APOBEC3F Determinants of HIV-1 Vif Sensitivity

Abstract

HIV-1 Vif counteracts restrictive APOBEC3 proteins by targeting them for proteasomal degradation. To determine the regions mediating sensitivity to Vif, we compared human APOBEC3F, which is HIV-1 Vif sensitive, with rhesus APOBEC3F, which is HIV-1 Vif resistant. Rhesus-human APOBEC3F chimeras and amino acid substitution mutants were tested for sensitivity to HIV-1 Vif. This approach identified the α3 and α4 helices of human APOBEC3F as important determinants of the interaction with HIV-1 Vif.

TEXT

Four APOBEC3 DNA cytosine deaminases have been demonstrated to restrict Vif-deficient HIV-1 in CD4-positive T cells (by many laboratories [reviewed in references 1 and 2]): APOBEC3D (A3D), APOBEC3F (A3F), APOBEC3G (A3G), and APOBEC3H (A3H). APOBEC3 proteins deaminate DNA cytosines to uracils, which during synthesis of the cDNA strand, provides a template for adenines, resulting in the characteristic G-to-A hypermutation observed in HIV-1 proviral sequences (3–5). These mutations can be detrimental to the virus (6, 7). HIV-1 Vif counteracts these host APOBEC3s by targeting them for ubiquitin-mediated proteasomal degradation (reviewed in references 2 and 8).

There is considerable interest in determining APOBEC3 motifs that govern the interaction with Vif. A3F provides a unique opportunity to advance this goal because the crystal structure of its catalytic domain, which also binds Vif, has been solved (9, 10). Previous studies have interrogated this interaction by manipulating human A3F (huA3F), which is normally sensitive to HIV-1 Vif-mediated degradation, to become resistant to Vif-mediated degradation (11–13). Here, the reciprocal approach is employed to determine the regions of A3F that mediate its sensitivity to HIV-1 Vif. We compared the A3F proteins from two primates: huA3F and rhesus macaque A3F (rhA3F). These proteins are 88% identical at the amino acid level and the same length, yet huA3F is sensitive to HIV-1 Vif-mediated degradation, whereas rhA3F is resistant (11, 14, 15).

We previously showed that replacing an amino acid in huA3F with the reciprocal residue found in rhA3F (E324K) is sufficient to render huA3F resistant to HIV-1 Vif-mediated degradation, but the reverse amino acid substitution in rhA3F does not endow sensitivity to HIV-1 Vif (Fig. 1A) (11). To elucidate a broader region in A3F involved in conferring sensitivity to HIV-1 Vif, a series of rhA3F (accession no. NM_001042373)-huA3F (accession no. NM_145298) chimeras was constructed with a C-terminal V5 tag (Fig. 1B) (11). The restriction potential and Vif sensitivity of these A3F constructs were determined by cotransfection into 293T cells with HIV-1 clone pIIIB proviral DNA (GenBank accession no. EU541617), either expressing wild-type Vif (Vif proficient) or tandem-stop Vif (Vif deficient), followed by immunoblot analyses and infectivity assays (16, 17).

FIG 1.

Vif-A3F interaction model and A3F experimental constructs. (A) Model illustrating that while only a single amino acid change (e.g., E324K) will convert huA3F from HIV-1 Vif sensitive to Vif resistant (11), multiple mutations may be required to sensitize rhA3F to HIV-1 Vif-mediated degradation. (B) To-scale amino acid alignment of rhA3F, huA3F, and each rhA3F-huA3F chimera construct used in this study. Each construct is depicted as a horizontal line, with amino acid differences relative to the rhA3F sequence labeled and the residue number indicated above the top sequence. To facilitate comparison, α helices are shaded dark gray, and β strands are shaded light gray. The zinc-coordinating and catalytic residues in the C-terminal domain (CTD) are indicated by black dots.

huA3F restricted the infectivity of Vif-deficient HIV-1 and to a lesser extent Vif-proficient virus (Fig. 2A). In contrast, rhA3F restricted both Vif-deficient and Vif-proficient HIV-1 with similar potencies. Related to these viral restriction phenotypes, huA3F was degraded in the presence of HIV-1 Vif, while rhA3F protein was largely unaffected. A series of four chimeras mapped the HIV-1 Vif susceptibility determinants to an huA3F region spanning residue 278 to the C terminus (Fig. 2A).

FIG 2.

Humanized rhA3F restricts single-cycle HIV-1 replication in a Vif-dependent manner. (A and B) Histograms depicting the relative (normalized to the vector control) infectivities of Vif-deficient and Vif-proficient HIV-1 clone pIIIB produced in the presence of the indicated controls or hu/rhA3F-V5 chimeric constructs (n = 3; mean plus standard deviation shown). Representative immunoblots for each series of infectivity data are shown below each histogram. The infected cells were blotted for V5 to detect A3F, Vif, and tubulin (TUB). Purified viral particles were blotted for V5 to detect A3F and p24 (Gag). A schematic of each construct is shown, with huA3F-derived residues represented in white and rhA3F residues represented in gray. (C) The ratio of Vif-proficient HIV-1 infectivity to Vif-deficient infectivity is plotted using the data from panel B. One-way analysis of variance (ANOVA) was used to compare these values, and Dunnett's method was employed for post hoc testing. Data for constructs significantly different from rhA3F (P < 0.05) are represented by black bars, and P values are shown to highlight the most important comparisons (ns, not significant).

To further delineate the Vif interaction region, a series of additional chimeras and amino acid substitution mutants were constructed based on known secondary structural elements (Fig. 1B and 2B). Although several constructs showed intermediate phenotypes, only the rhA3F-huα3,4-EK/QE chimera had a reproducible and full sensitivity to HIV-1 Vif. This was evident for both infectivity and immunoblot phenotypes (Fig. 2B) and was significant upon quantification of data from multiple independent experiments (Fig. 2C).

To further assess the restriction capabilities of key constructs, a panel of A3F-expressing SupT11 clones was generated using previously described methods (18). SupT11 cells do not express APOBEC3s and are therefore permissive for Vif-deficient HIV-1 replication [for examples, see Fig. 3 and references 17 and 18). Clones expressing low and high levels of each A3F construct were chosen for experiments (Fig. 3A). At least one clone per construct expressed A3F at H9-like levels, which resemble the amounts in primary CD4-positive lymphocytes (18). Vif-proficient or Vif-deficient HIV-1 was used to infect these clones, and virus replication was monitored over time by applying culture supernatants to CEM-GFP reporter cells, followed by green fluorescent protein-positive flow cytometry for quantification (17).

FIG 3.

HIV-1 replication kinetics in SupT11 T-cell clones stably expressing humanized rhA3F constructs. (A) A representative immunoblot showing A3F levels in the indicated clones, compared to endogenous levels of A3F in H9 cells. (B) Spreading-infection kinetics of Vif-proficient HIV-1 clone pIIIB (solid lines) in SupT11 clones stably expressing the indicated A3F expression constructs. Vif-deficient HIV-1 clone pIIIB had no detectable replication under any of the conditions, except in cells expressing low rhA3F levels or the empty vector (dashed lines and data not shown). All infections were initiated with a 1% multiplicity of infection.

In agreement with prior observations with this experimental system (17, 19), Vif-deficient HIV-1 was restricted by huA3F, while Vif-proficient HIV-1 was able to replicate in the presence of both low and high levels of huA3F (Fig. 3B, upper left panel). Both Vif-deficient and Vif-proficient HIV-1 were able to replicate on cells expressing low levels of rhA3F, although the Vif-deficient virus exhibited a considerable growth delay (Fig. 3B, upper middle panel). Neither virus was able to replicate on cells expressing high levels of rhA3F. These results highlight the relative resistance of rhA3F to HIV-1 Vif. The viral replication kinetics on cells expressing rhA3F-EK323-324QE were similar to those observed on rhA3F-expressing cells, indicating that rhA3F-EK323-324QE is also not fully sensitive to HIV-1 Vif (Fig. 3B, bottom middle panel). Most importantly, viral replication kinetics on SupT11 cells expressing rhA3F-huα3,4-EK/QE were similar to those of viruses on cells expressing huA3F (Fig. 3B, compare bottom and top left panels). These data further demonstrate that the differences in HIV-1 Vif sensitivity between huA3F and rhA3F map to the α3 and α4 helices of A3F.

The α3 and α4 helices of A3F define a negatively charged surface (Fig. 4). These surface electrostatics are consistent with the recently published structure of the five-membered Vif complex (Vif–CBF-β–CUL5–ELOB–ELOC), which implicates a positively charged region on Vif for interaction with the various Vif-sensitive APOBEC3 proteins (20). Our studies represent the first interspecies example of altering an A3F protein that is normally resistant to HIV-1 Vif-mediated degradation (rhA3F) such that it becomes sensitized. Our results extend prior studies of the A3F-Vif interaction (11–13) by highlighting the importance of an extensive surface on A3F defined by the α3 and α4 helices for HIV-1 Vif-mediated degradation of this restriction factor. With the recent publication of the crystal structure of HIV-1 Vif (20), these findings are a timely addition to the APOBEC3-Vif interaction knowledge portfolio and provide information that may be used to interpret future crystal structures of the A3F-Vif complex and also potentially inform the development of novel therapeutics.

FIG 4.

Vif interaction surface of A3F. (A) The residues identified as being important for converting rhA3F from HIV-1 Vif resistant to HIV-1 Vif sensitive are depicted in yellow and labeled on the huA3F crystal structure (PDB accession no. 4IOU). (B) The residues indicated define a negatively charged surface (red), as seen in an electrostatic potential surface map of A3F. (C) Alignment of the regions determined to be important for conveying sensitivity to degradation by HIV-1 Vif in huA3F and rhA3F. The residues in α helices three and four that differ between huA3F and rhA3F are highlighted in yellow. Active-site residues that coordinate the zinc atom are indicated by black dots.