3) family are potent inhibitors of human immunodeficiency virus type 1 (HIV-1) and a wide range of other viruses and endogenous retroelements (8, 9, 13, 22). The HIV-1 virion infectivity factor (Vif) is a 23-kDa viral accessory protein that counteracts the anti-HIV activity of APOBEC3G (A3G) and APOBEC3F (A3F). In the absence of Vif, A3G and A3F are packaged into HIV-1 virions and deaminate cytidines in viral minus-strand DNA during reverse transcription, resulting in G-to-A hypermutation and premature degradation of newly synthesized viral DNA (13, 22, 30, 39, 58). A3G and A3F are associated with high-molecular-weight ribonucleoprotein complexes in the cytoplasm, possibly localizing to P bodies and stress granules (5, 10, 19, 50), and may also inhibit viral replication via deamination-independent mechanisms that may include inhibition of reverse transcription and proviral DNA formation (1, 12, 15, 21, 31, 41). Vif overcomes the antiviral activity of A3G and A3F predominantly by forming an E3 ubiquitin ligase with cullin 5 (Cul5), elongin B (EloB), and elongin C (EloC) that targets these proteins for degradation by the ubiquitin-proteasome pathway (6, 20, 25-27, 40, 46, 56, 57). Vif may also inhibit APOBEC3 activity through mechanisms independent of proteasomal degradation (17, 18, 27, 36, 46). Vif associates with the Cul5-EloB-EloC complex by binding directly to EloC via a BC box motif at positions 144 to 153 and to Cul5 via hydrophobic residues at positions 120, 123, and 124 within a zinc-binding region (residues 100 to 142) formed by a conserved H-X₅-C-X_17-18C-X_3-5-H (HCCH) motif and a recently identified Vif cullin box (26, 28, 33, 45). Vif binding to A3G and A3F is essential for their degradation by the Vif-Cul5 E3 ligase (25, 27, 29). Posttranslational modifications regulate these events, since Vif-mediated degradation of A3G is regulated by phosphorylation of Vif at Ser144 and of A3G at Thr32 (26, 42).

The ability of Vif to block the antiviral activity of A3G is species specific. HIV-1 Vif binds to and inactivates human A3G (hA3G) and hA3F, but not APOBEC3 proteins derived from African green monkeys (AGM) and rhesus macaques (2, 23, 24, 37, 51). Conversely, AGM simian immunodeficiency virus (SIVagm) Vif inactivates AGM and rhesus macaque A3Gs but not hA3G. A single amino acid difference in A3G, aspartic acid at position 128 in hA3G versus lysine in AGM A3G (agmA3G), controls species specificity by influencing Vif-A3G binding (2, 23, 37, 51). The N-terminal region of HIV-1 Vif is important for the binding and neutralization of hA3G and hA3F and contributes to species-specific recognition (25, 29, 35, 38, 43, 48, 49), but the specific residues that mediate these interactions have not been defined.

Previous studies demonstrated that HIV-1 Vif residues ₄₀YRHHY₄₄ and ₁₂QVDRMR₁₇ are important for interaction with hA3G and hA3F, respectively (29, 35), and residues 54 to 71 (29) or 52 to 72 (14) are sufficient for hA3G binding. Charged residues in this region (i.e., Arg56, Asp61 and Arg63) are largely dispensable for hA3G degradation (29). To investigate the role of other highly conserved residues in this region (Fig. 1A) in Vif-A3G interaction and A3G degradation, we used site-directed mutagenesis. First, we tested the ability of Vif V55A, I57A, P58A, L64A, I66A, Y69A, and W70A mutants to induce degradation of hA3G. Wild-type or mutant HIV-1 Vif proteins were coexpressed with hA3G-Myc in 293T cells, and the ability of these proteins to induce hA3G degradation was determined by Western blot analysis (Fig. 1B). The Y69A mutation abolished Vif-mediated degradation of hA3G, similar to the effect of ΔSLQ and H108,139N mutations, which prevent Vif binding to EloC and Cul5, respectively (27, 28) (Fig. 1B). The W70A mutation had a minor inhibitory effect on the ability of Vif to induce degradation of hA3G, whereas V55A, I57A, L64A, and I66A mutations had no discernible effect. The Vif P58A mutant was not stably expressed in transfected cells and, therefore, was not further analyzed. The Vif Y69F mutant exhibited an increased capacity to degrade hA3G compared to that of the wild-type Vif (Fig. 1B). This finding suggests a possible requirement for a bulky aromatic residue at position 69 and implies that potential hydrogen bonding mediated by the hydroxyl group of Tyr69 is not required for hA3G degradation. Tyr69 is unlikely to be modified by phosphorylation based on previous work by Yang et al., which suggested that the tyrosine residues in Vif are not phosphorylated (53, 54).

FIG. 1.

A conserved YXXL motif in HIV-1 Vif is important for binding and neutralization of hA3G. (A) Sequence alignment of residues 35 to 73 of primate lentivirus Vif proteins. Identical and similar residues are highlighted in dark or light gray, respectively. Vif residues examined for their potential role in A3G neutralization are marked with an asterisk. (B) hA3G-Myc was expressed in 293T cells with wild-type and mutant Vif proteins. Expressed protein levels were determined by Western blot analysis with anti-Myc monoclonal, rabbit anti-Vif, or anti-β-tubulin antibodies. (C) 293T cells were cotransfected with hA3G-HA and wild-type or mutant HIV-1 Vif expression plasmids. Lysates were immunoprecipitated with anti-HA antibody and probed by Western blotting. Equivalent expression levels were confirmed by Western blot analysis of cell lysates. (D) Single-cycle viruses were produced from 293/A3G cells following transfection with a vif-deleted proviral plasmid (pNLX ΔEnv ΔVif-luc), vesicular stomatitis virus glycoprotein (VSVG), and wild-type or mutant pCDNA3.Vif expression plasmids. Infectivity of normalized virus was measured in Cf2 cells. Shown are the means of the results ± the standard deviation (n = 3). In the bottom panel, immunoblots show the expression level of wild-type and mutant HIV-1 Vif proteins in the corresponding producer cells. (E) The W70A Vif mutant induces A3G degradation via the proteasome pathway. Left panel: hA3G-Myc was expressed in 293T cells with wild-type or mutant Vif proteins, and transfected cells were incubated with or without 2.5 μM MG132 for the last 15 h of transfection. Expressed protein levels were determined by Western blot analysis. Right panel: 293T cells were cotransfected with hA3G-HA and wild-type or W70A mutant HIV-1 Vif expression plasmids, and cells were incubated with or without MG132, as in the left panel. Lysates were immunoprecipitated with anti-HA and probed by Western blotting of cell lysates. (F) Mutant HIV-1 Vif proteins bind to Cul5 and EloC. Wild-type or mutant HIV-1 Vif expression plasmids were cotransfected with pCDNA3.HA-Cul5 or pCDNA3.T7-EloC into 293T cells. Lysates were immunoprecipitated with antibodies recognizing epitope tags on the indicated proteins and were probed by Western blotting. Equivalent levels of protein expression were confirmed by Western blotting of cell lysates. WT, wild type; +, present; −, without.

Tyr69 and Leu72 in Vif localize to a highly conserved YXXL motif. In other retroviral proteins, YXXL motifs function to facilitate interactions with cellular proteins (7, 32, 34, 44). For example, the YPXL late domain in equine infectious anemia virus p9 recruits the cellular protein ALIX (also known as AIP1) to facilitate virus budding (4, 47). We investigated whether the highly conserved leucine at position 72 (Fig. 1A) is important for Vif-mediated degradation of hA3G. An L72A mutation resulted in a moderate reduction in Vif-mediated hA3G degradation (Fig. 1B). Coimmunoprecipitation (co-IP) with transfected hemagglutinin (HA)-tagged hA3G demonstrated a significant reduction in hA3G binding of Vif Y69A and Vif W70A, similar to that of the previously characterized Vif H42,43N mutant (29, 35), and a moderate reduction in hA3G binding of Vif L72A, compared with the wild-type Vif (Fig. 1C). To test the importance of these Vif residues for viral infectivity, we assessed the production of infectious virus particles in the presence of hA3G. Vif Y69A, similar to Vif H42,43N, was severely impaired for the production of infectious HIV-1, consistent with the defect in hA3G binding and degradation (Fig. 1D). Vif L72A exhibited a twofold decrease in virus infectivity. Vif W70A supported the production of infectious virus at levels similar to the wild type. Vif Y69F slightly increased (∼1.3-fold) the production of infectious HIV-1 (Fig. 1D), consistent with its slightly enhanced capacity to induce hA3G degradation.

The finding that the W70A mutation showed a modest reduction in A3G degradation and no effect on viral infectivity was unexpected, given the strong inhibitory effect on Vif-hA3G binding (Fig. 1C), but is consistent with the idea that Vif can inhibit APOBEC3 proteins through mechanisms independent of proteasomal degradation (17, 18, 27, 36, 46). To investigate this possibility, we evaluated the ability of the Vif Y69A, W70A, and L72A mutants to degrade hA3G in the presence or absence of the proteasome inhibitor MG132. hA3G levels in cells expressing no Vif or Vif Y69A mutant were similar when cells were cultured with or without MG132 (Fig. 1E, left panel). In contrast, the Vif L72A and W70A mutants reduced A3G levels by 30 and 50%, respectively, compared to the 60% reduction induced by the wild-type Vif, only in the absence of MG132. These results suggest that the Vif W70A mutant induces A3G degradation via the proteasome pathway and confirm our preceding observation that the W70A mutation causes only a modest reduction in A3G degradation. Additional co-IP experiments confirmed our preceding data (Fig. 1C) that interaction of the Vif W70A mutant with A3G is significantly reduced and showed these results were similar in the presence or absence of MG132 (Fig. 1E, right panel). Together, these results suggest that the W70A mutation exerts only a modest inhibitory effect on A3G binding and degradation in intact cells. The strong inhibitory effects of the W70A mutation on Vif binding to A3G in co-IP and glutathione S-transferase (GST) pull-down assays (Fig. 1C; see Fig. 5) are likely to be due to experimental conditions that reduce Vif-A3G binding efficiency compared to physiological conditions in intact cells. Nonetheless, our findings are consistent with the study by He et al. (14), which showed that the W70A mutant exhibits reduced binding to A3G in a cell-based assay and a modest reduction in the ability to induce A3G degradation. Their MAGI assay showed a 70% reduction in infectivity, whereas we observed no significant reduction in infectivity, which may reflect differences in experimental conditions.

FIG. 5.

Tyr69 and Trp70 residues in the YXXL motif in recombinant HIV-1 GST-Vif(1-94) are required for binding to hA3G in vitro. (A) A total of 2.5 μg of recombinant GST or wild-type or mutant GST-Vif(1-94) proteins were prebound to glutathione-Sepharose beads. Recombinant His-A3G (200 ng; Immunodiagnostics, Woburn, MA) was added to the prebound beads and incubated overnight at 4°C. Bound GST, GST-Vif, and His-A3G were detected by Western blotting with anti-GST and anti-A3G antibodies. (B) Peptide inhibition of binding between GST-Vif(1-94)-A3G in a homogenous FRET assay (Lance). 15 nM recombinant GST-Vif(1-94) was incubated with 500 nM biotinylated A3G peptide (amino acids 110 to 148) in a 384-well plate. Binding was detected using allophycocyanin-streptavidin and europium-anti-GST antibody and expressed as a FRET ratio [(fluorescence emission intensity at 665 nm/intensity at 615 nm) × 10⁴]. Peptide competition was carried out by adding 2% dimethyl sulfoxide (DMSO) or Vif 15-mer peptides P15 (amino acids 57 to 71) or mutant P15 (amino acids 57 to 71; Y69A/W70A) to Vif-A3G binding reactions at the indicated concentration. Data are presented as means of the results for triplicate samples ± the standard deviations.

To confirm that the defect in Vif Y69A and L72A binding and degradation of A3G was not due to a defect in protein folding, we tested the ability of these mutants to bind Cul5 and EloC. As expected, the Y69A mutant retained the ability to bind Cul5 and EloC (Fig. 1F). Similarly, the L72A mutant bound to Cul5 (Fig. 1F) and EloC at a level of efficiency similar to that of the wild-type Vif (data not shown), suggesting that these mutations do not induce significant conformational changes. Together, these findings demonstrate that the highly conserved Tyr69 and Leu72 residues in the YXXL motif are important for HIV-1 Vif-hA3G interaction and hA3G degradation.

Vif residues ₁₂QVDRMR₁₇ were previously shown to be essential for hA3F but not hA3G recognition (35). Therefore, we tested whether conserved amino acids within Vif residues 54 to 72 were important for hA3F regulation. The Y69A mutation abolished Vif-mediated degradation of hA3F, similar to the phenotype of the QV12,13AA mutant (35) used as a control, and reduced Vif interaction with hA3F in co-IP experiments (Fig. 2A and B). The W70A mutant had no significant effect on the ability of Vif to bind and degrade hA3F. The L72A mutation resulted in a modest reduction in hA3F degradation and a significant reduction in hA3F binding (Fig. 2A and 2B). Consistent with this finding, He et al. (14) showed that the L72S mutant has a reduced ability to exclude A3F incorporation into HIV-1 virions, which resulted in reduced production of infectious viruses by 80% compared to the 98% reduction in infectivity in the absence of hVif. These results suggest that Leu72 mutants have a significantly reduced capacity to bind and degrade A3F, which might explain why Vif-A3F binding was not detected under the assay conditions used for the co-IP experiment (Fig. 2B). Vif H42,43N showed a slightly increased capacity to bind hA3F (Fig. 2B). This finding is in agreement with the study by Russell and Pathak, which showed that a ₄₀YRHHY₄₄→AAAAA mutation in Vif increased the infectivity of HIV-1 virions in cells that express hA3F (35). Accordingly, our results support their hypothesis that the ₄₀YRHHY₄₄ motif may influence the function of the ₁₂QVDRMR₁₇ motif through steric hindrance; the H42,43N mutation may thus facilitate stronger interactions between hA3F and Vif. Single-round infectivity assays in 293T cells showed that the Y69A mutation resulted in a 60% reduction in the production of infectious single-round viruses in the presence of A3F (data not shown). Overall, these results demonstrate the involvement of a novel YXXL motif in HIV-1 Vif in the regulation of both hA3G and hA3F. These results are in agreement with a recent study by Yamashita et al., which identified critical amino acid residues (including Tyr69) in HIV-1 Vif for binding and exclusion of A3G/F (52), and a recent study by He et al., which demonstrated a requirement for the motif ₅₅VXIPLX_4-5 LXΦX₂YWXL₇₂ (and the importance of Tyr69 and Leu72) in HIV-1 Vif for hA3G/F degradation, hA3G virion exclusion, and suppression of hA3G antiviral activity (14). These studies demonstrated that residues ₇₄TGERXW₇₉ in HIV-1 Vif are selectively involved in hA3F regulation (14, 52).

FIG. 2.

Tyr69 and Leu72 of HIV-1 Vif are important for degradation of hA3F. (A) hA3F-V5 was expressed in 293T cells with wild-type or mutant HIV-1 Vif proteins. Expressed protein levels were determined by Western blot analysis with anti-V5 monoclonal, rabbit anti-Vif, or anti-β-tubulin antibodies. (B) HIV-1 Vif Tyr69 and Leu72 are important for interaction with hA3F. 293T cells were cotransfected with HA-hA3F and wild-type or mutant HIV-1 Vif plasmids. Lysates were immunoprecipitated with anti-HA antibody and probed by Western blotting. Equivalent levels of protein expression were confirmed by Western blotting of cell lysates. WT, wild type; +, with; −, without.

The preceding studies demonstrate that residues Tyr69 and Leu72 of HIV-1 Vif are important for hA3G and hA3F interaction (Fig. 1C and 2A) and degradation (Fig. 1B and 2B). These residues are highly conserved among primate lentiviruses (Fig. 1A). The YXXL motif as well as Tyr40, His43, and Tyr44 of the ₄₀YRHHY₄₄ motif is also present in SIVagm Vif (Fig. 3A). To test whether these residues are important for degradation of agmA3G, we made mutations at the tyrosine and leucine residues at positions 71 and 74 of SIVagm Vif. Y71A, L74V, and L74A mutations in SIVagm Vif abolished degradation of agmA3G (Fig. 3B), whereas the W72A mutation, analogous to the W70A change in HIV-1 Vif, had only a minor inhibitory effect. Surprisingly, unlike HIV-1 Vif, mutations of the conserved Tyr40 and His43 in SIVagm Vif had no effect on agmA3G degradation (Fig. 3B).

FIG. 3.

YXXL motif in SIVagm Vif is important for neutralization of agmA3G. (A) Sequence alignment of residues 35 to 73 of primate lentivirus Vif proteins from HIV-1 HXB2, SIVagm, and SIVmnd2. Identical and similar residues are highlighted in dark or light gray, respectively. SIVagm Vif residues examined for their ability to neutralize agmA3G are marked with asterisks. (B) Tyr71 and Leu74 residues in SIVagm Vif are important for agmA3G degradation. agmA3G-Myc was expressed in 293T cells with wild-type or mutant SIVagm Vif proteins. Expressed protein levels were determined by Western blotting. (C) Single-cycle viruses were produced from 293T cells following cotransfection with a proviral plasmid (pNLX ΔEnv ΔVif-luc) and plasmids that express VSVG, agmA3G, and wild-type or mutant SIVagmVif proteins. Infectivity of normalized virus was measured in Cf2 cells. Shown are the means of the results ± the standard deviation (n = 3). The bottom panel represents immunoblots that show expression levels of agmA3G and wild-type or mutant HIV-1 Vif proteins in the corresponding producer cells. WT, wild type; +, with; −, without.

To examine the biological relevance of these observations, we tested the ability of SIVagm Vif mutant proteins to support production of infectious virus in the presence of agmA3G. The SIVagm Vif Y71A and L74A mutants were both impaired for the production of infectious single-round viruses in the presence of agmA3G, consistent with their inability to induce agmA3G degradation (Fig. 3C). In contrast, Y40A, H43N, and W72A SIVagm Vif mutations resulted in minor reductions in the production of infectious single-round viruses in the presence of agmA3G. Together, these results demonstrate that the YXXL motif of SIVagm Vif is important for neutralization of agmA3G. In contrast to HIV-1 Vif, Vif Tyr40 and His43 in SIVagm Vif are not important for this activity.

The ₄₀YRHHY₄₄ and ₆₉YXXL₇₂ motifs in HIV-1 Vif are both important for counteracting hA3G (29, 35). We therefore examined whether there is evidence of cooperative interactions between these motifs. To address this question, we searched for evidence of covariation at specific residues within these two regions of Vif among primate lentiviruses. We identified leucines at positions 40 and 71 in SIV Vif sequences from wild Mandrillus sphinx (SIVmnd2) (Fig. 3A) in place of the tyrosines at these positions present in Vif proteins from other primate lentiviruses. We then tested the anti-agmA3G effect of an SIVagm Vif mutant that had tyrosines 40 and 71 replaced by leucines. The SIVagm Vif Y40L mutant had no effect on agmA3G degradation, while the Y71L mutant had a minor effect. The Y40,71L mutation inhibited agmA3G degradation to a similar extent as the Y71A mutant (Fig. 3B). We also examined the effect of SIVagm Vif Y40,71L mutations in single-round infection assays. The single mutation Y40L or Y71L in SIVagm Vif resulted in only minor decreases (20%) in the production of infectious single-round viruses in the presence of agmA3G, whereas a double mutation (Y40,71L) reduced the production of infectious virus by 40% (Fig. 3C). The ability of these Vif mutants to support viral infectivity correlates with agmA3G expression in the producer cells; A3G levels were 7%, 14%, and 35% in cells expressing SIVagm Vif Y40L, Y71L, and Y40,71L, respectively, compared to A3G levels in the absence of SIVagm Vif (100%). These results suggest that the Y40,71L double mutant exhibits a net additive effect in reducing A3G degradation, suggesting that the ₄₀YRHHY₄₄ and ₆₉YXXL₇₂ motifs both contribute to neutralizing A3G (Fig. 3C). Determining the exact nature of structural and/or functional interactions between these domains may lead to a greater understanding of Vif function.

HIV-1 Vif inactivates hA3G and hA3F, but not agmAPOBEC3 proteins. Conversely, SIVagm Vif inactivates AGM and rhesus macaque A3G but not human A3G (2, 23, 24, 37, 51). APOBEC3C (A3C) is another member of the APOBEC3 cytidine deaminase family that is expressed in lymphoid cells and has weak anti-HIV-1 activity compared to A3G and A3F (3, 16, 55). Unexpectedly, hA3C is a potent inhibitor of SIVagm that can be degraded by both HIV-1 and SIVagm Vif proteins (59). The determinants important for functional interaction of Vif with hA3C have not yet been fully characterized. To address this question, we first examined the levels of hA3C protein in 293T cells expressing HIV-1 Vif with point mutations in the ₁₂QVDRMR₁₇, ₄₀YRHHY₄₄ and ₆₉YXXL₇₂ motifs. QV12,13AA mutation resulted in only a minor reduction in HIV-1 Vif-mediated degradation of hA3C, whereas a ₁₄DRMR₁₇ to SEMQ mutation (38) abolished the degradation of hA3C (Fig. 4A). H42,43N mutation in the ₄₀YRHHY₄₄ motif had no effect on hA3C degradation, similar to the lack of effect on hA3F (29, 35). Y69A and L72A mutations in the ₆₉YXXL₇₂ motif abolished the capacity of Vif to induce degradation of hA3C, while W70A degraded hA3C as efficiently as did the wild-type Vif (Fig. 4A). L72I mutation in HIV-1 Vif had a minor effect on A3C degradation (Fig. 4A). In contrast, the corresponding L74V mutation in SIVagm Vif abolished agmA3G degradation (Fig. 3B), suggesting that leucine or isoleucine but not valine is tolerated at position 4 in the YXXL motif. Next, we examined hA3C protein levels in the presence of SIVagm Vif mutants. Mutations of Tyr40, His43, and Trp72 had no effect on hA3C degradation, while Y71A and L74A mutations abolished hA3C degradation. Together, these results suggest that the YXXL motif of HIV-1 and SIVagm Vif proteins is important for regulation of hA3C. The results also suggest that hA3C, like hA3F, is selectively regulated by the ₁₂QVDRMR₁₇ motif but not the ₄₀YRHHY₄₄ motif of HIV-1 Vif (29, 35).

FIG. 4.

Tyr and Leu residues of the conserved YXXL motif in HIV-1 and SIVagm Vifs are important for hA3C regulation. hA3C-V5 was expressed in 293T cells with wild-type or mutant HIV-1 Vif (A) or SIVagm Vif (B) proteins. Expressed protein levels were determined by Western blot analysis. WT, wild type; +, with; −, without.

To further examine the requirements for the YXXL motif in Vif binding to APOBEC3 proteins, we created a truncated GST-Vif (amino acids 1 to 94) that includes the binding sites for A3G and A3F, but not for Cul5, EloB, and EloC. We then used site-directed mutagenesis to create point mutations in the ₄₀YRHHY₄₄ and ₆₉YXXL₇₂ motifs of GST-Vif(1-94) and tested the ability of these mutants to bind recombinant His-A3G in a GST pull-down assay. The Y69A and W70A mutants bound only trace amounts of His-A3G, while the H42,43N mutant bound similar amounts of His-A3G as did the wild type. These results suggest that the YXXL motif in Vif is critical for direct binding of GST-Vif(1-94) to His-A3G (Fig. 5A). Both the in vitro binding assay (Fig. 5A) and cell-based co-IP assay (Fig. 1C) demonstrated that Tyr69 of the YXXL motif is important for hA3G binding. Given the modest effect of the W70A mutation in functional assays (Fig. 1B and D), the strong inhibitory effect of the W70A mutation on Vif binding to A3G in the GST pull-down assay (Fig. 5A) is likely to be due to experimental conditions that reduce Vif-A3G binding efficiency compared to physiological conditions in intact cells. The H42,43N mutation, which reduces Vif binding to A3G in co-IP assays, failed to interfere with Vif-A3G binding in vitro. This difference in requirement for the histidines at positions 42 and 43 in cell-based co-IP assays and direct in vitro binding assays is consistent with a recent paper which showed strong inhibition of Vif-A3G binding in vitro by a 15-mer Vif peptide (P15, derived from residues 57 to 71) compared to minor inhibition by a 15-mer peptide (P9, derived from residues 33 to 47) (29). Similarly, in another study, deletion of amino acids 23 to 43 in HIV-1 Vif did not affect interaction with hA3G (11). These findings, together with the results in Fig. 1C and D, suggest that these residues, along with Trp70, function differently in the physiological cellular environment than in the context of recombinant proteins produced in Escherichia coli, for reasons that remain unclear.

Next, we performed a homogenous time-resolved fluorescence resonance energy transfer (FRET) assay (Lance; Perkin-Elmer) using recombinant GST or GST-Vif(1-94) proteins (15 nM) and a biotinylated A3G peptide (residues 110 to 148) (500 nM) in a 384-well format as described previously (29). The A3G peptide sequence was based on modeling of the A3G sequence onto the Bacillus subtilis cytidine deaminase (ljkt) structure and is predicted to be composed of two α-helices flanking a loop containing the D128 residue that determines species-specific Vif-A3G binding (2, 23, 24, 37, 51). Briefly, GST-Vif and the biotinylated A3G peptide were incubated for 30 min at room temperature and then labeled with anti-GST-europium (2 nM) antibody and streptavidin-allophycocyanin (25 nM), respectively, and the FRET signal was measured as fluorescence emission intensity. GST-Vif(1-94) binding to the A3G peptide produced a significant increase in the FRET ratio over background levels measured in the presence of GST (Fig. 5B). To further demonstrate the specificity of binding in the FRET assay, we performed peptide competition assays using 15-mer peptides derived from regions of Vif that were previously shown to be important for A3G binding in vivo (29). GST-Vif and a 15-mer Vif peptide were incubated for 30 min and then underwent additional incubation with the biotinylated A3G peptide. Binding of Vif to the A3G peptide was efficiently competed by Vif peptide (P15, derived from residues 57 to 71), which includes Tyr69 and Trp70, whereas a mutant P15 peptide containing Y69A and W70A mutations failed to compete with Vif for A3G binding even at the highest concentration tested (Fig. 5B). These data confirm the specificity of the FRET binding assay and demonstrate that the in vitro binding assay recapitulates physiologically relevant interactions. These results also suggest that the YXXL motif in Vif binds A3G at the interface formed by residues 110 to 148 of A3G.

In summary, we identified YXXL as a novel and highly conserved motif in diverse Vif proteins that is important for the regulation of APOBEC3 proteins. The ₆₉YXXL₇₂ motif in HIV-1 Vif mediates interaction with hA3G and its subsequent degradation to support viral infection (Fig. 1). Tyr69 and Leu72 are important for degradation of hA3F and hA3C (Fig. 2A and 4A). Like hA3F degradation, hA3C degradation requires Vif residues ₁₂QVDRMR₁₇. SIV Vif was shown to bind and degrade agmA3G and hA3C (24, 55, 59). Unlike HIV-1 Vif, however, SIVagm Vif does not require Tyr40 and His43 for agmA3G degradation (Fig. 3). Nonetheless, we cannot exclude the possible involvement of other regions of SIVagm Vif in binding to agmA3G. Importantly, we have shown that Vif binding to A3G, A3F, and A3C is mediated via a common YXXL motif; a second motif confers specificity for recruiting specific APOBEC3 proteins (14, 29, 35). The YXXL motif is critical for direct binding of recombinant GST-Vif(1-94) to His-A3G in vitro and in a FRET assay (Fig. 5). These results identify the YXXL motif in Vif as a potential target for small molecule inhibitors to block Vif interaction with A3G, A3F, and A3C and thereby protect cells against HIV-1 infection.