Various plus unique: Viral protein U as a plurifunctional protein for HIV-1 replication

Introduction

In 1988, five years after the discovery of human immunodeficiency virus type 1 (HIV-1) as the causative agent of acquired immunodeficiency syndrome,^1,2 Strebel et al. ³ and Cohen et al. ⁴ identified a novel gene within the HIV-1 genome encoding a 16 kDa, 77–86 amino acid protein which they named Viral protein U (Vpu). Importantly, Vpu is unique to the HIV-1 lineage including HIV-1, simian immunodeficiency viruses (SIVs) in chimpanzee (SIVcpz) and gorilla (SIVgor), and four SIVs in four species of Old World monkeys (OWMs): greater spot-nosed monkey (SIVgsn), mona monkey (SIVmon), mustached monkey (SIVmus), and Dent’s mona monkey (SIVden).^5–8 In contrast, Vpu is not encoded by human immunodeficiency virus type 2 (HIV-2) and the majority of SIVs. ⁹

Vpu is translated from a bicistronic vpu/env mRNA in the late stage of the HIV-1 life cycle at levels similar to those of gag and env.^9,10 The vpu gene is located immediately downstream of the 3′ end of the first coding exon of tat/rev and runs 28 amino acids into the start of the first open reading frame of the HIV-1 envelope protein (Env) (Figure 1(a)). ⁹ The HIV-1 Rev protein regulates the expression of Env and Vpu from the same mRNA by binding to the Rev response element located in env and promoting their export to the cytoplasm. ¹⁰ Vpu is expressed late in the viral life cycle, due in part to its dependence on Rev. ¹⁰ OPEN IN VIEWER

HIV-1 encodes at least 109 different spliced RNAs, of which at least 12 encode bicistronic vpu/env mRNAs.^10,14,15 HIV-1 mRNAs use the modified scanning model of translation^16,17 for selecting AUG initiation sites. ¹⁰ Relative to the initiation site, if a purine and guanine are simultaneously found at positions −3 and +4, respectively, there is a much greater likelihood of successful translation. ¹⁷ Alternatively, ribosomal shunting can also occur, resulting in bypassing of the vpu ORF and efficient translation at the downstream env AUG. ¹⁸

Vpu codes for an integral membrane phosphoprotein that contains a 31-amino acid hydrophobic N-terminal transmembrane anchor. As shown by nuclear magnetic resonance, Vpu is predicted to have one transmembrane α-helix and two cytoplasmic α-helixes between amino acids 6–29, 32–51 and 52–72, respectively (Figure 1(b); as the length of Vpu differs among viral strains, the residue numbers are based on the consensus sequence of HIV-1 clones NL4-3 and WITO). Vpu also has putative phosphorylation sites at serines at positions 24, 53, 57, and 62, as well as tyrosine at position 30 (Figure 1(b)). ⁹ Moreover, it has been reported that Vpu associates with multiple cellular proteins and elicits various responses that potently affect HIV-1 replication, infection, and pathogenesis. In the following sections, we describe the roles of Vpu in both in vitro cell culture systems and in vivo.

CD4 degradation

CD4, a 55 kDa membrane glycoprotein, is the main receptor involved in HIV infection^19,20 and is expressed on the surface of CD4⁺ T lymphocytes and some myeloid cells such as macrophages and dendritic cells (DCs). ²¹ It is well known that CD4 molecules expressed on the surface of HIV-1-infected cells are severely down-regulated, and that this decrease is mediated by viral Env, Nef, and Vpu proteins. ²² Specifically, Vpu degrades CD4 molecules through a ubiquitin/proteasome-dependent pathway.^22,23

Vpu-mediated CD4 degradation depends on the interaction of the cytosolic domains of these two proteins. ²⁴ Vpu contains two phosphorylated serine residues in a canonical DpSGxxpS sequence within its cytoplasmic domain (residues 53 and 57; Figure 1(b)) that binds cellular beta-transducin repeat-containing E3 ubiquitin protein ligases (BTRCs, also known as β-TrCPs). ²⁵ This recruitment leads to the addition of ubiquitin to lysine residues of the cytoplasmic tail of CD4, which marks the protein for proteasomal degradation.^26,27 Vpu is not ubiquitinated or degraded itself, as it has an equivalent function to ubiquitin ligase adaptors. ²⁸ The formation of the Vpu–BTRC complex and the interaction between Vpu and CD4 through their transmembrane domains also increases the amount of time CD4 proteins are retained in the endoplasmic reticulum (ER). ²⁹

The molecular mechanism of Vpu-mediated degradation of CD4 molecules follows an ER-associated protein degradation (ERAD)-dependent pathway. Proteins in the secretory pathway that fail to acquire a native structure are usually degraded via the canonical ERAD pathway in a proteasome-dependent manner. Experiments using specific inhibitors of the proteasome pathway reported partial or complete blocking of CD4 degradation in a polyubiquitination-dependent manner. This system is, however, independent of the activity of the ER chaperone calnexin. ²⁷ The polyubiquitination pathway requires the joining of several ubiquitin moieties to a protein. The CD4 cytoplasmic tail is considered to be the main target for the activity of ubiquitin-activating enzyme E1. The CD4 cytoplasmic tail contains four lysine residues that are potential targets for ubiquitin, and substitution of these amino acids by arginine (which cannot be joined to ubiquitin) leads to a complete blockade of CD4 degradation. ²⁷

However, the function of Vpu in ERAD appears to be non-canonical. ³⁰ In 2010, Magadan et al. ²⁹ reported the convergence of Vpu-induced CD4 degradation with the valosin-containing protein (VCP)–ubiquitin fusion degradation protein 1 homolog (UFD1L)–nuclear protein localization protein 4 (NPL4) complex, a key component of the ERAD pathway. The main function of this complex is to remove ubiquitinated substrates from the ER, making them available for proteasome degradation.^31–34 This activity was reported to be dependent not only on lysine residues within the cytoplasmic tail of CD4 but also on serine/threonine residues. This was the first time a ubiquitin ligase activity was shown to rely on serine/threonine activity for the degradation of an ERAD substrate. The three proteins in the VCP–UFD1L–NLP4 complex have unique functions. VCP is a member of the AAA-ATPase superfamily and has an N-terminal domain that binds UFD1L and NPL4 and two AAA-ATPase domains. Both domains are required to degrade CD4. ²⁹ The VCP–UFD1L–NPL4 complex is composed of a heterodimer of UFD1L–NPL4 joined with a homohexamer of VCP. Although both UFD1L and NPL4 have sequences to join ubiquitin chains that have been K48-linked or K63-linked to a target, Magadan et al. ²⁹ confirmed that VCP provides energy to the complex, UFD1L binds K48 CD4 ubiquitin chains, and that NPL4 stabilizes UFD1L.

Tetherin antagonization

Previous studies indicate that Vpu potently enhances HIV-1 replication and production in certain cell lines (e.g. HeLa cells). A heterokaryon experiment conducted by Varthakavi et al. ³⁵ identified the presence of a restriction factor that is expressed in HeLa cells whose action is antagonized by Vpu. Then, Neil et al. ³⁶ have reported that the expression of this restriction factor is induced by type I interferon (IFN). In 2008, two groups identified this restriction factor as tetherin (also known as bone marrow stromal antigen 2).^37,38 Tetherin strictly impairs HIV-1 release from the surface of infected cells (i.e. this protein “tethers” budding virions). Conversely, some Vpu proteins antagonize tetherin-mediated antiviral action by down-modulating tetherin surface expression on HIV-1-producing cells (Van Damme ³⁸ ; also well summarized in the literature^7,29,39).

To determine the Vpu residues responsible for tetherin antagonization, Vigan and Neil ⁴⁰ conducted alanine scanning mutagenesis of HIV-1 Vpu (strain NL4-3) and determined that A15, W23, and, to a lesser extent, A19 are required for tetherin antagonism (Figure 1(b)). ⁴⁰ Additionally, the S53/S57 mutant partially lost its antitetherin activity, suggesting that the BTRC-mediated pathway is also important.^41–43 Moreover, Kueck et al. ⁴⁴ revealed that Vpu degrades tetherin via the endosomal degradation pathway dependent on endosomal sorting complexes required for transport machinery. ⁴⁴ In addition to the interaction with tetherin via its transmembrane domain, two motifs in Vpu, the DSGNES phosphoserine motif (residues 52–57 in Figure 1) and the adjacent ExxxLV acidic dileucine motif (residues 60–65 in Figure 1), are required for clathrin-dependent sorting, which leads to tetherin antagonism. ⁴⁴

Although the potency of HIV-1 Vpu to degrade CD4 is highly conserved in all HIV-1 groups and related SIVs, the activity of Vpu to antagonize tetherin is only conserved in HIV-1 group M, a global pandemic strain. ⁴⁵ In contrast to HIV-1 group M, the Vpu proteins of HIV-1 groups N, O, and P, which are endemic in Africa, either do not or only weakly antagonize tetherin.^45–49

The Vpu proteins of SIVcpz and SIVgor, precursors of HIV-1, do not possess antitetherin activity, ⁴⁵ implying that Vpu only obtained this function, once it infected humans as HIV-1. This change quite possibly contributed to its pandemic status. Indeed, a multilayered experimental–mathematical investigation has addressed this scenario and suggests that tetherin plays the role of a host restriction factor, providing “intrinsic herd immunity,” while Vpu has evolved in HIV-1 group M as a tetherin antagonist. ⁵⁰

Although SIVcpz Vpu does not antagonize tetherin, the Vpu proteins of SIVgsn/mon/mus, precursors of SIVcpz in OWMs, possess antitetherin activity. ⁴⁵ This fact suggests that SIVcpz Vpu has lost its antitetherin activity during its evolution. It appears that SIVcpz received its nef gene from the SIVrcm lineage, whereas its vpu gene originated from the SIVgsn/mon/mus lineage. ⁵¹ The viral protein Nef of SIVcpz, SIVgor, and SIVrcm also possesses the ability to antagonize the tetherin of its host.^45,52–54 This antitetherin activity by Nef extends to HIV-1 group O. ⁵⁵ It appears that the switching of viral tetherin antagonists between Vpu and Nef has frequently occurred during the evolution of primate lentiviruses including SIVs and HIV-1.

Down-modulation of cellular transmembrane proteins

In addition to CD4 and tetherin, Vpu potently degrades and/or down-modulates a variety of cellular transmembrane proteins. In this section, we summarize how Vpu degrades and/or down-modulates these cellular transmembrane proteins and the relevance of each down-regulation to HIV-1 infection.

Inhibition of nuclear factor kappa light chain enhancer of activated B-cells (NF-κB) signaling

NF-κB is a transcription factor that is ubiquitously expressed and essential for effective transcription of lentiviral genomes. ⁸³ Galao et al. ⁸⁴ have reported that tetherin acts as a virus sensor to induce NF-κB-dependent proinflammatory gene expression, while HIV-1 Vpu (strain NL4-3) potently suppresses the signal transduction triggered by tetherin. Recently, Sauter et al. ⁸⁵ have reported that Vpu contributes to the strong suppression of NF-κB activity through inhibiting tetherin-triggered signaling and inhibition of nuclear translocation of p65. This activity of Vpu to suppress NF-κB is conserved among the Vpu proteins of most primate lentiviruses. ⁸⁵ Vpu’s ability to inhibit tetherin-induced NF-κB activation was not associated with its ability to reduce tetherin-mediated antiviral action. ⁸⁵ Mutagenesis experiments revealed that Vpu-mediated recruitment of BTRC, which is mediated by serine residues at positions 53 and 57, was not sufficient for inhibition of NF-κB signaling. Further mutagenesis determined that adjacent residues, particularly R45, R49, E51, and G59 were required for efficient Vpu-mediated inhibition of NF-κB signaling (Figure 1(b)). Thus, this region of Vpu appears to affect NF-κB activity by both BTRC-dependent and -independent mechanisms. Based on these findings, the authors have emphasized that a tight interplay occurs between early and late viral products to control NF-κB activity ⁸⁵ : early HIV-1 gene products such Nef and Tat promote NF-κB activation, whereas late products such as Vpu suppress it. However, it should be noted that the relevance of Vpu-mediated NF-κB inhibition in viral replication remains unclear. This issue should be addressed in a future study.

Ion channel activity and Vpu inhibitors

Before the identification of its role in targeting cellular proteins (summarized above), Vpu was characterized primarily as a cation-selective ion channel. ⁸⁶ Vpu forms homo-oligomers ⁸⁷ and is designated a “viroporin.” ⁸⁸ As an ion channel, Vpu depolarizes the plasma membrane by inhibiting the background potassium current and the membrane depolarization contributes to viral release.^89–91 Also, the Vpu channel can be blocked by amiloride analogs, and importantly, an amiloride analog, called BIT225, blocks the ion channel activity of Vpu in cell cultures.^92,93 Thus, this Vpu inhibitor that entered into clinical trials targets its viroporin function. ⁹⁴ This compound was obtained after the finding that amiloride derivatives that inhibit ion channels could depress the release of HIV-1 virions in macrophages.^92,93 Furthermore, acylguanidine-containing compounds have been recognized as inhibitors against a broad range of viruses such as hepatitis C virus and coxsackievirus B3, ⁹⁵ and a acylguanidine-containing compound, called SM111, potently impairs the late stage of HIV-1 replication. ⁹⁶ More importantly, SM111-mediated inhibition of HIV-1 replication was partially overcome by a Vpu I17R substitution. ⁹⁶ These observations suggest that SM111 targets Vpu and that acylguanidine-containing compounds including SM111 provide promising opportunities to develop new treatments targeting Vpu.

The ion channel activity of HIV-1 Vpu is independent of tetherin antagonism, ⁹⁷ and the serine residue at position 24 is crucial for channel activity (Figure 1(b)). ⁹⁸ Vpu’s ion channel activity is conserved among HIV-1 groups M, N, and P and also among SIVs from chimpanzees and from greater spot-nosed monkeys, despite their sequence variability. ⁹⁹ Although this suggests an involvement of Vpu’s ion channel activity for viral replication, survival, and pathogenicity, its relevance in vivo has remained unclear. Interestingly, it has been reported that the viroporins encoded by some viruses such as influenza virus M2 protein ¹⁰⁰ and encephalomyocarditis virus 2B protein ¹⁰¹ harbor the ability to trigger inflammasome activation. Therefore, it is conceivable that HIV-1 Vpu has similar activity.

GBP5 Antagonism: Trade-off of antitetherin activity and resistance to GBP5

Recently, guanylate binding protein 5 (GBP5), a novel restriction factor that potently suppresses HIV-1 replication, was identified by a genome screening analysis. ¹⁰² GBP5 belongs to an IFN-inducible subfamily of guanosine triphosphatases (GTPases) with host defense activity against intracellular bacteria and parasites. ¹⁰³ Krapp et al. ¹³ revealed that GBP5 suppresses the infectivity of HIV-1 and other retroviruses by interfering with N-linked glycosylation of the viral Env, which is critical for correct processing and viral infectivity. GBP5 is endogenously expressed in human monocyte-derived macrophages (MDMs) and CD4⁺ T-cells and its expression is induced by IFN-α and IFN-γ. ¹³ Additionally, the authors revealed that GBP5 gene expression is up-regulated by HIV-1 replication in infected individuals and that its expression level is inversely correlated with the infectious HIV-1 yield in human MDMs. ¹³ Mutational analysis revealed that the cysteine residue at position 583 of GBP5 is critical for its antiviral activity. ¹³

Importantly, the authors demonstrated that mutation of the vpu initiation codon or the insertion of premature stop codons in vpu gene increased Env expression and conferred partial resistance to GBP5. ¹³ Particularly, HIV-1 strains NL4-3, CH058, CH198, CH293, and CH167 with vpu deletion enhanced Env expression and were significantly more resistant to GBP5-mediated antiviral effects than vpu-proficient viruses. ¹³ Altogether, these observations suggest an unusual trade-off mechanism involving enhanced Env expression at the cost of an intact vpu gene. It is particularly noteworthy that Vpu potently antagonizes tetherin-mediated anti-HIV-1 effect, as described above. Therefore, HIV-1 must choose either activity: antitetherin activity (by maintaining vpu gene) or resistance to GBP5 (by augmenting Env expression through the loss of the vpu gene). Since most of the clinically isolated HIV-1s encode the vpu gene,^13,104 it is assumed that Vpu-mediated antitetherin activity is more crucial for HIV-1 to establish person-to-person transmission than acquiring resistance to GBP5. However, it should be noted that the initiation triplet AUG of vpu in certain HIV-1 strains such as strains HXB2, ¹⁰⁵ BH8, ¹⁰⁶ MAL, ¹⁰⁷ and Zr6¹⁰⁸ has been primarily deleted. This fact further raises the possibility that the resistance to GBP5 is important for viruses to expand and disseminate in certain tissues or organs in infected individuals. Further investigations will shed light on the importance of the trade-off relationship between antitetherin activity and GBP5 resistance.

Impact of Vpu on HIV-1 replication and future direction

In addition to the multiple roles of Vpu described above, some virological experiments have suggested that the deletion of vpu enhances cell-to-cell HIV-1 infection.^109,110 Although a study has suggested the association of Vpu’s antitetherin activity with the efficiency of cell-to-cell viral spread, ¹¹⁰ this issue remains controversial, ¹¹¹ and how vpu deletion enhances cell-to-cell transfer remains unclear.

To address the role of Vpu in HIV-1 replication and pathogenesis in vivo, at least two small animal models have been utilized. First, Aldrovandi and Zack have used the SCID-hu mouse model, mice that have severe combined immunodeficiency (SCID) and are transplanted with human thymus and liver. The authors demonstrated that vpu is dispensable for viral replication in the SCID-hu mouse model. ¹¹² Second, another group as well as ours have used a human hematopoietic stem cell-transplanted “humanized” mouse model to show that Vpu is slightly associated with viral cytopathicity.^113,114 Moreover, we have revealed that Vpu particularly boosts viral expansion during the acute phase of infection. ¹¹⁴ However, it should be noted that all of the three studies described above have used vpu defective viruses. As described in previous sections, since Vpu possesses multiple functions (Figure 1(b)), it remains unclear which activities of Vpu are most important for viral replication in vivo. This should be addressed to elucidate Vpu’s precise role in HIV-1 replication and pathogenesis in vivo to hopefully provide useful knowledge for the development of novel anti-HIV-1 drugs.

Authors’ contributions

AS, GJ-F, HA, MM, EY, YN, YK, and KS reviewed the manuscript and wrote the paper.