Negative and ambisense RNA virus ribonucleocapsids: more than protective armor

SUMMARY

Negative and ambisense RNA viruses are the causative agents of important human diseases such as influenza, measles, Lassa fever, and Ebola hemorrhagic fever. The viral genome of these RNA viruses consists of one or more single-stranded RNA molecules that are encapsidated by viral nucleocapsid proteins to form a ribonucleoprotein complex (RNP). This RNP acts as protection, as a scaffold for RNA folding, and as the context for viral replication and transcription by a viral RNA polymerase. However, the roles of the viral nucleoproteins extend beyond these functions during the viral infection cycle. Recent advances in structural biology techniques and analysis methods have provided new insights into the formation, function, dynamics, and evolution of negative sense virus nucleocapsid proteins, as well as the role that they play in host innate immune responses against viral infection. In this review, we discuss the various roles of nucleocapsid proteins, both in the context of RNPs and in RNA-free states, as well as the open questions that remain.

INTRODUCTION

Negative RNA viruses (NSV) and ambisense RNA viruses (ASV) cause detrimental human diseases, including Ebola hemorrhagic fever, influenza, Lassa fever, and measles. The genomes of negative sense RNA viruses consist of single-stranded RNA that can be segmented or non-segmented. We will refer to this viral genomic RNA as the viral RNA (vRNA). The vRNA molecules are bound by viral nucleoproteins (abbreviated NP or N, depending on the nomenclature used in each field; we will not abbreviate nucleoprotein here for clarity) and a viral RNA polymerase, forming viral ribonucleoprotein (vRNP) complexes (1–4). Upon entry into the host cell, the vRNPs must first be transcribed into capped and polyadenylated mRNAs by the RNA polymerase before host cell ribosomes can produce viral proteins. Following viral protein synthesis, the process of viral replication generates new vRNA molecules via a positive sense, complementary RNA (cRNA) intermediate. cRNA molecules are also bound by newly translated nucleoproteins and a copy of the viral RNA polymerase, forming cRNPs.

The RNA polymerase of most NSVs and ASVs is called the large (L) protein. In addition to an RNA-dependent RNA polymerase (RdRp) domain, the L protein can contain combinations of cap-binding and endonuclease, or capping and methyltransferase domains and may bind oligomers of the viral phosphoprotein (P) (1). In orthomyxoviruses, the RdRp domain resides in a heterotrimer of the viral proteins polymerase basic 1 (PB1), polymerase basic 2 (PB2), and polymerase acidic (PA) or polymerase 3 (P3). The PB2 and PA/P3 subunits of this protein complex also contain cap-binding and endonuclease domains (5). In contrast to the host cell’s DNA-dependent polymerases, the above RNA polymerases lack proofreading and robust error-repair mechanisms. This results in high mutation rates, which in turn help RNA viruses evade host immune defenses and adapt to host environments more quickly.

The nucleoprotein-encoding gene is present in all NSV and ASV genomes (6, 7). Nucleoprotein gene transcripts are typically the most abundant mRNAs during a negative and ambisense RNA virus infection, making nucleoproteins one of the most abundant viral proteins (8). Upon expression, nucleoproteins can self-oligomerize into symmetric helical and double-helical complexes in the presence or absence of vRNA or cRNA (9–12). These helical nucleoprotein multimers are inherently multifunctional and have evolved to balance a protective role that shields the vRNA of some NSVs, while allowing the vRNA to be accessed and copied by the RNA polymerase (8). Moreover, nucleoproteins minimize RNA secondary structure, especially long-range interactions, within the vRNP to maintain a contiguous organization that supports efficient replication, while at the same time allowing the folding of packaging signals (13, 14). Without this function, a vRNA may form secondary structures, such as template-loops (t-loops), that stall viral RNA synthesis and affect the generation of aberrant RNA molecules that activate the innate immune system (15–19).

Although nucleoproteins vary in length and sequence among different negative and ambisense RNA viruses, the structural principles of nucleoprotein-nucleoprotein interactions, as well as the structural symmetry of helical chain formation are mostly conserved. However, at a biochemical level, different NSVs and ASVs have evolved different solutions to coordinate nucleoprotein multimerization with RNA synthesis, requiring nucleoproteins to interact with P in non-segmented NSV infections, or host cell proteins, such as ANP32A, importins, kinases, and phosphatases in infections with segmented NSVs (20–22). Finally, nucleoproteins can also affect the antiviral immune response by modulating the function of key proteins within the immune signaling pathways, like TBK1 and IRF3, or by triggering host cell apoptosis and autophagy (23–26). In this review, we will discuss the vast number of functions of nucleoproteins from a structural perspective.

NUCLEOPROTEIN ARCHITECTURE

The overall size of nucleoproteins ranges greatly, from 233 amino acids in Bunyamwera virus (BUNV) to 739 amino acids in Zaire ebolavirus (EBOV). However, the core of all nucleoproteins has a high structural homology among NSV families (Fig. 1). Nucleoproteins are composed of two core domains that are connected via a positively charged region that is critical for RNA binding. The N-terminal domain and C-terminal domain of nucleoproteins form a conserved structural pattern, with five alpha-helices in the N domain (5 hours motif) and three alpha-helices in the C domain (3 hours motif) (note that the orientation of the helices varies) (27, 28). Many nucleoproteins also have intrinsically disordered N- or C-terminal regions (or tails) that play a key role in oligomerization by enabling head-to-tail or side-by-side interactions between nucleoprotein monomers, such as the paramyxovirus nucleoproteins (29–32). Oligomerization between nucleoprotein molecules is further stabilized through other, more dynamic interactions.

Fig 1.

Nucleoprotein structures. Examples of nucleoprotein structures showing a conserved bi-lobed core (left row) and positively charged RNA binding regions (right row). The influenza A virus (IAV) NP has non-contiguous head/body domains, whereas other nucleoproteins have terminally contiguous head/body domains, like the Lassa fever virus (LASV), Zaire ebolavirus (EBOV), and vesicular stomatitis virus (VSV) nucleoproteins. The EBOV and VSV nucleoproteins also have extended N-terminal tails. PDB accession numbers used to generate the models are indicated on the left of the panels, and colors are explained in the bottom legend.

NUCLEOPROTEIN OLIGOMERIZATION

Nucleoproteins interact to form large oligomeric structures (Fig. 2) in the presence and absence of RNA. Non-viral RNA can be encapsidated if nucleoprotein oligomerization is not modulated by additional (weak) interactions, such as with P (20, 32, 33). Nucleoprotein oligomerization appears to occur via two mechanisms: head-to-tail interactions and side-by-side interactions, with one of the two mechanisms driving the reaction, depending on the virus. Recent structural data suggest that nucleoproteins molecules in non-segmented NSV RNPs interact via side-by-side oligomerization [EBOV, rabies virus (RABV), vesicular stomatitis virus (VSV), measles virus (MeV), respiratory syncytial virus (RSV), cetacean morbillivirus (CeMV)], whereas segmented NSV RNPs oligomerize via head-to-tail interactions (e.g., IAV and BUNV) (Fig. 3) (34–41). Additional structures of multimeric nucleoproteins, some in complex with RNA, have been solved for several additional segmented NSVs, including BUNV, Rift Valley fever virus (RVFV), Toscana virus (TOSV), Leanyer virus (LEAV), tomato spotted wilt virus (TSWV), and La Crosse virus (LACV). These new structures also point toward a head-to-tail oligomerization pattern in native RNPs (30, 42–47). However, nucleoprotein molecules of Hantaan orthohantavirus (HTNV), a segmented NSV, do not fall within either discrete oligomerization pattern. Instead, HTNV nucleoproteins undergo a unique conformational fold upon oligomerization that results in an interaction between the N-terminal and C-terminal arms, creating a closed crescent-like nucleoprotein monomer with little lateral separation between the “head” and “tail” (Fig. 2) (48).

Fig 2.

Oligomerization of nucleoproteins into RNPs. Non-segmented (top) and segmented (bottom) NSV RNP structures modeled in surface representation. Alternating nucleoproteins in each structure are colored according to the domains indicated in the corresponding sequence maps below each individual structure. The non-segmented virus RNPs (top) are each shown as one turn of the rod-like helix. From left to right, the non-segmented virus RNPs illustrated are from: vesicular stomatitis virus (VSV), rabies virus (RABV), Zaire ebolavirus (EBOV), measles virus (MeV), cetacean morbillivirus (CeMV), and respiratory syncytial virus (RSV). The segmented virus RNPs (bottom) are shown as portions of the rod-like helices. For the influenza A virus (IAV) RNP, only one twist of the double helix is shown. For both Bunyamwera virus (BUNV) and Hantaan orthohantavirus (HTNV), roughly two twists of the single helices are shown.

Fig 3.

IAV nucleoprotein oligomerization. IAV ribonucleoprotein complex (PDB: 4BBL) with (A) 12 nucleoprotein monomers shown from two opposing strands of the double helix and (B) nucleoprotein head-to-tail oligomerization via the tail loop between adjacent same-strand nucleoprotein subunits and (C) opposing-strand nucleoprotein interaction interface. Domains are colored according to Fig. 2.

Electron micrographs reported for NSV RNPs show that NSV RNP structures fall on a flexibility spectrum (Fig. 4); for an overview see Table S1 (31, 41, 49–60). The side-by-side oligomerization pattern of the nucleoprotein molecules in the RNPs of non-segmented NSVs represents the most rigid and tight structure, whereas the head-to-tail oligomerization pattern of the nucleoprotein molecules of segmented NSV RNPs represents a more flexible structure (Fig. 3 and 4). The side-by-side interactions have a larger interaction area, which limits the dynamics of the nucleoprotein chain. Limited nucleoprotein dynamics would likely deform the RNP, such that it would be energetically unfavorable for the RNA polymerase to bind to both ends of the vRNA. The head-to-tail interaction pattern is inherently more flexible, allowing for the formation of loop-like or double helical RNPs, in which both the 5′ and 3′ ends are held close together by the RNA polymerase.

Fig 4.

Flexibility of NSV RNPs. Cartoon illustrating RNP structural variability observed by negative stain electron microscopy (Table S1). The flexibility continuum is indicated at the top of the cartoon. The RNA polymerase is depicted in dark gray. The nucleoprotein monomers forming the 5’ consecutive half of the RNP are depicted in light green and the nucleoprotein monomers forming the 3’ consecutive half are depicted in light blue. The leftmost structure represents the typical bunyavirales RNP, the middle structure the typical articulavirales RNP, and the rightmost structure a typical mononegavirales RNP. In the latter structure, no RNA polymerase is depicted. Virus abbreviations are as follows: Bunyamwera virus (BUNV), tomato spotted wilt virus (TSWV), Rift Valley fever virus (RVFV), Crimean-Congo hemorrhagic fever virus (CCHRV), citrus Psorosis virus (CPsV), influenza A virus (IAV), freesia sneak virus (FSnV), human respiratory syncytial virus (RSV), vesicular stomatitis virus (VSV), Ebola virus (EBOV), measles virus (MeV), rabies virus (RABV), Newcastle disease virus (NDV), and Sendai virus (SeV).

Once nucleoproteins are in a RNP-like oligomeric state, it is likely sterically difficult for these nucleoproteins to subsequently bind vRNA, especially for the NSV nucleoproteins which sequester the vRNA on the interior of the RNP helix. Given the current data, it is likely that the oligomerization of nucleoproteins and the binding of nucleoproteins to vRNA occurs in tandem, rather than sequentially, and that both processes are highly regulated during the viral infection cycle. Indeed, in vitro experiments with IAV nucleoprotein have shown that oligomerization states can be controlled by adjusting the salt concentration (10, 61, 62): at 50 mM NaCl, IAV nucleoproteins remain monomeric, whereas at 300 mM NaCl IAV nucleoproteins forms trimers and tetramers. Interestingly, the trimers/tetramers bind to RNA with threefold higher affinity than monomeric nucleoprotein, but nucleoproteins remain in trimer/tetramer form after binding. When multimeric nucleoprotein is added to RNA in 300 mM NaCl and the salt concentration is decreased to 50 mM NaCl, RNP-like structures form spontaneously, comparable to when monomeric nucleoprotein is added to RNA during viral replication (62). These results suggest that IAV nucleoprotein oligomerization is in equilibrium between multimeric and monomeric states and that only monomeric nucleoprotein can form RNP structures.

Lassa virus (LASV) nucleoproteins form trimers in mammalian cells but these multimers do not bind to RNA (63). It has been hypothesized that the observed trimeric form of LASV nucleoproteins is a regulatory mechanism to keep nucleoprotein RNA-free until it is needed for nascent RNA binding (64). However, by contrast to IAV, regulation is more complex than reducing nucleoprotein to a monomeric state as mutant LASV nucleoprotein that cannot form trimers does not support viral replication and transcription (63). Molecular dynamics studies have shown that upon disruption of the LASV nucleoprotein trimer into nucleoprotein monomers, the individual nucleoprotein subunits undergo conformational changes that are conducive to RNA binding as well as RNP formation (65). Together, these findings suggest that the oligomerization patterns of LASV nucleoprotein have multiple functions, preventing premature RNA binding and regulating the viral infection cycle.

To maintain nucleoproteins in conformations conducive to RNA binding and RNP formation, non-segmented NSVs [RABV, VSV, MeV, EBOV, RSV, and Nipah virus (NiV)] utilize chaperone proteins (P or VP35), which phosphorylate oligomerization grooves and/or competitively block RNA binding or nucleoprotein oligomerization through interactions between adjacent monomers (9, 11, 66, 67). These chaperone proteins are also responsible for regulating the replicase and transcriptase complexes and are strongly associated with full length vRNPs and weakly associated with monomeric nucleoprotein via their disordered arms (1, 20).

EBOV VP35 chaperones nucleoprotein in the RNA-free state. The N-terminal 80 residues of VP35 bind to nucleoprotein and prevent oligomerization as well as premature RNA binding. VP35 competes with the N-terminal arm of a potential neighboring nucleoprotein monomer (9). Since oligomerization induces the necessary conformational change in a nucleoprotein to make it capable of RNA binding, VP35 is thought to modulate RNA binding indirectly by preventing oligomerization. A crystal structure of a nucleoprotein-VP35 fusion protein from which the VP35 disordered regions were removed shows that VP35 straddles an EBOV nucleoprotein-specific beta-hairpin in the C-terminal domain of the nucleoprotein. Removal of the VP35 peptide using TEV cleavage and anion exchange chromatography revealed that monomeric nucleoproteins form oligomers in the absence of VP35. Moreover, addition of VP35 can reverse nucleoprotein oligomerization in the absence of RNA, but not when RNA is present. In other words, once the EBOV RNP is formed, it is stable. These observations are in line with the hypothesis that RNA is not necessary for oligomerization of viral nucleoprotein but important for stabilizing the vRNP to prevent spontaneous nucleoprotein dissociation. VP35 must dissociate from the monomeric nucleoprotein in order for it to oligomerize and bind vRNA (9). The driving force for this handoff is currently unknown, but it may be nucleoprotein-concentration dependent. The interaction between the EBOV nucleoprotein and VP35 is structurally distinct from distantly related NSVs.

The P protein, analogous to EBOV VP35, is known to act as chaperone for VSV, MeV, and NiV nucleoproteins and to bind to the RNA polymerase as an oligomer. These P proteins are primarily alpha-helical and bind nucleoproteins via disordered arms, whereas the EBOV VP35 is primarily unstructured (9, 11, 68–70). This may be attributed to the unique beta-hairpin structure in EBOV nucleoprotein with which VP35 interacts. Despite the significant structural distinction, the general mechanism of regulation is similar. Solved structures of VSV nucleoprotein-P complexes show that P blocks both RNA-binding and NP oligomerization, whereas in NiV nucleoprotein-P complexes, P blocks nucleoprotein oligomerization (68, 69).

NSVs with segmented genomes do not rely on a chaperone protein for the regulation of nucleoprotein oligomerization. Instead, they use phosphorylation and/or conformational changes to control nucleoprotein oligomerization (3). As previously mentioned, IAV nucleoprotein is present in both monomeric and multimeric states, dependent on the salt concentration and must be monomeric before RNA binding can occur. Mechanistically, the phosphorylation and dephosphorylation of S165 of the IAV nucleoprotein is crucial for the regulation of oligomerization. S165 is located in the tail-loop insertion groove and the phospho-mimic mutant S165E remains mostly monomeric. This suggests that phosphorylation of S165 prevents oligomerization by blocking the oligomerization insertion groove, and that dephosphorylation needs to occur before nucleoprotein molecules can oligomerize (61). Recruitment of nucleoprotein may involve the intrinsically disordered tail of host protein ANP32, analogous to the non-segmented NSV P protein (22, 70, 71). ANP32 is bound by a replicase complex consisting of two RNA polymerases, which are held together by ANP32. How a phosphatase is recruited to this complex is presently unknown.

A structural exception to the above descriptions of NSV RNPs is the Newcastle Disease virus (NDV) RNP. This paramyxovirus forms self-capped “double-RNPs” that contain two single-stranded spirals packed with their 5′ ends held in close proximity (31). Minigenome analysis shows that elimination of the nucleoprotein residues that create the 5′ seam between two helical spirals results in non-functional, single-filament RNPs. This self-capped organization means that multiple genome copies can be present in a single RNP and therefore a virion particle with only one RNP can contain multiple genome copies. Paramyxoviruses can contain multiple genome copies within a single virion particle and some segmented viruses, like RVFV, have also been found to contain duplicated genome segments within a single virion particle (72–74). However, NDV is the only known case where the duplicate genome can be encapsidated within the same RNP structure (31). While this is the only known case of self-capping NSV RNPs, this molecular mechanism has been observed in other filamentous systems, such as microtubule assembly (75).

RNA BINDING PROPERTIES OF NUCLEOPROTEINS

The molecular manner by which nucleoproteins interact with vRNA varies greatly. Typically, the nucleoproteins have a positively charged RNA-binding groove and undergo a conformational change that effectively clasps the vRNA after or during binding. The basic structure of an NSV nucleoprotein is made up of a C-terminal domain (CTD) and an N-terminal domain (NTD) (Fig. 2). In VSV and RABV nucleoprotein-RNA ring-like structures, 9 nucleotides (nt) are tightly clasped between the CTD and NTD of individual nucleoprotein subunits (76, 77). The RSV nucleoprotein-RNA ring-like complex showed a similar interaction, with 7 nt of RNA clasped between the CTD and NTD of each individual NP subunit (38). The same mechanism of vRNA binding between the CTD and NTD of individual nucleoprotein subunits was observed in cryo-EM analyses of helical MeV RNP structures and helical EBOV RNP structures, suggesting that the previously observed nucleoprotein-RNA binding interactions in a ring-like formation are similar to the interactions present in the virion (34, 37).

The structure of IAV nucleoprotein-RNA deviates from the relatively conserved RNA-binding interactions observed in mononegavirales. IAV nucleoproteins hold the RNA between the CTD and NTD, but biochemical analysis suggests IAV encapsulates 24 nt per nucleoprotein subunit and that this interaction does not involve the extensive base-stacking observed in RABV and VSV nucleoproteins (78, 79). In the IAV nucleoprotein-RNA complex, the RNA is instead held more flexibly in the basic groove of nucleoprotein subunits through polar interactions and hydrogen bonding and the bases are accessible to nucleases. The flexibility of the RNA winding around each nucleoprotein is hypothesized to be the reason why crystal structures of complete IAV nucleoprotein-RNA interactions have not yet been resolved (80). The only structural evidence currently available for IAV nucleoprotein-RNA comes from a crystal structure of a monomeric H5N1 IAV nucleoprotein in complex with a 9-nt long synthetic RNA molecule, of which only 3-nt could be modeled in the observed density (79).

The only other segmented NSV for which the interactions between nucleoprotein and RNA have been revealed at a structural level is BUNV (47). A crystal structure for a tetrameric ring-like BUNV nucleoprotein-RNA complex shows that the bound RNA is buried deep in a narrow cleft on the internal surface of the ring. The number of bases that is covered by a single nucleoprotein is 11 nt. Unlike IAV nucleoprotein-RNA interactions, the RNA bases are completely isolated from solvent exposure and held in place more rigidly. While this may not necessarily illustrate native NP-RNA interactions within a BUNV RNP, electron micrographs of BUNV RNPs show that the width of the native RNP is comparable to that of the tetrameric BUNV nucleoprotein-RNA ring structure. This is noteworthy because the width of nucleoprotein tetramers of the closely related RVFV do not match the width of RVFV RNPs seen in electron micrographs, suggesting that NP multimers can adopt several different conformations distinct from RNA-bound NP multimers (42, 44).

The shortest estimated RNA-binding footprint is 6 nt for the nucleoproteins in both MeV and EBOV RNPs (34, 37). This is followed by a 7 nt RNA-binding footprint in RSV nucleoprotein, 9 nt in VSV and RABV nucleoproteins, and 11 nt in BUNV, LACV, and LEAV nucleoproteins (30, 38, 45, 47, 76). The longest RNA-binding footprint observed in NSV nucleoproteins is 24 nt observed for IAV (78, 79). Although the number of known nucleoprotein footprints is small, the current data indicate that the RNA-footprint is not correlated with nucleoprotein size or the number of amino acids that make up a nucleoprotein molecule.

The location of the vRNA in the RNP also varies among NSVs. In both RABV and VSV RNP structures, the vRNA is encapsidated in the center of a left-handed conical helix (81). Paramyxovirus RNPs also adopt left-handed helices, but in contrast to the Rhabdoviridae (RABV and VSV) RNP structures, the vRNA is encapsidated on the outside of the helix, as seen in RSV (27). The nucleoprotein-RNA binding mechanism influences the amount of protection that the RNP offers. Most paramyxoviruses form RNPs that can protect RNA from RNase digestion despite the vRNA being more exposed. However, EBOV RNA-nucleoprotein complexes are sensitive to RNase digestion (82). IAV RNPs have fully exposed RNA bases which are also susceptible to RNase cleavage (83). The location of vRNA is important because vRNA that resides on the outside of the RNP can form RNA secondary structures. Recent SHAPE-MAP experiments have shown that IAV vRNA has less RNA secondary structure in a virus particle (in complex with nucleoproteins) than it does as naked vRNA (13, 14). The remaining RNA secondary structures seen in virions involve short-range interactions (spanning less than 100 nt) and segment-segment interactions. These observations support the hypothesis that nucleoproteins can minimize RNA secondary structure and facilitate the selective formation of those structures that are critical, like in the packaging of segmented NSVs.

Several experimental approaches have been used to determine the strength of the nucleoprotein-RNA interactions in influenza viruses, including filter binding assays, surface plasmon resonance, and polarization of fluorescence. These analyses show that the RNA binding affinity of nucleoproteins at room temperature ranges from 4 nM to 70 nM, depending on the salt concentration and type/length of RNA used (84). Divergent affinity measurements were observed at colder temperatures (380 nM at 4°C) and with monomeric mutants (1–10 uM). In comparison to other known RNA binding proteins (RBPs), which have Kd values ranging from <10 nM to uM, viral nucleoprotein bind RNA with moderate strength (85). The RNA binding affinity of other NSV NPs has not been determined.

DYNAMICS OF RIBONUCLEOPROTEIN COMPLEXES

As depicted in Fig. 4, NSV RNPs fall on a flexibility spectrum, with nonsegmented NSVs adopting the most rigid RNP structures and bunyaviruses the most flexible RNP structures. However, during infection, RNPs likely adopt additional conformations, each conducive to a given stage of NSV replication or transcription. During NSV replication, these RNP structures must allow the RNA polymerase to access the vRNA. A still unanswered question is how the RNA polymerase gains access to the vRNA template that is otherwise protected by oligomerized viral NPs. NPs likely need to dissociate from the RNA to allow the RNA polymerase to copy the vRNA, and then rebind the vRNA as the RNA polymerase translocates to avoid exposure of naked vRNA to nucleases or other host proteins. Presently, the exact mechanisms of nucleoprotein dissociation and reassociation during replication and transcription remain unknown for all NSVs.

In paramyxoviruses, it has been shown that transcription is efficient only when the vRNA genome consists of a multiple of 6 nucleotides (i.e., 6 n nucleotides and known as the “rule of six”) (86). This correlates with the number of nucleotides that each nucleoprotein subunit binds to. More specifically, it has been shown that in the Sendai virus (SeV) nucleoprotein the exact position of each nucleotide is critical for recognition by the RNA polymerase (87). Essentially, the precise way that the nucleoprotein holds each nucleotide in the RNA binding groove is distinct and important for transcription. Furthermore, many paramyxoviruses are known to have bipartite promoters, which are non-contiguous regions in the vRNA that can be recognized by the RNA polymerase only when they are in complex with nucleoprotein and in close physical proximity (88). This highlights the importance of the nucleoprotein structure, the helical arrangement of the RNP, and the resulting interactions of individual nucleotides of the vRNA with the RNP (89).

Structural analysis of IAV RNPs suggests that the RNP retains a double helical RNP conformation throughout transcription, and that the RNA polymerase remains bound to both the 3′ and 5′ ends of the vRNA (90). Using nucleozin, a known influenza virus inhibitor that causes nucleoprotein aggregation that can stall IAV transcription, further insight into RNP dynamics was obtained (90). Cryo-EM analysis of the nucleozin-stalled RNPs showed that nucleozin binds to two adjacent nucleoproteins from opposite strands within the RNP, acting like a staple (90). This inhibition mechanism is cooperative: once one nucleozin molecule binds, it brings the adjacent strand nucleoproteins closer together, allowing for the crosslinking to spread throughout the entire helix and resulting in a more rigid RNP structure. Importantly, the stapled nucleoprotein strands impair viral RNA synthesis, which implies that the opposing nucleoprotein strands need to “slide” past each other to allow the RNA polymerase to copy the template (Fig. 5A) (90). In this ‘processive helical track’ model, IAV transcription occurs when the RNA polymerase “pulls” the vRNA template locally from the nucleoprotein chain, detaching the RNA from the nucleoprotein upstream of the RNA polymerase entry channel, but likely leaving the nucleoprotein attached to rest of the nucleoprotein chain. The RNA polymerase subsequently pulls the RNA into the active site of the RNA polymerase for template-dependent RNA synthesis. The template vRNA next reassociates with the RNA-free nucleoprotein in the nucleoprotein chain upon exit from the RNA polymerase active site (91, 92).

Fig 5.

RNP dynamics during NSV transcription. (A) Articulavirus RNPs form a double-helical complex in which both the 3′ and 5′ ends of the vRNA template (red line) are bound to the RNA polymerase (green). In the model shown, IAV transcription (mRNA product shown as orange line) occurs when the RNA polymerase “pulls” the vRNA template locally from the nucleoprotein (blue) chain and into the active site of the RNA polymerase (denoted as a light green oval). The template vRNA next reassociates with the nucleoprotein chain upon exit from the RNA polymerase active site. Black arrows denote the movement of the template vRNA and associated nucleoproteins. The 3′ and 5′ terminal nucleoprotein subunits are denoted in dark blue with dashed edges to illustrate the proposed handling by the RNA polymerase. (B) The bunyavirus RNP is a spiraling, circular RNP in which the 3′ and 5′ ends of the vRNA are attached to the RNA polymerase (green). Transcription and replication involve local dissociation of the template vRNA (red line) from the nucleoprotein (blue) chain as the vRNA enters the active site of the RNA polymerase (denoted as a light green oval) and reassociation of the vRNA to the nucleoprotein chain upon exit of the RNA polymerase. Black arrows denote the movement of the template vRNA and associated nucleoproteins. (C) The mononegavirus RNP forms a tight helical rod-like structure with no permanent attachment point for the RNA polymerase (green). Phosphoprotein (yellow, shown here as a dimer) binds to the RNA polymerase and forms the transcriptase complex which associates (dashed arrow) with the initiation sites on the template vRNA (red line). Phosphoprotein interacts with the nucleoproteins (blue) eliciting conformational change which induces the release of template vRNA from the nucleoprotein chain as it enters the RNA polymerase active site (denoted as a light green oval). The nucleoprotein re-associates with the template vRNA upon exit of the RNA polymerase. Exact handling of nucleoprotein during the transient RNA-free state is currently unknown, but nucleoprotein is shown here to remain in chain-like form. Black arrows denote the movement of the template vRNA and associated nucleoproteins. It is currently unknown whether the RNA polymerase dissociates from or remains attached to the template vRNA after producing the mRNA (orange line).

Similar to articulaviruses, bunyavirus RNA polymerases hold onto the 5′ and 3′ ends of the vRNA while transcribing and replicating the vRNA segments (93). However, unlike articulaviruses, the RNPs of bunyaviruses form a more flexible, spiraling loop, instead of a double helix (41). It is likely that the bunyavirus vRNA also dissociates from the nucleoprotein molecules in the RNP before entering the active site of the RNA polymerase (Fig. 5B). The hypothesized dynamics are therefore less complex than the double helical track, as “sliding” is likely unnecessary due to the open loop structure of the RNP. The illustrated mechanism in Fig. 5B follows the same nucleoprotein handling mechanism as in articulaviruses (Fig. 5A).

Structural data for articulavirus and bunyavirus transcription elongation and termination is limited, and the only data available for IAV transcription, elongation, and termination has been obtained in the absence of nucleoprotein (5, 92–94). Following polyadenylation near the terminal 5′ ends of the vRNA segments, termination likely involves a “reset” of the RNP complex. In the current model, the 5′ end remains bound, while the polyadenylation sequence is released from the active site of the RNA polymerase (93). Nucleoproteins, still in a chain-like oligomerization state, are proposed to reassociate with the released template vRNA upon exit from the RNA polymerase.

A nucleoprotein-vRNA dissociation mechanism similar to the segmented NSVs has been proposed to take place during replication and transcription of the genome of RABV and other members of mononegavirales. Oligomerized nucleoprotein molecules are thought to remain connected, like links in a chain, as the RNA polymerase (L protein) copies the vRNA. The template then reattaches to the nucleoprotein near the exit channel of the RNA polymerase. Unlike IAV and other orthomyxoviruses, this process requires the viral P protein. A P-dimer, -trimer, or -tetramer attaches to the L protein, creating an L-P transcriptase complex that can bind to the individual nucleoprotein subunits within the RNP and trigger a conformational change in the nucleoprotein subunits that releases the vRNA (Fig. 5C) (95). It is unknown if the RNA polymerase dissociates from the vRNA upon reaching the end of the transcript or remains attached.

The above models of NSV RNP dynamics during RNA synthesis assume that the vRNA can be released from the nucleoprotein chain and “snap-back” onto the nucleoprotein chain after being copied by the RNA polymerase. In EBOV, monomeric nucleoproteins cannot bind RNA and RNA-binding is dependent on nucleoprotein oligomerization (96). It has been hypothesized that nucleoprotein oligomerization induces conformational changes that facilitate RNA binding. This idea suggests that the “snap-back” model may also be used during EBOV RNA synthesis.

The above model is not fully consistent with experimental results showing that IAV nucleoproteins form RNP-like oligomers without vRNA present and that these IAV oligomers, once formed, cannot turn into functional RNPs by binding vRNA. IAV nucleoprotein trimers/tetramers can readily bind RNA but they cannot form native oligomeric RNP structures, likely due to the closed loop interactions (61). It is possible that a transient conformational change exists within a vRNA-free nucleoprotein in the context of an RNP, while the vRNA is being copied by the RNA polymerase. In HTNV, RNA binding to HTNV nucleoproteins is followed by a conformational change to completely surround the template, creating a protective tunnel (48). Other NSVs simply form a “jaw” between the head and body domains that effectively “clamps” down the RNA post binding (8). The force on the vRNA template, as it is pulled into the RNA polymerase, may be sufficient to “open” these clamp-like structures, but calculations have not been done to support this idea.

ROLE OF NUCLEOPROTEINS IN REGULATION OF TRANSCRIPTION AND REPLICATION

In addition to nucleoprotein’s role in vRNA protection and acting as a scaffold, nucleoprotein can function as a mediator of gene regulation in several viruses. A recent study of IAV replication found that a single mutation in nucleoprotein can selectively modulate neuraminidase (NA) gene segment expression in the A/Puerto Rico/8/1934 (PR8) strain. One amino acid substitution, F346S, decreased the ability of nucleoproteins to promote NA segment vRNA synthesis while not altering the RNA binding or oligomerization function of nucleoprotein (97). The strain-specific effect of F346S was determined to derive from a motif within the UTR of the PR8 NA segment. It was subsequently found that the neighboring aromatic residues in the IAV nucleoprotein, Y385 and F479, also affect NA gene expression. Together, the pi-stacking aromatic residues F346, Y385, and F479 stabilize the nucleoprotein tail-loop. When disrupted through the F346S mutation, the tail-loop may adopt a less stable, alternative conformation that affects how segments are replicated. Presently, the underlying mechanism is not understood.

In the nucleoprotein of human parainfluenza virus 2 (hPIV2), amino acid residue Q202 is 1 of 10 nucleoprotein residues that interact with RNA and the only residue that interacts directly with the base of a nucleotide (98). Experiments have shown that mutation Q202A results in increased levels of RNA polymerase activity without affecting nucleoprotein interactions with phosphoproteins, adjacent nucleoproteins or RNA binding (98). Furthermore, hPIV2 RNA templates with disrupted bipartite promoters were found to be rescued by the Q202A nucleoprotein mutation (98). Together, these fascinating observations suggest that Q202 is necessary for the selective initiation of bipartite promoters and for enforcing the rule of six in the hPIV2 genome and underlining the interplay among RNP structure, nucleoprotein-RNA interactions, and viral RNA synthesis.

Post translational modifications (PTMs) regulate cellular activities as well as NSV replication. Host ubiquitin ligase CNOT4 has been found to ubiquitinate IAV nucleoprotein K184. Ubiquitination of K184 leads to an increase in RNA polymerase activity, whereas lack of or removal of by host deubiquitinase USP11 reduces RNA polymerase efficiency (99). K184 resides in the nucleoprotein RNA-binding groove, suggesting that ubiquitination enhances the RNA binding efficiency of nucleoproteins or the processivity of the viral RNA polymerase. In many other molecular contexts, protein ubiquitination is known to trigger protein degradation, however the amount of nucleoprotein does not decrease upon CNOT4 expression, which suggests that this ubiquitination does not alter the stability of the nucleoprotein (99).

While non-segmented NSVs use the viral P protein to regulate nucleoprotein oligomerization and facilitate the formation of an active replicase complex, segmented NSVs do not encode a protein analogous to P. Influenza viruses rely on host protein acidic nuclear phosphoprotein 32 (ANP32) for some of the roles that P performs in the replicase complex of other NSVs (22). The C-terminal region (LCAR) of both human and chicken ANP32 mediates a direct interaction with the IAV nucleoprotein. It is hypothesized that this flexible tail-like region of ANP32 recruits monomeric nucleoprotein to newly formed RNA, thereby facilitating encapsidation. ANP32-LCAR is acidic and interacts with the basic RNA-binding groove of the IAV nucleoprotein. This prevents any nucleoprotein already bound to RNA from getting recruited to the replicase complex. Furthermore, mutant ANP32 proteins that did not contain an LCAR domain did not support replication of full-length IAV genome segments, suggesting that LCAR is critical for the replication of IAV.

Lymphocytic choriomeningitis virus has a bi-segmented ambisense genome. Each segment has two genes encoded in opposite orientations and separated by structured intergenic regions (IGRs) (100). Early in infection, the RNA structure of the IGR functions as a terminator for the RNA polymerase, leading to the transcription of only the gene upstream of the IGR in the vRNA. As infection progresses, replication of the vRNA results in an encapsidated cRNA. It has been proposed that increasing levels of intracellular nucleoprotein concentrations may help unfold the secondary structure within the IGRs, allowing the RNA polymerase to replicate past the IGR (100). However, in this model, it is unclear if these nucleoproteins would only help melt the RNA structure or also insert into the existing nucleoprotein chain. Alternatively, in replication mode, triggered by the production of new nucleoprotein and RNA polymerase subunits, the RNA polymerase may be more processive and pass through the structure, creating a cRNP. Indeed, once the cRNP has been formed, transcription of the cRNA template occurs upstream of the IGR but not downstream of the IGR, whereas replication of the cRNP leads to the production of new vRNPs (100). An increase in nascent nucleoprotein levels may thus be a key signal to initiate the switch from transcription to replication. Interestingly, this simple regulatory switch from transcription to replication also occurs in many paramyxoviruses, like VSV and SeV (101). In IAV, the switch is dependent on nascent RNA polymerase levels, RNA polymerase dimer formation, and stabilization of newly formed RNA by nucleoprotein (71, 102, 103).

Recent work has shown that nucleoprotein concentration is vital for viral replication across many NSVs (including IAV, RSV, VSV, EBOV, LASV, and MeV) (17, 18). Limiting the availability of the viral nucleoprotein by gene knock-down induces a strong host immune response despite the reduced amount of vRNA present in infected cells. The lack of nucleoprotein reduces the processivity within IAV replication and as a result more aberrant defective mRNA particles are formed, which act as a trigger for the host interferon (IFN) response. This novel discovery is foundational for new antiviral drug design for potentially all NSVs. By targeting nucleoproteins, not only will viral replication be inhibited but the heightened immune response will prime neighboring uninfected cells, creating an antiviral that could be more effective than most existing therapeutics that target surface glycoproteins and the RNA polymerase.

INTERACTIONS WITH OTHER VIRAL PROTEINS

The viral matrix protein (M) is a crucial protein for enveloped NSVs. The M protein interacts with viral RNPs, envelopes glycoproteins, and facilitates virion assembly during the late stages of infection. In many enveloped NSVs, nucleoproteins directly interact with M to mediate assembly and release of virion particles. In human parainfluenza virus 3 (hPIV3), nucleoproteins interact directly with M. Data suggest that hPIV3 nucleoprotein regulates virion assembly via a single interaction with M protein residue L305 (104). Similarly, in EBOV infection, the matrix protein, VP40, has been found to directly interact with the C-terminal 50 amino acids of the EBOV nucleoprotein (105). In contrast, the RSV M protein does not directly interact with vRNPs, but is still connected by an interaction with the viral protein M2-1 (106).

Apart from involvement in the release of new virion particles, the M protein of Borna disease virus (BoDV) has also been found to colocalize with RNPs. Minireplicon assays have shown that this interaction surprisingly does not inhibit RNA polymerase activity (107). In most other enveloped NSVs, the binding of the RNPs to the M proteins inhibits RNA polymerase activity, likely to ensure that virus particles can efficiently form and egress from the host cell. Given that the interaction between M and RNPs within BoDV does not inhibit RNA polymerase activity, it is likely that the interaction is also important for infection steps prior to virion assembly and egress.

The viral life cycle in IAV is distinct from other NSV life cycles in that replication occurs in the host cell nucleus. This is advantageous because it allows the virus to utilize the nuclear environment as protection and a resource for efficient replication and transcription. On the other hand, the import of vRNPs to the nucleus adds a layer of complexity and the need for additional virus-host interactions (108). The segmented nature of the IAV genome likely facilitates the nuclear import/export processes, as smaller RNPs may transition through the nuclear pore complex more efficiently. Import of vRNPs is initiated by the recognition of the nuclear localization sequences of the IAV nucleoproteins by host protein importin-ɑ (109). Once bound to the vRNP, importin-ɑ is recognized by host protein importin-β, which traffics the vRNP through the nuclear pore complex. This complex process can occur rapidly (within about an hour of infection) suggesting that this process has evolved to be highly efficient and effective (108).

After the IAV genome is delivered to the nucleus and replicated, new vRNPs must exit the nucleus and be transported to the cell membrane for virion budding (110). IAV genome segment eight encodes two non-structural proteins (NS1 and NS2/NEP), which perform a variety of functions during infection. NS1 has been shown to inhibit both interferon production and the nuclear export of host mRNA, likely to ensure that viral mRNA does not have to compete with host mRNA for translation by host ribosomes (111). NS2/NEP is a small 121 amino acid protein that is produced from a spliced genome segment eight mRNA. Although NS2/NEP does not appear to have a significant structural role within the virion, it has been observed to regulate the export of vRNPs from the host nucleus and it is therefore also referred to as IAV nuclear export protein (NEP) (110, 112). Evidence suggests that IAV uses NEP to aid in hijacking the host protein chromosome region maintenance 1 (Crm1) for nuclear export of vRNPs. In this model, NEP acts as a linker between Crm1 and IAV matrix 1 protein (M1)-bound viral RNPs. This complex is then shuttled out of the nucleus via the host nuclear export pathway. Other data show that IAV nucleoprotein can interact directly with Crm1 through a Crm1-specific nuclear export signal, independent of both M1 and NEP (113). Interestingly, M1 also contains a nuclear export signal for Crm1, which suggests that it is possible that M1 plays a regulatory role in nuclear export of vRNPs independent of NEP (114).

The above discussion describes several known interactions between NSV nucleoproteins with other viral proteins of the same species. A more complete list of known interactions can be found in Table S2. The core interactions are visually summarized in Fig. 6 (68, 105–107, 110, 115–124). While the mononegavirales encode a viral P protein to chaperone nucleoproteins, it is notable that segmented NSVs (both in articulavirales and bunyavirales) do not encode such a protein. As previously mentioned, it has been observed that human host protein ANP32 acts as a P protein analog in influenza virus infections. No host analog has been identified for bunyavirales, which suggests that the regulation of nucleoprotein multimeric states in bunyaviruses is performed via an alternative mechanism or is not needed at all due to currently unidentified structural differences within the nucleoproteins. Furthermore, there have been no direct interactions observed between mononegavirus nucleoproteins and the associated RNA polymerase (L protein) to date. Interactions are instead coordinated through the encoded P protein, maintaining indirect attachment. It is possible that the lack of an encoded P protein in the segmented orders of NSVs (articulavirales and bunyavirales) is compensated for by the direct interactions of the nucleoproteins with the RNA polymerase.

Fig 6.

Key nucleoprotein interactions with other viral proteins by phylogenetic order. (Left) Mononegavirales NP interactions, (middle) articulavirales, (right) bunyavirales.

ROLE OF NUCLEOPROTEINS IN THE DISRUPTION OF HOST CELL IMMUNE PATHWAYS

The host innate immune system uses multiple mechanisms to detect and inhibit a viral infection. One method is the recognition of pathogen-associated molecular patterns (PAMPS) by human host pattern recognition receptors (PRRs). Recognition of such PAMPS activates antiviral signaling cascades, including type I and III IFN expression. Retinoic acid-inducible gene I (RIG-I) is a host pathogen receptor that binds di- or tri-phosphorylated double-stranded RNA (dsRNA) molecules in the host cell cytoplasm (125). The (partially) complementary 5′ and 3′ terminal ends of segmented NSV vRNAs are substrates for RIG-I and activation of RIG-I induces IFN production which limits the spread of NSV particles (126). NSV nucleoproteins are critical for protecting viral RNA from the detection by RIG-I, and reduced nucleoprotein expression has been shown to increase RIG-I activation (17, 18).

Arenavirus nucleoproteins suppress the IFN response by degrading PAMP RNAs through a C-terminal DEDDh-like exonuclease domain (64, 127). While seemingly counterintuitive, this exonuclease activity is pro-viral and required for the anti-IFN activity observed in arenavirus infections. To prevent virus detection by the host innate immune system, dsRNA formed during a LASV infectious cycle is digested by the LASV nucleoprotein. Only one other subfamily of mammalian RNA viruses is known to encode a protein with exonuclease activity: coronaviruses. For instance, severe acute respiratory syndrome coronavirus nsp14 is also a member of the DEDDh family and targets both ssRNA and dsRNA during proofreading, but nsp14 is not known to be directly involved in IFN suppression (128).

MeV infection is followed by severe immunosuppression, a process that involves many viral proteins. The MeV nucleoprotein plays a critical role in inhibiting antibody production by binding to host B cells, although the structural specifics of this interaction are still unclear (129, 130). Additionally, the MeV nucleoprotein binds to eukaryotic translation initiation factor 3 (eIF3) and inhibits host cell translation of host mRNA by upwards of 90% at 36 hours post infection (131, 132).

In many NSVs, nucleoprotein-P complexes can form cytoplasmic inclusion bodies (IBs), which serve as hubs for viral replication and shield the virus from detection by the host innate immune system. This tactic has been observed in many non-segmented NSVs including RABV, MeV, and EBOV (133–135). NiV is distinct from other non-segmented NSVs in that it forms two types of IBs, neither of which require the viral P protein (136). NiV can induce formation of perinuclear IBs with expression of NP alone as well as plasma membrane IBs when nucleoprotein is in complex with the viral M protein. This is the first observed instance of an NSV forming two IB populations for two separate functions. Interestingly, the RSV nucleoprotein can sequester immunostimulatory proteins MDA5 and MAVS into small cytoplasmic IBs (118). Contrary to the IBs discussed above, these are not viral replication hubs, but rather these IBs serve to hold the host immune system hostage while RSV takes over the cell. Alternative roles of NSV nucleoprotein-induced IBs are highlighted in the recent work in NiV and RSV, which suggests that NSV IBs may contribute to the viral life cycle in more ways than originally postulated.

There are likely many more viral-host interactions that are modulated by viral nucleoproteins, and the above discussion only represents the current understanding of the most well studied NSVs. A list of NSV nucleoprotein interactions with host proteins that was assessed for this review can be found in Table S3 (22, 25, 26, 30, 99, 113, 116, 118, 122, 130, 132, 137–143).

DISCUSSION AND OUTLOOK

NSV nucleoproteins are key structural and scaffolding components of NSV RNPs and critical to NSV infection. Recent advances in structural biology techniques and analysis procedures have made it possible to obtain new insight in NSV RNP structures and their dynamics during viral replication. The structural homology of the nucleoprotein core may offer an opportunity to connect the conserved and unique functions of nucleoproteins to the evolution of NSVs. In addition, NSV nucleoproteins are now considered possible antiviral targets and our understanding of how the function of nucleoproteins can be inhibited is better understood.

However, many key questions concerning nucleoprotein function and inhibition remain. In particular, it is unclear why NSVs that employ similar RNA synthesis mechanisms use nucleoproteins of a different size and with different RNA binding properties. It is also unclear why they form different RNP structures with different dynamics. At another level, it remains poorly understood how RNPs are assembled during viral RNA synthesis, whether differences exist between the RNP dynamics during NSV replication or transcription, and whether a dysregulation in RNP dynamics contributes to detection of viral RNA by the host innate immune response. Finally, the role of PTMs on nucleoproteins and RNP dynamics has only just begun to be explored and their exact function in the regulation of viral RNA synthesis, or how they may potentially perturb it, will require further attention in the future.

Overall, the studies discussed in this review highlight our current knowledge of the many mechanisms by which NSV nucleoproteins are active players in the NSV infection cycle. Further investigations into the role of nucleoproteins in lesser studied NSVs will undoubtedly uncover even more fascinating and diverse aspects. Studying the complex functions of viral nucleoproteins offers the benefits of finding novel therapeutic targets while also fortifying our understanding of viral evolution as it pertains to structural conservation.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/mmbr.00082-23.

Table S1: NSV Taxonomy and structure accession data. mmbr.00082-23-s0001.pdf.

Supplementary Table 1: NSV Taxonomy and structure accession data.

Table S2: NSV NP interactions with other viral proteins. mmbr.00082-23-s0002.pdf.

Supplementary Table 2: NSV NP interactions with other viral proteins.

Table S3: NSV NP interactions with host proteins. mmbr.00082-23-s0003.pdf.

Supplementary Table 3: NSV NP interactions with host proteins.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.