How SARS-CoV-2 and Other Viruses Build an Invasion Route to Hijack the Host Nucleocytoplasmic Trafficking System

The host nucleocytoplasmic trafficking system is often hijacked by viruses to accomplish their replication and to suppress the host immune response. Viruses encode many factors that interact with the host nuclear transport receptors (NTRs) and the nucleoporins of the nuclear pore complex (NPC) to access the host nucleus. In this review, we discuss the viral factors and the host factors involved in the nuclear import and export of viral components. As nucleocytoplasmic shuttling is vital for the replication of many viruses, we also review several drugs that target the host nuclear transport machinery and discuss their feasibility for use in antiviral treatment.


Introduction
Despite advancements in science and technology, humans are still plagued by communicable diseases, and the development and production of effective antiviral drugs and vaccines remains challenging. Pathogenic viruses, one major category of infectious agents, have caused not only substantial morbidity and mortality, but also devastating socioeconomic impacts. Within the past century, there have been two particularly severe pandemics, as follows: the 1918 influenza pandemic and the currently ongoing coronavirus disease 2019 (COVID- 19) pandemic.
As part of their lifecycle, many viruses hijack the host transcription and translation machinery while evading the host immune responses [1]. The host nucleoplasm and cytoplasm are segregated by a nuclear envelope (NE), a lipid bilayer embedded with numerous nano-gates known as nuclear pore complexes (NPCs) (reviewed in [2]). Nuclear transport is important for mediating numerous cellular activities, such as cell division [3,4], cell metabolism [5,6], gene regulation (review in [7]), and innate immune response activation [8]. As exploitation of the host nuclear transport is important for viral replication, the mechanisms that are used by viruses to hijack the host nuclear trafficking are potential targets for antiviral drugs; such drugs might halt viral genome transcription, viral protein synthesis, and viral assembly. In this review, we focus on the various mechanisms used by viruses, including severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), to hijack the host nucleocytoplasmic trafficking machinery. We describe various viral factors, and their target host factors, including importins and nucleoporins (Nups). Finally, we discuss the feasibility of using drugs that target the host nuclear transport machinery as antiviral therapies. Cargoes with leucine-rich NES Snurportin HuR (ELAV-like protein 1) [62] Nxf3 (Nuclear RNA export factor 3) [63] LRPPRC (Leucine-rich PPR motif-containing protein) [64] Nmd3 (60S ribosomal export protein) [ Importins α: adapted from a review by Pumroy et al. [20]. Importins β: adapted from a review by Mosammaparast et al. [21] and other studies [64][65][66][67]. Yellow indicates nuclear import, orange indicates nuclear export, green indicates bidirectional shuttling, and blue indicates unknown cargo. Only human cargoes are mentioned here.

Mechanisms of the Host Nuclear Transport Machinery Hijacking by Viruses
The dominant NTRs, including heterodimeric importin-α/β, Crm1, and Nxf1, together with the Nups involved in nuclear transport (Nup358, Nup214, Nup98, and Rae1), are the common targets that viruses hijack for shuttling viral factors between the cytoplasm and the nucleus. The viral replication site determines the purpose of the host nuclear transport subversion by the virus. For example, viruses that replicate in the cytoplasm tend to hijack the host nuclear transport for suppressing the interferon (IFN)-inducing antiviral responses. Conversely, viruses that replicate in the nucleus and assemble their products in the cytoplasm generally exploit the host nuclear transport for much more complicated activities, such as the nuclear import of a viral genome and the nuclear export of a newly synthesized viral genomic RNA to the cytoplasm. Below, we discuss the hijacking mechanisms employed by viruses in various viral families, grouped according to their site of replication, that is, cytoplasm (the Coronaviridae, Filoviridae, Flaviviridae, and Togaviridae families) or nucleus (the Orthomyxoviridae, Retroviridae, Papillomaviridae, Hepadnaviridae, and Herpesviridae families). We summarize these interactions in Table 2 and illustrate them in Figure 1.   Coronaviruses (CoVs; members of the Coronaviridae family) are enveloped, positivesense, single-stranded (+ss) RNA viruses. There are the following four genera in the Coronaviridae family: Alphacoronavirus (α-CoV), Betacoronavirus (β-CoV), Gammacoronavirus (γ-CoV), and Deltacoronavirus (δ-CoV) (reviewed in [68]). Although most human CoVs (HCoVs) are associated with mild upper respiratory diseases and enteric diseases (reviewed in [69]), the recently emerged zoonotic β-CoVs, such as severe respiratory acute syndrome CoV (SARS-CoV) and Middle East respiratory syndrome CoV (MERS-CoV), have caused severe lower respiratory diseases with high mortality (reviewed in [70]). Furthermore, a highly infectious novel β-CoV, SARS-CoV-2, that was first reported in Wuhan, China in 2019 has caused the greatest pandemic in the 21st century [71]. All CoVs, including SARS-CoV-2, replicate in cytosolic double-membrane vesicles (DMVs) [72,73].
A protein-protein interaction (PPI) analysis of the viral factors and the host factors of SARS-CoV-2 revealed that the viral factors NSP9, NSP15, and Orf6 can interact with the host nuclear transport machinery [74]. The associations suggested by these PPI analyses need to be validated by further experiments. To date, only Orf6 has been shown to interrupt the host nucleocytoplasmic trafficking [75][76][77][78] and to induce an aberrant distribution of Nups (Nup98 and Rae1) [76]. Xia et al. found that Orf6 binds to importin-α1 to block the nuclear translocation of IRF3, resulting in an impaired type I IFN production in the HEK293T cell line [78]. Miorin and colleagues reported that type I and type II IFNs failed to induce the transcription of IFN-stimulated genes (ISGs) in SARS-CoV-2-infected Vero E6 cells owing to an impediment of the STAT1 and STAT2 nuclear translocation by Orf6 [75]. SARS-CoV-1 Orf6 plays a similar role in antagonizing IFN; it binds to importin-α1 and tethers importin-β1 to the endoplasmic reticulum (ER)/Golgi membrane, sequestrating the NTRs from activated STAT1 (pY-STAT1) [79], thus blocking the nuclear translocation of a pY-STAT1 [79,80]. Miorin and colleagues proposed two mechanisms to explain this blockade. First, Orf6 binds to Nup98 via its C-terminal domain to interrupt the docking of cargo-importin-α5/β1 ternary complexes at the NPCs. The Orf6-Nup98 interaction does not affect the Nup98-Rae1 interaction, which suggests that Orf6 binding to Nup98 targets only the nuclear import pathway. Second, Orf6 competes with STAT1 and STAT2 for α-importins (α5 and α1) to block the STAT1/2 nuclear ingress. Nevertheless, these α-importins may not be the direct key players because their overexpression failed to rescue an Orf6-dependent blockade of green fluorescent protein (GFP)-tagged STAT1 nuclear import. Interestingly, Addetia et al. reported that the SARS-CoV-2 Orf6 interactions with Nup98 and Rae1 were much stronger than those of SARS-CoV-1 Orf6, suggesting the strong IFN antagonism triggered by SARS-CoV-2 contributes to the high prevalence of asymptomatic cases of SARS-CoV-2 infection [77].
We found that, along with being able to impair nuclear import, SARS-CoV-2 Orf6 can also block nuclear export [76]. Our results show mislocalizations of Nup98 and Rae1 in Orf6-overexpressing cells [76]. In addition, nuclear accumulation of the nuclear RNAbinding protein hnRNPA1 was observed in Orf6-overexpressing cells. For mRNA export, hnRNPA1 is needed [81,82], and the saturation of hnRNPA1 in the nucleus inhibits mRNA export [81]. These findings, that is, the aberrant localization of Nup98 and Rae1 and the nuclear saturation of hnRNPA1, suggest that the Orf6-Nup98-Rae1 interaction can block mRNA export. Addetia et al. later found that SARS-CoV-2-infected cells exhibited a nuclear mRNA accumulation that could be mediated by Orf6 [77]. Recently, Zhang et al. reported that SARS-CoV-2 non-structural protein 1 (Nsp1) halted the host gene expression by blocking an Nxf1-Nxt1-mediated mRNA export [83]. Mechanistically, Nsp1 did not impair the RNA-binding ability of Nxf1; rather, it interrupted Nxf1 interacting with mRNA export adaptors, including Aly/REF and UAP56. Furthermore, Nsp1 reduced the interactions between Nxf1 and key Nups for nuclear export (Nup358, Nup214, and Nup98). SARS-CoV-1 Nsp1 was previously found to disrupt the localization of Nup93 (from the NE to the nucleoplasm) without affecting the NPC structure [84]. Furthermore, SARS-CoV-1 Nsp1 specifically promotes the cytoplasmic accumulation of the nuclear RNA-binding protein nucleolin with unknown consequences. These findings may indicate a role for SARS-CoV-1 Nsp1 in disrupting the host mRNA export [84].
SARS-CoV-1 Orf9b lacks an NLS and thus is localized mainly in the cytoplasm [85]. However, a small amount of Orf9b enters the nucleus passively, and nuclear Orf9b activates the host caspase-3-mediated apoptosis. As Orf9b degradation occurs in the cytoplasm, the nuclear export of Orf9b can protect the host and sustain viral replication [85]. Orf9b possesses an NES at aa residues 46-54 (LRLGSQLSL) that allows it to bind to Crm1 [85]. MERS-CoV Orf4b possesses an NLS (aa residues 2-38), and its nuclear localization is important for inhibiting IRF3-and IRF7-induced IFN-β production [86]. Canton et al. later demonstrated that MERS-CoV Orf4b has a strong affinity for importin-α3, which blocks the NF-κB p65 subunit from entering the nucleus [87].

Ebola Virus (Zaire Ebolavirus)
Ebola virus (EBOV), a member of the Filoviridae family, is a filamentous, enveloped, single-stranded, negative-sense (−ss) RNA virus [88]. At present, there are six recognized EBOV species. The Zaire ebolavirus species has caused two outbreaks in Africa with 40 and 66% fatality, respectively, [89]. As EBOVs replicate in the cytoplasm, they hijack the host nuclear transport to suppress IFN-mediated antiviral responses. The EBOV VP24 protein interacts with importin-α5, -α6, and -α7 [90], but does not interact with importin-α1, -α3, or -α4 [91]. A more recent study found that EBOV VP24 recognizes a non-classical NLS-binding site in importin-α6 [92]. These interactions allow EBOV to suppress the host IFN signaling by preventing the nuclear translocation of pY-STAT1 [91]. The VP24importin-α interaction also inhibits IFN-λ production [93]. VP24 can directly bind to the inactive form of STAT1, unphosphorylated STAT1 (U-STAT1) [94]. The nuclear import of U-STAT1 is mediated by Nup153 and Nup214, independently of importins [95]. Within the nucleus, U-STAT1 activates and prolongs the expression of a set of IFN-induced genes that are distinct from those mediated by pY-STAT1 [96]. The VP24-importin-α5 interaction causes a cytoplasmic accumulation of the nuclear protein hnRNP C1/C2 [97]. During mitosis, hnRNP C1/C2 is exported to the cytoplasm for internal ribosomal entry site (IRES)dependent c-myc translation [98]. Several viruses similarly re-localize hnRNP C1/C2 to the cytoplasm, either for viral replication or IRES-dependent viral protein translation [99][100][101][102], which suggests that VP24-dependent hnRNP C1/C2 cytoplasmic accumulation is essential for EBOV replication [97]. Together, nucleoproteins (NPs) and VP35 form EBOV inclusion bodies (IBs) for viral replication [103]. A study conducted by Gabriel et al. showed that importin-α7 is also involved in IB formation [104]. In addition to creating new copies of EBOV genomic RNA, sub-genomic EBOV RNAs (viral mRNAs) are also produced for viral protein synthesis. The EBOV NP recruits Nxf1 by interacting with the RNA-binding domain of Nxf1. In the presence of mRNA, the binding preference of the Nxf1 RNAbinding domain shifts from NPs to mRNA, and then it delivers mRNA to the cytoplasm for translation [105].

Dengue Virus (DENV) and Zika Virus (ZIKV)
Flaviviruses (members of the Flaviviridae family) are small, enveloped +ssRNA viruses that replicate in vesicle packets located at the ER [106,107]. The flavivirus Dengue virus (DENV) causes mosquito-borne dengue fever in tropical and sub-tropical countries. Two nonstructural proteins (NSs), the DENV NS3 helicase and the DENV NS5 RNA-dependent RNA polymerase (RdRp), form a complex to mediate the viral replication [108,109]. NS5 promotes STAT2 degradation to suppress the host IFN-mediated antiviral responses [110]. An early study reported that NS5 has a bipartite NLS composed of an importin-β1-binding site (βNLS) [111] and a heterodimeric importin-α/β1-binding site (cNLS) [112]. NS5 nuclear import is predominantly mediated by importin-α/β1 receptors [112]. However, later studies suggested that these mapped NLSs are not accessible to the host importins [113]. The subcellular localization of NS5 varies among the four DENV serotypes (DENV 1, 2, 3, and 4); the NS5 proteins of DENV2 and DENV3 reside in the nucleus, but the NS5 proteins of DENV1 and DENV4 are located in the cytoplasm [114,115]. A new monopartite NLS of NS5 has been identified at its C-terminus [115]. An in vitro assay revealed that the NS5 proteins of DENV2 and DENV3 have similar and strong affinities for importin-α2, whereas those of DENV1 and DENV4 have weak affinities for importin-α2 [115]. The NES in NS5 allows for Crm1-dependent NS5 nuclear egress, and this egression correlates with elevated IL-8 production and impaired viral replication, the mechanism of which has not yet been defined [116]. An NE abnormality together with deregulated NPC components were found in DENV-infected cells [117]. NS3 and its cofactor, NS2B3, disrupt the NPC integrity by inducing the proteolytic degradation of FG-Nups, including Nup62, Nup153, and Nup98 [117]. Palacios et al. reported that NS3 has a putative NLS and a putative NES. NS3 was localized within the nucleus during the early stages of infection, whereas it remained predominantly in the cytoplasm during the later stages of infection [118].
Another member of the Flaviviridae family, Zika virus (ZIKV), has posed serious health concerns owing to the rapid increase in cases of neonatal microcephaly linked to ZIKVinfected mothers in Brazil [119,120]. Similar to DENV, ZIKV also exploits the host NTRs and Nups during infection via its NS3 helicase and NS5 RdRp. A study conducted by De Jesús-González et al. showed that ZIKV NS3, in concert with NS2B3, promotes the proteolytic degradation of several FG-containing Nups, including TPR, Nup153, and Nup98, and also disrupts the NE structure [117]. ZIKV NS5, similar to DENV NS5, has a bipartite NLS (a βNLS and a cNLS), which is recognized by importin-α7 [121]. The nuclear import of ZIKV NS5 protects it from cytoplasmic degradation, and this strategy helps to sustain viral replication in the host cells [122]. The cNLS was initially thought to be the primary site of NTR binding [123], but a later study revealed that both the βNLS and the cNLS are required for NS5 nuclear import [122]. Nuclear NS5 sequesters various α-importins (α1, α3, and α4) in nuclear bodies [123]. Intriguingly, the NS5 accumulated in nuclear bodies was incorporated with STAT1 in a glioblastoma cell line (LN229) but not in a hepatocellular carcinoma cell line (Huh-7), suggesting a role for NS5 in the tissue-specific activation of inflammatory responses [123]. Unlike the NS5 protein in Japanese encephalitis virus (JEV), which competes with the host IRF3 and NF-κB for α-importins (α1, α3, and α4) [124], ZIKV NS5 inhibits the activation of TANK-binding kinase 1 to prevent IRF3 activation [125], a strategy used to inhibit the IFN production by infected cells. Furthermore, ZIKV NS2A promotes the chaperone-mediated autophagy (CMA) of importin-α1 [126], possibly in an effort to suppress the host antiviral response.

Chikungunya Virus (CHIKV)
Chikungunya virus (CHIKV), a member of the Togaviridae family, is an enveloped +ssRNA virus that replicates in the host cytoplasm [127]. A CHIKV infection is associated with chronic inflammatory arthritis and other musculoskeletal diseases [128]. The CHIKV capsid protein (CP) has been reported to have two NESs and one NLS [129,130]. Thomas and colleagues demonstrated that the CP bound specifically to the C-terminal NLS-binding site of importin-α3 for its nuclear translocation [129]. The CP binds to Crm1 via its Crm1mediated NES, which was mapped to a leucine-rich region between aa residues 143 and 155 [129]. Thus, importin-α3 and Crm1 are the host factors targeted by CHIKV in its disruption of the host nucleocytoplasmic trafficking [129]. A mutation of the CHIKV CP's NES near the N-terminus (aa residues 44-53) caused the retention of viral CPs in the nucleus and also blocked the host nuclear import system for unknown reasons [130]. CHIKV nsP2 inhibits the host IFN-induced antiviral response [131][132][133][134]. Although nsP2 lacks an NLS (reviewed in [135]), the nuclear import of nsP2 is necessary for suppressing the host antiviral response [134]. Interestingly, similar to that of DENV NS3, the nuclear localization of CHIKV nsP2 also occurs temporarily during early infection, after which this protein resides in the cytoplasm [136]. IFN antagonism by nsP2 can be achieved by several mechanisms, including via a reduction in the cGAS level by a global translational inhibition [131,136], an inhibition of STAT1 activation and/or block of pY-STAT1 nuclear import [133], and a promotion of STAT1 nuclear export [132].

Human Immunodeficiency Virus (HIV)
Human immunodeficiency virus (HIV), a member of the Retroviridae family, is an enveloped +ssRNA virus that was first recognized in 1981 as the causative agent of a new disease affecting T lymphocytes [137]. HIV-1 is more virulent and infectious compared to HIV-2; thus, HIV-1 is the leading cause of acquired immunodeficiency syndrome (AIDS) in the ongoing AIDS pandemic [138]. Upon viral entry, the viral replication complex undergoes reverse transcription followed by integration to form a pre-integration complex (PIC) before entering the nucleus (reviewed in [139]). Mutations in HIV proteins and the silencing of some NPCs can change the HIV integration pattern, which suggests that nuclear import is closely associated with the selection of a viral cDNA integration site for integration into the host genome (review in [140]).
The HIV-1 viral capsid was initially postulated to rapidly disassemble upon viral entry, releasing PICs into the cytosol, from which they subsequently translocate to the nucleus through NPCs [141]. A group of RNA interference (RNAi) screening studies have identified several Nups that are required for a HIV-1 infection, which include Nup98, Nup85, Nup133, Nup107, Nup160, Nup153, Nup214, Nup358, Nup155, Crm1, and the nuclear import receptor transportin 3 (TNPO3) [142][143][144]. In addition, a HIV-1 infection induces a downregulation in Nup50 and an upregulation in Nup62 for currently unknown reasons [145]. Some HIV-1 proteins, such as integrase (IN) [146], the viral protein R (Vpr) [147], and the matrix protein (MA) [148], contain NLSs. Notably, the width of the broad end of the HIV-1 capsid is approximately 60 nm [149], whereas the NPC pore size has long been thought to be only approximately 40 nm wide [19]. Collectively, these findings suggested that capsid disassembly is a prerequisite for PIC nuclear import.
More recent findings showing that the HIV-1 capsid remained assembled for at least the reverse transcription process [150] and that capsid uncoating can occur in the nu-cleus [151][152][153] have challenged the model described above. With a combination of different imaging tools, including cryo-electron tomography (cryo-ET) and subtomogram averaging, Zila and colleagues have visualized the entry of a whole HIV-1 capsid into a host nucleus through an NPC [154]. The diameter of an NPC measured in intact human cells (i.e., in a transporting state) was in fact larger than the diameter of an NPC measured in an isolated nuclear envelope (i.e., in a constricted state). The diameter of an NPC in the transporting state is larger than the broad end of the HIV-1 capsid; therefore, it can accommodate the import of a HIV-1 capsid into the nucleus. The team proposed a three-step process for the nuclear import of a HIV-1 capsid, during which the intact capsid interacts with different Nups in each stage. First, the HIV-1 capsids travel along microtubules to dock onto NPCs and interact with the FG-repeats and cyclophilin (Cyp) domain of Nup358 at the NPC cytoplasmic sites. Second, the intact capsids move deep into the central channels of the NPCs, which contain high local concentrations of FG-containing Nups within the Nup62 complex. Last, the capsids bind to Nup153 and cleavage and polyadenylation specificity factor subunit 6 (CPSF6). The HIV capsid-Nup153-CPSF6 interaction triggers capsid disassembly, which releases PICs into the nucleoplasm.
HIV Vpr suppresses the host antiviral response by interacting with α-importins (preferentially α5, but also, to a lesser extent, α1 and α4) to inhibit the IRF3 activation and to block the nuclear import of IRF3 and NfκB [155]. HIV Vpr has an N-terminal NLS and a C-terminal NLS [156]. Its C-terminal NLS was initially thought to be non-functional [156], but later studies have shown that it binds α-importins and mediates nuclear import without importin-β1 [157,158].
HIV-1 Rev has an arginine-rich NLS site in its N-terminal domain and a leucine-rich NES sequence in its C-terminal domain (reviewed in [159]). HIV-1 Rev enters the nucleus without binding RNA because only the RNA-free form of Rev can bind to importin-β1 for nuclear import [160]. The Rev-importin-β1 interaction is highly specific and can be blocked by importin-α5 [160]. The Rev NES is required for the nuclear export of unspliced viral RNA (vRNA). Rev binds to vRNA via the Rev response element (RRE) (reviewed in [161]). Rev contains an NES that allows it to bind to Crm1 and subsequently bring vRNA, in the form of a viral ribonucleoprotein (vRNP) transport complex, out into the cytoplasm in a RanGTP-dependent manner [162]. Several studies have suggested that Nup214, Nup153, Nup98, and Nup62 interact indirectly with vRNP through different Rev co-factors [163][164][165]. For example, the Rev-vRNP complex disturbs NPC integrity and causes a Nup62 cytoplasmic localization and subsequent encapsidation into a progeny virus [166]. A downregulation in Nup62 induced a vRNA nuclear accumulation, suggesting the importance of Nup62 in vRNA export [166]. The host Nxf1-dependent RNA export always involves active pre-mRNA splicing to prevent aberrant protein translation. Rev suppresses Nxf1-mediated vRNA export to secure a viral gene expression in the host [167].
McCauley et al. reported that an HIV-1 provirus used Crm1 and Rev to export its unspliced HIV for translation, which triggered chronic inflammation in patients with HIV who were receiving anti-retroviral therapy [168].

Influenza A Virus (IAV)
Influenza A virus (IAV), a member of the Orthomyxoviridae family, is an enveloped, segmented −ssRNA virus that causes an epidemic respiratory disease (reviewed in [169]). IAV genomic RNAs are packed together with a viral nucleoprotein (NP) and a heterotrimeric RNA-dependent RNA polymerase (RdRp) complex (composed of PA, PB1, and PB2) into a rod-like vRNP (reviewed in [169]). Upon viral entry, free vRNPs translocate to the nucleus for viral RNA transcription and replication. The IAV NP and three components of the RdRp complex possess at least one NLS [170]. An IAV NP has two NLSs, a non-classical NLS in the N-terminus [171] and a classical bipartite NLS in the middle [172], and these are sufficient for vRNP nuclear import [170]. A mutation of the bipartite cNLS of the NP did not block its nuclear import, which suggests that the non-classical NLS of the NP is the predominant site [173] for its interaction with importins α5 and α7 [169]. Therefore, the non-classical NLS of an NP is essential for IAV replication [173]. Donchet et al. reported the binding affinity of an NP to different α-importins; an NP had the highest affinity for importin-α7 and the lowest affinity for importin-α1 [174].
The Crm1-mediated nuclear export of newly synthesized IAV vRNPs is crucial for new viral synthesis. IAV vRNP, M1, and NES-bearing NS2 form a complex for vRNP export [182,183]. IAV NS1 exerts a negative regulatory effect on the host Nxf1-dependent mRNA export by downregulating Nup98 expression and forms an inhibitory complex with several of the host mRNA export factors (Nxf1, Nxt1, Rae1, and E1B-AP5) [184]. IAV NS1 binds at the FG-repeat binding site of Nxf1, where it interacts with the FG-domain of Nup98 during mRNA export [185]. In addition to disrupting the host mRNA export, IAV NS1 also interrupts the host mRNA processing, which it achieves by interacting with CPSF30 [186] and PABII [187]. Collectively, these activities suppress the host protein biogenesis, particularly the factors required for an effective antiviral response [185]. Interestingly, in severe cases of influenza, a reduction in the protectin D1 (PD1) level enables IAV vRNA to bind directly to Nxf1, which then interacts with Nup62 for nuclear export. This mechanism does not require other mRNA export factors (e.g., Nxt1 or Crm1) or FG-containing Nups (e.g., Nup98 or Nup214) [188].

Human Papilloma Virus (HPV)
Human papilloma virus (HPV), a member of the Papillomaviridae family, is a small, icosahedral, non-enveloped, double-stranded (ds)DNA virus [189]. In 1995, the International Agency for Research (IARC) classified HPV16 and HPV18 as human carcinogens because infection with these viruses increases the risk of developing cervical cancer [190]. HPV replicates in the host nucleus, and the entry of its viral genome into the host nucleus requires NE breakdown (during mitosis) rather than passage through an NPC [191,192].
The nucleocytoplasmic shuttling of E6 and E7, two oncoproteins found in high-risk HPVs, promotes carcinogenesis in HPV-infected cells. Only high-risk HPV E6 (E6 high ) can translocate to the nucleus [193] because its C-terminal contains three NLSs [193][194][195] that interact with importin-α1/β1, importin-β1, and importin-β2 [194]. Low-risk HPV E6 (E6 low ) predominantly resides in the cytoplasm, but E6 low acquires nuclear import activity when it is conjugated with an E6 high NLS [193]. The nuclear import of E6 high requires RanGDP, but GTP hydrolysis is not necessary [194]. Nuclear E6 high is critical for E6-mediated p53 degradation, independent of MDM2 [196]. E6 high mediates the polyubiquitination of p53, making it susceptible to proteosome degradation [197]. The E6 high -p53 interaction promotes p53 export, and the p53 NES is required for a Crm1-mediated nuclear export [197]. The nuclear import of E6 high is compulsory for E6 high -mediated p53 degradation; however, the p53 degradation can occur in both the nucleus and the cytoplasm [197].
To support viral DNA amplification, high-risk HPV E7 (E7 high ) must travel to the nucleus to hijack the host cell-cycle machinery, driving the host cell to re-enter the Sphase (reviewed in [198]). HPV16 E7 was initially thought to lack an NLS, and its nuclear import is Ran-dependent but independent of classical importin-α/β and importin-β2 receptors [199]. Knapp et al. later identified two NLSs in the N-terminal domain and CR3 domain of the HPV16 E7 protein [200]. Eberhard et al. found an explanation for the ability of the HPV16 E7 protein to undergo importin-independent nuclear import, despite having NLSs [201]. The zinc-binding domain within the E7 CR3 domain contains a hydrophobic patch ( 65 LRLCV 69 ) that enables E7 to accomplish its nuclear import by interacting with the FG-domain of Nup62 via a hydrophobic interaction [201]. Furthermore, this hydrophobic patch also facilitates the interaction of HPV16 E7 with Nup153 [202]. The hydrophobic interaction between HPV E7 and FG-containing Nups is conserved in other HPV serotypes (HPV8 and HPV11) [202,203]. A functional NES has been identified in HPV16 E7, which suggests that E7 nuclear egress is Crm1-dependent [200].
HPV encodes the E1 DNA helicase and the E2 original recognition protein and uses them to hijack the host DNA replication machinery for viral replication [204]. Therefore, the nuclear localization of E1 and E2 are important for efficient viral replication. The E1 proteins of several HPV serotypes (HPV11, HPV31, and HPV16) have bipartite NLSs [205], and these NLSs behave in a similar way to the bipartite NLS of the bovine HPV E1 protein [206], which can interact with multiple importins α (α1, α3, and α5) [207]. In addition, the phosphorylation of HPV E1 by the host ERK and/or JNK is needed for its nuclear import [205,206]. Regarding the HPV E1 nuclear export, this protein has a functional NES site, suggesting it undergoes a Crm1-dependent export [204,208]. This NES has a cyclindependent kinase (CDK) phosphorylation site, and NES phosphorylation by CDK causes HPV E1 to be retained in the nucleus, which is essential for viral replication [206,208].
The subcellular localizations of E2 proteins from low-risk (HPV11 and HPV6) and high-risk (HPV16 and HPV18) HPVs differ; low-risk HPV E2 is predominantly localized in the nucleus, whereas high-risk HPV E2 can be found in both the cytoplasm and the nucleus [209]. The NLS of the HPV11 E2 protein, but not those of the HPV16 and HPV18 E2 proteins, was found to have a dominant function [209]. Interestingly, cNLSs are found in high-risk HPV E2 but not in low-risk HPV E2; however, the NTRs for high-risk HPV E2 are still unknown [210]. The NES in high-risk HPV E2 enables the nucleocytoplasmic shuttling of this protein. Cytoplasmic accumulation of the high-risk HPV E2 proteins promotes caspase-8-mediated cell apoptosis [209].
The HPV L1 major capsid protein (L1) and L2 minor capsid protein (L2) are delivered to the nucleus for virion production. HPV16 L1 possesses both a monopartite and a bipartite NLS, and its nuclear import requires importin-α1/β1, RanGDP, and free GTP, but occurs independently of GTP hydrolysis [211]. The L1 protein also binds importin-β2, but RanGTP is unable to dissociate the complex [211]. This interaction inhibits the importin-β2-mediated nuclear import of hnRNP A1 [211]. An L2 protein nuclear import is required for HPV viral assembly [212]. For nuclear import, the HPV16 L2 protein can interact with importin-β2, importin-5, or the importin-α1/importin-β1 heterodimer. An NES has been identified in the L2 protein, but its function remains unclear.

Hepatitis B Virus (HBV)
Hepatitis B virus (HBV), a member of the Hepadnaviridae family, is a small, enveloped dsDNA virus that propagates in hepatocytes [213]. Acute infection with HBV causes acute hepatitis, whereas chronic infection with HBV increases the risk of hepatocellular carcinoma (HCC) (reviewed in [214]). The HBV capsid disassembles in the nucleus, releasing relaxed circular DNA (rc)DNA, which is subsequently repaired by the host DNA repair machinery to form a covalently closed circular DNA, (ccc)DNA. The HBV cccDNA serves as a template for viral RNA replication (reviewed in [215]). Phosphorylation of the C-terminal HBV core or the capsid protein (Cp) exposes the Cp NLS, allowing the Cp to recruit classical importinα/β1 receptors to mediate the nuclear import of the HBV capsid [216,217]. In addition to phosphorylation, the completion of the (+) DNA and removal of viral DNA polymerase from rcDNA triggers a conformational change of the Cp, exposing the C-terminal NLS site [218]. During its translocation through an NPC, RanGTP triggers the dissociation of the Cp-importin-α/β1 ternary complex, which allows for an interaction between the Cp and Nup153 [219]. Notably, Nup153 is the sole FG-containing Nup that interacts with the Cp, which suggests that this interaction is not mediated by a hydrophobic interaction [219]. Finally, mature HBV capsids disintegrate in a Ran-independent manner [217], releasing viral rcDNA; capsid disassembly is halted in immature capsids [219].
The Cp has the following four arginine-rich domains (ARDs): ARD I and III each behave in a similar way to an NLS, and ARD II and IV each behave in a similar way to an NES [220]. The Cp NES is needed for an Nxf1-dependent viral pre-genomic RNA (pgRNA) export [220]. The nuclear export of pgRNA is a critical step for new virion synthesis. The subcellular localization of the Cp is rather complex, depending on both the intrinsic factors (NLS and NES) and the extrinsic factors (importins, Nxf1, and cellular kinase) [220]. A recent study showed that increasing the nuclear Cp concentration induces Cp export in a Crm1-dependent manner, which suggests that the Cp concentration may manipulate different export signals [221]. An empty Cp can interact with importin-β1 independent of importin-α via its IBB domain, but the exact biological function of this interaction is yet to be elucidated [222]. Mitra and colleagues reported that the cytosolic HBV e antigen (HBeAg), also known as the precore protein intermediate (p22), interacts with importin-α5 through its C-terminal ARD. This interaction blocks the pY-STAT1 nuclear import and subsequently suppresses the host IFN response [223].

Herpes Simplex Virus Type-1, Human Cytomegalovirus, and Epstein-Barr Virus
Herpes simplex virus type-1 (HSV-1) is a dsDNA α-herpesvirus that causes cold sores in human beings [224]. After internalization, the viral capsid travels along microtubules to reach the nucleus to release a viral genome into the nucleoplasm for transcription and translation [225]. The tegument proteins VP1/2 [226] and pUL25 [225] interact with Nup358 and Nup214 to dock the viral capsid on an NPC. VP1/2 has an efficient NLS at its N-terminal, but the protein is mainly located in the cytoplasm [227]. The NLS of VP1/2 is specific to importin-β1 rather than importins α [227,228], and VP1/2-importin-β1 interaction is important for directing the viral capsid to dock on an NPC [227]. Using mouse embryonic fibroblast (MEF) cells, Döhner et al. comprehensively demonstrated the distinct importin-α isoforms (α1, α3 and α4) in mediating HSV-1 gene expression, nuclear localization of viral proteins, capsid assembly, and capsid egress [229]. Importins are not needed for the NPC docking of an HSV-1 capsid, for the nuclear import of incoming HSV-1 genomes, or for the nuclear translocation of the HSV-1 VP16 protein. Importins α indirectly regulate the HSV-1 protein expression, and within this regulation importin-α1 is facilitative but importin-α4 is restrictive. Importin-α1 and Importin-α3 orchestrate the nuclear localization of immediateearly (ICP4 and ICP0) and early (ICP8 and pUL42) HSV-1 proteins. The HSV-1 DNA polymerase subunits pUL30 and pUL42 utilize several nuclear import mechanisms. HSV-1 pUL30 has a non-canonical and a classical bipartite NLS, and interacts with importin-α5, but interactions with other importin-α isoforms remain unknown [230][231][232]. HSV-1 pUL42 has a bipartite NLS, and the NLS binds mainly to importin-α7 and, to a lesser extent, to importin-α1, but not to importin-α3 [233]. Interestingly, the mutation of the NLS in either pUL30 or pUL42 does not impede the nuclear import of both proteins under the holoenzyme form [233]. The holoenzyme only resides in cytosol if the NLSs in both pUL30 and pUL42 are mutated [233]. Importin-α1 is crucial for an HSV-1 infection because importin-α1 is required for efficient capsid assembly and egress to make new virions in cytosol. The team also illustrated that silencing either importin-α1 or importin-α3 is sufficient to suppress HSV-1 gene expression in terminally differentiated cells, neurons for example, but HSV-1 gene expression remains unperturbed in MEF cells [229]. MEF is not a terminally differentiated cell type [234]. Such a discrepancy indicates that importins α repertoire in MEF is sufficient to compensate for the absence of importin-α1 or importin-α3 to sustain HSV-1 gene expression.
HSV-1 ICP27 is an immediate-early protein that is required for enhancing HSV-1 gene expression and for exporting intronless HSV-1 mRNAs (review in [235,236]). Different conformations of ICP27 confer different specificities/preferences for nuclear export. Viral mRNA-bound ICP27 interacts with Aly/REF to recruit Nxf1 to export intronless HSV-1 mRNA for translation [237]. Nonetheless, Aly/REF knockdown does not significantly dampen viral mRNA export [238], suggesting that ICP27 could utilize other host export factors to compensate for Aly/REF. Free ICP27 does not require both Nxf1 and Crm1 for its export [239]. Instead, its N and C termini interact with Nup62 to enable ICP27 shuttles between the cytoplasm and the nucleus. The nuclear import of ICP27 inhibits both classical, importin-α/β1-dependent, and transportin-dependent nuclear import [240]. Beside the main export receptor (Nxf1), a recent finding reveals that the HSV-1 integral protein, glycoprotein M (gM), binds to exportin-6 (XPO6) for nuclear export to the trans-Golgi network (TGN) [241].
Human cytomegalovirus (HCMV) is a ds-DNA β-herpesvirus that is associated with mononucleosis syndrome (review in [242]), venous thromboembolism (VTE) (review in [242]), and cytomegalovirus encephalitis in immunocompromised patients (review in [243]). HCMV pUL97 is a multifunctional protein kinase that phosphorylates CDK1 and modifies the host G2/M cell cycle checkpoint regulators to promote viral replication [244]. There are two isoforms of pUL97, large and small isoforms. The large isoform has two bipartite NLSs (NLS1 and NLS2), whereas the small isoform only has NLS2, located at their N-terminals [245]. Therefore, the large isoform has a higher nuclear localization efficiency compared to the small isoform [245]. Only importin-α1 has been shown to interact with pUL97, and other importin-α isoforms have not been tested [245]. HCMV pUL79 is an elongation factor of RNA polymerase II for viral gene transcription during the late stages of HCMV infection [246]. HCMV pUL79 possesses a hydrophobic PY-NLS, which enables it to translocate to the nucleus through an importin-β2-mediated pathway [247]. The multifunctional protein, HMCV pUL84, is believed to initiate lytic viral DNA synthesis [248]. Lischka et al. performed in vitro transport assays and found that the nuclear import of pUL84 relies on the classical importin-mediated import pathway [249]. Although pUL84 has a putative NLS, the NLS does not have the NLS activity [249]. Instead, a large domain with 282 amino acids is needed for an importin-α-pUL84 interaction [249]. A team has shown that pUL84 interacted with several importin-α isoforms, including α1, α 3, α 4, and α5 [249]. Interestingly, the pUL84-importin-α interaction domain also contains two leucin-rich NESs, and this region allows the nuclear export of pUL84 via the Crm1-dependent pathway [250]. Gao and his coworkers conducted RNA pulldown assays to reveal pUL84 s viral mRNA export activity [248]. The team reported that pUL84 interacted with a viral-encoded transcript known as IRS1 and mediated the cytoplasmic localization of an IRS1 transcript [248].
Similar to HSV-1 ICP27, HCMV encodes the pUL69 regulatory protein to mediate a viral mRNA export. The arginine rich RNA-binding domain and the DExD/H-box helicase UAP56-binding motif in pUL69 [251] allows pUL69 to interact with the mRNA export factor UAP56/URH49 [252], transcription elongation factor hSPT6, and intronless viral mRNA to form a messenger ribonucleoprotein (mRNP) for an Nxf1-Nxt1-dependent mRNA export (reviewed in [253]). An NES is required for the nuclear export of free pUL69 but independent of Crm1, and the exact export mechanism remains elusive [254]. Likewise, in silico analysis failed to determine a classical NLS within pUL69, the NTRs associated with pUL69 nuclear import are also unknown [252]. CDK9 phosphorylation of pUL69 is crucial for a pUL69-mediated viral mRNA export because CDK inhibition triggers the nuclear accumulation of pUL69 and suppresses the mRNA export activity of pUL69 [255].
Non-functional NESs have been found in EBV BFLF2 [267] and HSV-1 UL31 [264,269], and no NESs have been found in HCMV UL53 [270]; these findings suggest that these proteins do not require Crm1 for their export. EBV BFLF2 interacts with Nxf1 in the absence of RNA for its export. The nuclear export receptors for HSV-1 UL31 and HCMV UL53 are yet to be determined [267]. Funk et al. conducted a comprehensive analysis on the EBV and HSV-1 tegument proteins using an in vitro assay called Nuclear EXport Trapped by RAPamycin (NEX-TRAP) [271] and found that two EBV (pBTRF1 and pBGFL3) and nine HSV-1 tegument proteins showed nuclear export activity. The group compared the nuclear export activity between the EBV and HSV-1 tegument orthologs; EBV pBTRF1 behaved in a similar way to its HSV-1 orthologue pUL21, exhibiting active export activity. Conversely, EBV pBGFL3 (exported) displayed the opposite behavior compared to its HSV-1 orthologue pUL14 (non-exported). The NES of pBTRF1 matches the Rev NES consensus, and that of pBGFL3 matches the PKI NES consensus. The NESs of both Rev and PKI are recognized by Crm1, which suggests that EBV pBTRF1's and pBGFL3's export activities are Crm1-dependent. A leptomycin B (LMB) assay revealed six HSV-1 tegument proteins (pUL4, pUL11, pUL13, pUL21, pUL37d11, and pUL48) that are exported in a Crm1-dependent manner. Table 2. Interactions between viral factors and host factors in the hijacking of the host nuclear transport machinery.

Nxf1
Binds to FG-repeat binding site of Nxf1 to block host mRNA export.
vRNA Nxf1 In severe influenza infection, vRNA binds to Nxf1 alone and interacts only with Nup62 for nuclear export.

Virus Name Viral Factors Host Factors Effect of Viral-Host Factor Interaction References
Papillomaviridae (dsDNA) Nucleus HPV

Microtubule
Reorganizes nucleus shape to allow nuclear import of non-NLS-bearing viral proteins. EGR1 Importin-7 (IPO7) Negatively regulates importin-7 expression to a level that is optimal for EBV-transformed cell growth.

Potential Antiviral Drugs That Target the Host Nuclear Transport Machinery
The exploitation by viruses of the host nucleocytoplasmic trafficking is crucial for viral genome replication, viral component assembly, suppression of the host antiviral response, and cellular state alteration. To date, viral polymerase inhibitors are actively used in clinics against various RNA and DNA viruses including EBOV [273], HIV [274], and HPV [275]. In addition to viral polymerase inhibitors, nuclear transport inhibitors can also be considered to reduce the viral load and/or ameliorate clinical symptoms, either in monotherapy or in combination with other antiviral agents. In this section, we discuss the host-specific and the viral-specific nuclear transport inhibitors and relevant clinical trials.
The key points are summarized in Tables 3 and 4.

Host-Specific Nuclear Export Inhibitors
As Crm1 is crucial for the nuclear export of a vast number of cargo types, it is a main target of viruses for the export of their proteins. LMB and its derivatives [288,289], as well as selective inhibitor for nuclear export (SINE) [290,291] compounds, target a common site of Crm1, Cysteine-528, to block Crm1 from interacting with the NES-bearing cargo. After it binds to Crm1, LMB undergoes hydrolysis and subsequently forms a salt bridge in LMB-Crm1, irreversibly inhibiting Crm1's function [292]. Conversely, SINEs do not undergo hydrolysis after they bind to Crm1; therefore, there is no salt bridge formation [293,294]. However, the SINE-Crm1 interaction promotes Crm1 degradation, and Crm1 re-synthesis occurs after SINE administration is discontinued [295]. Clinical trials of LMB were halted because the treatment caused profound adverse effects, presumably owing to the irreversible LMB-Crm1 interaction [296]. Conversely, SINEs have shown much lower toxicity in clinical trials because their inhibition of Crm1's function is reversible (reviewed in [297]). The SINEs Selinexor and Verdinexor, which have been actively tested in cancer patients, are now under evaluation for their antiviral activity in in vitro and in vivo settings (Table 3).
Two potent HIV capsid inhibitors, GS-6207 and GS-CA, have been developed by Gilead Science [301,302]. These small molecule inhibitors bind to the HIV capsid protein, impeding the delivery of viral PIC into the host nucleus and thus interrupting capsid assembly and viral particle production. The binding site of the inhibitors is the site where Nup153 and CPSF6 bind during the HIV capsid's nuclear entry.
Mohl et al. developed a model for an IAV PB1/PA-RanBP5 complexation, and then used PPI software-based virtual screening to identify potential inhibitors that target IAV PB1 and block its interaction with importin-5 (RanBP5) [303]. In vitro studies have shown that five compounds may inhibit the nuclear localization and in vitro polymerase activity of PB1/PA. Some groups have applied in silico screening to identify inhibitors of EBOV VP24 and importin-α isoforms. Tanaka et al. used a CE-SELEX system to select oligonucleotide-based VP24-binding aptamers that inhibit the VP24-importin-α1 interaction [304]. Two of these inhibitors, VPKS-2 and VPKS-5, were found to inhibit the VP25-importin-α5 interaction in vitro. Another research team, Song et al., utilized the Random non-standard Peptides Integrated Discovery (RaPID) system to design high-affinity macrocyclic peptides that compete with importin-α6 for VP24 binding [305]. These strategies protect importins α from VP24, and thus restores the host antiviral defense system during infection.
The therapeutic effects of interferon β (type-I interferon) on recurrent HPV infection lesions have been clinically proven, especially in patients with cervical intraepithelial neoplasia (CIN) [306]. Type I IFN strongly induces p56 expression, and p56 binds the E1 protein of several HPV serotypes [307]. In vivo experiments demonstrated that cytoplasmic p56 promotes the nuclear export of the E1 protein [307]. Other in vitro experiments revealed that recombinant p56 inhibits the DNA helicase activity of E1, consequently halting HPV DNA replication [307].

Clinical Translation of Nuclear Transport Inhibitors against Viral Infections
Regarding the importin-α inhibitors, only Ivermectin has been assessed in clinical studies for its antiviral efficacy. Three clinical studies have been performed to evaluate the clinical benefits of Ivermectin in patients with COVID-19 [325][326][327]. Two of these clinical studies (one randomized clinical trial (RCT) and one non-RCT) found that Ivermectin, either in monotherapy [327] or in combination therapy [326], significantly improved the viral clearance in patients with mild cases of COVID-19. However, another clinical study on Ivermectin monotherapy found that it did not shorten the time required for symptom resolution in patients with mild cases of COVID-19 [325]. The observed clinical and immunological spectra of patients with COVID-19 indicate that viral clearance may not directly correlate with symptom resolution because symptoms that are induced by a dysregulated immune response can persist after successful viral clearance [328]. The synergism between Ivermectin and Remdesivir for treating COVID- 19 has not yet been evaluated.
In patients with dengue fever, a phase two RCT trial showed that Ivermectin monotherapy promoted NS1 antigenemia clearance but not viral clearance [329]. The serum NS1 level is positively correlated with dengue hemorrhagic fever (DHF) [330]. The role of Ivermectin in reducing the risk of DHF development in patients with dengue fever has not yet been investigated.
Two SINE candidates, Selinexor and Verdinexor, are actively being tested as therapeutics in cancer patients. Currently, one clinical trial (NCT04349098) has completed to determine the clinical benefits of Selinexor in treating moderate-to-severe cases of COVID-19. Additionally, a recently completed phase one clinical trial (NCT02431364) reported that Verdinexor exhibited an improved adverse effect profile in healthy volunteers. However, the study has been terminated for administrative reasons.
Viral factor-specific nuclear transport inhibitors, such as GS-6207, REBACIN ® , and CAM (NVR 3-778 and GLS4), are currently being tested in clinical trials. A phase one clinical trial of GS-6207 (NCT07379866) found that a single dose of subcutaneously administered GS-6207 effectively reduced the plasma HIV viral load and that the protection could persist for more than six months [301]. In two parallel RCTs, REBACIN ® was found to induce a strong viral clearance in HPV-positive patients [331]. Mechanistically, REBACIN ® inhibits the mRNA transcription of high-risk HPV E6 and E7 oncoproteins in mice [331]. In addition, Yang et al. performed a retrospective analysis and a meta-analysis and found that REBACIN ® showed better efficacy than IFN for treating a persistent high-risk HVP infection [332]. Yuen et al. performed a phase one RCT to test the anti-HBV activity of an SBA class CAM, NVR 3-778, in patients with chronic HBV infection without cirrhosis [333]. It revealed that NVR 3-778 was well-tolerated by patients and that this CAM compound could reduce the viral load (as measured by viral DNA and RNA levels); furthermore, its administration with pegylated IFN (pegIFN) produced a synergetic antiviral effect [333]. Zhang et al. reported that a HAP class CAM, GLS4, demonstrated antiviral activity in patients with a chronic HBV infection in a phase 1b study, and that this drug was welltolerated by patients [334].

Conclusions
Viral proteins masquerade as host factors to exploit the host nuclear transport machinery for viral replication, and they reprogram the host cellular environment to enhance their replication and evade the host immunity. Nuclear transport inhibitors can not only interrupt viral replication but also disrupt the viral assembly and restore the host immunity. Targeting host factors to block nuclear transport could exert a broad-spectrum antiviral effect, as reported for Ivermectin. Nonetheless, the feasibility of using inhibitors of host factors as antiviral drugs could be limited when they target important NTRs, such as importin-β1 and Crm1, owing to likely adverse effects. Furthermore, many infectious diseases, especially COVID-19, have a complicated clinical spectrum. The administration timing of the host-specific nuclear transport inhibitors needs to be carefully evaluated to maximize clinical benefits. In silico drug screening and in vitro and in vivo experiments could accelerate the identification of potent viral-specific nuclear transport inhibitors. As certain viral factors not only exploit the host NTRs but also have their own biological functions, viral-specific nuclear transport inhibitors could "kill two birds with one stone", that is, not only improve the viral clearance but also preserve/restore the host nucleocytoplasmic trafficking. Lastly, on the basis of our experience applying HS-AFM imaging to human NPCs [17,336], DNA-DNA-binding protein interactions [337], and viral proteins [338,339], we believe it will be worthwhile to observe the interaction between the viral factors and the host nuclear transport factors in real-time.