Demystifying Cell Cycle Arrest by HIV-1 Vif

Abstract

Although APOBEC3 degradation is the canonical function of HIV-1 Vif, this viral protein also induces potent cell cycle arrest through a newly defined mechanism. Here, we review recent advances in this area and propose that the scope of this activity may go beyond subversion of the host cell cycle.

Vif Degrades Essential Phospho-Regulators to Induce G2/M Cell Cycle Arrest

The best-characterized function of HIV-1 Vif is counteraction of the APOBEC3 (A3) DNA cytosine deaminases [1]. In the absence of Vif, A3s package into nascent HIV-1 particles and generate C-to-U lesions in the viral cDNA, causing genomic strand G-to-A hypermutation and inhibiting reverse transcription. Vif prevents viral restriction by nucleating the formation of a CBF-β-ELOB/C-CUL5-RBX2 E3-ubiquitin ligase complex to target A3 enzymes for proteasomal degradation [2]. Because of the multi-log potency of this viral restriction mechanism, A3 degradation was assumed to be the sole function of Vif.

However, Vif also induces a strong G2/M-phase cell cycle arrest [3]. The first major advance toward understanding this mechanism was showing that the E3-ubiquitin ligase complex required for A3 degradation is also utilized for G2/M arrest [4-6]. However, the mechanisms are not identical because several amino acids required for A3 recognition are distinct from those used for arrest. For example, Vif residues ⁷⁴TGERDW⁷⁹ are required for A3D/F degradation but are dispensable for G2/M arrest, and ³¹ISR³³ residues are required for G2/M arrest but not for A3 degradation (Fig. 1).

Fig. 1. Vif recognition of PPP2R5 substrates.

Left, depiction of a PPP2R5 (tan) and Vif (pink) co-complex based on protein-protein docking simulations [13]. PPP2R5 residues required for Vif-mediated degradation are highlighed (red). Additionally, key Vif residues required for PPP2R5 (G2/M), A3D, A3F, and A3G degradation are depicted. These substrates were selected to emphasize the partially overlapping, but distinct, nature of substrate recognition (shared = residue required for degradation of 2 or more substrates). Right, model of PPP2R5C (PDB: 2IAE) docked onto the x-ray structure of Vif in complex with the CBF-β/E3-ubiquitin ligase complex (PDB: 4N9F).

The second breakthrough emerged from quantitative proteomics of HIV-1 infected CD4⁺ T cells to identify additional cellular proteins degraded by Vif [7, 8]. Top hits included A3 enzymes and, surprisingly, multiple members of the PPP2R5 family of protein phosphatase 2A (PP2A) regulators. PP2A holoenzymes are heterotrimeric complexes that are comprised of a phosphatase enzyme, a scaffolding protein, and a regulatory subunit [9]. PPP2R5A-E proteins are classified as regulatory subunits and, interestingly, are the only heterotrimer component degraded by Vif. Quantitative proteomics and live-cell degradation assays demonstrated that steady-state levels of the PP2A enzyme and scaffolding proteins remain unchanged in the presence of Vif ([8] and unpublished). Importantly, PP2A/PPP2R5 complexes have been implicated in regulating the G2/M transition [9].

This connection led to the third major advance that directly linked the PPP2R5 phospho-regulators to Vif-induced G2/M arrest [10-12]. Analyses of both single amino acid substitution mutants and large-scale mutagenic libraries demonstrated that Vif residues required for PPP2R5 degradation are identical to those required for inducing G2/M arrest [10-12]. For example, substitutions within the motif ³¹ISRKAK³⁶ abolish both G2/M arrest and PPP2R5 degradation activities, but not A3G degradation activity [10, 11]. In addition, siRNA-mediated knock-down of specific combinations of PPP2R5 transcripts revealed a functional redundancy in which depletion of at least two different PPP2R5s are required for inducing arrest in the absence of Vif [10, 11].

Additional studies revealed that the surface recognized by Vif overlaps with the surface used by PPP2R5s to bind cellular substrates (Fig. 2) [13]. For instance, single amino acid substitutions (²⁵⁵S or ²⁵⁸N) prevent PPP2R5A from dephosphorylating cellular substrates and being degraded by Vif [10]. Moreover, overexpression of a substrate-mimicking peptide, LxxIxE, blocks Vif from degrading PPP2R5 proteins and has no effect on A3G degradation. Together with high-resolution structures and computational simulations, these results are explained by a model in which Vif and cellular substrates compete for an electrostatic surface on PPP2R5 proteins that includes the LxxIxE substrate-binding cleft (Fig. 2) [10, 13]. Additionally, this electrostatic surface is conserved among PPP2R5 family members, which explains how Vif degrades multiple PPP2R5 proteins and clarifies genetic evidence demonstrating that G2/M arrest occurs following degradation of at least two family members [9, 10]. This working model also suggests that Vif engages one substrate at a time, either a PPP2R5 or an A3, but turnover is likely to be rapid as both substrates have half-lives of less than 6 hours [13]. Taken together, these observations indicate that Vif uses both a competitive binding mechanism and a degradation mechanism to alter the phosphorylation landscape of infected cells.

Fig. 2. Vif competes with physiological PPP2R5 substrates.

Left, depiction of the PPP2R5 surfaces bound by either Vif (brick red) or cellular substrates (light blue). Amino acid residues that are shared between substrates are depicted in purple to emphasize the partial overlap of these interfaces. Right, the substrate-mimicking LxxIxE peptide (green) is depicted bound to the PPP2R5 protein through a well-defined substrate binding cleft. This model demonstrates how the LxxIxE peptide could disrupt Vif-mediated degradation of PPP2R5 proteins as the peptide occludes several amino acid residues required for Vif recognition.

Vif may also induce G2/M arrest by inhibiting MDM2-mediated ubiquitination and nuclear export of TP53 [14]. Nuclear-localized Vif may directly shield TP53 from MDM2 recognition and subsequent nuclear export. This possibility suggests that Vif could exist in two distinct locations, a nuclear fraction that antagonizes PPP2R5C/D and protects TP53 from MDM2 and a cytoplasmic fraction that antagonizes PPP2R5A/B/E and APOBEC3s. Additional studies are needed to determine whether these mechanisms are separate or connected through shared components.

HIV-1 Pathogenicity and Cell Cycle Arrest

Although the importance of PPP2R5 degradation has yet to be established in vivo, several observations imply that cell cycle modulation may benefit HIV-1 pathogenesis. First, Vpr potently induces G2/M arrest by hijacking a different ubiquitin ligase complex (CUL4/DDB1/DCAF1) to degrade >40 cellular proteins and induce a systems-level remodeling of the proteome [15]. While several substrates have been implicated, the relative importance of each is still being investigated; however, two broad phenotypes have been linked to this activity. First, Vpr-induced degradation of MUS81 and EME1 simultaneously leads to G2/M arrest and premature activation of the SLX4 complex, which suppresses HIV-1-mediated induction of type I interferon responses and may contribute to immune evasion (reviewed by [16]). Second, G2/M arrest has also been linked to enhancing HIV-1 gene expression through CCDC137 depletion, which may boost nascent particle production and enhance HIV-1 pathogenicity [17].

Second, Vif-induced arrest has been linked to increased production of HIV-1 particles from both primary and immortalized CD4⁺ T cells [14]. We speculate that increased particle production may stem from antagonizing PP2A/PPP2R5 complexes as they regulate phosphorylation-dependent activities of S6-kinase and 4E-BP1, two major regulators of protein translation efficiency. This possibility is attractive given that host cell transcription and translation are diminished significantly during G2/M [18, 19]. Taken together, Vif- and Vpr-induced arrest may serve a dual purpose to stall host cell transcription and translation, and simultaneously hijack cellular machineries to enhance production of HIV-1 transcripts and proteins to boost particle production. In addition, HIV-1 sequence analyses have suggested that arrest-proficient Vif variants are prominent in global populations [10, 11]. However, it is worth noting that many patient-derived Vif variants are incapable of inducing cell cycle arrest, which demonstrates that this activity is dispensable for infectivity and transmission [13]. Nevertheless, further investigation is warranted to determine whether Vif and Vpr may act synergistically and/or provide compensatory functions.

Finally, live cell imaging experiments demonstrated that the half-life of PPP2R5A is only ~2 hours slower than that of APOBEC3G following infection, which is striking given that APOBEC3G counteraction is essential for HIV-1 pathogenesis [13]. This raises the possibility that antagonizing PP2A/PPP2R5 holoenzymes may somehow benefit viral infectivity prior to the establishment of G2/M arrest. For instance, PPP2R5C has been shown to suppress IL-2 production, which promotes T cell activation [20]. Thus, because HIV-1 only infects activated T cells, it is possible that PPP2R5C degradation leads to enhanced IL-2 production and a more permissive environment for virus replication.

Concluding Remarks

Subversion of the cell cycle is a conserved mechanism used by diverse viral pathogens to create a favorable environment for replication, and the discovery of the Vif/PPP2R5 axis is a major step toward understanding this process in HIV-1 pathogenesis. Interestingly, PP2A and PPP2R5 proteins appear to be a common target, given that several viruses antagonize these proteins to facilitate replication (e.g., SV40, HTLV-1, Adenovirus, and Ebola) [21-23]. Altogether, these observations support a model in which PPP2R5 modulation and corresponding global changes in the phospho-proteome are likely to be advantageous for the pathogenesis of multiple viruses.