Proteostasis in Viral Infection: Unfolding the Complex Virus–Chaperone Interplay

Abstract

Viruses are obligate intracellular parasites that rely on their hosts for protein synthesis, genome replication, and viral particle production. As such, they have evolved mechanisms to divert host resources, including molecular chaperones, facilitate folding and assembly of viral proteins, stabilize complex structures under constant mutational pressure, and modulate signaling pathways to dampen antiviral responses and prevent premature host death. Biogenesis of viral proteins often presents unique challenges to the proteostasis network, as it requires the rapid and orchestrated production of high levels of a limited number of multifunctional, multidomain, and aggregation-prone proteins. To overcome such challenges, viruses interact with the folding machinery not only as clients but also as regulators of chaperone expression, function, and subcellular localization. In this review, we summarize the main types of interactions between viral proteins and chaperones during infection, examine evolutionary aspects of this relationship, and discuss the potential of using chaperone inhibitors as broad-spectrum antivirals.

Most proteins must fold properly before they can perform their functions, and many require assistance of molecular chaperones to achieve a native conformation. Considering the complexity of viral proteins, it is not surprising that they, too, depend on chaperones for proper folding and function. Although some viruses encode their own chaperones, the vast majority—from bacterial, plant, and invertebrate to human viruses—rely on chaperones expressed by the host cell, most notably members of the so-called heat shock protein (HSP) family (e.g., Hsp70, Hsp90, and Hsp60) (Table 1). As clients, viral proteins require chaperones for synthesis, folding, trafficking, and assembly of replication complexes (RCs) and viral particles; as modulators, they regulate the activity and subcellular localization of chaperones, affecting other interactors involved in pathogenesis, immune response, and apoptosis. These complex interactions have likely evolved as a result of the unique features of viral proteins, which are often expressed as multifunctional multidomain precursors. Positive-strand RNA viruses produce a single polyprotein that requires co- and posttranslational processing into mature individual proteins, rendering it prone to misfolding and aggregation (Nagata et al. 1987; Mah et al. 1990; Hung et al. 2002; Geller et al. 2007). Capsid proteins that enclose the viral genome in virions are particularly sensitive to misassembly, as they must be folded into soluble conformations that form structures rigid enough to protect the genome against harsh extracellular environments, yet flexible enough to readily disassemble upon entry into the cell and allow replication (Rossmann 1984). Additionally, the high mutation rate of viruses inevitably leads to frequent emergence of protein variants with compromised function or stability, and chaperones can help buffer the deleterious effects of such mutations (Crowder and Kirkegaard 2005; Lauring et al. 2013).

Table 1.

Major chaperone systems in mammalian cells

Chaperones	Cellular compartment	Major functions	Cofactors and functions
HSP70 system
Hsc70, constitutive (HSPA8)	Cytosol	Folding and stabilization of newly synthesized proteins; assembly and disassembly of multimeric complexes; import into ER and mitochondria	Hsp40s (DnaJs) stimulate Hsp70 ATPase activity; nucleotide exchange factors, for example, Bag1-5, Bap (SIL1), and Grp170 (HYOU1) stimulate ADP release; HSPBP1 inhibits chaperone activity by interfering with ATP binding
Hsp70, inducible (HSPA1A/B)	Cytosol
BiP/Grp78 (HSPA5)	ER
Mortalin (HSPA9)	Mitochondria
HSP90 system
Hsp90 (HSP90AA1/AB1)	Cytosol	Stabilization, maturation, and activation of enzymatic complexes (e.g., kinases, receptors); mediates intracellular signaling	Hop (STIP1) mediates interaction of Hsp70 and Hsp90; p23 (PTGES3), Cdc37 stabilize Hsp90 interactions with clients
Grp94/endoplasmin (HSP90B1)	ER
Chaperonins
TRiC/CCT (TCP1, CCT2-8)	Cytosol	Folding; prevention of aggregation	Prefoldin cofactor guides clients to TRiC/CCT
Hsp60 (HSPD1)	Mitochondria	Folding of proteins imported into mitochondria	Hsp10 (HSPE1) sequesters substrates in Hsp60 cavities
Others
Calnexin (CANX)	ER	Folding and refolding of secretory proteins
Calreticulin (CALR)
Protein disulfide isomerase (PDI)	ER	Rearrangement of disulfide bonds
Peptidyl-prolyl cis–trans isomerase (PPI)	Cytosol, ER, mitochondria	Catalysis of energetically unfavorable cis-to-trans isomerization

ER, Endoplasmic reticulum; Bap, BiP-associated protein.

The rate of protein production in virus-infected cells is another source of pressure on the proteostasis machinery. Rapidly replicating lytic viruses reprogram their hosts to produce large amounts of a small number of viral proteins within a short period of time, likely taxing the capacity of chaperones required to fold them. This may explain why so many viruses induce a shutoff of host translation (Stern-Ginossar et al. 2018): not only to curtail antiviral responses, but also to minimize competition over a limited chaperone pool. In contrast, viruses associated with chronic infection replicate slowly over a long time and may, thus, have lower demand for chaperones. In these, chaperones are still used for folding but also to dampen antiviral responses, suppress premature apoptosis, and remodel the cellular environment to ensure persistent infection. Both infection strategies involve induction of chaperone expression either as a direct consequence of the febrile response (Mayer 2005; Kim and Oglesbee 2012) or because of more selective mechanisms encoded by viral genomes. When unfolded proteins accumulate, Hsp70 (Shi et al. 1998) and Hsp90 (Zou et al. 1998) are titrated away from heat shock factor 1 (HSF1), allowing it to activate chaperone transcription (Kijima et al. 2018). Therefore, one of the ways viruses can induce chaperone expression is simply by mass production of nascent or misfolded proteins. Alternatively, viral proteins can directly activate specific promoters or otherwise regulate HSF1-independent stress signaling. For example, herpes and polyomaviruses express viral proteins that interact with TATA- or CCAAT-binding transcription factors (Lum et al. 1992; Damania et al. 1998), whereas adenoviruses encode for a protein that allows Hsp70 messenger RNA (mRNA) to escape the virus-induced block on nuclear export (Moore et al. 1987).

Taken together, these observations suggest that the biogenesis of viral proteins imposes a significant and uniquely regulated burden on the cellular proteostasis network, highlighting the importance of chaperones in infection. The next few sections review the interplay between viral proteins and chaperones at distinct steps of the replication cycle, from entry and disassembly to synthesis and release of viral particles (summarized in Fig. 1). As cells express multiple constitutive and stress-inducible isoforms of most chaperones, generic family names (e.g., Hsp70) are used herein for the sake of simplicity, unless the mention of a specific isoform is mechanistically or functionally informative (e.g., endoplasmic reticulum [ER]-resident Hsp70, BiP).

Figure 1.

Roles of chaperones in the major steps of the viral replication cycle. (A) Cell surface chaperones interact with viral envelope or capsid proteins and facilitate internalization. Intracellular chaperones can then destabilize the nucleocapsid conformation to release the viral genome. By binding to internal ribosome entry sites (IRESs) or nascent polypeptide chains, chaperones can stimulate translation, prevent aggregation and proteasome-mediated degradation, and facilitate folding into a protease-competent conformation for subsequent processing by viral proteases. Through direct interactions with viral structural or nonstructural proteins, chaperones can maintain an active conformation of reverse transcriptases (RTs) and support nuclear import of viral proteins and genomes. (B) To prevent premature apoptosis of host cells, viruses hijack nuclear and mitochondrial chaperone-modulated prosurvival pathways. Chaperones facilitate transcription and replication by stabilizing viral polymerases and activating promoter regions either directly or indirectly. Some viruses remodel the endoplasmic reticulum (ER) membrane or the nucleoplasm to form replication compartments or specialized virus-induced chaperone-enriched (VICE) domains that serves as hubs of viral protein quality control. Finally, chaperones assist in the multimeric assembly of nucleocapsids and can be packaged into and released with infectious particles.

ROLES OF CHAPERONES IN THE REPLICATION CYCLE

Entry and Disassembly

Most viruses enter host cells through either receptor-mediated endocytosis or plasma membrane fusion (Yamauchi and Helenius 2013). Multiple reports have shown that chaperones can localize to the cell surface and act as coreceptors aiding virus entry, probably by mediating conformational changes in the viral capsid. Nonenveloped double-stranded RNA (dsRNA) viruses of the reovirus family are internalized and disassembled through ATP-dependent interactions between Hsp70 and their capsid proteins. Preincubation of viral particles with recombinant Hsp70 or antibodies against Hsp70 prevented rotavirus internalization (Guerrero et al. 2002; Zarate et al. 2003; Perez-Vargas et al. 2006), and Hsp70 depletion impaired reovirus disassembly into transcriptionally active particles (Ivanovic et al. 2007). In cucumber necrosis virus, Hsp70 was found to directly alter capsid conformation, as determined by electron microscopy (Alam and Rochon 2017). A similar extracellular interaction with Hsp70 was reported for positive-strand RNA viruses of the flavivirus family, for example, hepatitis C virus (HCV) (Parent et al. 2009; Khachatoorian et al. 2014), Japanese encephalitis virus (JEV) (Das et al. 2009), and dengue virus (DENV) (Reyes-Del Valle et al. 2005). DENV also uses Hsp90, BiP, and several DnaJ proteins for internalization, demonstrating that this process is not exclusive to Hsp70 (Reyes-Del Valle et al. 2005; Cabrera-Hernandez et al. 2007; Taguwa et al. 2015).

One of the best-studied examples for chaperone-mediated virion disassembly is that of simian virus 40 (SV40), a nonenveloped dsDNA polyomavirus that is endocytosed and transported intact into the ER. There, it undergoes uncoating to allow a smaller viral particle to enter the cytoplasm and then the nucleus. In the ER, host-induced conformational changes render the virus more hydrophobic and allow its membrane penetration and subsequent translocation. First, capsid protein VP1 pentamers are destabilized by the protein disulfide reductase/isomerase function of DNAJC10/ERdj5, PDIA1/PDI, and PDIA3/ERdj57, exposing the more hydrophobic minor coat proteins (CPs) VP2/3 (Schelhaas et al. 2007; Inoue et al. 2015). Then, BiP and DNAJB11 hold the virus in a translocation-competent state (Goodwin et al. 2011), which is further stabilized by binding to ER membrane complex 1 (EMC1), preventing premature uncoating (Bagchi et al. 2016). The stabilized intermediate interacts with a cytosolic extraction complex comprised of Hsp70 and cochaperone SGT1/SGTA, which is recruited to membrane penetration sites through interactions with DNAJB12, DNAJB14, and DNAJC18 (Bagchi et al. 2015). For translocation to be completed, nucleotide exchange factors (NEFs) HYOU1/Grp170 and Bag2 or Hsp105/HSPH1 release the capsid from BiP (Inoue and Tsai 2015) and Hsp70 (Dupzyk and Tsai 2018), respectively.

After crossing the cellular membrane barrier during cell entry, most animal viruses must undergo further disassembly before initiating viral gene expression. In many cases, these disassembly mechanisms remain poorly defined but likely also involve the proteostasis machinery. For instance, Hsp70 plays a key role in the final step in disassembly of reovirus outer capsid during or soon after membrane penetration. ATP and Hsp70 are required for cytoplasmic release of the central, δ fragment of membrane penetration protein μ1 to yield the transcriptionally active viral core particle to prepare the entering particle for gene expression and replication (Ivanovic et al. 2007).

Nuclear Import

Following internalization and disassembly, nuclear-replicating viruses must deliver viral particles or individual proteins into the nucleus. In influenza, a negative-strand RNA virus, both the RNA-dependent RNA polymerase and viral genome require chaperone-mediated nuclear import. Direct interaction of Hsp90 with polymerase subunits PB1 and PB2 is needed for both import and assembly of the polymerase complex (Momose et al. 2002; Naito et al. 2007). Hsp90 inhibition prevented nuclear relocalization and enhanced the degradation of polymerase subunits (Chase et al. 2008). In addition, viral genome import is supported by Hsp70 cofactor DNAJB1/Hdj1, which bridges between the nucleoprotein and importin α (Batra et al. 2016). By binding to DNAJB1, the nucleoprotein also displaces and activates DNAJC3/P58(IPK), which in turn inhibits dsRNA-dependent protein kinase (PKR), a key regulator of the antiviral immune response (Sharma et al. 2011).

Both Hsp90 and Hsp70 are also involved in the nuclear localization of DNA-dependent DNA polymerases of multiple herpesviruses. Epstein–Barr virus (EBV) BMRF1, herpes simplex (HSV) UL30, and varicella zoster virus (VZV) ORF29p, three proteins involved in viral replication, were shown to depend on Hsp90 for nuclear import (Burch and Weller 2005; Kyratsous and Silverstein 2007; Kawashima et al. 2013). In addition to Hsp90, the latter two require a complex consisting of Hsp70 and cochaperone Bag3 (Li et al. 2004; Burch and Weller 2005; Kyratsous and Silverstein 2007, 2008). In human immunodeficiency virus (HIV), which undergoes cytoplasmic reverse transcription followed by formation of preintegration complexes (PICs), DNAJB6/Hsj2 facilitates nuclear import of PIC through interactions with viral protein Vpx. Overexpression of DNAJB6/Hsj2 enhanced PIC import, whereas depletion or expression of a nuclear-localization mutant reduced it (Cheng et al. 2008). Furthermore, Hsp70 can compete with viral protein Vpr for binding to importin α and either block import in wild-type virus or facilitate it in Vpr-deficient virus (Agostini et al. 2000).

Viral proteins can also drive the nuclear relocalization of chaperones. During early stages of HSV infection, viral protein ICP22 is responsible for the formation of discrete nuclear foci called virus-induced chaperone-enriched (VICE) domains near viral RCs. These chaperone-rich domains consist of protein quality control components, including Hsp90, Hsp70, small HSPs, and active 20S proteasomes, and are thought to facilitate viral replication by sequestering misfolded or unwanted proteins that would otherwise interfere with infection (Mathew et al. 2009; Bastian et al. 2010).

Folding and Assembly of Viral Polymerases and Other Nonstructural Proteins

One of the better-characterized roles of chaperones in viral infection is stabilization and activation of viral polymerase components, influencing both transcription and replication (Fig. 2A). The human papillomavirus (HPV) encodes two proteins, E2 and E1, which bind and unwind the origin of replication (oriP), respectively. E2 recruits E1 but also inhibits unwinding; by displacing E2, Hsp70 and cofactors DNAJA1 and DNAJB1 enhance E1 activity and promote viral replication (Liu et al. 1998; Lin et al. 2002). Similarly, interaction of HSV origin-binding UL9 with Hsp70 and DNAJA3/hTid1 supports its multimeric assembly on oriP (Eom and Lehman 2002; Tanguy Le Gac and Boehmer 2002). In Kaposi's sarcoma–associated herpesvirus (KSHV), Hsp70 inhibition prevented RC formation and recruitment of RNA polymerase II to the viral genome, leading to reduced transcription (Baquero-Pérez and Whitehouse 2015). Hsp70 also maintains the stability of RCs in HCV (Chen et al. 2010) and JEV (Ye et al. 2013) through direct interactions with polymerase components. In measles, Hsp70 acts as a polymerase processivity factor by loosening binding and allowing the polymerase to move to the next template, and its overexpression in mice was associated with higher viral load and mortality rates (Zhang et al. 2005; Carsillo et al. 2006). In HIV, transcription and replication are stimulated by mitochondrial Hsp70 and DNAJB1 through stabilization of the viral Nef protein (Kumar and Mitra 2005; Kumar et al. 2011; Shelton et al. 2012). They are further supported by Hsp70, Hsp90, and the Hsp90 cochaperone Cdc37 through stabilization of CDK9/Cyclin T1, which is required for phosphorylation and activation of transcription by HIV Tat protein (O'Keeffe et al. 2000). Hsp90, Hsp70, and Bag1 were even found to independently interact with HIV and human cytomegalovirus (HCMV) viral promoters, potentially regulating viral transcription (Niyaz et al. 2003; Vozzolo et al. 2010).

Figure 2.

Distinct functions of chaperones in viral replication and capsid assembly. (A) Chaperones (e.g., Hsp70, TRiC/CCT, and Hsp90) assist the folding of nascent viral polymerases into native conformations and prevent their proteasomal degradation. Native polymerases are then guided by chaperones to the appropriate cellular compartments and assembled into a replication complex (RC). Following assembly, some polymerases require chaperones for both activation and continued conformational maintenance. (B) Role of Hsp90 in picornavirus capsid maturation exemplifies chaperone-assisted virion assembly. The poliovirus capsid precursor is bound cotranslationally by Hsp70 and then folded by Hsp90 and cochaperone p23 into a conformation that allows proteolytic processing by the viral protease. Following cleavage, capsid subunits can be assembled in successive steps leading to the mature virion containing the viral genome.

Many viral polymerases depend on Hsp90 and its cochaperones Cdc37 and p23 to prevent their degradation. Hsp90 inhibition resulted in the proteasomal degradation of RNA-dependent RNA polymerase of vesicular stomatitis virus (VSV) and rabies virus (RABV) (Connor et al. 2007; Xu et al. 2016), HCMV (Basha et al. 2005), and RSV (Munday et al. 2015). In contrast, Hsp90 was shown to be required for synthesis, but not stability or activity, of flock house virus (FHV) polymerase (Castorena et al. 2007). In paramyxoviruses, for example, mumps and measles (Katoh et al. 2017), Nipah (Bloyet et al. 2016), human parainfluenza (HPIV), and SV5 (Connor et al. 2007), Hsp90 is only required for stabilization of the large polymerase subunit (L) until it forms a complex with the viral phosphoprotein. When Hsp90 is inhibited, monomeric L is ubiquitinated and targeted for proteasomal degradation by E3 ligase carboxyl terminus Hsp70-interacting protein (CHIP) (Katoh et al. 2017). HCV NS5A, an essential part of the viral RC, is stabilized and recruited to membranes by a complex consisting of Hsp90, peptidyl-proline isomerase (PPI) FKBP8, and human butyrate-induced transcript 1 (hB-ind1/HACD3), which serves as a cochaperone through its p23-like domain (Taguwa et al. 2008, 2009). Both NS5A and polymerase NS5B also depend on cyclophilin PPI activity for maturation of their disordered domains (Hanoulle et al. 2009; Liu et al. 2009; Verdegem et al. 2011). Cyclosporine, a cyclophilin inhibitor, suppressed RNA binding by NS5A (Nag et al. 2012), and blocked the incorporation of NS5B into RCs without affecting its protein levels (Liu et al. 2009).

In the reverse-transcribing hepatitis B (HBV) and duck hepatitis B (DHBV) viruses, chaperones are required for reverse transcriptase (RT) activation. First, Hsp70 and DNAJB1 expose a carboxy-terminal region of the RT, enabling binding to the RNA template (Beck and Nassal 2003; Stahl et al. 2007). This is enough to support in vitro reverse transcription (Beck and Nassal 2003), but activity is further enhanced and maintained by recruitment of Hsp90, p23, the Hsp70-Hsp90 organizing protein Hop (Hu and Seeger 1996; Hu et al. 1997, 2002), Cdc37 (Wang et al. 2002), and Hsp60/HSPD1 (Park and Jung 2001). Components of the Hsp90 complex can remain associated with the polymerase, get packaged, and then released from the host cell inside virions (Hu et al. 1997).

Finally, some viral nonstructural proteins rely on the ring-shaped chaperonin TRiC/CCT for their function. TRiC/CCT is recruited to Negri bodies, sites of transcription and replication in RABV-infected cells, and knockdown of CCT1 and CCT3 inhibits RABV replication (Zhang et al. 2013, 2014). CCT5 was found to mediate TRiC/CCT interactions with nonstructural proteins of EBV (Kashuba et al. 1999) and HCV (Inoue et al. 2011). In influenza, polymerase subunit PB2 associates with TRiC/CCT as a monomer, and silencing of CCT2 resulted in reduced replication, suggesting a function in folding and assembly (Fislova et al. 2010). Furthermore, interaction of HCV F protein with TRiC/CCT cochaperone prefoldin interferes with microtubule organization; because HCV replication requires microtubule polymerization, F protein may limit HCV replication and contribute to viral persistence in chronic infection (Tsao et al. 2006).

In sum, the complex structure of most viral polymerases makes them critically dependent on chaperone assistance for folding, assembly, and activation, as well as conformational maintenance to support continued function. Accordingly, chaperone inhibition results in the misfolding and degradation of many viral polymerases.

Regulated Transition to Replication

Some RNA viruses regulate the transition between translation, transcription, and replication of their genome in a process that involves Hsp70. In RABV, a small regulatory sequence called leader RNA supports transcription by delaying RC formation. Leader RNA function is inhibited by Hsp70, which is initially down-regulated to allow sufficient transcription and later induced to shift the balance toward replication (Lahaye et al. 2012; Zhang et al. 2017). In beet black scorch virus, Hsp70 interacts with viral protein p23 to support RC formation. The viral CP competes with p23 for binding to Hsp70 thus reducing replication rates, suggesting that accumulation of CP serves as an Hsp70-mediated negative feedback loop to limit replication rates (Wang et al. 2018). In the closely related potato virus A, CP binds to viral RNA and blocks replication; phosphorylation by creatine kinase stimulates CP degradation by Hsp70 and CHIP, allowing initiation of replication (Lõhmus et al. 2017).

Synthesis, Folding, and Assembly of Viral Structural Proteins

As mentioned above, chaperones play a key role in the folding and assembly of viral capsids and envelopes, the so-called structural proteins that coat the genome and form the virion. Capsid and envelope monomers must assemble into large regular structures, and often tend to rapidly misfold or oligomerize into insoluble aggregates. In influenza, folding intermediates of the hemagglutinin glycoprotein are first bound by BiP and then transferred to calnexin for continued folding, repair of misfolded intermediates, and prevention of premature oligomerization (Singh et al. 1990; Hogue and Nayak 1992; Tatu et al. 1995). A similar dependence on BiP, and in some cases calnexin/calreticulin, was also reported for glycoproteins of measles and SV5 (Ng et al. 1989; Roux 1990; Watowich et al. 1991), VSV and RABV (de Silva et al. 1990, 1993; Hammond and Helenius 1994), as well as HIV (Earl et al. 1991; Otteken et al. 1996), HCMV (Buchkovich et al. 2008), and rotavirus (Mirazimi and Svensson 2000; Maruri-Avidal et al. 2008). In DENV (Limjindaporn et al. 2009; Wati et al. 2009) and HCV (Dubuisson and Rice 1996; Choukhi et al. 1998; Liberman et al. 1999), BiP, calnexin, and calreticulin levels are induced; calnexin and calreticulin recognize an early glycosylation intermediate of HCV E1/E2 and recruit PDIA3/ERp57 to catalyze disulfide bridge formation, associating and dissociating until the glycoprotein is either properly folded or degraded (Vieyres et al. 2014). BiP and calreticulin also interact with misfolded aggregates of E1/E2 and are likely involved in their repair (Choukhi et al. 1998).

Interestingly, BiP also serves as a switch between the two functions of HBV large envelope glycoprotein (L). HBV, which only replicates in the liver, enters its host cells through interactions between the preS domain of L and cell surface heparin sulfate proteoglycans (HSPGs). However, HSPGs are found on many cell types, and so to minimize nonproductive docking outside the liver, HBV particles are released as a mixed population of both active and inactive forms, with preS either exposed or hidden inside the envelope (Seitz et al. 2016). During translation of L, Hsp70 interacts with the preS domain on the cytosolic side of the ER membrane and actively prevents its translocation while the S domain is cotranslationally inserted into the membrane (Lambert and Prange 2003). BiP is then responsible for the subsequent posttranslational translocation of the preS domain into the lumen, which can be inhibited by overexpression of BiP cochaperones such as DNAJB9/ERdj4 or BiP-associated protein (BAP/SIL1) (Awe et al. 2008). Thus, chaperone action mediates a switch between the two functions of preS: virion assembly in the pretranslocated state or receptor binding in the posttranslocated state (Bruss and Vieluf 1995; Le Seyec et al. 1999). Given that BiP was identified as a liver cell receptor for DENV (Jindadamrongwech et al. 2004), it is tempting to speculate that inactive HBV particles could use cell surface BiP to mature into an HSPG-binding form that supports entry.

While many envelope proteins are clients of BiP, some capsid proteins are folded and assembled by cytosolic Hsp70 (e.g., polyomavirus, SV40, and HCV) (Cripe et al. 1995; Chromy et al. 2003, 2006; Khachatoorian et al. 2014, 2015, 2016). In HCV, Hsp70 also interacts with DNAJA2/Dj3 and NS5A to support translation through stimulation of the viral IRES (Gonzalez et al. 2009). Other DnaJs, however, can have opposite effects on viral proteins; for example, DNAJA3/hTid1 and DNAJB1/Hdj1 accelerate the degradation of HBV core protein, and their overexpression or knockdown resulted in decreased or increased HBV replication, respectively (Sohn et al. 2006). TRiC/CCT, on the other hand, facilitates both folding and stabilization of HBV core (Lingappa et al. 1994) and HIV Vif (Luo et al. 2016). In reovirus infection, TRiC/CCT folds the viral σ3 major outer-capsid protein into a conformation capable of forming the nucleocapsid (Knowlton et al. 2018). Furthermore, biogenesis of reovirus σ1 homotrimeric complex, which mediates binding to host receptors, involves cotranslational ATP-independent trimerization of the amino terminus followed by sequential binding of Hsp70 and Hsp90/p23, which facilitates posttranslational ATP-dependent trimerization of the carboxyl terminus (Leone et al. 1996; Gilmore et al. 1998). The importance of chaperone-mediated assembly for viral transmission cannot be overemphasized, as misassembled virions will not infect cells and may have a dominant negative effect on transmission (Geller et al. 2007; Sokolskaja et al. 2010).

Proteolytic Processing-Competent Folding

For RNA viruses that synthesize a polyprotein precursor, proteolytic processing is intimately linked to protein folding and maturation. Although most viral proteases recognize linear sequence elements, chaperone-mediated folding of the polyprotein is still required to achieve a cleavable conformation. In picornaviruses, the capsid consists of 60 copies of four subunits that are generated by cleavage of their precursor P1. Hsp70, Hsp90, and p23 associate with newly synthesized P1 but not its cleavage products (Macejak and Sarnow 1992; Geller et al. 2007), and Hsp90 inhibition prevented P1 proteolysis without affecting protease activity, suggesting the chaperone complex directs P1 into a proteolysis-competent conformation (Fig. 2B; Geller et al. 2007; Newman et al. 2018). Similarly, Hsp90 and cyclophilin A are required to render HCV polyprotein amenable to cleavage by the viral protease (Waxman et al. 2001; Kaul et al. 2009). DNAJC14/Hdj3 plays a role in formation of yellow fever virus (YFV) RCs by promoting a protease-competent conformation of the viral polyprotein, allowing cleavage between nonstructural proteins NS3 and 4A (Yi et al. 2012; Bozzacco et al. 2016). As such, DNAJC14 is an important regulator of the ratio between cleaved and uncleaved products, which could drive different viral functions. Interestingly, both silencing and overexpression of DNAJC14 inhibits replication of multiple flaviviruses, including HCV, YFV, and DENV (Yi et al. 2011; Taguwa et al. 2015), suggesting a dose-dependent role in maintaining optimal cleavage rates.

Extracellular Chaperones

Mature extracellular virus particles were found to harbor components of the chaperone machinery. Hsp70 was detected in virions of influenza (Sagara and Kawai 1992), Crimean-Congo hemorrhagic fever virus (CCHFV) (Surtees et al. 2016), HIV (Gurer et al. 2002), RABV and VSV (Sagara and Kawai 1992), and ZIKV (Khachatoorian et al. 2018). HCV was shown to use exosomes for receptor-independent transmission; these exosomes contain replication-competent viral RNA associated with Ago2, miR-122, and Hsp90, and their transmission can be suppressed by Hsp90 inhibition. Ago2:miR-122 can support infection by protecting HCV genomes from 5′ exonuclease host mRNA decay, whereas Hsp90 may stabilize the RISC-loading complex (Bukong et al. 2014) and possibly mediate rapid release of viral genomes inside the host. Finally, extracellular release of chaperones, for example, Hsp70 can drive type I interferon (IFN)-dependent antiviral immunity. Mouse neuronal cells infected with measles released Hsp70, which induced IFN-β signaling in microglial cells through toll-like receptors 2 and 4 (Kim et al. 2013).

ANTIAPOPTOTIC SIGNALING

Regulated cell death can have either anti- or proviral effects, depending on the stage of infection. It can be detrimental to virus biogenesis early during infection, but also beneficial for virion release and spread at a later stage when infectious particles have already been formed. Hsp70, Hsp90, and Hsp27/HSPB1 are known to regulate apoptotic signaling. Hsp70 inhibits proapoptotic Bax activation, cytochrome c release, and formation of the apoptosome complex; Hsp90 stabilizes and prevents the degradation of the prosurvival survivin/BIRC5, PI3K/Akt kinase, and NF-κB kinase IKK; and Hsp27/HSPB1 binds and sequesters cytochrome c and caspase-3 and activates Akt (Kennedy et al. 2014; Wang et al. 2014). Viruses can therefore prevent apoptosis by inducing chaperone expression and modulating their function in apoptotic signaling. Early in HIV infection, expression of Hsp70 and Hsp27/HSPB1 is up-regulated and the proteins interact with and suppress Vpr nuclear import, thus, preventing premature Vpr-induced cell cycle arrest and apoptosis (Iordanskiy et al. 2004; Liang et al. 2007). Later, expression of Hsp70 is down-regulated to support viral spread through apoptosis (Kumar et al. 2011). In another retrovirus, human T-lymphotropic virus (HTLV), apoptosis is blocked by viral protein Tax, which binds and sequesters proapoptotic DNAJA3/hTid1 in the cytoplasm, preventing its infection-induced relocalization to mitochondria (Cheng et al. 2001). HIV and HTLV reactivation from latency also requires Hsp90 and Cdc37; Hsp90 inhibition in transformed cells destabilized IKK and resulted in apoptosis (Gao and Harhaj 2013; Anderson et al. 2014). In EBV, Hsp90 inhibitors both prevented transformation and induced apoptosis of transformed cells by reducing the amounts of EBNA1, the only viral protein expressed during latency (Sun et al. 2010). Hsp90 and DNAJB11/ERdj3 also promote antiapoptotic signaling in KSHV by stabilizing glycoprotein K1 (Wen and Damania 2010) and facilitating complex formation between the viral FLIP protein and IKK (Field et al. 2003), thus, stimulating PI3K/Akt and NF-κB survival signaling, respectively. HCV E2 was shown to block apoptosis and facilitate persistent infection by inducing the overexpression of ER-resident Hsp90 (Grp94/HSP90B1), which can also activate NF-κB. Overexpression of Grp94/HSP90B1 resulted in increased expression of antiapoptotic proteins, whereas knockdown abrogated E2-induced prosurvival activity (Lee et al. 2008). Interactions of HCV core protein with either the mitochondrial chaperone prohibitin, which is induced during infection, or Hsp60, led to inhibition of mitochondrial function, triggered the production of reactive oxygen species, and enhanced TNF-α-mediated apoptosis (Kang et al. 2009; Tsutsumi et al. 2009; Fujinaga et al. 2011). A similar effect was reported for HBV in which Hsp60 colocalized with HBx protein in the mitochondria and enhanced HBx-induced apoptosis (Tanaka et al. 2004).

VIRUS-ENCODED CHAPERONE-LIKE PROTEINS

Although the vast majority of viruses rely on their ability to induce and compete for host chaperones, some have evolved to encode their own specialized viral chaperones. It is thought that by providing their own chaperones, viruses can better control and adapt their function and improve the recruitment of other host components (Mayer 2005). For example, bacteriophages have been shown to repurpose the host GroEL/Hsp60 chaperonin system by encoding a GroES-like cofactor protein that forms a complex with host GroEL to promote capsid protein folding (Ang et al. 2000); others even encode a stand-alone chaperonin-like protein required to fold a phage protein (Kurochkina et al. 2012; Molugu et al. 2016). Members of the plant closterovirus family encode Hsp70-like proteins, which share very little sequence homology with that of their hosts, and are essential for both virion assembly and microtubule-mediated cell-to-cell transmission (Verchot 2012). SV40 large T antigen (TAg) contains an amino-terminal J-domain that binds Hsp70, stimulates its ATPase activity, and is thought to confer specificity to viral targets (Spence and Pipas 1994; Wright et al. 2009). HSV also encodes two J-domain proteins: UL14, which is required for the maturation and efficient nuclear transport of viral protein VP16 and capsid proteins (Yamauchi et al. 2002, 2008; Ohta et al. 2011), and ICP10PK, an Hsp27/HSPB1 homolog that stabilizes Bcl2, up-regulates the expression of Hsp70, Hsp27, and Bag1, and enhances the phosphorylation of TGF-β-activated kinase 1 (TAK1), leading to activation of the extracellular signal-regulated kinase (ERK) survival pathway (Chabaud et al. 2003; Gober et al. 2005; Aurelian et al. 2012). Finally, some viruses express proteins that show no homology with known chaperones, such as the African swine fever virus (ASFV), whose capsid-associated protein 80 (CAP80) facilitates cotranslational folding of the aggregation-prone capsid protein p73 (Cobbold et al. 2001). Similarly, the closely related Nipah and chandipura viruses both encode a phosphoprotein that binds the viral nucleoprotein, suppresses its aggregation, and maintains a conformation that accommodates RNA binding (Majumdar et al. 2004; Yabukarski et al. 2014).

EVOLUTIONARY ASPECTS OF VIRUS–CHAPERONE INTERACTIONS

Viruses are characterized by high mutation rates, which fuel their evolutionary plasticity and adaptation to varying environments and host pressures (Lauring et al. 2013). However, most mutations destabilize protein folding and structure (Tokuriki and Tawfik 2009) and may, therefore, compromise viral evolution. Chaperones, which recognize nonnative protein conformations, can have a significant impact on organismal fitness as buffers of phenotypic variation. In the event of destabilizing mutations leading to misfolding or aggregation, chaperones can recognize the misfolded state, suppress aggregation, and either facilitate refolding or promote protein clearance (Hartl et al. 2011). Therefore, chaperones can improve structural and functional plasticity and support expanded complexity such that is present in numerous viral proteins. Genetic and pharmacological perturbations of chaperone function have shed light on the possible roles of chaperones in evolution. In bacteria, proteins that depend on chaperonins for folding evolve faster than chaperonin-independent proteins through buffering of deleterious misfolding-related mutations (Bogumil and Dagan 2010). Endogenous up-regulation or ectopic overexpression of bacterial chaperonins can also restore fitness under conditions of increased mutational load (Fares et al. 2002; Maisnier-Patin et al. 2005). In contrast, inhibition of Hsp90 in Drosophila (Rutherford and Lindquist 1998) and Arabidopsis (Queitsch et al. 2002) led to the emergence of detrimental phenotypic variation that was encoded in the genome but masked by chaperone activity.

During viral infection, when adaptation and survival depend on population diversity, chaperones are thought to play a fundamental role. The progeny population of highly mutable RNA viruses always consists of an ensemble of closely related genotypes called “quasispecies,” which enable survival under constantly changing environmental pressures (Vignuzzi et al. 2006). It is, therefore, likely that acquisition of chaperone dependence allows viral genomes to explore a broader sequence space. This may be particularly beneficial for viral capsid proteins that must consistently mutate surface-exposed epitopes because of immune surveillance pressure. Indeed, a computational analysis of HIV evolution showed that mutations related to immune evasion and drug resistance tend to destabilize viral proteins (Olabode et al. 2017). Furthermore, a study of influenza revealed that a common mutation conferring resistance to a host restriction factor loses its fitness benefits when chaperone induction is inhibited at febrile temperatures (Phillips et al. 2018). These results confirm that viruses use chaperones to uncouple the deleterious effects of mutations from the benefits of immune evasion.

In poliovirus, Hsp90 was recently shown to facilitate the emergence of sequence variants with increased hydrophobicity, potentially leading to both higher stability and aggregation propensity (Fig. 3). As such, Hsp90 mediates the trade-off between stability and aggregation by maintaining solubility of these variants. Inhibition of Hsp90 was not only associated with accumulation of variants that reduced aggregation at the expense of stability, it also promoted local codon deoptimization, leading to slower translation and allowing more time for either chaperone-independent or chaperone-mediated cotranslational folding. In HCV, inhibition of cyclophilin A, which is thought to function cotranslationally, was associated with emergence of mutations that slow down the kinetics of polyprotein processing (Kaul et al. 2009). Another study showed that fixation of nonsynonymous mutations in influenza was slower with Hsp90 inhibition and faster with elevated chaperone activity induced by HSF1 activation. Although moderately destabilizing mutations in the polymerase were tolerated regardless of the proteostatic intervention, strongly destabilizing mutations were only observed when chaperone levels were induced by HSF1 activation. Interestingly, some polymerase mutations rendered the virus significantly more fit with Hsp90 inhibition and less fit with HSF1 activation, suggesting that chaperones can increase the fitness cost of specific mutations (Phillips et al. 2017). These experiments indicate that chaperones profoundly affect viral evolution and adaptation by modulating trade-offs between protein stability and aggregation as well as translation rate and cotranslational processing (Geller et al. 2018).

Figure 3.

Role of Hsp90 in shaping poliovirus evolution. The low fidelity viral polymerases continuously generate sequence variants in the viral population. The proteostasis machinery modulates the fitness of these sequence variants and thus affects the direction of viral evolution. In poliovirus, Hsp90 modulates the energy landscape of protein folding and balances trade-offs between protein stability and aggregation for the capsid protein variants. By protecting against aggregation, Hsp90 allows the emergence of more stable variants because their higher hydrophobic character has the drawback of increased aggregation propensity.

CHAPERONES AS BROAD-SPECTRIUM ANTIVIRAL TARGETS

Viral infections are a major cause of human morbidity and mortality, imposing a heavy economic burden worldwide. Despite major efforts, only a handful of antivirals are currently available, largely because of the rapid acquisition of resistance to drugs directed against viral enzymes. Any drug targeting a protein that is under the replicative control of a virus may be rendered ineffective through mutations that easily and frequently generate escape mutants. Therefore, targeting host factors involved in proteostasis may be an attractive alternative, given their widespread requirement for the replication of many, if not most, viruses. The unique constraints on chaperone-mediated folding in infected cells could render viruses hypersensitive to inhibitors of these chaperones with relatively low toxicity to the host. Because numerous viruses rely on various chaperones, as described above, such inhibitors have the potential to become broad-spectrum antivirals.

Both Hsp90 and Hsp70 are highly druggable, with multiple structurally diverse inhibitors identified to date showing therapeutic potential in cancer and other pathologies (Neckers and Workman 2012; Kumar et al. 2016). As of 2017, several Hsp90 and Hsp70 inhibitors were being evaluated in clinical trials, but a first-in-class drug has yet to be approved (Chatterjee and Burns 2017). Small molecule Hsp90 inhibitors geldanamycin (GA) and 17-AAG show antiviral activity in tissue cultures against a diverse range of viruses, including herpesviruses (HSV, KSHV, and EBV), arboviruses (HCV, chikungunya virus [CHIKV]), retrotranscribing viruses (HBV and HIV) and influenza, and in animals against poliovirus and HCV (reviewed in Geller et al. 2012; Wang et al. 2017). In the case of poliovirus, development of drug resistance was observed for most antivirals tested but never for Hsp90 inhibitors (Geller et al. 2007). Allosteric Hsp70 inhibitors, such as JG40 and HS-72, were also shown to potently block infection with flaviviruses, for example, DENV, YFV, and JEV in tissue cultures, without host toxicity or emergence of drug resistance. While HS-72 was found to inhibit DENV entry through disruption of Hsp70 association with the DENV receptor complex (Howe et al. 2016), JG40 was found to affect multiple steps of the virus life cycle, including entry, replication, and assembly (Taguwa et al. 2015). JG40 also suppressed the production of proinflammatory cytokines and chemokines, which contribute to the pathology of some viral infections.

Members of the oxidative folding pathway, PPIs (cyclophilins), and protein disulfide isomerases (PDIs) have also been implicated in multiple steps of the flavivirus life cycle, including RC formation, polyprotein processing, envelope protein folding, and virion assembly (Kaul et al. 2009; Weng et al. 2012; Langsjoen et al. 2017). Cyclophilin inhibitors, including cyclosporine and its nonimmunosuppressive derivatives, inhibit HCV infection both in tissue cultures and animals by blocking the interaction between NS5A and viral RNA (Nag et al. 2012). Similarly, cyclosporine blocks the incorporation of HIV envelope protein into virions (Sokolskaja et al. 2010), and its therapeutic potential is currently being evaluated in clinical trials (Nicolas et al. 2017). Direct and indirect PDI inhibitors can also inhibit folding and assembly of the envelope proteins of CHIKV and ZIKV (Langsjoen et al. 2017). The inability of viruses to elicit resistance to drugs targeting chaperones, together with their favorable therapeutic window, low toxicity, and efficacy against broad families of related viruses provide hope that these types of compounds could be used both therapeutically and prophylactically against acute viral infections and emerging epidemics.

CONCLUDING REMARKS

Essentially, all steps of the viral replication cycle depend on interactions between chaperones and viral proteins. Studying the intricacies of viral proteostasis continues to illuminate basic principles of chaperone function in both healthy and infected cells, and may lead to novel types of resistance-free antiviral drugs. Considering their panviral activity, involvement in multiple steps throughout the viral life cycle, favorable safety and tolerability profiles, and lack of observed drug resistance, chaperone inhibitors could become a valuable tool in our fight against the global burden of viral infection.