Tat and Rev

The viral protein Tat is a critical transactivator that promotes transcriptional elongation. It binds to the TAR (trans-activation-responsive) element, an RNA stem-loop structure located in the R region of the LTR and recruits cellular co-factors Cyclin T1 and CDK9 in a complex called P-TEFb (Roy et al. Reference Roy, Delling, Chen, Rosen and Sonenberg1990; Wei et al. Reference Wei, Garber, Fang, Fischer and Jones1998; Bieniasz et al. Reference Bieniasz, Grdina, Bogerd and Cullen1999). This complex, also known as TAK (Tat-associated kinase), is crucial in phosphorylating the COOH terminal domain of RNA polymerase II, enhancing transcriptional elongation (Emerman, Reference Emerman1998).

The Rev protein binds a short RNA segment called Rev responsive element (RRE) located in the 3′ region of the viral mRNA. Rev recruits the nuclear export receptor CRM1, which, upon multimerization at the RRE, facilitates nuclear export of unspliced and incompletely spliced mRNAs (Pollard and Malim, Reference Pollard and Malim1998; Tazi et al. Reference Tazi, Bakkour, Marchand, Ayadi, Aboufirassi and Branlant2010). Thus Rev, by recruiting a host cell factor, overcomes the block to nuclear export of unspliced mRNAs in human cells.

Assembly, budding and maturation

The Gag polyprotein (which encodes for p17 matrix (MA), p24 capsid (CA), p7 nucleocapsid (NC) and p6) is the main driving force for the assembly of immature HIV (Ivanchenko et al. Reference Ivanchenko, Godinez, Lampe, Kräusslich, Eils, Rohr, Bräuchle, Müller and Lamb2009). Upon translation, most of the Gag proteins accumulate and multimerize at the plasma membrane, creating an assembly site. The N-terminus MA domain is tightly associated with the inner surface of the lipid bilayer while the C-terminus p6 projects away from the phospholipid membrane. Contacts between the nucleocapsid p7 domain and full length viral RNA act as a scaffold that recruits viral RNA towards the cell membrane for assembly. At this stage, Env is incorporated into the host membrane in the form of gp120 and gp41 (cleaved by host furine protease) (Freed and Martin, Reference Freed, Martin, Knipe and Howley2001). Subsequently, the Endosomal Sorting Complexes Required for Transport (ESCRT) is recruited to the assembly site to facilitate the budding of nascent immature virion (Garrus et al. Reference Garrus, von Schwedler, Pornillos, Morham, Zavitz, Wang, Wettstein, Stray, Côté, Rich, Myszka and Sundquist2001; Martin-Serrano et al. Reference Martin-Serrano, Zang and Bieniasz2001). The early interaction between the P(T/S)AP motif of the p6 protein and TSG101 (Tumor Susceptibility Gene 101) of the ESCRT-I complex initiates the formation of budding viruses This then activates ESCRT-II and consequently triggers the formation of ESCRT-III (Sundquist and Kräusslich, Reference Sundquist and Kräusslich2012). Another p6 domain, LXPX_nL, serves as a docking site for the ESCRT-III binding partner: Alix [ALG-2 (apoptosis-linked gene 2)-interacting protein X] (Stuchell et al. Reference Stuchell, Garrus, Müller, Stray, Ghaffarian, McKinnon, Kräusslich, Morham and Sundquist2004; Usami et al. Reference Usami, Popov, Popova, Inoue, Weissenhorn and Göttlinger2009; Briant, Reference Briant2011). Lastly, the ESCRT complex recruits CHMPs (charged multivesicular body proteins) to promote closure of the membrane neck and VPS4 (vacuolar protein sorting) ATPase to complete the process of membrane fission (Sundquist and Kräusslich, Reference Sundquist and Kräusslich2012). As a result, an immature and non-infectious virion buds off from the host cell. Finally, the immature virion undergoes maturation by activating viral protease p11 (PR) through autocleavage. Together with major conformational changes and structural rearrangements, PR cleaves Gag and Gag-Pol polyproteins into mature proteins: MA, CA, NC, p6, IN and RT, producing a fully infectious virion (Briant, Reference Briant2011; Sundquist and Kräusslich, Reference Sundquist and Kräusslich2012).

Latency

Combined antiretroviral therapy (cART) is the standard treatment for HIV infection and targets different steps in the HIV replication cycle. Typically, there will be three or more antiviral drugs taken in combination (e.g. a nucleoside RT inhibitor, a protease inhibitor and a non-nucleoside RT inhibitor) to reduce the likelihood of developing viral resistance. cART is very effective (Palella et al. Reference Palella, Delaney, Moorman, Loveless, Fuhrer, Satten, Aschman and Holmberg1998) however, upon therapy suspension, the latent viral reservoir is activated and viral load rebounds to pre-treatment levels in a few weeks (Finzi et al. Reference Finzi, Hermankova, Pierson, Carruth, Buck, Chaisson, Quinn, Chadwick, Margolick, Brookmeyer, Gallant, Markowitz, Ho, Richman and Siliciano1997).

How the latent reservoir is established and maintained is an issue not fully resolved. It is accepted that there is very low viral production even under cART (Palmer et al. Reference Palmer, Maldarelli, Wiegand, Bernstein, Hanna, Brun, Kempf, Mellors, Coffin and King2008), but it is unclear if this residual viraemia is due to ongoing replication in specific sites (mainly the gastrointestinal lymphatic system – GALT) where drugs may diffuse poorly, or to a long-lived reservoir that is randomly activated, such as resting memory T cells (Hakre et al. Reference Hakre, Chavez, Shirakawa and Verdin2012). In latently infected cells, the virus remains transcriptionally silent. Latently infected cells are typically few, estimated at 0·03–3 per million resting CD4⁺ T cells, and are often undetectable by even the most sensitive clinical assays (Siliciano et al. Reference Siliciano, Kajdas, Finzi, Quinn, Chadwick, Margolick, Kovacs, Gange and Siliciano2003). These memory T cells can also survive for several years (Michie et al. Reference Michie, McLean, Alcock and Beverley1992; Chun et al. Reference Chun, Carruth, Finzi and Shen1997) and can be activated under appropriate stimulation, causing viral rebound. Therefore, the long-lived latent viral reservoir prevents HIV-1 eradication and a cure.

HIV-1 latency is a multifactorial process involving chromatin modifications, low levels of specific transcription factors, integration site selection and cell activation (Chan and Greene, Reference Chan and Greene2011; Hakre et al. Reference Hakre, Chavez, Shirakawa and Verdin2011; Van Lint et al. Reference Van Lint, Bouchat and Marcello2013) and it can be broadly divided in pre-integration latency and post-integration latency (Chun et al. Reference Chun, Carruth, Finzi and Shen1997; Donahue and Wainberg, Reference Donahue and Wainberg2013). Pre-integration latency occurs when the viral DNA failed to be integrated into the host genome. This could happen either because reverse transcription was impaired or PIC nuclear import was blocked (Zack et al. Reference Zack, Arrigo, Weitsman, Go, Haislip and Chen1990; Bukrinsky et al. Reference Bukrinsky, Sharova, Dempsey, Stanwick, Bukrinskaya, Haggerty and Stevenson1992). Inefficient reverse transcription and PIC nuclear import and integration in resting T cells is in part determined by the limited nucleotide pool, possibly due to host restriction factor SAMHD1 (Baldauf et al. Reference Baldauf, Pan, Erikson, Schmidt, Daddacha, Burggraf, Schenkova, Ambiel, Wabnitz, Gramberg, Panitz, Flory, Landau, Sertel, Rutsch, Lasitschka, Kim, König, Fackler and Keppler2012; Descours et al. Reference Descours, Cribier, Chable-Bessia, Ayinde, Rice, Crow, Yatim, Schwartz, Laguette and Benkirane2012) and in part by other host co-factors (Zack et al. Reference Zack, Arrigo, Weitsman, Go, Haislip and Chen1990). On the other hand, post-integration latency represents a more stable form of latency (Chun et al. Reference Chun, Finzi and Margolick1995). This post-integration latency may occur when infected, activated T cells return to resting G0 state (Chun et al. Reference Chun, Finzi and Margolick1995), or due to integration of viral DNA into repressed chromatin regions of the host genome, or due to interference with a flanking cellular promoter (Colin and Van Lint, Reference Colin and van Lint2009). Nonetheless, these latently infected cells can be activated and start producing virus again, but the signal transduction pathways that regulate this switch from latent to productive infection are understood only in part. It is generally believed that if the host cell is exposed to the appropriate antigens or activating stimuli, latency may be reversed and the cell can start producing new viruses (Siliciano and Greene, Reference Siliciano and Greene2011).

Hsp90 and HIV transcriptional control

Surprisingly, despite its abundance and critical functions in human cells, Hsp90 has been little studied in the context of HIV-1 infection, and its relevance has started to become apparent only recently.

Proteomics and microarray analyses on HIV-1 infected cells detected an increased expression of Hsp90 upon HIV-1 infection (Ringrose et al. Reference Ringrose, Jeeninga, Berkhout and Speijer2008; Boukli et al. Reference Boukli, Shetty, Cubano, Ricaurte, Coelho-Dos-Reis, Nickens, Shah, Talal, Philip and Jain2012), although it remained unclear if Hsp90 up-regulation was specifically induced by HIV-1 or it was simply a cellular response to infection-induced stress. Early studies showed that latently infected cells can reactivate HIV-1 when incubated in heat shock condition (42 °C) without any added cytokines (Stanley et al. Reference Stanley, Bressler, Poli and Fauci1990; Hashimoto et al. Reference Hashimoto, Baba, Gohnai, Sato and Shigeta1996) and this hyperthermia-induced activation has been exploited for rapid detection of HIV in clinical samples (Re et al. Reference Re, Furlini and La Placa1989). However, it remained unclear if HIV-1 reactivation upon heat shock actually depended on Hsp90.

More recently, hyperthermia (39·5 °C) was reported to enhance HIV-1 replication by stimulating viral gene expression (Roesch et al. Reference Roesch, Meziane, Kula, Nisole, Porrot, Anderson, Mammano, Fassati, Marcello, Benkirane and Schwartz2012). While both Hsp70 and Hsp90 were up-regulated in CD4⁺ T cells grown at 39·5 °C, selective inhibition of Hsp90 with small chemical compounds was sufficient to negate this enhanced HIV-1 replication (Roesch et al. Reference Roesch, Meziane, Kula, Nisole, Porrot, Anderson, Mammano, Fassati, Marcello, Benkirane and Schwartz2012). Furthermore, confocal microscopy experiments showed that hyperthermia promoted localization of Hsp90 at the viral transcriptional site (Roesch et al. Reference Roesch, Meziane, Kula, Nisole, Porrot, Anderson, Mammano, Fassati, Marcello, Benkirane and Schwartz2012). Therefore, the study demonstrated a direct link between enhanced replication of HIV-1 in conditions of hyperthermia and Hsp90. These findings have implications for the management of HIV-1-positive individuals, who often suffer from co-infection by other pathogens (tuberculosis, malaria) that cause high fever episodes.

An interesting question prompted by this study is why Hsp90 localizes at the HIV-1 transcriptional site. Indeed chromatin immunoprecipitation (ChIP) experiments showed that Hsp90 localizes on chromatin at the HIV-1 promoter (Vozzolo et al. Reference Vozzolo, Loh, Gane, Tribak, Zhou, Anderson, Nyakatura, Jenner, Selwood and Fassati2010). The functional significance of this observation was supported by the fact that selective Hsp90 inhibitors repressed HIV-1 gene expression in acutely infected cells (Vozzolo et al. Reference Vozzolo, Loh, Gane, Tribak, Zhou, Anderson, Nyakatura, Jenner, Selwood and Fassati2010). Interestingly, Hsp90 was required for efficient HIV-1 gene expression in acutely infected but not in chronically infected cells, suggesting that the chaperone is important for some specific early event regulating viral transcription.

One hypothesis to explain the role of Hsp90 in the early events post-HIV-1 integration is that the chaperone promotes chromatin remodelling to facilitate viral gene expression. The evidence supporting this hypothesis is growing. During the process of LTR activation, Tat recruits the SWI/SNF chromatin remodelling complex to the HIV promoter, facilitating nucleosomal reorganization (Mahmoudi et al. Reference Mahmoudi, Parra, Vries, Kauder, Verrijzer, Ott and Verdin2006; Tréand et al. Reference Tréand, Du Chéné, Brès, Kiernan, Benarous, Benkirane and Emiliani2006). Moreover, Hsp90, along with its co-chaperone p23, promotes disassembly of transcriptional regulatory complexes (Freeman and Yamamoto, Reference Freeman and Yamamoto2002). In yeast, Hsp90 has been shown to interact with components of the chromatin remodelling machinery such as The2 proteins and Rvb1 and Rvb2 helicases to promote transcription (Wong and Houry, Reference Wong and Houry2006). Hsp90 cofactors such as Tah1 and Pih1 are also suggested to mediate the folding and assembly of Rvb1/2 into the INO80 and SWR-C chromatin remodelling complexes (Zhao et al. Reference Zhao, Davey, Hsu, Kaplanek, Tong, Parsons, Krogan, Cagney, Mai, Greenblatt, Boone, Emili and Houry2005; Wong and Houry, Reference Wong and Houry2006).

In yeast, deletion of Hsp90 was shown to delay rapid transcriptional induction in response to galactose by inhibiting the removal of nucleosomes from the promoter DNA and recruitment of the transcriptional machinery (Floer et al. Reference Floer, Bryant and Ptashne2008). In Drosophilia, Hsp90 was shown to cooperate with Trithorax at Polycomb Response Element (PRE) to recruit Polycomb group proteins (PcG) for chromatin remodelling (Tariq et al. Reference Tariq, Nussbaumer, Chen, Beisel and Paro2009). After integration, the HIV-1 provirus can actually be considered like a cellular gene that must be rapidly activated. Therefore, the role of Hsp90 in regulating rapid chromatin remodelling to stimulate transcription in response to external stimuli (Floer et al. Reference Floer, Bryant and Ptashne2008) may be of particular relevance to HIV-1 replication. This is also important in HIV-1 reactivation from latency, where a degree of chromatin remodelling is required for induction of viral gene expression (Van Lint et al. Reference Van Lint, Bouchat and Marcello2013). Remarkably, a global ChIP analysis in Drosophila showed that Hsp90 maintains RNA polymerase II pausing, and is required for maximal and rapid activation of paused genes in response to environmental stimuli (Sawarkar et al. Reference Sawarkar, Sievers and Paro2012).

An alternative and complementary hypothesis to explain the role of Hsp90 in HIV-1 gene expression is that the chaperone is necessary for the function of specific transcription factors. For example Hsp90 is involved in the regulation of the NF-kB pathway, which is critical for early HIV-1 gene expression (Nabel and Baltimore, Reference Nabel and Baltimore1987; Bouwmeester et al. Reference Bouwmeester, Bauch, Ruffner, Angrand, Bergamini, Croughton, Cruciat, Eberhard, Gagneur, Ghidelli, Hopf, Huhse, Mangano, Michon, Schirle, Schlegl, Schwab, Stein, Bauer, Casari, Drewes, Gavin, Jackson, Joberty, Neubauer, Rick, Kuster and Superti-Furga2004; Pittet et al. Reference Pittet, Lee, Pespeni, O'Mahony, Roux and Welch2005; Chan and Greene, Reference Chan and Greene2011). Furthermore, Hsp90 was shown to aid the assembly of an active P-TEFb complex, which is recruited by Tat to phosphorylate the C-terminal domain of RNA Pol II and promote transcriptional elongation (O'Keeffe et al. Reference O'Keeffe, Fong, Chen, Zhou and Zhou2000).

In summary, evidence is accumulating that Hsp90 may regulate HIV-1 gene expression at multiple levels and may play a significant role in HIV-1 reactivation from latency (Fig. 4).

Fig. 4.

Multiple roles of Hsp90 in regulating HIV-1 gene expression. Hsp90 is shown promoting activation of the NF-kB pathway, activation of the P-TEFb complex and the SWI/SNF chromatin remodelling complex.

Hsp90 and the stability of the viral capsid core

As a result of its protein-stabilizing role, Hsp90 has also been implicated in buffering genetic variations (Rutherford and Lindquist, Reference Rutherford and Lindquist1998). In other words, Hsp90 significantly contributes to the robustness of an organism phenotype in the presence of genetic variations caused by environmental insults or stochastic changes (Queitsch et al. Reference Queitsch, Sangster and Lindquist2002). If Hsp90 is inactivated, accumulated genetic variations become manifest, promoting stepwise evolution by adaptation (Queitsch et al. Reference Queitsch, Sangster and Lindquist2002). Importantly, evidence indicates that, at least in Drosophila, Hsp90 acts by buffering epigenetic variations as well (Pigliucci, Reference Pigliucci2003; Sangster et al. Reference Sangster, Queitsch and Lindquist2003). If Hsp90 is inactivated, changes in chromatin state become phenotypically expressed and heritable (Sollars et al. Reference Sollars, Lu, Xiao, Wang, Garfinkel and Ruden2003). Hsp90 may also prevent phenotypic variation by suppressing the mutagenic activity of transposons. Transposons are small DNA sequences that are able to ‘jump’ from one chromatin position to another, potentially creating mutations or genetic variability. Hsp90 regulates mobilization of these transposable elements by affecting Piwi-interacting RNA (piRNA) (Specchia et al. Reference Specchia, Piacentini, Tritto, Fanti, D'Alessandro, Palumbo, Pimpinelli and Bozzetti2010), a molecule associated with gene silencing (Grimson et al. Reference Grimson, Srivastava, Fahey, Woodcroft, Chiang, King, Degnan, Rokhsar and Bartel2008). Reduction in Hsp90 influences piRNA silencing activity, which causes stress-response-like activation of transposons, generating new variation (Specchia et al. Reference Specchia, Piacentini, Tritto, Fanti, D'Alessandro, Palumbo, Pimpinelli and Bozzetti2010).

Within this conceptual framework, one might propose that Hsp90 plays an important role in buffering genetic variations of viruses that have a high mutation rate. It is well-established that the mutation rate of HIV-1 is very high at 3·4×10⁻⁵ per base pair per cycle in vivo (Mansky and Temin, Reference Mansky and Temin1995), compared with human DNA polymerase with an error rate of 1×10⁻⁹ per nucleotide. HIV-1 genetic variability is under additional pressure in the presence of cART, because viral strains that become drug resistant often carry mutations that reduce viral fitness (Götte, Reference Götte2012; Kouyos et al. Reference Kouyos, Leventhal, Hinkley, Haddad, Whitcomb, Petropoulos and Bonhoeffer2012; Quashie PK et al. Reference Quashie, Mesplède and Wainberg2013). Hence the ability to ‘buffer’ such fitness-reducing mutations would be advantageous to HIV-1.

Support for this idea has come from two recent publications showing that Hsp90 rescues a defect of the viral core of mutated, protease-resistant HIV-1 strains, improving viral fitness and efficient replication (Joshi and Stoddart, Reference Joshi and Stoddart2011). Furthermore, Hsp90 was shown to rescue core stability defects of several HIV-1 mutants (Joshi et al. Reference Joshi, Sloan, Torbett and Stoddart2013). Hsp90 was found to help folding and stabilization of the capsids of hepatitis B virus (HBV) and poliovirus (Geller et al. Reference Geller, Vignuzzi, Andino and Frydman2007; Shim et al. Reference Shim, Quan, Yi and Jung2011). Not only did inhibition of Hsp90 impair viral replication, it also prevented the emergence of drug-resistant viruses (Geller et al. Reference Geller, Vignuzzi, Andino and Frydman2007), in agreement with the potential role of Hsp90 in buffering genetic variation (Fig. 5).

Fig. 5.

Role of Hsp90 in buffering genetic changes. Mutations in the capsid p24 protein (shown as an ‘X’) allow HIV-1 to escape protease inhibitors but reduce fitness, possibly by making the viral core too stable. Hsp90 restore, in part, viral fitness presumably by chaperoning the core.

Therefore, there is evidence supporting the exciting possibility that Hsp90 may be a capacitor for the evolution of HIV-1 and other viruses, by buffering genetic variations that reduce fitness. If this is the case, then Hsp90 may also be implicated in the evolution of drug-resistant variants, which represent a major problem of cART.

Immunological function

HIV-1 infection induces a broad range of immune responses, including innate responses leading to IFN and TNFα production. Indeed, during the acute phase of infection there is significant up-regulation of inflammatory cytokines (Borrow, Reference Borrow2011). While such responses may be beneficial in the short term, chronic immune activation is a feature of HIV-1 infection that best correlates with progression to AIDS (Giorgi et al. Reference Giorgi, Hultin, McKeating, Johnson, Owens, Jacobson, Shih, Lewis, Wiley, Phair, Wolinsky and Detels1999). Inhibition of Hsp90 has been shown to result in anti-inflammatory effects through several pathways, including the phosphatidylinositol-3 kinase (PI3K)/Akt and the NF-kB pathway (Shimp et al. Reference Shimp, Chafin, Regna, Hammond, Read, Caudell, Rylander and Reilly2012a , Reference Shimp, Parson, Regna, Thomas, Chafin, Reilly and Nichole Rylander b ). Interestingly, Hsp90 with its co-chaperone Cdc37 have also been implicated in intracellular sensing of retroviral DNA by stabilizing TBK1 and conversely chemical inhibition of Hsp90 abrogated retroviral DNA sensing in mouse embryonic cells lacking the TREX nuclease (Lee et al. Reference Lee, Roy, Ong, Mertins, Villani, Li, Dotiwala, Sen, Doench, Orzalli, Kramnik, Knipe, Lieberman, Carr and Hacohen2013).

HIV-1 infection also induces robust, albeit ultimately ineffective, CD8T cell responses. HIV-1-derived peptides have variable half-lives and some HLA-associated HIV-1 peptides obtained from clinical isolates have a shorter half-life relative to consensus sequence. The shorter half-life of the HLA-associate peptides impairs proper presentation and CTL recognition. Hsp90 in complex with histone deacetylase HDAC6 enhanced the stability of the peptides whereas chemical inhibition of Hsp90 dramatically reduced peptide stability (Lazaro et al. Reference Lazaro, Kadie and Stamegna2011). These results suggest that the chaperone may be important in vivo to mount an immune response to HIV-1 (Fig. 6).

Fig. 6.

Multiple roles of Hsp90 in regulating the immune response against HIV-1. Hsp90 is shown to promote processing of antigenic peptides for loading onto HLA-I, to facilitate sensing of viral DNA and to induce inflammation in response to HIV-1 infection.