Sending Mixed Signals: Polyomavirus Entry and Trafficking

Abstract

Polyomaviruses are mostly non-pathogenic, yet some can cause human disease especially under conditions of immunosuppression, including JC, BK, and Merkel cell polyomaviruses. Direct interactions between viruses and the host early during infection dictate the outcome of disease, many of which remain enigmatic. However, significant work in recent years has contributed to our understanding of how this virus family establishes an infection, largely due to advances made for animal polyomaviruses murine and SV40. Here we summarize the major findings that have contributed to our understanding of polyomavirus entry, trafficking, disassembly, signaling, and immune evasion during the infectious process and highlight major unknowns in these processes that are open areas of study.

Graphical Abstract

Introduction

Polyomaviruses (PyVs) successfully overcome cellular barriers and defenses to cause infection. This entails localizing to and usurping cellular attachment and entry receptors and hijacking intracellular signaling networks. All of these events are necessary for proper arrival of PyVs at destinations within the cell to disassemble the virion and transcribe and replicate the viral DNA genome [1]. In doing so, PyVs select the cellular factors critical for accomplishing this arduous task while also avoiding detection by the host immune system, as activation of an immune response results in the clearance of viral infection. Largely, PyVs accomplish this challenging task as most human PyVs establish a persistent infection in the majority of the population [2].

Of the 14 polyomaviruses of human origin, four have been associated with human disease and are of particular interest: polyomaviruses BK (BKPyV), affiliated with nephropathy and hemorrhagic cystitis [3], JC (JCPyV), the cause of progressive multifocal leukoencephalopathy (PML) [4], Trichodysplasia spinulosa (TSPyV), associated with Trichodysplasia spinulosa [5], and Merkel cell (MCPyV), capable of causing Merkel cell carcinoma (MCC) [6]. It’s important to note that this is a subset of known human polyomaviruses; further investigation is required to understand the association of other human PyVs and disease. However, many advancements in our understanding of these human viruses are attributed to research findings for murine (mPyV) and simian virus 40 (SV40) polyomaviruses. Discussed in this review are the commonalities and striking differences in the infectious entry of BKPyV, JCPyV, MCPyV, mPyV, and SV40. Identifying the similarities and differences in the initiation of the infectious cycle is critical to the development of antiviral therapies to prevent polyomavirus-associated disease.

Polyomavirus Attachment

Most polyomaviruses have been demonstrated to engage gangliosides for attachment to the cell surface [7–11]. Gangliosides are glycosphingolipids with a ceramide anchored in the lipid-rich cell membrane and an oligosaccharide chain with one or more sialic acids (SA) conducive for polyomavirus binding, including α2,3, α2,6, and α2,8 conformations [12–18]. Interestingly, most gangliosides and glycoproteins expressed in humans terminate in a N-acetylneuraminic acid (Neu5Ac)-SA linkage [19]. The Neu5Ac moiety serves as the chief recognition site for polyomavirus-glycan interactions [20]. Although SA branching patterns differ, the surface of PyV viral protein 1 (VP1) exhibits plasticity, accommodating differing Neu5Ac-SA linkages for different PyVs [19]. Exploration of interactions between polyomaviruses and host cells identified gangliosides GT1a, GD1a, and GT1b as attachment factors for mPyV and ganglioside GM1 for SV40 [7–9, 11, 21]. Ganglioside GT1b serves as an attachment factor for MCPyV and BKPyV, while BKPyV can also bind GD1b, and b-series gangliosides GD2 and GD3 [22, 23]. JCPyV has also been shown to bind gangliosides [15, 23–26], though attachment occurs with low affinity and ganglioside expression does not impact viral infection [27]. The sequence homology among VP1 proteins of PyVs (74% homology for JCPyV, BKPyV, and SV40) is key to the similarities in PyV engagement of glycan structures with terminal SAs for attachment [19]. Although MHCI has been reported to serve as an alternative receptor for SV40 [28–31], X-ray crystallography and complementary techniques have identified GM1 as the primary receptor for SV40 [11, 21]. mPyV has been demonstrated to also bind α4 integrins following ganglioside engagement [32]. Interaction with GM1 and α4 integrins activates downstream signaling mechanisms important for infection [33].

Aside from gangliosides, PyVs have been identified to interact with other cell surface molecules including glycoproteins and glycosaminoglycans (GAGs). Both MCPyV and JCPyV can interact with GAGs [34, 35], while JCPyV requires α2,6 SA-containing lactoseries tetrasaccharide c (LSTc) [18, 27], and MCPyV engages an additional sialylated co-factor [34]. Interestingly, several studies suggest that JCPyV may use alternative attachment factors, as LSTc expression across permissive cell types remains unclear [36]. Recently, Adipocyte Plasma Membrane Associated Protein (APMAP), an N-linked glycosylated cellular receptor widely distributed throughout the brain, was identified as a co-factor for JCPyV infection of glial cells, though the importance of this receptor for steps of the viral lifecycle including attachment and/or entry is unclear [37]. Collectively, interactions with cellular receptors or co-receptors used by PyVs are essential for virus entry, activate signaling networks critical for infection, and often determine host and tissue tropism [1, 38].

Polyomavirus Entry

With the exception of JCPyV, all polyomaviruses, for which internalization has been investigated, enter host cells through caveolin-mediated endocytosis (Cav-ME) (Figure 1) [39–44]. Polyomaviruses induce the formation of flask-shaped, lipid raft-rich invaginations [45]. These structures are stabilized by caveolins, most often caveolin-1, and are anchored by actin [45, 46]. Internalization of SV40 by Cav-ME leads to deposition within an early endosome (EE) [29, 39, 47, 48]. Utilization of Cav-ME by PyVs has further been described for BKPyV, mPyV, and MCPyV [40, 49–53]. However, in some permissive cell types, BKPyV entry occurs independently of caveolins [52]. Caveolin-independent internalization has also been described for SV40, albeit in murine embryonic fibroblasts (MEFs), an abortive cell model [42]. Activation of Cav-ME for entry requires intracellular cholesterol, receptor tyrosine kinase (RTK) activity, and cellular GTPases and phosphatases [53–56].

Fig. 1. Cellular internalization and trafficking by polyomaviruses.

Polyomavirus (PyV) activation of host cell receptors results in the internalization of the viral particle by either caveolin-dependent or independent endocytosis (SV40, BKPyV, mPyV, and MCPyV, dark blue virions), clathrin-mediated endocytosis (JCPyV, green virion), or may occur through macropinocytosis of extracellular vesicles (BKPyV and JCPyV, gold virions). Caveolin-dependent endocytosis involves the activation of either integrins or gangliosides (PyV specific) in lipid raft-rich regions (yellow), resulting in the recruitment of caveolin, cavins, receptor tyrosine kinase activity, and the utilization of cellular phosphatases. Clathrin-mediated endocytosis (CME) of JCPyV requires endocytic proteins clathrin, AP2, dynamin, and β-arrestin. CME occurs following JCPyV activation of cellular receptors, LSTc and/or GAGs embedded in lipid raft-rich regions (yellow), and is mediated by the 5-HT subfamily 2 serotonin receptors (5-HT₂Rs). Macropinocytosis of PyVs occurs independent of attachment factors. Upon activation of entry pathways, PyVs are deposited into early endosomes, trafficked through the endosomal-lysosomal system, and then enter the endoplasmic reticulum (ER), where conformational capsid changes occur. Following conformation changes within the ER, PyVs are transported through the cytosol to the nucleus for replication (light blue virions). Figure created with BioRender.com.

Distinct from other PyVs, JCPyV does not use Cav-ME, instead requiring clathrin-mediated endocytosis (CME) for cellular uptake (Figure 1) [57–59]. Hijacking many of the key players involved in CME including clathrin, β-arrestin, AP2, and dynamin, JCPyV is likely internalized into an early endosome [59]. It is speculated that this dichotomy in JCPyV dependence on CME for entry stems from secondary receptor utilization. JCPyV entry requires the 5-hydroxytryptamine (5-HT) subtype 2 receptor family for internalization, yet expression of 5-HT₂Rs does not impact viral attachment [59–62]. Moreover, the critical nature of β-arrestin in JCPyV utilization of 5-HT₂Rs for entry has been well described; inhibition of β-arrestin or mutagenesis of β-arrestin-binding domains within 5-HT₂Rs impairs entry and infection [59, 62].

In recent years reports of alternative, receptor-independent entry, through the utilization of extracellular vesicles (EVs), has gained traction. This mechanism involves the transport of multiple infectious particles or genomes in membranous extracellular vehicles [63]. Utilized by RNA and DNA viruses alike [64], EVs can be advantageous for infection in multiple ways. For instance, (1) packaging in an extracellular vesicle could allow for circumventing immune system detection, and (2) entry of EVs appears to occur independently of cellular attachment and entry factors, thereby providing an alternative mechanism to broaden tissue tropism [64–66]. EVs have been described for BKPyV and JCPyV (Figure 1), each involving the packaging of whole particles into vesicles derived from either the plasma membrane or multivesicular bodies [67, 68]. JCPyV usurpation of this transport mechanism utilizes an alternative cellular uptake pathway, macropinocytosis, suggesting that both mechanisms may occur simultaneously in the infected host [68]. Although EVs have not been illustrated for other PyVs, it is expected that similar reports could emerge in the future given the many parallels in PyV entry.

Trafficking of Polyomaviruses to the Endoplasmic Reticulum

Before reaching the nucleus, PyVs traffic to the ER with the assistance of microtubules, aided by motor proteins dynein and kinesin [69–71]. Interactions with SA-containing receptors directs MCPyV, mPyV, and SV40 to the ER [53]. Moreover, ganglioside GD1a sorts mPyV from endolysosomes to the ER, suggesting a role for ganglioside binding as a general ER-targeting mechanism [72]. Both JCPyV and SV40 rely on COPI-mediated transport [73, 74]. SV40 was previously thought to traffic through caveosomes to the ER [75]. However, increasing evidence suggests that caveosomes are not independent structures but rather modified late endosomes or endolysosomes [48].

MCPyV, BKPyV, mPyV, SV40, and likely, JCPyV, enter EEs shortly after internalization (Figure 1) [48, 53, 76]. Interestingly, MCPyV acquires a membrane envelope within endosomes – a novel finding among PyVs [53]. While the majority of mPyV enters EEs to then be carried by recycling endosomes to the ER [72, 76], two populations of EE exist [77] – static EE and rapidly maturing endosomes, dynamic EEs. It is proposed that sorting of mPyV to dynamic EEs may represent the infectious route, while mPyV within static EE might result in nonproductive infection [72]. Although it is well-accepted that acidification of endosomes and lysosomes is due to translocation of H⁺, Ca²⁺ channels are also essential in MCPyV and SV40 infections, with MCPyV additionally requiring K⁺ flux [78]. Although JCPyV utilizes CME to gain entry, the virus traffics through caveolin-1-dependent vesicles [57, 59, 79]. This trafficking pattern is still poorly understood. However, caveolae-derived vesicles may dock onto the EE, releasing cargo into the EE in a pH-dependent, bidirectional manner [80], representing crosstalk between two transport mechanisms thought to be distinct for JCPyV trafficking [79]. Polyomaviruses discussed herein also accumulate within late endosomes and most are reliant on low pH, before arrival at the ER [49, 53, 70, 71]. Although SV40 also traffics to late endosomes [81, 82], infection occurs without a requirement for low pH [71]. Cystic fibrosis transmembrane conductance regulator (CFTR) has also been linked to BKPyV trafficking, in line with previous implications of CFTR in endosome trafficking and fusion [83].

Following trafficking of PyVs within the endosomal-lysosomal system, PyVs are deposited within the ER, a process that is not well understood and likely complex. As described for mPyV, ganglioside GD1a may sort mPyV from endolysosomes to the ER [72]. There are two proposed mechanisms for this occurrence; clustering of GD1a may cause formation of a hydrophobic platform within the bilayer, recruiting cytoplasmic factors that mediate budding from the endosomal membrane, or clustering may physically alter the membrane beneath the virus, resulting in membrane invagination and budding [72]. Ganglioside-mediated sorting has been demonstrated for BKPyV as well [84]. Importantly, though undescribed, a similar phenomenon may occur with non-ganglioside, SA-dependent, receptor sorting of other PyVs to the ER.

Polyomavirus Hijacking of the Endoplasmic Reticulum and Nuclear Arrival

PyV trafficking to the ER is critical for progression of the infectious cycle. BKPyV infection required syntaxin18, a SNARE (“SNAP REceptor”) protein, and likely uses this SNARE to enter the ER via a vesicle fusion event [84]. A vesicle containing BKPyV, Rab18, and an unidentified v-SNARE buds from the endosome, allowing Rab18 to interact with ZW10 of the NRZ tethering complex located on the ER [84]. Post-tethering, interactions between ER-bound syntaxin 18 and v-SNARE mediate vesicle fusion [84]. SV40 docking at the ER also utilizes syntaxin18, along with ER membrane protein complex (EMC) subunits EMC4 and EMC7 [82]. These EMC subunits directly interact with Rab7, presumably to dock the LE at the ER for stabilized transfer of SV40 [82]. The requirement of vesicle docking machinery for other PyVs is poorly understood.

Following deposition within the ER, PyVs undergo conformational changes, facilitated by ER-resident redox proteins (reviewed in [85]). PyVs usurp PDI, ERdj5, and ERp57 (SV40, JCPyV), ERp29 (mPyV, JCPyV), or ERp72 (JCPyV), thereby acting as chaperones for capsid destabilization (Figure 2) [86–89]. Although MCPyV arrives in the ER, similar to other PyVs studied, the necessity of the redox environment is unclear [53].

Fig. 2. Polyomavirus disassembly within the ER and nuclear localization.

(1) Upon arrival of PyVs (dark blue virions) within the ER, conformational capsid changes of PyVs occurs through the utilization of endoplasmic reticulum-associated degradation (ERAD) machinery. Partial disruption of the capsid results in the exposure of hydrophobic minor capsid proteins. The hydrophobic particle is embedded within the ER membrane and extracted into the cytosol by ER- and cytosolic-chaperones including derlins, sel, BAP31, BiP, EMC1, and DNA J proteins (the necessity of these proteins may differ among the specific PyVs). (2) After arrival in the cytosol PyV interacts with Golgi apparatus membrane-associated cargo adaptor subunits BICD2 and BICDR1 for disassembly of the viral capsid (light blue virions). (3) The disassembled PyV is maintained in the cytosol by chaperones Hsc70, Hsc105, SGTA, and Bag2. (4) The particle enters the nucleus, assisted by importins α/β, which recognize the nuclear competent particle through the exposed nuclear localization sequences (NLS) of the capsid components. (5) Within the nucleus, replication and transcription of the viral genome occur. Figure created with BioRender.com.

While export of PyVs from the ER into the cytoplasm largely remains enigmatic, key components of this process have been described for SV40, likely serving as a platform for the investigation of ER usurpation by other PyVs. Disassembly of SV40 in the ER-redox environment results in a destabilized hydrophobic particle [86]. The hydrophobic capsid protein VP2/VP3 is then embedded in the ER membrane via the N-terminus, similar to that of misfolded proteins [90–92]. This is likely attributed to a cascade of events that occur in rapid succession: (1) proteins affiliated with the export of misfolded proteins from the ER, BAP31 and BiP, (ER associated degradation proteins) are recruited to the site of penetration [93]; (2) SV40 interacts with the membrane chaperone EMC1 preventing further capsid disassembly [94]; (3) DNA J proteins are recruited, including B12, B14, and C18, and membrane proteins derlin1 (SV40, BKPyV, JCPyV) [74, 95–98], derlin 2 (mPyV) [99], or sel1 (JCPyV) [74]; (4) these proteins function in concert with cytosolic chaperones Hsc70, Hsc105, Bag2, and SGTA, to extract SV40 through the ER membrane into the cytosol (Figure 2) [90, 100, 101]. Once in the cytoplasm, interactions between SV40 and cytosolic chaperones likely prevent targeting of SV40 to the proteasomes for degradation. It has been reported that SV40 requires molecular motor dynein-1 [102], which was originally thought to contribute to cytosolic disassembly through mechanical stress. However, it is now appreciated that SV40 disassembly in the cytosol occurs through Golgi-associated dynein cargo adaptors bicaudal-D (BICD) BICD2 and BICDR1, which directly engage cytosolic SV40 resulting in disassembly, independent of dynein-1 function [103]. Therefore, the requirement of dynein-1 for SV40 infection remains poorly understood, though dynein-1 may be necessary for intracellular localization of dynein BICD cargo proteins [103]. However, aside from BICD adaptors, Hsc70-dependent disassembly has also been described for SV40 and mPyV [104].

Although PyVs have undergone significant destabilization of the particle within the ER and cytoplasm, the virus is still too large to utilize passive diffusion into the nucleus [105]. Localization of these particles within the nucleus thereby relies on nuclear transport machinery, importins [105]. Exposed within the ER, nuclear localization sequences (NLS) of VP1, 2, and 3 are recognized by importins within the cytoplasm, as demonstrated for SV40, BKPyV, and JCPyV (Figure 2) [106–110]. Although direct ER-to-nucleus transport has been proposed [111], this possibility lacks sufficient evidentiary support; it is thought that exposed VP2 and VP3 may act as viroporins on the inner leaflet of the nuclear membrane [112, 113]. However, numerous reports describe the necessity of PyV arrival in the cytoplasm prior to nuclear localization [96–98, 100, 101, 114]. Aside from the requirement of importins, the nuclear transport of PyVs is not well characterized, and as a critical component in the lifecycle of DNA viruses, this warrants further exploration.

Hijacking of Intracellular Signaling during Polyomavirus Entry

Viruses can activate multiple signaling cascades to drive specific steps of the viral lifecycle [1]. Polyomaviruses have been shown to activate or utilize an array of signaling pathways including the phosphoinositide 3-kinase (PI3K) (mPyV, MCPyV, and JCPyV), FAK/SRC (mPyV) and mitogen-activated protein kinase (MAPK) (mPyV, MCPyV, SV40, BKPyV) pathways. PyVs may differentially usurp MAPK signaling during infection. Although mPyV interaction with cellular receptors promotes the activation of MAPK, this pathway is dispensable for infection [33, 115, 116]. Meanwhile, JCPyV, BKPyV, and SV40 require MAPK signaling for infection, likely executing different functions in the viral lifecycle [116–119].

A unique role for the MAPK cascade has been demonstrated for PyVs [120]. JCPyV induces the MAPK-extracellular regulated kinase (ERK) cascade immediately following entry in immortalized glial cells, though ERK is not required for viral internalization [119, 121]. Viral activation of this pathway leads to the differential expression of numerous genes downstream of the MAPK pathway [122], including transcription factors (TFs) NFAT4, c-MYC, and SMAD4, that support JCPyV replication [121, 123]. Upon activation, NFAT is translocated to the nucleus by the phosphatase, PP2B [124]. Sustained phosphorylation of this transcription factor can be accomplished by various cellular signaling cascades, including MAPK [124, 125]. During JCPyV infection, nuclear translocation is required to upregulate both early and late gene transcription for successful viral infection [123]. Similarly, BKPyV also activates the MAPK-ERK cascade and requires NFAT4 for viral replication [117, 118]. On the contrary, not all polyomaviruses trigger this canonical signaling pathway [126]. MAPK-ERK signaling is not required for SV40 or MCPyV replication [126, 127]. SV40 instead activates NFAT4 independently of MAPK for replication [128], as well as induces JNK phosphorylation, needed for virus release [116]. It is thought that activation of signaling networks during infection is intimately linked to receptor specificity for attachment and entry [38]. However, the implications of the dichotomy in PyV attachment and entry and subsequent activation of signaling pathways, including the MAPK network, are not well characterized.

Polyomaviruses have also been demonstrated to differentially utilize other signaling pathways such as the PI3K-AKT pathway. mPyV binding to GD1a and GT1a enhances PI3K signaling during endocytosis [33]. In addition to PI3K, mPyV also activates FAK/SRC signaling to allow for mPyV trafficking along microtubules, yet, FAK/SRC are not required for mPyV internalization [33]. It was recently demonstrated that inhibition of PI3Kγ and regulatory protein PIK3R5 reduces JCPyV and JC pseudovirus infection of glial cells as well as BKPyV, MCPyV, and SV40 pseudovirus infection [129]. PI3Kγ inhibition did not affect JCPyV viral attachment or internalization, yet a possible connection between engagement of viral receptors and the dissemination of intracellular signaling to mediate trafficking has been hypothesized and requires further study [129]. In addition to virus-receptor signaling cascades, further signaling networks can be activated through immune detection of viruses [1].

Subversion of the Immune Response by Polyomaviruses During Entry

PyVs have evolved strategies to circumvent the host immune response to infection. On a cellular level, PyVs employ tactics to evade immune detection. While trafficking within endosomes, MCPyV acquires a membrane-derived envelope [53]. While the nature of this finding has not been determined, it is thought that this may contribute to evasion of the immune system early in infection [53]. In addition, MCPyV infection results in downregulation of MHCI expression on the cell surface, reducing antigen presentation during infection [130]. Upon arrival in the ER for uncoating, the hydrophobic inner capsid components of PyVs are exposed [74, 90]; detection of those hydrophobic components may activate cellular unfolded protein response pathways (UPR) and ER stress [131]. As a result, ER stress can activate TFs that induce an inflammatory response through the production of pro-inflammatory cytokines [132], likely negatively influencing progression of infection. To avoid this fate, PyV hydrophobic capsid proteins are embedded in the membrane of the ER, preventing activation of UPR [90]. Additionally, PyV use of EVs as a potential mechanism of immune avoidance must be considered [66–68]; PyVs may hijack EVs for viral persistence within the host as well as immune system avoidance [63]. The release of infectious particles contained within EVs may avoid cellular lysis, which could otherwise trigger clearance of the infection [66].

Conclusions and Perspectives

Investigation of mechanisms by which viruses hijack host cells is critical for the development of targeted antivirals. Many similarities can be illuminated from the comparison of PyV entry; aside from use of EVs, all PyVs undergo receptor-mediated endocytosis, traffic within the endosomal-lysosomal system, and are deposited in the ER, resulting in nuclear translocation, all while subverting the immune system. However, the differences that occur between PyVs in this process cannot be overlooked. Regardless of the high degree of conservation, for the most part, PyVs engage distinct cellular receptors. Utilization of these receptors activates select intracellular signaling pathways likely tailored to individual viruses. While common pathways, like MAPK, are activated, further research is required to define how signaling at the cell surface results in similar downstream viral targeting and activation of cellular factors responsible for replication. Continued research into PyV-receptor interactions will undoubtedly provide clarity on their impact on downstream signaling critical for the progression of infection.

In addition, the recent determination that a subset of PyVs can utilize EVs to overcome the necessity of a cell-surface receptor, yet still activate intracellular signaling for infection, is unclear. EVs enable PyVs to overcome the limitation of receptor expression, most likely directly influencing tissue tropism and viral spread. Collectively, continued experimentation will serve to improve our understanding of the complex processes that mediate entry of polyomaviruses, likely resulting in the generation of improved antiviral therapeutics for pathogenic human viruses.