The most studied cytoplasmic RNPs are SGs and PBs. These structures are highly dynamic centers of mRNA sorting, storage and degradation, where the processes of splicing, non-sense mediated RNA decay (NMD), translation, turnover and RNA silencing intersect ( Figure 1 ) (22–27). SGs are transient foci enriched in translation initiation factors and 40S ribosomal subunits, while PBs are enriched for RNA decay machinery. SGs and PBs each contain unique marker proteins although many proteins can be found in both SGs and PBs, such as eukaryotic translation initiation factor 4E (eIF4E), AU-rich RNA binding protein tristetraprolin (TTP), argonaute RISC catalytic component 2 (AGO2), apolipoprotein B mRNA-editing complex 3 (APOBEC3), poly(rC) binding protein 2 (PCBP2) and others (23, 28, 29). In fact, apart from having shared protein components, PBs and SGs dynamically exchange RNP cargo, often form co-aggregates and have been proposed to serve as nucleation sites for SGs formation (30, 31). SGs, PBs and other cytoplasmic foci are highly dynamic structures, although PBs are quite stable over the time (32). They are in a dynamic steady state with other RNPs, such as polysomes, in response to the translational state of the cell (23) ( Figure 1 ).

SGs are 200-400 nm dynamic structures quickly formed when cells encounter external stresses and translation rates decline, and disperse when translation conditions are restored (33, 34). The most commonly described triggers of SG formation are oxidative, starvation, heat stresses and infection with different pathogens. These stimuli activate eukaryotic translation initiation factor 2 alpha (eIF2α) kinases: heme-regulated kinase (HRI), general control non-depressible 2 kinase, (GCN2), dsRNA-activated protein kinase R (PKR) and PKR-like endoplasmic reticulum kinase (PERK). The eIF2α phosphorylation blocks translation, forcing an accumulation of the stalled 43S and 48S ribosomal preinitiation complexes. Inhibition of the function of eukaryotic translation initiation factors eIF4G or eIF4A (DDX2A) is also linked to SG formation (35) and some mechanisms of SG formation can proceed in the absence of eIF2α phosphorylation (36–38). The components of SGs can be classified into three main groups: core components, such as stalled initiation complexes; RNA-binding proteins associated to silencing and transcript stability, such as the scaffold proteins T-cell-restricted intracellular antigen 1 (TIA-1) or TIA-1-related protein (TIAR), and RNA-binding proteins associated to mRNA metabolism (23, 39–42). The key function of SGs is to protect the mRNAs during cell stress, altering the composition of the RNPs in a reversible manner (43). Moreover, SGs are generally believed to have an antiviral activity upon viral infection and many viruses manipulate SGs to evade host responses (2, 44–47).

PBs act as temporary reservoirs for non-translated mRNAs which may further enter translation or be degraded. PBs contain components of the 5’ to 3’ decay machinery, NMD pathway and RNA-induced silencing machinery (48). This enrichment on mRNA decay machinery components led to propose that significant RNA decay occurs within PBs. However, this is controversial since some activators of mRNA decay pathways found in PBs are translational regulators also present in SGs (29, 49). PBs are constitutively present in the eukaryotic cells but increase in size and number when translational arrest occurs, coincident with SG formation, resulting in physical interactions among SG and PB structures (23, 30). It has been described that mRNA degradation occurs in PBs and depends on the existence of degradation enzymes and mRNA degradation intermediates (50, 51). Many of the PB components are not restricted to these foci and are also present in the soluble cytoplasm and nucleus, suggesting that the different processes might start before the mRNAs entry into PBs.

Virus infection activates cell stress responses at many levels that modulate RNA granules. In general, RNP granules can represent an obstacle for virus replication and also serve as sensors to mount the innate immune response. Therefore, viruses have evolved different mechanisms to control the assembly and functions of RNP granules and, in some cases the components of RNPs are co-opted into novel virus-specific structures required for virus replication (52).

Mammalian orthoreovirus (MRV), a dsRNA virus, induces SGs at an early stage of the infection, correlating with reduced translation of both cellular and viral mRNAs and increased phosphorylation of eIF2α (53). Poliovirus inhibits SGs assembly at the late stages of infection, by a mechanism involving cleavage of SG protein Ras GTPase-activating protein-binding protein 1 (G3BP1), which is mediated by the viral 3C protease (54). In addition, TIA-1-containing granules persisted even at late stage of poliovirus infection, but those RNPs are remnants of normal SGs that do not correlate with translational repression (55, 56). Unlike MRV and poliovirus, Influenza A virus (IAV) prevents SG formation at all stages of the infection, by a mechanism depending on NS1 protein expression, since SGs inhibit IAV replication (57).

As an example of viral co-opting of RNP components, RNAs from flaviviruses, such as West Nile virus and Dengue virus, bind TIA-1 and sequester many PB components near or within perinuclear viral replication centers while the number of PBs was decreased (58). Human immunodeficiency virus (HIV)-1 TAT protein interacts with the PB component DDX3 to facilitate viral mRNA translation (59), and HIV-1 TAT protein is antagonized by antiviral factors APOBEC3 and Moloney leukemia virus 10 (MOV10), that are PB constituents (60–63).

CoVs are single-stranded RNA viruses with large genomes of around 30 Kb, which life cycle occurs in the cytoplasm of the infected cell. CoV RNA synthesis, including replication and transcription, is a complex process involving host cell membrane rearrangements, viral and cellular proteins (64, 65). Regarding the role of cytoplasmic RNPs in CoV infection, it has been shown that mouse hepatitis virus (MHV) replication induces host translational shutoff and mRNA decay, with concomitant formation of SGs and PBs (66). Porcine transmissible gastroenteritis virus (TGEV) triggers SG formation and interferes with PB formation, correlating with viral replication and transcription (67). Severe and acute respiratory syndrome (SARS)-CoV, SARS-CoV-2 and avian infectious bronchitis virus (IBV) nucleocapsid (N) proteins have been found to interact with the SG component G3BP1, and this has been proposed as a mechanism to avoid SG formation (68–72). In addition, porcine epidemic diarrhea virus (PEDV) promotes G3BP1 cleavage by caspase-8 (73) and Middle East respiratory syndrome (MERS)-CoV 4a accessory protein interferes with SG formation by inhibiting the activation of PKR binding to dsRNA, thereby inhibiting the formation of SGs and ensuring viral protein translation and efficient virus replication (74, 75). It has been recently reported that SARS-CoV-2, HCoV-229E and HCoV-OC43 cause PB disassembly. In the case of SARS-CoV-2, this process may be mediated by N protein (76).

RNA helicases share conserved helicase domains, all containing ATP binding motifs, providing the basis for helicases classification ( Figure 2 ). Most cellular RNA helicases belong to the SF2 superfamily and only a few of them, the up-frameshift suppressor 1(Upf1)-like helicases (i.e., UPF1, also known as RENT1, and MOV10 helicase), belong to SF1. In contrast, many viral RNA helicases, such as CoV nsp13, belong to SF1 superfamily (77). The SF1 and SF2 helicases contain seven to nine conserved motifs that constitute the helicase core. In addition, SF2 RNA helicases, generally referred to as DExD/H box RNA helicases, are divided in five different subgroups ( Figure 2 ): DEAD box (for the conserved amino acid residues Asp-Glu-Ala-Asp), DEAH (for the conserved amino acid residues Asp-Glu-Ala-His)/RHA, Ski2 (Superkiller-like 2)-like, retinoic acid-inducible gene I (RIG-I)-like, and Viral DExH proteins, named after one of the consensus amino acid sequence motifs (16, 78–80). The DEAD box (DDX) helicases are the largest subgroup within SF2, most of them involved in RNA metabolism, from transcription to degradation and in establishment of larger RNA-protein complexes, such as ribosomes (81, 82).

Cellular DDX RNA helicases play essential roles in a broad array of biological processes and serve multiple roles at the virus-host interface (83–86). Specifically, DDX RNA helicases are hijacked by CoVs and participate in essential DDX-mediated viral replication steps (87). During CoV infection, these RNA helicases interact with viral proteins, as described below ( Table 1 ). In addition, many of them interact with viral RNA (88–90) ( Table 1 ), raising the possibility that these RNA helicase-CoV protein interactions are mediated by RNA.

DDX1, interacts with the nsp14 protein both from SARS-CoV and IBV, suggesting that this interaction may be conserved for other CoVs. DDX1-nsp14 interaction contributes to efficient CoV RNA replication in cell cultures (91). DDX1 also interacts with the nsp14 protein from TGEV, inducing interferon (IFN) production (92) and suggesting that nsp14 would be the viral component sensed by DDX1-DDX21-DHX36 cytoplasmic sensor, described in dendritic cells (93). Supporting CoV interaction with the DDX1-DDX21-DHX36 complex, N proteins of SARS-CoV-2 and IBV interact with DDX21, although the consequences of this interaction are still unclear (71, 94). In addition, phosphorylation of MHV N protein allows the recruitment of DDX1 to the CoV replication transcription complex (RTC), increasing the synthesis of longer viral RNAs, suggesting that this is one of the viral strategies to support the transition from discontinuous to continuous transcription (95).

The interaction between DDX3X and N proteins from SARS-CoV, SARS-CoV-2, or IBV was described (71, 96). Since DDX3X is involved in the activation of the innate immune response at different levels (97–99), it was suggested that N protein can modulate the immune response by binding to DDX3X and inhibiting these antiviral pathways, as demonstrated for other viruses (100). SARS-CoV nsp13 interacts with DDX5 and inhibition of DDX5 results in the suppression of viral replication. A proviral function for DDX5 has been suggested, maybe acting as nsp13 co-activator during RNA synthesis (101). In addition, since DDX5 is involved in nuclear factor kappa-light-chain-enhancer of activated B cells (NF-kB) response, the binding between nsp13 and DDX5 can also have the role of evading host inflammatory response (102). Interestingly, DDX5 interacts with other DEAD box helicases, such as DDX3X and its close homolog DDX17. The phosphorylation-dependent interaction between DDX5 and DDX3X was proposed as a combined mechanism of action for DEAD box helicases involved in RNP remodeling and splicing, by forming a complex that functions in shuttling RNP export from the nucleus to the cytoplasm (103, 104).

DDX6, a helicase located in SGs and PBs, is downregulated during PEDV infection to reduce endoplasmic reticulum (ER) stress, facilitating viral replication (105). The accumulation of DDX6 is decreased by HCoV-OC43 infection, while it does not change with HCoV-229E or SARS-CoV-2 infection (76). DDX6 interacts with RIG-I to increase antiviral signaling and IFN induction (106). However, the potential effect of DDX6 innate immunity modulation during PEDV infection was not analyzed as IFN incompetent Vero cells were used.

There is also recent evidence of other RNA helicases being involved in the immune response during SARS-CoV-2 infection. DHX16 has been described as a novel sensor for SARS-CoV-2 replication, triggering IFN response (107). The levels of DHX9 expression in effector TCD8+ cells have been correlated with better COVID-19 outcome (108).

MOV10 protein is a UPF1-like RNA helicase (109, 110). MOV10 helicase has more than 1060 interactors, according to BioGRID database (111), some of them linking MOV10 to RNA metabolism pathways such as the NMD pathway or the siRNA gene silencing pathway. In fact, depletion of MOV10 mRNA by using siRNAs interferes with RNAi activity (110, 112). In addition, MOV10 is an IFN-stimulated gene (ISG) (113) and MOV10 protein is involved in IFN induction after viral infection (114).

Both proviral and antiviral functions have been reported for MOV10. As a proviral factor, it is required for hepatitis delta virus (HDV) replication but not for the translation of its mRNA (115) and facilitates enterovirus replication (116). As an antiviral factor it inhibits: (i) HIV-1 and other retroviruses replication at multiple steps (63), (ii) nuclear import of influenza virus nucleoprotein (117), (iii) porcine reproductive and respiratory syndrome virus (PRRSV) replication by avoiding nucleocapsid protein trafficking to the nucleus (118), (iv) hepatitis C virus (HCV) and Dengue virus replication by a partly unknown mechanisms (119, 120), and (v) bunyavirus replication by blocking several nucleoprotein functions (121). In the case of hepatitis B virus (HBV), contradictory data has been reported for both, its proviral (122) and antiviral activities (123).

The role of MOV10 during CoV replication has recently been analyzed by our group. The interaction between endogenous MOV10 and N protein during infection was demonstrated for porcine TGEV virus, and both mild (HCoV-229E) and highly pathogenic CoVs (SARS-CoV, MERS-CoV and SARS-CoV-2) (124). This interaction was also reported by other authors, outside the infection context using N protein overexpression (71, 94, 125). During MERS-CoV infection, both MOV10 and N protein co-localized in cytoplasmic RNPs, not related with CoV RTCs, that may eventually contain also other cellular proteins, such as TIAR, AGO2 or UPF1 (124) ( Figure 3 ). Functional analyses indicated that MOV10 has antiviral activity during MERS-CoV and SARS-CoV-2 infection, but not during HCoV-229E infection, suggesting a relationship with CoV pathogenesis that will be further explored (124). In agreement with other previously described MOV10 interactions with cellular proteins, MOV10 interaction with CoV N protein was RNA-dependent (124). Interestingly, MOV10 interaction with SARS-CoV-2 RNA was recently described (88, 90, 126–128) and it was conserved for other human CoVs (127). This issue opens the possibility of N protein having an active role in the MOV10 antiviral activity or of its non-specific recruitment to MOV10 RNPs, mediated by its binding to viral RNAs. In contrast with other viral infections, MOV10 helicase activity was required for antiviral function during MERS-CoV infection (124). The binding of MOV10 and CoV N protein was conserved independently of MOV10 antiviral activity, i.e., with a MOV10 mutant without helicase activity, or in HCoV-229E infection (124), suggesting that N binding to MOV10 is not related with MOV10 antiviral function. In addition, CoV N protein has been proposed to subvert SGs or counteract NMD pathway to facilitate viral replication (129, 130). Therefore, additional experimental evidence is needed to determine whether N protein may have an active role in the formation or function of MOV10 RNPs.