The hnRNP family is comprised of 20 major hnRNPs, named from hnRNP A1 to U, with molecular weights ranging from 34 to 120 kDa (5, 11, 12). Most hnRNPs are commonly known by their designated hnRNP names, but few of them, are also known by their alternative names, as summarized in Table 1 . Among these hnRNPs, the hnRNP A/B sub-family of proteins (comprised of A0, A1, A2/B1, and A3) contributes to the regulation of the majority of cellular RNAs (15). In addition to the major hnRNPs, there are many minor hnRNPs such as HuR, PABP1, RNPS1, SRm160, and SRp20 [reviewed in (12)]. Unlike major hnRNPs, these minor hnRNPs are characterized either as unstably binding hnRNAs, or only associated with a subset of hnRNAs (5, 12, 13, 16).

The binding of hnRNPs to nucleic acids is carried out by multiple protein domains (12). Most hnRNPs contain classical RNA-Binding Domains (RBDs) such as RNA recognition motif (RRM), glycine-arginine-rich (GAR) domains or arginine-glycine-glycine rich (RGG) boxes, hnRNP K-homology (KH) domain, DEAD-box helicase domain, dsRNA motif (DSRM), and zinc fingers (ZF) domain (17). RRMs play a crucial role in binding the RNA sequences (18). The RRM (the second RRM in some proteins), the external β-strands, the loops, and the C- and N-termini, can have a high RNA-binding affinity and specificity, as well as extreme structural versatility when they interact with RNA (19). This makes the RRM-containing proteins capable of performing numerous biological functions (19). In addition to RRM-containing hnRNPs, five hnRNPs (hnRNP E 1-4, hnRNP K) contain KH domains (5). The KH domains cooperatively bind polypyrimidine or C-rich regions in RNA, ssDNA, or dsDNA with high affinity and specificity (20). Moreover, there are numerous protein partners that interact specifically with the KH domain proteins (20). As a result, hnRNPs can influence many events in the cell at the transcriptional, post-transcriptional, and translational levels (21). hnRNP A/B, hnRNP P, hnRNP D, hnRNP H, and hnRNP U also contain a low-complexity domain (LCD) typically in the C-terminus region, which is also known as an intrinsically disordered region (IDR) or prion-like domain (PrLD) (22, 23). The PrLD encourages liquid-liquid phase separation as well as fibrillation by self-association (24). The activity of these interactions is thought to facilitate transient and diverse multivalent de-mixing of proteins and/or RNA (25). PrLDs/IDRs have also been shown to bind RNA (26–28). Since hnRNPs contain either RBDs or both RBDs and IDRs, they can interact with RNA in both specific and nonspecific ways.

hnRNPs have been reported to undergo various post-translational modifications (PTMs), such as phosphorylation, glycosylation, s-nitrosylation, methylation, acetylation, ubiquitination, and SUMOylation. These PTMs are essential for cellular localization, protein-protein interactions, and nucleic acid binding of hnRNPs, in addition to regulating viral RNA processing and virus replication (29–31). Some hnRNPs are strictly nuclear [e.g., hnRNP C and U (32)], whereas others remain associated with the mRNA as it passes through nuclear pores and therefore undergo constant nucleocytoplasmic movement (e.g., members of hnRNP A, B, E, D, I, and K group) (12). The subcellular localization is controlled by discreet sequence signals within hnRNPs. For example, the hnRNP C and U contain nuclear localization signal (NLS) sequence and are confined to the nucleus. hnRNP A1 harbours noncanonical nucleocytoplasmic shuttling (NS) domains called M9, and hnRNP K contains both an NLS (N-terminal) and an NS (termed KNS, K nuclear shuttling) domain (32–34). Of note, hnRNP A1 was identified as the first cargo of Transportin-1 (Trn-1), which is a well-characterized nuclear import receptor (35, 36). Trn-1 also imports other proteins of the hnRNP family such as hnRNP A/B, hnRNP D, hnRNP F, hnRNP H, hnRNP M (37). A schematic structural diagram of major hnRNPs, along with their size, major RNA binding domains, and localization signals are shown in Figure 1 and Table 1 .

Pathogenic viruses continuously emerge and re-emerge, which have become a major threat to public health for millennia. Understanding the molecular mechanism of the viral life cycle is important for the development of diagnostic tests and effective antiviral therapies. The process of viral life cycle can be divided into six major stages: attachment, penetration, uncoating, genome replication and expression (including vRNA transcription and viral protein synthesis), assembly, and virion release from the host cell ( Figure 2 ) (39). Viruses generate new infectious particles by assembling preformed components. Following entry into host, viruses must rapidly generate numerous copies of genome and package them into virions (39). To do so, viruses must express their genes as functional mRNAs as early as possible during infection, and subsequently instruct the cellular translation machine to create viral proteins (40).

Based on the type of viral genome and mode of transcription, viruses are categories into RNA viruses, DNA viruses, and retroviruses. These viruses are further subcategorised into double-stranded (ds) or single-stranded (ss), and positive-strand (+) or negative strand (-) viruses, as shown in the Figure 2 (41). In RNA viruses, (+) ssRNA acts as mRNA, and one of the RNA strands of the dsRNA or (-) ssRNA acts as a template for viral mRNA or (+) ssRNA synthesis. The mRNA or (+) ssRNA is then recognized by the host cell translation machinery (42). Similarly, in DNA viruses, dsDNA is transcribed into viral mRNA. However, ssDNA is first converted to dsDNA intermediate, which is further transcribed into mRNA. For efficient viral replication, a high degree of regulation is required to switch from DNA transcription to replication when virion packaging occurs. The transcription begins with early genes, such as polymerase and regulatory proteins, followed by the initiation of viral DNA replication once these genes are transcribed. Finally, the transcription ends with late genes that code for the capsid and envelope (38). In retroviruses, a reverse transcription step precedes the replication of the genome and the transcription of DNA. In RNA retroviruses, virion enzyme reverse transcribes (+) ssRNA into dsDNA intermediates upon infection (43). Once a virus genome has been integrated into the host genome, that integrated virus genome is transcribed into mRNA and is translated by the host synthesis machinery (44). Like in DNA viruses, the transcription and genome replication are also highly regulated in retroviruses for efficient viral replication (45).

Viruses depend on the host to supply tools for their DNA/RNA replication and transcription. New viral RNA complementary synthesis from an RNA template is described by the term “vRNA synthesis”, whereas mRNA synthesis from an RNA or DNA template is described by the term “vRNA transcription”. In this text, the term “viral genome replication” includes both vRNA synthesis as well as viral DNA replication. In this section, we review the roles of hnRNPs in viral genome replication and vRNA transcription in RNA, DNA, and retroviruses.

RNA viruses: One of the best-studied RNA viruses is human poliovirus (PV). During PV infection, hnRNP C binds to the 3′ end of the PV (-) ssRNA through its RRM domain and interacts with the PV replication protein 3CD (52). This binding further recruits 3CD to the (-) ssRNA template (52). The chaperone activity of hnRNP C helps to fold (-) ssRNA into a replication-compatible conformation, which promotes RNA synthesis (52). Additionally, in the replication initiation complex at the 5′ end, PV encoded RNA-dependent RNA polymerase (RdRp) is activated by the cleavage of 3CD. hnRNP E2 directly interacts with PABP bound to the 3’ poly(A) sequence and carries activated RdRp to its 3’end (53, 54). Further, hnRNP C interacts with both 5’- and 3’-end sequences of PV (-) ssRNA, enabling the convergence of 5’- and 3’-end sequences, which facilitates (+) ssRNA synthesis (55). On the other hand, upon the binding to viral RNAs, hnRNP E2 is cleaved by the PV-encoded RdRp precursor (3CD) or protease (3C) proteins and, as a consequence, loses one of its three RNA-binding sites (as explained in the translation section below) (56, 57). In the cleaved form of hnRNP E2, two RNA-binding sites are retained, allowing the protein to bind to the cloverleaf (CL) structure located at the 5’ untranslated region (UTR) of PV (+) ssRNA (56, 57). This makes PV RNA replication possible, as the poly(A)-binding proteins (PABPs) unite the 3’ and 5’ ends of the viral RNA ( Figure 3B ) (56, 57).

For other RNA viruses, including coronavirus and its prototype mouse hepatitis virus (MHV), hnRNP A1 specifically binds to the leader and intergenic sequences in MHV (-) ssRNA (58, 59). MHV’s RNA synthesis is stimulated by overexpression of wild-type hnRNP A1 while its replication is inhibited by the expression of a dominant-negative mutant hnRNP A1 (carrying a 75 amino acid deletion from the C-terminus) (60). It is evident from an in vitro reconstitution assay that the hnRNP A1 interacts with the leader and intergenic sequences and form a RNP complex which may participate in MHV RNA transcription (61). Further, hnRNP A1 interacts with MHV’s nucleocapsid protein (N), which suggests that hnRNP A1 might recruit the viral N protein to MHV transcription sites (62). Moreover, N protein recruitment to replication-transcription complexes (RTCs), via binding to non-structural protein 3 (nsp3), is a critical step for optimal MHV viral RNA synthesis (63). hnRNP I is also strongly associated with (+) ss leader sequence and with the 5’ UTR of (-) ssRNA (64). RNA affinity purification identified the interaction between hnRNP Q and the 5’ NCR of MHV genomic RNA and 5’ UTR of antigenomic RNA (65). This interaction is not limited to in vitro, but also to MHV-infected cells that overexpressed hnRNP Q (65). Moreover, an hnRNP Q mutant lacking the C terminus interferes with viral replication but does not affect viral translation which indicates that the hnRNP Q participates in the synthesis of RNA (65). In SARS CoV infection, hnRNP A1’s C terminal region has a high affinity and directly binds to the SARS’s N protein, suggesting that the interaction facilitates viral transcription and replication (66–68). In addition, hnRNP A2/B1 along with ILF3, QKI, and SFPQ interact with the SARS-CoV-2 genome and promote RNA synthesis (69). There are other examples of the involvement of hnRNPs in viral RNA synthesis used by other RNA viruses, and they are summarized in Table 2 .

DNA viruses: Unlike RNA viruses where there are many interactions between hnRNPs and viruses, there are only two main examples in DNA viruses. In a hepatitis B virus (HBV), APOBEC3B (A3B) protein inhibits HBV expression by inhibiting the binding of hnRNP K to the enhancer II of HBV (84). A3B also directly suppresses HBV’s S gene promoter activity (84). Enhancer II of HBV plays an important role in activating the transcription of the HBV gene (85). HBV expression may be inhibited by A3B at the transcription level by reducing hnRNP K’s interactions with the enhancer II of the virus (84). According to this model, A3B and hnRNP K may form a super-complex that may affect hnRNP K’s binding to the enhancer II of HBV by changing its conformation (84). In another DNA virus, Kaposi’s sarcoma-associated herpesvirus (KSHV), the interaction of RNA polymerase II (RNAP II) with hnRNP P negatively affects serine-2 phosphorylation of RNAP II’s C-terminal domain at the replication and transcription activator gene (RTA) of KSHV, resulting in decreased nascent RNA synthesis (86). Further, knockdown of hnRNP P increases transcription of RTA, and enhances expression of KSHV lytic genes (86).

hnRNPs are also associated with viral DNA replication. Lytic replication is a major step of herpesvirus pathogenesis (87). Immunocompromised individuals are often at risk of life-threatening diseases from the lytic phase of human cytomegalovirus (HCMV) infection (88), and also the lytic phase contributes to cancer cell evolution during chronic infections of KSHV (87). Several hnRNPs are associated with lytic DNA replication of herpesviruses. A herpesvirus’s lytic DNA replication begins at an origin (oriLyt) and requires trans-acting elements (89). In addition to its role in promoting lytic DNA synthesis, oriLyt is also thought to play a role in regulating the transition between lytic and latent phases (89). In case of HCMV, hnRNP K is required for the efficient lytic DNA replication by interacting with the oriLyt DNA and virally encoded UL84 protein (89). RNAi knockdown of hnRNP K prevents amplification of oriLyt in a transient assay (89). Similarly, several hnRNPs such as hnRNP A1, hnRNP A1/B1, hnRNP A3, hnRNP C, hnRNP D, hnRNP G, hnRNP K and hnRNP U were shown to be associated with the KSHV oriLyt DNA in a mass spectrometry study (90).

Retroviruses: Human immunodeficiency virus (HIV) is the well-studied retrovirus with many demonstrated interactions between hnRNPs and the viral RNA synthesis. HIV’s pathogenicity factor, Nef forms a protein complex with Lck kinase and nPKC subfamily (PKC δ and θ) proteins that form a complex called NAKC (Nef-associated kinase complex), which is required to increase viral replication and infectivity (91). The hnRNP K interacts with Nef and nucleates Nef-interacting kinases, including Lck, PKC δ, and PI-3 kinase, leading to the activation of Lck and Erk1/2 (92). Thus, hnRNP K may mediate transcriptional de-repression by coordinated membrane signaling with Erk1/2, and HIV targets it to enable Tat-dependent transcription (92). Another hnRNP, hnRNP G regulates HIV-1 repressive trimethylation of histone H3 lysine 9 (H3K9me3) by binding to proviral DNA at the LTR downstream region (93). This results in the blockage of the recruitment of the phosphorylated RNA polymerase II and substantially impediment of transcription, which helps the virus to maintain its latency (93). Another study showed hnRNP A2/B1 as a HIV’s long terminal repeat (LTR)-G-quadruplex-interacting protein that functions as a transcription activator (94). Moreover, silencing of the hnRNP A2/B1 in cells decreases the LTR activity (94).

hnRNP proteins control viral mRNA splice site selection, promote viral RNA polyadenylation, and stabilize viral RNA in a variety of ways.

RNA viruses: hnRNP K helps to splice influenza A virus (IAV)’s M mRNA segments by binding M mRNA downstream of the M2 5’ splice sites and promoting U1 snRNP recruitment (95). A mutation in either or both of the hnRNP K and NS1-BP-binding sites results in a mis-spliced M segment and diminished replication of the IAV (95). Further, M1 mRNA segment is properly spliced to yield M2 mRNA by NS1-BP, without affecting either M4 or NS mRNA segments. hnRNP K plays a role in mediating the interaction of NS1-BP and M RNA (96). Another study showed that nuclear speckles of hnRNP K increase in IAV-infected cells, which could influence the accessibility of hnRNP K to host transcripts, thereby enhancing the replication of IAV (97).

In case of a Rous sarcoma virus (RSV), a negative regulator of splicing (NRS) element is located in the pre-mRNA of the gag gene, which inhibits RNA splicing (98). hnRNP H, serine/arginine-rich (SR), and U1/U11 snRNPs bind to the NRS element and inhibit splicing by causing a decoy 5′ splicing site selection, subsequently promoting polyadenylation (99, 100).

hnRNPs also stabilize PV mRNA. PV mRNA does not possess a 5’-terminal cap structure (101), and uncapped mRNAs are rapidly degraded by 5’ exonuclease (102). Yet, PV mRNA is very stable despite the absence of a 5’-terminal cap in HeLa S10 translation-replication reaction (103). PV’s genome possesses a 5’ cloverleaf structure that binds to hnRNP E, and this binding stabilizes PV (+) ssRNA ( Figure 3A ) (103, 104). The mutant hnRNP E protein that is unable to bind to the cloverleaf structure causes PV (+) ssRNA to degrade rapidly (103). Since in-vitro 5’ 7-methylguanosine capped PV (+) ssRNA can bypass the requirement for hnRNP E in stability assays, it is likely that binding of hnRNP E to PV (+) ssRNA blocks XRN1 5’-to-3’ exonuclease-mediated RNA decay (103, 104). In another example, rabies virus glycoprotein mRNA transcripts steal host hnRNP E2, which leads to increased transcript abundance and stability (105).

DNA viruses: hnRNP A1 is shown to regulate human papillomavirus (HPV) splicing and polyadenylation ( Figure 4 ). HPV life cycle progression occurs in conjunction with host epithelial cell differentiation (106). The expression of nuclear hnRNP A1 increases during differentiation of HPV-16 infected epithelial cells, suggesting that hnRNP A1 is necessary for appropriate alternative splicing for late transcript synthesis (107). Further, hnRNP A1 also directly interacts with the HPV-16 late regulatory element (LRE) in the differentiated HPV-16-infected cells (107). Several post-transcriptional mechanisms have been identified as part of the LRE’s regulation of late gene expression. The HPV-16 genome is known to contain four 5’ splice donor sites (SDs) (e.g., SD3632) and seven 3’ splice acceptor sites (SAs) (e.g., SA409, SA742, SA2709, SA5639) (108, 109). A splicing silencer element (SSE) interacts with hnRNP A1 and hnRNP A2 to inhibit splicing at the SA409 to modulate E6 and E7 mRNA transcripts ( Figure 4 ) (108, 109). Another hnRNP, hnRNP G interacts with splicing enhancer element (SEE), and this interaction promotes splicing to SA2709 and enhances E2 mRNA production through splicing factor U2AF65 ( Figure 4 ) (110). The splice sites SD3632 and SA5639 are responsible for balancing the production of HPV 16 L1 and L2 transcripts. A splicing silencer element suppresses SD3632, allowing polyadenylation at an early polyadenylation signal (pAE). In particular, SD3632 and SA5639 are found to be “hot spots” for RBPs including hnRNPs ( Figure 4 ) (111). The hnRNP A2/B1, hnRNP D, and hnRNP L bind to the SSE of the SD3632 to modulate L1 and L2 transcripts (112–114). In addition, overexpression of hnRNP C reduces suppression of the splice site SD3632 in an HPV-16 early 3′-UTR-dependent manner (115). Similarly, to modulate L1 and L2 mRNA, hnRNPs such as hnRNP A1, A2/B1, and hnRNP L bind to the SSE at the SA5639 that is located immediately upstream of the AUG sequence at Late 1 (L1) mRNA (114, 116, 117). Further, binding of hnRNP A1 to the SSE suppresses the use of the SA5639, and results into the inhibition of late mRNAs splicing by influencing SSE, preventing L1 expression prematurely (118, 119). This hnRNP A1 inhibition of HPV 16 mRNAs makes the virus infection undetectable by the host immune surveillance and allows for viral persistence (116). Moreover, hnRNP L binds to SD3632 and SA5639 in response to Akt-dependent phosphorylation (114).

In addition to the above mechanism, hnRNP D inhibits the splicing of HPV-16 E1/E2 and E6/E7-mRNAs, thereby causing E1-and E6-mRNAs to contain introns, respectively (120). As a result, the cytoplasmic levels of intron retained HPV-16 mRNAs increases in the presence of hnRNP D (120). hnRNP A1 promotes HPV-16 E6 exon exclusion, whereas Brm and Sam68 promote exon inclusion (121). hnRNP I also blocks splicing inhibitory elements located at the downstream and upstream of HPV-16’s major late 5’ splicing site (SD), thereby enhancing gene expression in the late stage (122). Moreover, some of the examples of hnRNPs on the viral RNA splicing used by other DNA viruses are summarized in Table 3 .

In case of HPV polyadenylation, an early polyadenylation signal (pAE) plays a significant role in the transition from the early to late stages of viral lifecycles (125). An GGG motif located 174 nucleotides downstream of the pAE is known to act as a signal for hnRNP H to downregulate the late gene L2 (125). As a result of this hnRNP H binding, polyadenylation at pAE is enhanced, inhibiting L2 mRNA synthesis (125). It is necessary to read-through to the late region of L2 mRNA and polyadenylate the late transcripts at the late polyadenylation signal (pAL) in order to synthesize L2 mRNA ( Figure 4 ) (125). L1 mRNA captures hnRNP H and releases the inhibitory effects on L2 at the late infection stage (125, 126). This late gene autoregulation mechanism promotes the production of viral capsid protein (125, 126). In addition, in the splice site SD3632, activation of the DNA damage response increases binding of hnRNP C and inhibits polyadenylation at pAE ( Figure 4 ) (127).

Similarly, in the case of Epstein-Barr virus (EBV), hnRNPs helps to compensate the inefficient processing of the mRNA through polyadenylation process (128). Inefficient cleavage and polyadenylation of EBV DNA polymerase (pol), which contains a noncanonical polyadenylation signal, UAUAAA, is thought to be regulated by the EBV early proteins SM and M (128). hnRNP C1/C2 and hnRNP A1/A2 interact with SM suggesting that the hnRNPs participate in the pre-mRNA 3′ processing of the intronless pol transcript and compensate the inefficient processing of the mRNA (128).

Retroviruses: Several hnRNPs regulate HIV mRNA splicing. Following infection, short spliced mRNAs are produced that encode the viral regulatory proteins Tat, Rev, and Nef, which are required in the early stages of viral infection (129, 130). In response to their expressions, the partially spliced mRNA coding for Vif, Vpr, Vpu, and Env are produced and exported to the nucleus (129, 130). Lastly, the 9 kb transcript encodes both viral structural proteins, Gag and Pol (129, 130). Virus maturation and multiplication depend heavily on the control and coordination of these splicing activities. This post-transcriptional regulation is mediated by proteins Rev and Tat, and by four alternative 5’ splice sites and eight alternative 3’ splice acceptor sites ( Figure 5 ) (130).

A few hnRNPs modulate Tat protein expression by interfering with enhancer splicing elements (ESEs) or interacting with the exonic and intron splicing silencer (ESS and ISS) elements. hnRNP A/B bind the Tat exon 2 ESE to repress splicing, while SC35 binds to the ESE to activate it (131). Both hnRNP A1 and SC35 have overlapping binding sites within the ESE/ESS9 (131). The hnRNP A1 binds to the ESS and directly covers the SC35 binding site, inhibiting the splicing of the upstream intron (131, 132). Additionally, hnRNP A1 has also been shown to compete at ESE3/(GAA)3 with ASF/SF2 (131). Consequently, the ratio of ASF/SF2 to hnRNP A1 determines whether ESE3/(GAA)3 is activated or repressed at site A7 (133, 134). Inhibition of A3 splicing occurs when hnRNP A1 and hnRNP K bind to ESS2 (131, 135). The UAG triplets in ESS2 are required for hnRNP A1 binding (136). The hnRNP H also involves in splicing inhibition by binding to the ESS2p element, which results in the inhibition of splicing at 3′-splice site A3 (137). In addition, hnRNP A1 inhibits Tat splicing by binding with ISS, which is not a component of ESS2 and prevents entry of U2 snRNP (133, 138). hnRNP A/B proteins also inhibit splicing at 3′ splice site A2, which is responsible for the production of Vpr mRNA (139). hnRNP D also binds to ESSV and inhibits splicing (140). The hnRNP A/B protein binding at ESSV inhibits U2AF65 binding to the repressed 3′ splice site and interferes with the splicing efficiency of the 3′ splice site (139).

Interestingly, hnRNP H acts positively on the splicing of exon 6D of HIV mRNA, by interacting with CGGA and enabling the assembly of the U1 snRNP onto exon 6D (135, 141). Another study revealed that hnRNP H1 plays an important role in splicing enhancement by binding to its target elements (142). Depleted or mutated hnRNP H1 leads to diminished Vif expression and decreased replicative fitness, resulting from the uncontrollable splice acceptor A1 (142). hnRNP F also serves as a modulator of splicing events, indicating that hnRNPs are crucial factors for the regulatory pathways of each viral mRNA (143).

Viruses are fully dependent on the translation system of their host cells to produce the proteins that are essential for viral replication (40). Therefore, viruses recruit host ribosomes to translate viral mRNAs, which also concurrently suppresses host translation. Virus families are characterized by mRNAs that have a range of structures at their 5’ (capped, not capped, cap-substituted) and 3’ (polyadenylated or not) ends and carry various cis-acting sequences (such as internal ribosome entry sites, IRES) features that determine many aspects of viral mRNA translation (144).

RNA viruses: Most hnRNPs positively regulate virus mRNA translation. The IRES sequences of (+) ssRNA viruses drive translation of viral proteins (145). IRESs were initially discovered in picornaviruses where they have been shown to directly recruit ribosomes to viral mRNA ( Figure 3A ) (146). Several hnRNPs bind to the IRES of different viruses and act as IRES-trans acting factors (ITAFs) that assist with the translation of viral mRNAs.

During hepatitis C virus (HCV) virus infection, the majority of hnRNPs promote translation of HCV viral mRNA by binding to the 5′ UTR sequence (6). hnRNP D interacts with the stem-loop II, and its overexpression promotes HCV IRES-dependent translation, while hnRNP D knockdown inhibits the HCV IRES activity (147, 148). Additionally, hnRNP D sequestration with an interacting DNA polymer inhibits HCV mRNA translation in an in vitro system, reinforcing its role in the translation process (148). Other hnRNPs, such as hnRNP A1, E, I, L, and Q also stimulate the translation of HCV RNAs. hnRNP A1 binds to the cis-acting elements which are present in both the 5’ NTRs and the 3’ NTRs of HCV RNA (149). Similarly, hnRNP I interacts with NS5B, blocking the binding of NS5B to the 3’NCR, which positively affects the translation of HCV RNAs in vivo (150–152). hnRNP E plays a role in the circularization of HCV RNA through interactions with the 5’- and 3’UTRs for RNA translation and replication (153). hnRNP L bind to the 3’ border region of the HCV IRES and enhances mRNA translation (154). hnRNP Q assists the proper positioning of the 40S ribosomal subunit on HCV mRNA in order to facilitate IRES-dependent translation, which occurs through the interaction(s) between hnRNP Q and r-proteins of the 40S ribosomal subunit (155, 156). Additional roles of hnRNPs in the viral protein synthesis of RNA viruses are summarized in Table 4 .

DNA viruses: In HPV-16, hnRNPs control the translation of the late genes (HPV-16 genome is shown in Figure 4 ) (111). The hnRNPs E1, E2, and K bind to the HPV-16 L2 mRNA in a sequence-specific manner, inhibiting its translation (173). During the L2 translation inhibition, hnRNP K is phosphorylated on multiple tyrosines by c-Src, which inhibits its binding to RNA (174). In another DNA virus, KSHV, hnRNP I interacts with an IRES RNA within a vCyclin (latently expressed gene) that controls the expression of a downstream ORF encoding an inhibitor of apoptosis viral FLICE inhibitory protein (vFLIP) (175).

Retroviruses: In the HIV infected cells, hnRNP A1 is redistributed into the cytoplasm during late HIV infection and increases IRES-mediated translation (176). Knock-down of hnRNP A1 and A2 in HeLa cells transfected with a bicistronic dual-luciferase construct containing Gag under the control of the HIV-1 IRES decreases Gag synthesis (176). Similarly, depletion of hnRNP D expression decreases the synthesis of HIV-1 Gag and Env structural proteins (177). In addition, hnRNP D relocalizes to the cytoplasm upon HIV-1 infection and is associated with Gag protein (177). Moreover, the four isoforms of hnRNP D have differential effects on HIV-1 Gag expression, in which p45 and p42 isoforms increase viral Gag synthesis while p40 and p37 suppress it (177). It is subsequently confirmed by selective deletion of p45 and p42 that relative abundance of hnRNP D isoforms likely contribute to cell types’ ability to replicate HIV-1 (177). In another study, overexpression of hnRNP E1, but not hnRNP E2, lead to a reduction in Rev production, primarily through the suppression of viral RNA translation, and the C- terminal end of hnRNP E1 is required for this effect (178).

A retroviral IRES’s activity is not only influenced by the recruited ITAFs, but also by the post-translational modifications to these proteins (179). hnRNP A1 serves as an ITAF for HTLV-1 and MMTV IRESs, with MMTV IRES showing a higher dependence on hnRNPA1 than those of HTLV-1 or HIV-1 (179). In addition, HIV-1 and HTLV-1 IRES exhibit equally responsive IRES-activity to hnRNP A1 and its phosphorylation mutants S4A/S6A, S4D/S6D, and S199A/D (179). It appears, however, that the S4D/S6D mutant stimulates MMTV-IRES activity significantly higher than the wild type hnRNP A1 (179). The PRMT5-induced symmetrical di-methylation of arginine residues of hnRNP A1 enhances the stimulation activity of the hnRNP A1 on HIV-1 and HTLV-1 IRES, while decreasing the stimulation activity on MMTV IRES. (179).

When cells are infected with (+) ssRNA viruses, the viral genomic (+) ssRNA serves not only as template for translation but also for genome replication (180). Although the two processes, translation and replication, rely on the same strand of (+) ssRNA, their path to completion is opposite (180). The viral replication proteins may repress translation and switch to viral RNA synthesis for a variety of reasons, one being that viral genome (+) ssRNAs must avoid degradation by mRNA decay pathways after translation in order to continue replication (180). The host RBPs, including hnRNPs, contribute to the regulation of this process by interacting with cis-elements in the viral (+) ssRNA (54, 181).

In PV genome replication and translation, hnRNP E2 has two functions, which makes it an ideal mediator of the switch from translation to replication ( Figures 3A, B ) (54, 181–183). During mid-to-late infection, the viral protease 3C/3CD cleaves hnRNP E2 to facilitate PV’s genome replication (57; 184). A full-length hnRNP E2, which contains three RNA binding domains (KH-domains) ( Figure 1 ), binds selectively to the stem-loop IV of IRES (as shown in Figure 3A ) (185, 186). However, proteolytic cleavage of the linker sequence between KH2 and KH3 results in the truncated hnRNP E2 protein being devoid of a KH3 domain, which blocks its binding to the IRES and inhibits PV translation (57; 184). Interestingly, in hnRNP E2-depleted extract, truncated hnRNP E2 still rescues PV (+) ssRNA replication (57). With the similar mechanism, hnRNP K plays dual role in foot-and-mouth disease virus (FMDV) translation and RNA replication (187). hnRNP K’s KH2 and KH3 domains bind directly to FMDV IRES domains II, III, and IV, inhibiting IRES-mediated translation by interfering with the recognition of another positive ITAF, hnRNP I (187). The viral protease, 3C, however, antagonizes the inhibition of hnRNP K through cleavage at the Glu-364 residue of hnRNP K during FMDV infection (187). This result indicates that the N-terminal cleavage product is capable of inhibiting IRES activity while the C-terminal cleavage product promotes FMDV replication (187).

hnRNP A2/B1 binds NS1 and NS2 mRNAs of IAV and inhibits the nuclear export of the NS1 mRNA (188). Conversely, knockdown of hnRNP A2/B1 promotes the transport of NS1 mRNA to the cytoplasm in infected cells and results in increased levels of NS1 (vRNA, mRNA, and protein) consequently leading to enhanced virus replication (188). Therefore, the hnRNP A2/B1 inhibits the replication of IAV in host cells possibly by suppressing nucleocytoplasmic translocation of NS1 mRNA (188). In addition, nucleoprotein (NP) cooperates with hnRNP A/B to further restrict viral mRNA nuclear export by acting as an RNA retention factor for IAV mRNA (189). An additional consequence of IAV NP’s interaction with hnRNP A/B is that it interrupts the interaction between ALY and UAP56 (RNA helicase), leading to repression of the ALY-viral mRNA binding, followed by inhibition of viral mRNA binding for NXF1 (mRNA transport receptor) (189).

For retroviruses, viral RNA trafficking from nucleus to cytoplasm is the major event in the viral lifecycle. In HIV, hnRNPs facilitate translocation of viral RNAs. Several spliced and unspliced RNA molecules of HIV exist in the nucleus after transcription and splicing (6). hnRNP D may assist in HIV RNA translocation, because loss of hnRNP D reduces the accumulation of HIV-1 unspliced and singly spliced RNAs in the cytoplasm but does not affect viral RNA abundance (177). HIV-1 Vpr and Gag mRNAs contain cis-acting RNA trafficking sequences (hnRNP A2RE-1 and A2RE-2) that are responsible for cytoplasmic RNA trafficking (190, 191). hnRNP A2 and both A2REs bind to form a complex in vitro, which is required for the nuclear export of RNA in oligodendrocyte (190, 191). RNA transport mediated by the A2RE requires both microtubules and hnRNP A2 (190, 191). As a consequence, the replication of A2RE-2 mutant viruses is significantly compromised and genome RNA encapsidation is reduced (191). hnRNP A2 is involved in carrying HIV-1 RNA to the microtubule-organizing center (MTOC) and then transporting it to the plasma membrane for virion assembly (177, 192, 193). Although inhibiting the localization of genomic RNA at the MTOC as a result of hnRNP A2 knockdown does not impact Gag synthesis, it adversely affects viral production and infection (192).

Similarly, hnRNP A2B1 and hnRNP C can retain the genomic RNA in the nucleus. hnRNP A2B1 participates in the nuclear retention of the HIV-1 genomic RNA in the absence of Rev as well as in the release of the genomic RNA from cytoplasmic ribonucleoprotein (RNP) complexes (194), but the exact mechanism is unknown. hnRNP C binds to the nuclear retention sequence (NRS) within the intron of unspliced RNA and mediates nuclear retention of the RNA (195). hnRNP U also controls HIV-1 mRNA nuclear export by targeting the 3′ long terminal repeat (3′LTR) and blocking HIV-1 mRNA accumulation in the cytoplasm (196). Interestingly, HIV-1 also increases hnRNP A1 expression and promotes its relocation to the cytoplasm. The cytoplasmic translocation of hnRNP A1 is escorted by the viral genomic RNA and is synergistically induced by Rev (197). During the late phases of HIV-1 replication, hnRNP A1 and viral genomic RNA colocalize in the cytoplasm, supporting a post nuclear function of hnRNP A1 (176). Additional examples of hnRNPs involvement on viral RNA trafficking and packaging are summarized in Table 5 .

Viral egress involves capsid assembly and release, where hnRNPs play a role in either one or both steps. In dengue virus (DENV), vimentin interacts with hnRNP C1/C2, hnRNP K, and NS1 protein to form a complex, and the dissociation of the complex by acrylamide treatment reduces viral release (203). The vimentin-hnRNP K-NS1 complex was also studied in Japanese encephalitis virus (JEV) infection, in which knocking down vimentin and hnRNP K reduces viral replication, whereas over-expression of vimentin and hnRNP K enhances viral replication (204). Moreover, vimentin was also shown to interact with hnRNP K following JEV infection or after overexpression of NS1 (204).

In contrast, hnRNP E1 normally limits virion assembly and secretion, which prevents the escape of viral genomes packaged into virions that contribute to viral RNA accumulation in infected cells (205). Similarly, hnRNP K suppresses HCV particle production. The hnRNP K protein is recruited to the virion assembly sites in proximity of lipid droplets and the hnRNP K colocalizes with core protein and HCV plus-strand RNA (206).

Viruses also utilize hnRNPs for controlling the apoptosis process for the release of mature virions as an indirect mechanism. EV71 releases its virus through hnRNP A1 via an apoptotic axis involving 3C protease, hnRNP A1, Apaf-1, and caspase 3 (207). When EV71 infection is present, or 3C protease is expressed, hnRNP A1 is cleaved, which affects its binding to apaf-1 IRES (207). This allows IRES-dependent synthesis of Apaf-1, and subsequent activation of caspase-3, and apoptosis (207). Conversely, hnRNP K suppresses bovine ephemeral fever virus (BEFV) replication by degrading viral α3 gene and thus inhibits apoptosis induced by α3 gene (208). Moreover, BEFV infection degrades hnRNP K protein in BHK-21 cells, which is mediated by viral activation of caspase 3 (208). However, the hnRNP K has a supportive role in the HSV-1 virion egress (209). Infectious particles are released by the exocytic pathway after newly assembled HSV-1 particles acquire a secondary envelope at the trans-Golgi network (209). The downregulation of hnRNP K leads to a blockage of virus particles leaving the cell, and the virion release is crucial for viral propagation (209). Similarly, the hnRNP A2B1 supports the egress of exosomes and HSV-1 from infected cells (210).

Viral nucleic acids and/or proteins encounter hnRNPs in these cellular compartments and alter their subcellular localization to promote viral replication. The Ebola virus (EBOV) prevents hnRNP C1/2 from entering the nucleus (211). VP24 from the EBOV binds to NPI-1 subfamily karyopherin α (KPNA) nuclear import proteins, which prevents KPNA from interacting with phospho-STAT1 (211). This results in the inhibition of nuclear import of the hnRNP C1/C2-KPNA-VP24 complex (211). Similarly, hnRNP A1, hnRNP K, and hnRNP C1/C2 are relocalized to the cytoplasm during vesicular stomatitis virus (VSV) infection (212). Recently, it was shown that open-reading frame 6 (ORF6) protein of SARS-CoV-2 can directly manipulate localization and functions of nucleoporins by disrupting nuclear rim of nucleoporins, RAE1 and NUP98 (213). This disruption results into aberrant nucleocytoplasmic trafficking and leads to nuclear accumulation of hnRNP A1 (213). Similarly, in case of junin virus (JUNV), hnRNP A1 interacts with the N protein of JUNV in acutely infected cells, and upregulating N protein induces cytoplasmic re-localization of hnRNP A1 (214). JUNV as well as DENV-2 induce hnRNP K cytoplasmic translocation to favor viral particle production (214, 215).

Many hnRNPs relocalize to the cytoplasm upon rotavirus infection. Rotavirus replicates in the cytoplasm of infected cells in viroplasms (VMs, unique inclusion bodies induced by the virus), which are nucleated by two essential virus nonstructural proteins, NSP2 and NSP5 (216). NSP2 and NSP5 interact directly with the hnRNPs (which are translocated to cytoplasm from the nucleus) in an RNA independent manner and sequester themselves in the viroplasms of infected cells, thus creating favorable conditions for virus replication (216). In another example, Seneca Valley virus (SVV) 3C protease (SVV 3C) cleaves hnRNP A1 and this cleaved hnRNP A1 translocates to the cytoplasm, which is utilized by SVV to aid viral replication (217).

hnRNPs interact with the viral proteins to control the viral gene expression or translocation of the viral proteins. hnRNP A1 interacts with multiple proteins of different viruses. hnRNP A1 interacts with p17 protein of avian reovirus (ARV) to facilities its nuclear import and export (218). hnRNP A1 also interacts with nucleoprotein (NP) of IAV and co-localize in both nucleus and mitochondria of mammalian cells that are either transfected with NP or infected with IAV (219). Similarly, hnRNP A1 interacts with nucleocapsid (N) protein of porcine epidemic diarrhea virus (PEDV) and enhances the PEDV infection (220). During infection with DENV, hnRNP C1/C2 and hnRNP I accumulate in the cytoplasm and colocalizes with NS1 and NS2 viral proteins (221, 222). DENV core protein is also shown to interact with hnRNP K by coimmunoprecipitation analysis (221). Moreover, silencing of hnRNP I inhibits virus replication, while overexpression of hnRNP I enhances DENV replication (223). hnRNP H also interacts with DENV non-structural 1 protein to promote the viral replication (224). During IAV infection, the central BACK domain of cellular protein NS1-BP interacts directly with hnRNP K and hnRNP I and with the influenza virulence factor, NS1 protein (225). In addition, proteomic analysis revealed that hnRNP M and hnRNP H1 are part of the complexes formed by the IAV polymerase (226). hnRNP H interacts with Newcastle disease virus (NDV) V protein to promote viral replication (227). Rev protein, which is a key regulator of HIV-1 gene expression, also interacts with hnRNP A1, hnRNP Q, hnRNP K, hnRNP R, and hnRNP U (228). This interaction of Rev with functionally diverse hnRNPs is proposed to be involved in HIV replication and virus-host cell interactions (228).

The presence of infection triggers innate immune responses against viruses in mammalian cells. The cells detect viral dsRNA by means of cytoplasmic receptors, such as the retinoic acid-inducible gene I (RIG-I) and the melanoma differentiation-associated gene 5 (MDA5), both of which interact with the adaptor mitochondrial antiviral signaling (MAVS) (229). The MAVS integrates signaling from RIG-I-like receptors (RLRs) and mediates the downstream activation of TANK-binding kinase 1 (TBK1)/IκB kinase ε, eventually leading to the induction of type I IFNs, pro-inflammatory cytokines, and multiple IFN-stimulated genes (229–232). This results in pacing cells in an antiviral state (229). hnRNP E1 and E2 are synergetic feedback inhibitors of MAVS, which facilitate the degradation of the MAVS protein by recruiting Atrophin 1-interacting protein 4 (AIP4, E3 ubiquitin ligase) to MAVS-containing complex after viral infection (233). This mechanism is considered as the fine tuning of antiviral innate immunity for the prevention of excessive harmful immune responses (232, 233). In addition, hnRNP M was shown to act as a negative regulator of RLRs-mediated signaling, where hnRNP M interacts with RIG-I and MDA5 and impairs the binding of the RLRs to viral RNA, leading to the inhibition of RNA virus-triggered innate immune responses (234).

hnRNP U (a nuclear matrix protein functions as a nuclear viral dsRNA sensor for both DNA and RNA viruses. As a response to viral dsRNA, hnRNP U oligomerizes and activates the distal enhancers of antiviral genes, including Interferon beta 1 (IFNB1) (235). Also, hnRNP U is required for the activation of super-enhancers, which initiate robust immune gene expression for antiviral defense (235, 236). Another study showed NLS domain of hnRNP U interacts with bunyavirus NP and recognizes the bunyavirus genomic RNA in the cytoplasm, which activates STING-TBK1 signaling axis against the virus infection (237). In addition, hnRNP U was shown to interact with a IVRPIE (Inhibiting IAV Replication by Promoting IFN and ISGs Expression, an lncRNA) (238). The expression of IVRPIE significantly inhibits IAV replication in A549 cells (238). Additionally, IVRPIE positively regulates interferon β1 transcription and several vital interferon-stimulated genes (ISGs) (238).

hnRNP A2B1 recognizes and binds to viral DNA in the nucleus during herpes simplex virus-1 (HSV-1) infection (239). The binding is followed by homodimerization and demethylation of hnRNP A2B1 which leads to the release of HSV-1 DNA (239). Subsequently, hnRNP A2B1 translocates to the cytosol and activates TBK1-IRF3 signaling pathway to initiate cytokine production, thus enabling innate immunity through type I interferon signaling for virus elimination, ultimately leading to the inhibition of HSV-1 replication (239). During an innate immune response to HSV-1 infection, hnRNP A2B1 dissociates from fat mass and obesity-associated protein (FTO), leading to the promotion of N6-methyladenosine (m6A) modification, nucleocytoplasmic trafficking, and translation of cGAS, STING, and IFI16 mRNAs to enhance IFN-I production (239).