Mutational analysis of Rift Valley fever phlebovirus nucleocapsid protein indicates novel conserved, functional amino acids

Rift Valley fever phlebovirus (RVFV; Phenuiviridae, Phlebovirus) is an important mosquito-borne pathogen of both humans and ruminants. The RVFV genome is composed of tripartite, single stranded, negative or ambisense RNAs. The small (S) segment encodes both the nucleocapsid protein (N) and the non-structural protein (NSs). The N protein is responsible for the formation of the viral ribonucleoprotein (RNP) complexes, which are essential in the virus life cycle and for the transcription and replication of the viral genome. There is currently limited knowledge surrounding the roles of the RVFV nucleocapsid protein in viral infection other than its key functions: N protein multimerisation, encapsidation of the RNA genome and interactions with the RNA-dependent RNA polymerase, L. By bioinformatic comparison of the N sequences of fourteen phleboviruses, mutational analysis, minigenome assays and packaging assays, we have further characterised the RVFV N protein. Amino acids P11 and F149 in RVFV N play an essential role in the function of RNPs and are neither associated with N protein multimerisation nor known nucleocapsid protein functions and may have additional roles in the virus life cycle. Amino acid Y30 exhibited increased minigenome activity despite reduced RNA binding capacity. Additionally, we have determined that the N-terminal arm of N protein is not involved in N-L interactions. Elucidating the fundamental processes that involve the nucleocapsid protein will add to our understanding of this important viral protein and may influence future studies in the development of novel antiviral strategies.

Introduction RVFV is a member of the Phlebovirus genus of the Phenuiviridae family of viruses, in the Bunyavirales order [1]. The Phlebovirus genus can be broadly split into two distinct groups. One group pertains to the viruses transmitted by mosquitos or sandflies, commonly referred to as the Phlebotomine group, of which RVFV is a member. The second group pertains to viruses transmitted primarily by ticks [2]. The RVFV genome consists of two negative and one ambisense RNA segment. The large (L) segment encodes the RNA dependent RNA polymerase. The medium (M) segment encodes two non-structural proteins (p78 and NSm) and two glycoproteins, Gn and Gc [3]. The small (S) segment encodes the nucleocapsid protein (N) and the non-structural protein (NSs) using an ambisense coding strategy [4]. Each genome segment is flanked by untranslated regions (UTRs) that are involved in regulating transcription and replication of the genome segments. Due to terminal sequence complementarity, the 3' and 5' UTRs base pair and generate panhandle structures resulting in the recruitment of the viral polymerase (L), thus allowing the formation of viral ribonucleoprotein complexes (RNPs) [5,6] as is the case for all bunyaviruses [2,7,8]. RVFV N protein is a key protein within the RVFV proteome. It is characterised by a protruding N-terminal arm, an RNA binding cleft and a multimerisation groove. It has several essential functions that allow the replication and transcription of the viral genome segments. The N proteins of viruses belonging to the Bunyavirales order have been shown to function to encapsidate the viral genome which protects the genetic information from harsh conditions found in the intracellular environment, such as RNase degradative enzymes [9]. This encapsidation function of the N protein and the formation of the viral RNP complex allow the binding of the RNA-dependent RNA-polymerase thereby allowing transcription and replication to take place. Furthermore, N protein forms multimeric structures in infected cells [10][11][12]; in the case of RVFV the binding of the N-terminal arm to adjacent N monomers results in the formation of ring-shaped oligomers and allows the creation of filamentous RNPs required for replication of the viral genome [10].
N proteins of other families in the Bunyavirales order have also been investigated and have added to our understanding of N protein function. For the related genus Orthohantavirus, the N protein is thought to have RNA chaperoning activity where it dissociates the viral RNA duplexes, allowing sequestration of the 5' end of the RNA and the binding of the L protein [13,14]. Additionally, Sin Nombre virus (of the genus Hantavirus) N has been shown to bind mRNA caps, an essential process in cap-snatching [15]. The Bunyavirales order is very broad and viral N protein structures from different families are often unrelated, as evidenced by the Crimean-Congo haemorrhagic fever orthonairovirus (CCHFV) N protein. CCHFV is part of the Orthonairovirus genus within the Nairoviridae family and has a distinct N protein structure more closely related with the Arenaviridae family compared to other members of the Bunyavirales [16]. In the case of the Peribunyaviridae, the order prototype bunyavirus Bunyamwera orthobunyavirus (BUNV) nucleocapsid protein has also been shown to encapsidate the viral genome and carries out largely the same functions as demonstrated for phleboviruses. A previous mutagenesis study carried out using BUNV identified several residues that impact the replication efficiency of the virus [17].
Several studies examining RNA binding have identified residues within the core of the RVFV N protein that are essential for the formation of viral RNPs [10,18]. The bound RNA interacts with 18 conserved residues within the protein core and the hinge region between the N-terminal arm and the core. In particular, amino acid Y30 is located at the hinge region of the N-terminal arm and has been shown to stack with the 5' most RNA base when binding RNA [19]. Residues R64D, K67D and K74D were predicted to form the RNA binding cleft and a triple substitution mutant resulted in the loss of RNA binding [10]. The formation of RNPs requires N monomers to form hexameric ring structures through the binding of the N-terminal arm of one monomer to an adjacent monomers oligomerisation groove [10]. Residues Y3, L7, I9, P11, V16, I21, W23, V25, F28 and Y30 present on the N terminal arm of the RVFV N protein are predicted to interact with the oligomerisation groove. Disruption of the oligomerisation groove prevents the formation of higher order N protein structures and thus stops the formation of the viral RNPs [20]. N protein also has a dimeric closed conformational state which is predicted to occur in the absence of viral or host RNA [10]. In the presence of RNA, the conformation opens the N-terminal arm allowing the binding to adjacent subunits forming higher order structures [10]. Several RVFV N protein studies have identified the functions of named amino acid residues and these are summarized in Table 1 [10,[18][19][20]. These were found to be involved in either N multimerisation or RNA binding, and were used as a reference for selecting the mutants in this study that have been previously unexplored. This study aimed to widen RVFV N protein research by informing on a number of non-characterized, yet conserved amino acids found across multiple members of the Phlebovirus genus. As we believed these to be essential nucleocapsid residues with regards to involvement in important protein-protein interactions and other functions, we generated a panel of uncharacterised N protein mutants based on conservation data. These mutant proteins were investigated by utilising minigenome assays, packaging assays as well as biochemical techniques to assess their relevance in the RVFV life cycle. This analysis revealed that there are still fundamental underlying questions regarding the formation of phlebovirus RNP complexes, in particular, regarding residues P11 and F149 that confer loss of function that were not previously described. Our data add to the understanding of RVFV N and the role(s) of individual conserved amino acids in replication.
bacterial expression and purification. The virus-like particle assays were carried out using the pTM1-based plasmids above, with the addition of pTM1-M.

Sequence alignments
Sequence alignments were performed using CLC-Genomic Workbench using the following sequences: Punta Toro phlebovirus (Adames) (ABD92922.1); sandfly fever Naples phlebovirus Table 1. Summary of known RVFV N protein functions. Compiled predicted and known functions of RVFV N protein from studies focused on uncovering the RNA binding properties of RVFV N. Functional information was determined through varied methods, including analysis of RVFV N crystal structure and mutagenesis studies [10,18,19].

Protein expression and purification in Escherichia coli
The p14-N protein expression plasmids for each mutant N protein were transformed into Rosetta2 (DE3) cells (Novagen). Subsequent colonies were grown in LB Broth to OD0.6-0.8

DSP crosslinking assay
The purified mutant RVFV N proteins were concentrated to 0.2 μg/μl in 25 μl PBS/glycerol using Vivaspin centrifugal concentrators (Sartorius). Dithiobis(succinimidyl propionate) (DSP)-Lomants reagent (Thermo Scientific) was added to a final concentration of 1 mM and gently mixed, then incubated at 20˚C for 40 min. The reaction was stopped by directly adding protein loading buffer under non-reducing conditions and then analysed on a 4-12% Bis-Tris plus (Novex) gradient gel.

RNA binding assay
Recombinant parental virus (rMP12) and mutant N proteins purified by Ni-NTA were examined for RNA-binding activity. 10 μg of purified protein was mixed with 2x RNA gel loading buffer (Thermo Scientific) containing 95% formamide. The RNA was separated on the 2% TopVision Agarose gel (Thermo Scientific) in TBE buffer stained with Gel Red (Biotium) and visualised by UV.
Immunoprecipitation BSRT-7/5 CL21 cells were seeded into 6-well plates at 4.8 x10 5 cells per well. Each well was transfected with 3 μg pTM1-N or a mutant pTM1-N and 2 μg pTM1-L3V5 expression plasmids. After 24 h, cells are lysed with lysis buffer containing 20 mM Tris pH 7.5, 150 mM NaCl, 5 mM MgCl 2 , 0.5% Triton X-100, and Halt Protease inhibitor cocktail (Thermo Scientific) and 10 μg RNase (Thermo Scientific). The lysate was kept on ice for 20 min then centrifuged at 16000 x g for 20 min to pellet the cellular debris. Protein A beads were bound with anti-V5 antibody (#27671, Abcam). Lysates were rotated with beads for 2 h, after which, the beads were washed three times with lysis buffer and eluted by heating the sample to 90˚C in LDS Sample buffer and 2-mercaptoethanol reducing agent (Thermo Scientific). Samples were analysed by western blotting with primary anti-RVFV N antibody (1:5000) or anti-V5 antibody (1:2000) and secondary HRP-conjugated anti-rabbit (1:1000) or Veriblot anti-IgG (1:1000) and processed as described for immunoblotting.

N protein conservation and mutagenesis
Fourteen Phlebovirus nucleoprotein sequences available from Genbank were aligned (Fig 1) and conserved residues were identified between all phleboviruses, between phleboviruses thought to be transmitted by an insect vector or phleboviruses transmitted by ticks. Of the conserved residues identified, five amino acids were selected for alanine substitution studies. Residues Y30 (previously analysed), D34, F149 and N181 in N protein sequences were found to be conserved across the whole Phlebovirus genus, derived from viruses transmitted by both insect and ticks. Residue F11, also previously analysed, is highly conserved between the mosquitoborne phleboviruses, with the exception of CDUV, but often substituted with an isoleucine within the UUKV-like tick-borne virus group.
The ambisense coding strategy of the Phlebovirus S segment results in differential expression of both the N protein and the non-structural protein NSs [9]. To evaluate the importance of these highly conserved N residues in isolation, without any interference of NSs on N protein or minigenome activity [22], an expression plasmid system containing only the N protein open reading frame (pTM1-N) was used over plasmids that express cDNA copies of the viral S segment RNA. The two N-arm mutants, a deletion of the amino acids 1-14 and 1-31, their location modelled in Fig 2, were also introduced to our panel of N protein mutants to examine the impact of N-N multimer interactions on N protein functions.
Functional ribonucleoprotein complexes mediate the transcription and replication of the viral genome. This study aimed to assess the importance of the conserved residues identified from our phlebovirus N protein alignments. This was achieved using a previously described minigenome system, in which a virus genome segment analogue, where the viral coding sequence has been replaced with Rluc leaving the viral M segment UTRs intact, is transfected into T7 RNA polymerase expressing cells along with plasmids expressing the L protein and our N protein mutants [6,24]. As seen in Fig 3A, the two N-arm mutants, delN1-14 and delN1-31 showed no activity in the minigenome system. Interestingly, the residues examined that were highly conserved within the Phlebovirus genus showed a range of activities. Mutants F11A and F149A had no replication activity, mutant D34 showed a decreased activity compared to wildtype (WT) N and mutants Y30A and N181A showed a significantly increased activity. Previous mutagenesis studies of BUNV N protein suggested that inconsistent protein expression levels may have significant effects on minigenome system activity [17]. As such a western blot for N protein was carried out using the cell lysates from the minigenome assay to control the expression levels of N protein. Expression levels of N protein were largely consistent between different transiently transfected N mutants (Fig 3B), except arm mutant delN1-31 which showed a greatly reduced expression level and was detected by using a BSR-T7/5 clone with higher T7 RNA polymerase activity produced during this project (S1 Fig).

RNA binding properties of N protein mutants
RVFV N protein has two key functions; encapsidation of the viral RNA and multimerisation to allow for the formation of the viral RNPs [10,18,19]. To determine if the RNA binding capacity of the mutant panel was impaired, we carried out an RNA binding assay on purified mutant N proteins (Fig 4A). For this we used the viral N protein's ability to bind non-specifically bacterial RNA, which occurs before and during the protein purification process, as shown previously [18,25]. To determine whether this function was impaired, RNA was dissociated from the purified N protein using formamide-containing RNA loading buffer [25]. As shown in Fig 4B, RNA binding activity was observed in the majority of mutant N proteins examined. However, neither the arm mutant delN1-31 nor mutant Y30A had detectable RNA bound to mutant N protein when purified, which was consistent with our measurement of 260/280 ratio for all these proteins.

Multimerisation properties of N proteins
The N protein of the Smithburn vaccine strain of RVFV was shown to form higher order structures: tetramers, pentamers and hexamers in multiples of the 27 kDA monomer [10]. Here we assessed whether the minigenome activities observed previously resulted from changes in the multimerisation. The multimerisation properties of the mutant N proteins panel were determined through a DSP crosslinking assay on purified N protein from a bacterial expression system. Only the delN1-31 mutant showed a reduced capacity to multimerise and all other mutants showed multimerisation capacity similar to WT N protein (Fig 4C).

Interactions of N and L proteins
The ability for the N protein to interact with the viral L protein is important for the formation of replication-active RNP complexes. By using a construct expressing a V5-tagged L protein (L3V5) [22], we were able to perform co-immunoprecipitation (co-IP) studies to assess the interaction of L protein with our panel of mutant N proteins. The co-IP was performed under RNase conditions to reduce the effect of bound RNA influencing the interaction, though RNA bound within N protein prior to co-IP would not be removed in these conditions. Following co-IP with an anti-V5 antibody, western blotting with anti-N and anti-V5 sera was performed to identify if N protein co-immunoprecipitated with the L protein. The mutants found to be functional in the minigenome system (Y30A, D34A and N181A) interacted with the L protein. Intriguingly, N mutants' delN1-14, delN1-31 and F11A and F149A could also bind to the L protein despite not being active in the minigenome assays (Fig 5). This suggested that any functional deficiencies observed are not due to absence of direct interactions between the replication complex proteins, though L polymerase processivity could be affected. Cross-linking was used to determine the formation of N multimers, and purified mutant proteins were crosslinked as described in Materials and methods, and analysed by western blot with RVFV anti-N antibodies. βmercaptoethanol was added to the control (WT R) to reduce the di-sulfide bonds after multimerisation. This image is representative of three repeats.

Impact of N mutations on packaging and virus-like particle (VLP) formation
Amongst other functions, RVFV N protein has been implicated in an interaction with the first 30 amino acids of the cytoplasmic tail of the viral glycoprotein Gn to mediate efficient packaging of viral RNP complexes into newly forming virus particles [26]. To determine if any of the mutated amino acids of N protein abrogated packaging, we employed an assay based on packaging of a minigenome segment into VLPs. Briefly, BSR-T7/5 CL21 cells were transfected with a RVFV M segment based humanised Rluc minigenome-encoding plasmid (pTVT7-GM: hRen), plasmids encoding the L protein and WT N protein or one of the mutant N proteins, and a FFluc expressing plasmid as a transfection control. The transfection mixes were also supplemented with expression plasmid encoding the RVFV M segment glycoprotein precursor. At 48 h post transfection, the cell culture media of the donor cells (cells used to generate VLPs) was clarified by centrifugation and nuclease treated to prevent plasmid carry over. Subsequently the nuclease-treated VLP containing supernatant was used to inoculate recipient BSR-T7/5 CL21 cells transiently transfected with plasmids expressing the RVFV L protein and WT N protein. In the case of functional packaging of the minigenome segment into virions, the encapsidated pTVT7-GM:hRen is delivered to the recipient cell by the VLP where it is replicated, transcribed and translated resulting in Rluc expression (Fig 6). The VLP assay data corroborated with the minigenome data presented in Fig 3A, showing that only mutants Y30A, D34A and N181A that were functional in the minigenome system also could form functional, minigenome containing VLPs.

Discussion
In this study, we sought to gain a greater understanding of the biology and functions of the RVFV N protein through the identification of amino acids that are highly conserved among phleboviruses and that have yet unknown functions. RVFV N is a very highly conserved protein in nature as evidenced by multiple sequence alignment of different isolates, and as such there is a higher propensity of conserved residues to confer significant function. Of the conserved residues identified, both F11A and F149A had unidentified roles which appeared essential for the replication or transcription of viral RNPs and/or formation of RNP complexes. The F149A mutant showed no activity in minigenome assay and has not been previously described in the literature. We hypothesise that due to this amino acid being surface exposed within a protein cleft found in the N protein structure, it may be involved in the binding of host factors Interaction of L3V5 viral polymerase with WT and mutant N proteins. BSR-T7/5 CL21 cells were transfected with mutant or WT pTM1-N and pTM1-L3V5. Negative control was transfected pTM1-N without pTM1-L3V5 (Con). After 24 hours, cells were lysed and the lysate was applied to magnetic beads carrying V5 antibody. The bound L protein was dissociated from the beads and analysed by western blot. The proteins were fractionated on a 4-12% Bis-Tris plus (Novex) gel and transferred to a nitrocellulose membrane. Subsequently the blot was probed with anti-V5 antibody and RVFV anti-N antibodies and visualised using LI-COR. This image is representative of three repeats.
https://doi.org/10.1371/journal.pntd.0006155.g005 required for these processes. A F11G mutant has previously been shown to be involved in loss of N-N dimer formation, by a different assay system and using GST-fused N protein, through the potential misfolding of the N-terminal region and disruption of the N-N interaction [20]. Interestingly here we show that purified F11A mutant protein of RVFV N was still able to form dimers and higher order structures within our multimerisation assay system. This discrepancy may be explained by the glycine substitution resulting in a less stable structure. While it is clear that the F11 residue is essential for N protein function, it may be involved in other interactions relevant for replication other than the ascribed N-N dimer formation or interactions. These could involve, for example, processivity of RNA synthesis or specific interactions with host factors that are required for virus replication.
We also identified two residues, Y30 and N181, which showed increased activity in the minigenome system (Fig 3). The residue N181 has no previously described function, but the N181A mutant was also able to form functional virions in the VLP assay system (Fig 6). Y30A was considered to be essential in a previous study [19].
Despite evidence that Y30 is an essential residue within the N-terminal arm, involved in the multimerisation of N protein and the base stacking of RNA within the RNA binding groove, the Y30A mutant used in this study only appeared to have reduced RNA binding capacity but retained the ability to form multimers. This corroborated with a previous study indicating that the Y30A mutant did not disrupt N-N interactions [20]. This reduced RNA binding capacity of Y30A did not negatively affect the overall expression of the minigenome reporter gene or packaging into VLPs. There may be an inverse relationship between minigenome activity and RNA binding, it has been proposed that N may non-specifically bind mRNA preventing translation however luciferase transfection control levels showed no significant change in the presence of N (S2 Fig). A mutational analysis of UUKV, also of the Phlebovirus genus, observed that the Y30A mutant still had activity in a minigenome reporter assay [27]. However, in the Katz et al. study, further analysis of RNA binding was not carried out. It is possible that the reduced RNA binding capacity may positively affect RNA replication at the loss of protection from degradation, however this requires further study. The formation of functional VLPs also shows that the Y30A mutant still interacts with the Gn glycoprotein to successfully package the reporter construct.
Both the delN1-14 and delN1-31 N protein mutants showed no activity in the minigenome assay. UUKV N N-terminal mutants also showed no activity in minigenome-based assays analysed by the detection of CAT signal. These data support the hypothesis that the N-terminal arm of phleboviral N proteins are functionally essential [27]. However, the delN1-14 mutant N protein was still able to form higher order multimers and bound RNA. Based on our RNAbinding and multimerisation analysis data, it suggests that N protein is still able to efficiently multimerise and bind RNA without the 1 st helix of its N-arm, these functions are impaired if the second helix of the arm is also removed. However, if the whole arm was removed, oligomeric structures were still formed, although at a very low efficiency, suggesting that there is another mechanism or region involved in N-N binding. Comparatively, a delN1-19 mutant in UUKV N showed approximately 25% N-N binding capacity indicating that binding was still occurring at reduced levels [27]. It is possible that the loss of function is due to the impaired ability of the RNA-dependent RNA-polymerase L to track along the viral RNP and/or due to the formation of N protein monomers in an incorrect configuration. The lack of activity found in minigenome assays with delN1-31 N mutant is likely due to its inability to form higher order structures and bind viral RNA, though the ability for delN1-31 mutant to interact with L also indicates that the N-terminal arm is not involved in the direct interaction between RNAdependent RNA polymerase and viral nucleocapsid protein.
In summary, this study expands the number of known essential residues of RVFV N protein that are conserved across phleboviruses, while also informing that other residues may have additional or different functions than previously documented in RVFV. The availability of protein residue information is an important resource for further studies into the functions and potential therapeutic targets of RVFV N protein. The conserved nature of these residues in N proteins also may indicate conserved functions or interactions across the whole Phlebovirus genus and thus provide important information beyond what we currently understand for RVFV.