The N Terminus of Rift Valley Fever Virus Nucleopr
http://www.100md.com
病菌学杂志 2005年第18期
Unité de Génétique Moléculaire des Bunyaviridés, Institut Pasteur, Paris, France
ABSTRACT
Rift Valley fever virus (RVFV) is a Phlebovirus in the Bunyaviridae family. The nucleoprotein N is the most abundant component of the virion; numerous copies of N associate with the viral RNA genome and form pseudohelicoidal ribonucleoproteins (RNPs) circularized by a panhandle structure formed by the base-paired RNA sequences at the 3' and 5' termini. These structures play a central role in transcription and replication. We investigated the intermolecular interactions of the RVFV N protein and found that after chemical cross-linking treatment, the nucleoprotein from purified RNPs migrates mainly as dimers. The N-N interaction was studied using the yeast two-hybrid system, the GST pull-down method, and mutational analysis. We demonstrated that the N terminus from residue 1 to 71, and particularly Tyr 4 and Phe 11, which are conserved among phlebovirus N sequences, are involved in the interaction. The C-terminal region did not seem to be essential for the N-N interaction. Moreover, we showed that NTOS, the N protein of the related Toscana phlebovirus, interacts with itself and forms heterodimers with NRVF, suggesting that the dimeric form of N may be a conserved feature in phlebovirus RNPs.
INTRODUCTION
Rift Valley fever virus (RVFV) is an arthropod-borne virus infecting humans and animals which is transmitted mostly by Aedes sp. and Culex mosquitoes (22, 28). This virus is endemic in Africa and spread recently outside the continent as illustrated by the recent outbreaks in the Arabian Peninsula (2, 3). RVFV can persist in infected mosquito eggs, and epidemics are often associated with ecological changes or high rainfall that leads to vector amplification. During human infections, various symptoms are observed, ranging from benign fever to severe encephalitis and fatal hepatitis with hemorrhagic fever. Among domestic animals, sheep are very sensitive; infected lambs may die within 24 h, and in pregnant females, infection induces teratogenic and abortogenic effects.
RVFV belongs to the Bunyaviridae family (genus Phlebovirus), a family of spherical enveloped viruses with a tripartite RNA genome of negative or ambisense polarity (36). The L and M segments are of negative polarity and code, respectively, for the L RNA-dependent polymerase and the glycoprotein precursor, whereas the S segment utilizes an ambisense strategy and codes for the nucleoprotein N and the nonstructural protein NSs (13). Following virus entry and uncoating, the L polymerase mediates primary transcription in the cytoplasm, using the cap-snatching mechanism first described for myxoviruses (6, 21) and utilized by bunyaviruses (5, 17, 32, 38) and arenaviruses (11). Whereas primary transcription leads to the synthesis of mRNAs, secondary transcription is involved in replication and leads to the synthesis of cRNA representing replicative intermediates for the synthesis of viral RNA. Unlike naked viral mRNAs, viral RNA and cRNA are associated with numerous copies of the nucleoprotein N (27 kDa, 246 amino acids) and a few molecules of the L polymerase to form pseudohelicoidal ribonucleoproteins (RNPs). These RNPs appear circular, due to complementary sequences of the 5' and 3' noncoding extremities which allow the formation of stable panhandle structures (15, 33-35). The RNPs play a central role in the viral cycle, since they are the functional templates for transcription and replication. It is speculated also that the cytoplasmic concentration of nucleoprotein, which is the most abundant viral component in infected cells, regulates the transition between primary transcription and replication.
Because the capsid proteins of several viruses, such as hepatitis C virus, human immunodeficiency virus, Marburg virus, Sendai virus, and hantaviruses, are able to dimerize or oligomerize (1, 4, 18, 23, 25, 29, 30), we examined intermolecular interactions of the RVFV nucleoprotein. Homodimers stabilized by chemical cross-linking reagents were found in purified RNPs. Using the yeast two-hybrid system, the glutathione S-transferase (GST) pull-down method, and mutational analysis, we mapped the interacting domain to the N-terminal 71 residues, in which the Tyr 4 and Phe 11 residues, which are conserved among phleboviruses, play an important role. In addition we showed that the nucleoprotein of Toscana virus (TOSV), a related phlebovirus, is able to form dimers with itself or with the RVFV nucleoprotein, suggesting that the dimeric form of N may be a conserved feature in phlebovirus RNPs.
MATERIALS AND METHODS
Cell lines and virus strain. HeLa and Vero cells were grown in Dulbecco's modified Eagle's medium containing, respectively, 10% and 5% fetal calf serum. Antibiotics (5 U/ml penicillin and 5 μg/ml streptomycin) were added to the maintenance media.
Strain MP12 (9) of RVFV was grown in Vero E6 cells in Dulbecco's medium containing 2% fetal calf serum and antibiotics. Vaccinia virus expressing T7 polymerase (vTF7-3) (kindly provided by B. Moss) was produced in HeLa cells as described previously (10).
Plasmids. The cDNAs coding for the RVFV strain MP12 nucleoprotein and its truncated forms or for TOSV nucleoprotein were obtained, respectively, by reverse transcription-PCR from MP12-infected cell RNA and by PCR from pGEM4Z-S-TOSV expressing the S segment of TOSV (a kind gift from C. Giorgi). The primer sequences used for amplification and containing the BglII restriction sites at the termini are available on request. The cDNA fragments were digested with BglII and cloned at the BamHI site of pTM1-GST (a generous gift from S. Moyer), pACT2, or pGBKT7 vector (Clontech). To generate pTM1-NRVF and pTM1-NTOS, the coding sequences of RVFV or TOSV nucleoprotein were amplified with specific primers containing the NcoI or BglII restriction sites at the 5' and 3' termini, respectively, and the cDNAs digested with NcoI and BglII were cloned at the NcoI and BamHI sites of the pTM1 vector (10).
The mutants pTM1-GST-NY4G, -NF11G, -ND17G, and -NW24G were constructed using PCR-directed mutagenesis from pTM1-GST-NRVF plasmid and with primers containing the mutated sequence. All the constructs described were sequenced with an ABIPRISM sequencer (Applied Biosystems).
Yeast two-hybrid assays. Saccharomyces cerevisiae strain SFY526 (MATa trp1 leu2 ura3 his3 lys2 ade2 gal4 gal80 URA3::GAL1UAS-GAL1TATA-lacZ) (14) was cotransformed with pGBKT7 and pACT2 plasmids by the lithium acetate method (12) and selected for tryptophan and leucine prototrophy on appropriate minimal media. Liquid ?-galactosidase assays were performed as described previously (7, 8).
RVFV RNP purification. Vero cells were infected by RVFV strain MP12 at a multiplicity of infection of 5 PFU per cell. At 24 h postinfection, infected cells were lysed in TNE buffer (Tris, 25 mM [pH 7.5]; NaCl, 50 mM; EDTA, 2 mM) containing 0.6% NP-40 and protease inhibitors (Roche Molecular) for 5 min at 4°C. After centrifugation, cytoplasmic extracts were treated with 6 mM EDTA. RNPs were then purified by ultracentrifugation through a 20 to 40% (wt/vol) cesium chloride gradient run for 16 h at 32,000 rpm at 12°C in a Beckman SW41 rotor. The opalescent band corresponding to a density of 1.32 was collected and dialyzed against phosphate-buffered saline through nitrocellulose membranes (Millipore).
Cross-linking of purified RNPs. Purified RNPs untreated or treated with RNase (250 μg per ml), 500 mM NaCl, or both for 30 min at room temperature were incubated with a solution of glutaraldehyde (25%; Merck) diluted to a final concentration of 0.05% for 30 min at room temperature or with suberic acid bis(N-hydroxy-succinimide ester) (SAB) (Sigma) at concentrations ranging from 1.0 μM to 333 μM for 16 h at 20°C. The samples were then denatured with an equal volume of 2x dissociation buffer (Tris-HCl [pH 6.8], 10 mM; glycerol, 25%; ?-mercaptoethanol, 10%; sodium dodecyl sulfate [SDS], 10%; bromophenol blue, 0.02%), and after 5 min of heating at 95°C, the proteins were separated in SDS-10% polyacrylamide gels. Immunoblotting was carried out with RVFV N-specific murine polyclonal antibodies. The N protein was detected using the enhanced chemiluminescence system (Pierce).
GST pull down. To assess protein-protein interactions in cells, subconfluent HeLa cells were infected with vT7-3 at a multiplicity of infection of 5 PFU per cell and at 1 h postinfection were transfected with pTM1-GST or pTM1-GST-N (1.5 μg) and pTM1-N constructs (1 μg), using DOTAP (Roche Molecular) and OptiMEM medium (GIBCO). The cytoplasmic extract was prepared at 24 h posttransfection in lysis buffer (Tris-HCl, pH 7.5; NaCl, 140 mM; EDTA, 1 mM; NP-40, 0.5%; protease inhibitor [Roche Molecular]), and membranes and cell debris were pelleted at 13,000 rpm for 5 min at 4°C. The GST complexes were selected on Sepharose beads coupled to glutathione (Pharmacia). After extensive washing with phosphate-buffered saline, the complex retained on the beads was treated with the dissociation buffer. The proteins, denatured for 5 min at 95°C, were analyzed by polyacrylamide gel electrophoresis (PAGE) and immunoblotting with RVFV or TOSV N-specific murine polyclonal antibodies.
RESULTS
RVFV nucleoprotein forms dimers in purified RVFV RNPs. To determine the oligomerization state of the RVFV N protein, RNPs from infected Vero cells were purified to equilibrium in a cesium chloride gradient and treated with two chemical cross-linking reagents: (i) glutaraldehyde, which reacts with Lys, His, Trp, and Tyr residues with a spacer arm length of 6 angstroms, and (ii) SAB, which is a bifunctional reagent of fixed size (13.1 angstroms) reacting only with lysine residues. When the RVFV RNPs were treated with glutaraldehyde (0.05%) prior to protein analysis by SDS-PAGE and immunoblotting with RVFV N-specific murine polyclonal antibodies (Fig. 1A), the N protein migrated as a band of 55 kDa, a size consistent with the migration of a dimeric form, compared to the monomeric form of 27 kDa observed in the absence of cross-linker (Fig. 1A, lane 1). Bands corresponding to higher molecular mass, including multimers which did not penetrate in the gel, were also detected (Fig. 1A, lanes 2 to 5). When the treatment was carried out with SAB, dimers were almost the only form of oligomers observed (Fig. 1B). The proportion of oligomers increased with the concentration of SAB. Of note, after treatment with both agents, the relative proportion of dimers compared to other oligomers suggests that the dimer represents the major oligomeric form of N in RNPs.
To eliminate the possibility that what we identified as dimers/oligomers might be an adduct between N and the viral RNA, purified RNPs were treated with RNase in the presence or absence of 500 mM NaCl prior to glutaraldehyde treatment. After this treatment, which would disrupt RNA-protein structures and some protein-protein interactions, dimers and multimers were still detected (Fig. 1A, lanes 3 and 5), suggesting that in RNPs, two N molecules are kept in close vicinity, allowing an efficient cross-linking, and that the interaction is relatively stringent.
The N-terminal 71 residues are involved in the N-N interaction. To ascertain the existence of N-N interactions, we used the yeast two-hybrid system. When the yeast strain SFY526 was transformed with pGBKT7-N and pACT2-N, expressing, respectively, the Gal4 DNA binding domain (DBD) and activating domain (AD) fused to the RVFV N protein, the homotypic interaction of RVFV N protein was clearly detected as monitored by histidine prototrophy (not shown) and ?-galactosidase activity (Fig. 2A and B). As controls, we checked by Western blotting that the fusion proteins DBD-N and AD-N containing the entire RVFV N sequence were expressed with the proper size (data not shown). We verified also that expression of DBD-N and AD-N was not toxic for the yeast and that these fusion proteins did not bind or transactivate nonspecifically the reporter gene when expressed alone (Fig. 2B). To localize the region involved in dimerization, N- and C-terminally deleted mutants were constructed in the pACT2-N plasmid (Fig. 2A), and the mutated AD-N proteins were tested for their ability to interact with wild-type DBD-N. Except for pACT2-N71-246, which transactivated the ?-galactosidase gene when transformed alone and was not assayed further in this system, the others had no transactivating activity (not shown) and could be tested for their interaction with DBD-N. Two C-terminally truncated forms, AD-N1-133 and AD-N1-152 (encoding residues 1 to 133 and 1 to 152, respectively), coexpressed with the wild-type DBD-N activated the expression of ?-galactosidase almost as efficiently as the wild-type NRVF (Fig. 2B). In contrast, the N-terminally truncated AD-N133-246 did not activate the reporter gene expression (Fig. 2B). These results indicate that the N-terminal moiety is essential for the N-N interaction.
To better delineate the domain of interaction, we coexpressed the native NRVF protein with GST-NRVF, the glutathione S-transferase fused to wild-type or mutated N, and assessed the interaction by the ability of nucleoprotein to copurify with GST-NRVF on glutathione-Sepharose beads. Expression was carried out via the T7-based plasmids pTM1-NRVF and pTM1-GST-NRVF after transfection in HeLa cells infected with vT7-3, a vaccinia virus expressing the T7 RNA polymerase (Fig. 3). When the wild-type NRVF and GST-NRVF were coexpressed, the two proteins were found to copurify as visualized in the immunoblot by the bands of 27 and 54 kDa, which represent NRVF and GST-NRVF, respectively (Fig. 3B, lane 1). As a negative control, GST and NRVF were coexpressed, and, as expected, the nucleoprotein was not retained on the glutathione-Sepharose beads (data not shown). These data confirmed the N-N interaction and the specificity of the assay. We then analyzed the interaction of the native NRVF with GST-N1-133, GST-N1-152, GST-N133-246, and GST-N71-246, in which the nucleoprotein is truncated from the C- or N-terminal region (Fig. 3A). When NRVF was coexpressed with GST-N1-133 or GST-N1-152, it copurified with the GST fusion protein (Fig. 3B, lanes 2 and 3), whereas it did not when it was coexpressed with GST-N133-246 or GST-N71-246 (Fig. 3B, lanes 4 and 5). These results clearly indicated that the N-terminal first 71 amino acids contain residues crucial for the N-N interaction. Of note, since some mutations reduced significantly the level of expression of the native form (not shown) but not that of the GST-NRVF mutants (Fig. 4B), the mutational analysis was carried out with the sequence of the GST-NRVF protein but not with the N native form.
Tyrosine 4 and phenylalanine 11 are crucial for the N-N interaction. Comparison of the nucleoprotein sequences of several phleboviruses, i.e., RVFV, TOSV, Punta Toro virus, and Sicilian fever sandfly virus, using the Clustal W algorithm indicated the presence of several conserved sequences (Fig. 4A) within the 71 first N-terminal amino acids. Four residues are conserved and aligned, i.e., Tyr 4, Phe 11, Asp 17, and Trp 24 (numbers refer to the RVFV sequence) (Fig. 4A), as well as two motifs (residues 28 to 34 and 46 to 68). The latter two are also conserved in the most divergent Uukuniemi virus sequence. Mutational analysis was performed to investigate the putative role of the conserved amino acids in the N-N interaction. A series of Ala substitutions were made within the two conserved motifs of GST-NRVF; neither the single mutations F28A, Y30A, D34A, R64A, N66A, K67A, and P68A nor the double mutations F28A Y30A, Q31A D34A, R64A N66A, and K67A P68A disrupted the interaction (not shown). Regarding the conserved amino acids at positions 4, 11, 17, and 24, each of these amino acid residues was mutated into a Gly residue in the N sequence of pTM1-GST-NRVF, generating plasmids pTM1-GST-NY4G, -GST-NF11G, -GST-ND17G, and -GST-NW24G, and the ability of the mutants to interact with NRVF was tested using the GST pull-down assay. When GST-ND17G or GST-NW24G was coexpressed with the wild-type NRVF protein, NRVF copurified with GST-ND17G or GST-NW24G (Fig. 4B, lanes 5 and 6), but coexpression of NRVF and GST-NY4G or GST-NF11G drastically impaired the complex formation (Fig. 4B, lanes 3 and 4), indicating that substitutions of Tyr 4 and Phe 11 impeded the interaction with nucleoprotein.
The crucial role of phenylalanine 11 in N-N interaction is conserved in TOSV. Because of the amino acid conservation between phlebovirus nucleoproteins, we investigated whether assembly of N into dimers is a general property among phleboviruses and tested the ability of the nucleoprotein from another phlebovirus, Toscana virus, to interact with itself. Thus, pTM1-NTos and pTM1-GST-NTos plasmids were generated. After coexpression of the proteins in HeLa cells infected with vT7-3 and analysis by GST pull-down assay, NTOS was found to interact with the GST-NTOS protein, indicating that NTOS interacts with itself (Fig. 5, lane 2).
When coexpressed with GST-NRVF, NTOS copurified with the GST fusion protein (Fig. 5, lane 3), showing that RVFV and TOSV nucleoproteins can form heterodimers. To estimate the role of the conserved residues in the heterotypic interaction, we assayed the interaction between NTOS and the substitution mutants in GST-NRVF. NTOS copurified with GST-NY4G and GST-ND17G (Fig. 5, lanes 4 and 6), but the N-N interaction was abolished when NTOS was coexpressed with GST-NF11G or GST-NW24G (Fig. 5, lanes 5 and 7). These results show that in both homotypic and heterotypic interactions, Phe 11 in RVFV nucleoprotein plays an important role. In contrast with the homotypic interaction, Tyr 4 in NRVF is not essential for the heterotypic interaction with NTOS, but instead Trp 24, another conserved residue, becomes crucial in the interaction with NTOS.
DISCUSSION
In this paper, we demonstrate that the RVFV N protein interacts with itself and forms homodimers. These structures were observed in purified RNPs treated with glutaraldehyde and SAB (Fig. 1). The interaction was not affected after treatments that disrupt the association between viral RNA and N proteins. Similar dimeric forms were detected with cross-linked recombinant nucleoprotein (data not shown). Although we cannot ruled out the possibility that the nucleoprotein is also associated with cellular proteins contaminating the RNPs, these complexes must be in minor amounts, since only trace amounts of contaminating proteins were detected in the purified RNP fraction after Coomassie blue staining (data not shown). The N-N interaction was analyzed using the yeast two-hybrid system and GST pull-down method from proteins expressed in HeLa cells (Fig. 2B and 3B). It must be noted that the two-hybrid system could not be utilized extensively to delineate the interaction domain, because some of the deletion mutants (e.g., AD-N71-246) transactivated expression of the reporter gene. Using the GST pull-down assay, we tested the N-N interaction with the wild-type protein and the wild-type or mutated NRVF fused to GST. The mutations were introduced into the GST fusion protein and in the native NRVF, but the latter ones were not used because N-terminal truncations led to drastically reduced expression levels, probably due to the encephalomyocarditis virus internal ribosome entry site present in pTM1 and the sequence downstream of the initiating ATG (16). Using this system, we demonstrated that the C-terminal region did not seem to be essential for the N-N interaction but that the N terminus, and Tyr 4 and Phe 11 in particular, plays an important role. (Fig. 3 and 4). Except for Uukuniemi virus N protein, these two amino acid residues are conserved within phlebovirus N proteins (Fig. 4A). In fact, it could well be that homodimer formation is a property of the phlebotomine and mosquito-borne phlebovirus nucleoproteins, as illustrated by the data showing that the N protein of TOSV is able to interact with itself and with NRVF (Fig. 5). However, it seems that this does not apply to N protein of the tick-borne Uukuniemi virus, in which both N and C termini are necessary for dimerization (A. Plyusnin, personal communication). For the formation of RVFV homodimers and heterodimers with TOSV, Phe 11 in the RVFV sequence appeared to be crucial for both interactions. A second amino acid, Tyr 4 or Trp 24, is also important in, respectively, homo- or heterodimer formation. The N-terminal region was predicted to contain alpha-helices. Thus, being separated by five or six residues, the conserved amino acid residues at the N terminus (Tyr 4, Phe 11, Asp 17, or Trp 24) are likely localized on the same face of the alpha-helix. Thus, except for Asp 17, for which no role was found, changes at one of these positions might disrupt the folding of the N-terminal region and impair the interaction. At the present time, the three-dimensional structure of bunyavirus nucleoproteins is still unknown. This information would be of value to better understand dimer formation.
To our knowledge, no data are available for the N protein of nairoviruses, but, interestingly, the N proteins of tospoviruses, hantaviruses, and probably orthobunyaviruses (31) have been shown to assemble into oligomers. However, the N-N interactions appear to differ from those in the present situation. In the case of the tomato spotted wilt tospovirus, the nature of the oligomers in the RNPs is not known, but using two-hybrid and plasmon resonance analyses, Uhrig et al. showed that both the N and C termini are crucial for the N-N interaction, which implies a head-to-tail interaction forming a multimeric chain (37). In the case of hantaviruses, the N protein present in viral RNPs or expressed as a recombinant protein in Escherichia coli assembles into dimers and trimers, the latter being the major form (1, 18). Our results suggest that the structure of RVFV (and probably phlebovirus) nucleoprotein oligomers is different from that of hantavirus nucleoprotein oligomers. These differences are evidenced by several observations: (i) among the different bunyavirus genera, nucleoproteins are poorly conserved in size and sequence, with the hantavirus N being more complex, with a size twice that of the phlebovirus N; (ii) gel analysis of cross-linked N oligomers clearly showed that trimers are the major forms in hantavirus RNPs (1, 18), whereas N dimers are the main oligomers in RVFV RNPs (this paper); and (iii) the domains important for the hantavirus N trimer assembly are conserved only among hantaviruses, and not phleboviruses or other bunyaviruses, and vice versa. For hantaviruses, the N-terminal residues assemble into trimeric coiled coils. A prediction for such a motif in the N terminus of RVFV N protein, using the algorithm of Lupas et al. (24), has a rather low score of 0.2. Notwithstanding, we explored this possibility and changed Ala-Ala at positions 12 and 13 to Gly-Gly to disrupt this putative domain. However, when tested in the GST pull-down assay, the double mutant was still able to interact with N (data not shown), indicating that if a coiled-coil motif exists in the N terminus, it does not contribute significantly to the interaction. Together these results point out a major difference in the structures of RVFV and hantavirus nucleoproteins, leading, respectively, to dimers and trimers. In RVFV nucleoprotein, only the N-terminal region is involved in N-N interaction whereas in the hantavirus protein, the N and C termini both contribute to the interaction, with the primary domain being located at amino acids 393 to 398 in Tula virus N and the domain from amino acid 1 to 43 representing a secondary interaction (19, 20). The mono- and dimeric forms may exist transiently, while the trimeric form is highly stable (27).
Finally, the question which remains is whether oligomers have specific functions in the viral cycle. The N proteins of bunyaviruses must have at least two intrinsic functions, one for binding to the viral RNA and the other for the formation of a complex active in transcription and replication, since only encapsidated RNA can serve as template for transcription and replication. Investigations of the N proteins of viruses in the Orthobunyavirus, Hantavirus, and Tospovirus genera have shown that N binds to RNA essentially in a nonspecific manner, but it seems to preferentially bind double-stranded RNA or sequence present at the 5' end of the genomic RNA (31, 39). Recently, Mir and Panganiban (26, 27) reported that the mono- and dimeric forms of Sin Nombre hantavirus N bind RNA, forming a complex that is semispecific and salt sensitive. In contrast, purified N trimers are able to discriminate between viral and nonviral RNA molecules and recognize and bind with high affinity the panhandle structure composed of the 5' and 3' ends of the viral RNA. The discrimination of N trimers for RNA binding between viral and nonviral RNA molecules could be very important during the early steps of encapsidation by acting as assembly intermediates during RNP formation, and therefore oligomerization may be involved in replication. In the case of RVFV, further work will be required to determine if dimers play a similar role in RNP formation.
ACKNOWLEDGMENTS
N.L.M. is a recipient of fellowships from Ministère de la Recherche, Association de la Recherche Contre le Cancer (ARC), and Fondation de France. N.G. is a recipient of a "Contrat Jeune Chercheur" from DGA and a fellowship from Académie de Médecine. This work was supported in part by grant MIC0310 from INSERM.
REFERENCES
Alfadhli, A., Z. Love, B. Arvidson, J. Seeds, J. Willey, and E. Barklis. 2001. Hantavirus nucleocapsid protein oligomerization. J. Virol. 75:2019-2023.
Anonymous. 2000. Outbreak of Rift Valley fever, Yemen, August-October 2000. Wkly. Epidemiol. Rec. 75:392-395.
Anonymous. 2000. Rift Valley fever, Saudi Arabia, August-October 2000. Wkly. Epidemiol. Rec. 75:370-371.
Becker, S., C. Rinne, U. Hofsass, H. D. Klenk, and E. Muhlberger. 1998. Interactions of Marburg virus nucleocapsid proteins. Virology 249:406-417.
Bouloy, M., N. Pardigon, P. Vialat, S. Gerbaud, and M. Girard. 1990. Characterization of the 5' and 3' ends of viral messenger RNAs isolated from BHK21 cells infected with Germiston virus (Bunyavirus). Virology 175:50-58.
Bouloy, M., S. J. Plotch, and R. M. Krug. 1978. Globin mRNAs are primers for the transcription of influenza viral RNA in vitro. Proc. Natl. Acad. Sci. USA 75:4886-4890.
Campbell, K. S., A. Buder, and U. Deuschle. 1995. Interaction of p56lck with CD4 in the yeast two-hybrid system. Ann. N. Y. Acad. Sci. 766:89-92.
Campbell, K. S., A. Buder, and U. Deuschle. 1995. Interactions between the amino-terminal domain of p56lck and cytoplasmic domains of CD4 and CD8 alpha in yeast. Eur. J. Immunol. 25:2408-2412.
Caplen, H., C. J. Peters, and D. H. Bishop. 1985. Mutagen-directed attenuation of Rift Valley fever virus as a method for vaccine development. J. Gen. Virol. 66:2271-2277.
Fuerst, T. R., E. G. Niles, F. W. Studier, and B. Moss. 1986. Eukaryotic transient-expression system based on recombinant vaccinia virus that synthesizes bacteriophage T7 RNA polymerase. Proc. Natl. Acad. Sci. USA 83:8122-8126.
Garcin, D., M. Lezzi, M. Dobbs, R. M. Elliott, C. Schmaljohn, C. Y. Kang, and D. Kolakofsky. 1995. The 5' ends of Hantaan virus (Bunyaviridae) RNAs suggest a prime-and-realign mechanism for the initiation of RNA synthesis. J. Virol. 69:5754-5762.
Gietz, D., A. St Jean, R. A. Woods, and R. H. Schiestl. 1992. Improved method for high efficiency transformation of intact yeast cells. Nucleic Acids Res. 20:1425.
Giorgi, C., L. Accardi, L. Nicoletti, M. C. Gro, K. Takehara, C. Hilditch, S. Morikawa, and D. H. Bishop. 1991. Sequences and coding strategies of the S RNAs of Toscana and Rift Valley fever viruses compared to those of Punta Toro, Sicilian sandfly fever, and Uukuniemi viruses. Virology 180:738-753.
Harper, J. W., G. R. Adami, N. Wei, K. Keyomarsi, and S. J. Elledge. 1993. The p21 Cdk-interacting protein Cip1 is a potent inhibitor of G1 cyclin-dependent kinases. Cell 75:805-816.
Hewlett, M. J., R. F. Pettersson, and D. Baltimore. 1977. Circular forms of Uukuniemi virion RNA: an electron microscopic study. J. Virol. 21:1085-1093.
Hunt, S. L., A. Kaminski, and R. J. Jackson. 1993. The influence of viral coding sequences on the efficiency of internal initiation of translation of cardiovirus RNAs. Virology 197:801-807.
Jin, H., and R. M. Elliott. 1993. Characterization of Bunyamwera virus S RNA that is transcribed and replicated by the L protein expressed from recombinant vaccinia virus. J. Virol. 67:1396-1404.
Kaukinen, P., V. Koistinen, O. Vapalahti, A. Vaheri, and A. Plyusnin. 2001. Interaction between molecules of hantavirus nucleocapsid protein. J. Gen. Virol. 82:1845-1853.
Kaukinen, P., V. Kumar, K. Tulimaki, P. Engelhardt, A. Vaheri, and A. Plyusnin. 2004. Oligomerization of Hantavirus N protein: C-terminal alpha-helices interact to form a shared hydrophobic space. J. Virol. 78:13669-13677.
Kaukinen, P., A. Vaheri, and A. Plyusnin. 2003. Mapping of the regions involved in homotypic interactions of Tula hantavirus N protein. J. Virol. 77:10910-10916.
Krug, R. M., M. Bouloy, and S. J. Plotch. 1980. RNA primers and the role of host nuclear RNA polymerase II in influenza viral RNA transcription. Philos. Trans. R. Soc. London B 288:359-370.
Linthicum, K. J., A. Anyamba, C. J. Tucker, P. W. Kelley, M. F. Myers, and C. J. Peters. 1999. Climate and satellite indicators to forecast Rift Valley fever epidemics in Kenya. Science 285:397-400.
Luban, J., K. B. Alin, K. L. Bossolt, T. Humaran, and S. P. Goff. 1992. Genetic assay for multimerization of retroviral Gag polyproteins. J. Virol. 66:5157-5160.
Lupas, A., M. Van Dyke, and J. Stock. 1991. Predicting coiled coils from protein sequences. Science 252:1162-1164.
Matsumoto, M., S. B. Hwang, K. S. Jeng, N. Zhu, and M. M. Lai. 1996. Homotypic interaction and multimerization of hepatitis C virus core protein. Virology 218:43-51.
Mir, M. A., and A. T. Panganiban. 2005. The hantavirus nucleocapsid protein recognizes specific features of the viral RNA panhandle and is altered in conformation upon RNA binding. J. Virol. 79:1824-1835.
Mir, M. A., and A. T. Panganiban. 2004. Trimeric hantavirus nucleocapsid protein binds specifically to the viral RNA panhandle. J. Virol. 78:8281-8288.
Morrill, J. C., and D. J. McClain. 1996. Epidemiology and pathogenesis of the Rift Valley fever and other phleboviruses, p. 281-293. In H. Fraenkel-Conrat and R. R. Wagner (ed.), The viruses. Plenum Press, New York, N.Y.
Myers, T. M., A. Pieters, and S. A. Moyer. 1997. A highly conserved region of the Sendai virus nucleocapsid protein contributes to the NP-NP binding domain. Virology 229:322-335.
Nolandt, O., V. Kern, H. Muller, E. Pfaff, L. Theilmann, R. Welker, and H. G. Krausslich. 1997. Analysis of hepatitis C virus core protein interaction domains. J Gen. Virol. 78:1331-1340.
Osborne, J. C., and R. M. Elliott. 2000. RNA binding properties of bunyamwera virus nucleocapsid protein and selective binding to an element in the 5' terminus of the negative-sense S segment. J. Virol. 74:9946-9952.
Patterson, J. L., and D. Kolakofsky. 1984. Characterization of La Crosse virus small-genome transcripts. J. Virol. 49:680-685.
Pettersson, R. F., and C. H. von Bonsdorff. 1975. Ribonucleoproteins of Uukuniemi virus are circular. J. Virol. 15:386-392.
Samso, A., M. Bouloy, and C. Hannoun. 1975. Circular ribonucleoproteins in the virus Lumbo (Bunyavirus). C. R. Acad. Sci. D 280:779-782.
Samso, A., M. Bouloy, and C. Hannoun. 1976. Demonstration of circular ribonucleic acid in the Lumbo virus (Bunyavirus). C. R. Acad. Sci. D 282:1653-1655.
Schmaljohn, C., and J. W. Hooper. 2001. Bunyaviridae: the viruses and their replication, 4th ed. Lippincott Williams & Wilkins, Philadelphia, Pa.
Uhrig, J. F., T. R. Soellick, C. J. Minke, C. Philipp, J. W. Kellmann, and P. H. Schreier. 1999. Homotypic interaction and multimerization of nucleocapsid protein of tomato spotted wilt tospovirus: identification and characterization of two interacting domains. Proc. Natl. Acad. Sci. USA 96:55-60.
Vialat, P., and M. Bouloy. 1992. Germiston virus transcriptase requires active 40S ribosomal subunits and utilizes capped cellular RNAs. J. Virol. 66:685-693.
Xu, X., W. Severson, N. Villegas, C. S. Schmaljohn, and C. B. Jonsson. 2002. The RNA binding domain of the hantaan virus N protein maps to a central, conserved region. J. Virol. 76:3301-3308.(Nicolas Le May, Nicolas G)
ABSTRACT
Rift Valley fever virus (RVFV) is a Phlebovirus in the Bunyaviridae family. The nucleoprotein N is the most abundant component of the virion; numerous copies of N associate with the viral RNA genome and form pseudohelicoidal ribonucleoproteins (RNPs) circularized by a panhandle structure formed by the base-paired RNA sequences at the 3' and 5' termini. These structures play a central role in transcription and replication. We investigated the intermolecular interactions of the RVFV N protein and found that after chemical cross-linking treatment, the nucleoprotein from purified RNPs migrates mainly as dimers. The N-N interaction was studied using the yeast two-hybrid system, the GST pull-down method, and mutational analysis. We demonstrated that the N terminus from residue 1 to 71, and particularly Tyr 4 and Phe 11, which are conserved among phlebovirus N sequences, are involved in the interaction. The C-terminal region did not seem to be essential for the N-N interaction. Moreover, we showed that NTOS, the N protein of the related Toscana phlebovirus, interacts with itself and forms heterodimers with NRVF, suggesting that the dimeric form of N may be a conserved feature in phlebovirus RNPs.
INTRODUCTION
Rift Valley fever virus (RVFV) is an arthropod-borne virus infecting humans and animals which is transmitted mostly by Aedes sp. and Culex mosquitoes (22, 28). This virus is endemic in Africa and spread recently outside the continent as illustrated by the recent outbreaks in the Arabian Peninsula (2, 3). RVFV can persist in infected mosquito eggs, and epidemics are often associated with ecological changes or high rainfall that leads to vector amplification. During human infections, various symptoms are observed, ranging from benign fever to severe encephalitis and fatal hepatitis with hemorrhagic fever. Among domestic animals, sheep are very sensitive; infected lambs may die within 24 h, and in pregnant females, infection induces teratogenic and abortogenic effects.
RVFV belongs to the Bunyaviridae family (genus Phlebovirus), a family of spherical enveloped viruses with a tripartite RNA genome of negative or ambisense polarity (36). The L and M segments are of negative polarity and code, respectively, for the L RNA-dependent polymerase and the glycoprotein precursor, whereas the S segment utilizes an ambisense strategy and codes for the nucleoprotein N and the nonstructural protein NSs (13). Following virus entry and uncoating, the L polymerase mediates primary transcription in the cytoplasm, using the cap-snatching mechanism first described for myxoviruses (6, 21) and utilized by bunyaviruses (5, 17, 32, 38) and arenaviruses (11). Whereas primary transcription leads to the synthesis of mRNAs, secondary transcription is involved in replication and leads to the synthesis of cRNA representing replicative intermediates for the synthesis of viral RNA. Unlike naked viral mRNAs, viral RNA and cRNA are associated with numerous copies of the nucleoprotein N (27 kDa, 246 amino acids) and a few molecules of the L polymerase to form pseudohelicoidal ribonucleoproteins (RNPs). These RNPs appear circular, due to complementary sequences of the 5' and 3' noncoding extremities which allow the formation of stable panhandle structures (15, 33-35). The RNPs play a central role in the viral cycle, since they are the functional templates for transcription and replication. It is speculated also that the cytoplasmic concentration of nucleoprotein, which is the most abundant viral component in infected cells, regulates the transition between primary transcription and replication.
Because the capsid proteins of several viruses, such as hepatitis C virus, human immunodeficiency virus, Marburg virus, Sendai virus, and hantaviruses, are able to dimerize or oligomerize (1, 4, 18, 23, 25, 29, 30), we examined intermolecular interactions of the RVFV nucleoprotein. Homodimers stabilized by chemical cross-linking reagents were found in purified RNPs. Using the yeast two-hybrid system, the glutathione S-transferase (GST) pull-down method, and mutational analysis, we mapped the interacting domain to the N-terminal 71 residues, in which the Tyr 4 and Phe 11 residues, which are conserved among phleboviruses, play an important role. In addition we showed that the nucleoprotein of Toscana virus (TOSV), a related phlebovirus, is able to form dimers with itself or with the RVFV nucleoprotein, suggesting that the dimeric form of N may be a conserved feature in phlebovirus RNPs.
MATERIALS AND METHODS
Cell lines and virus strain. HeLa and Vero cells were grown in Dulbecco's modified Eagle's medium containing, respectively, 10% and 5% fetal calf serum. Antibiotics (5 U/ml penicillin and 5 μg/ml streptomycin) were added to the maintenance media.
Strain MP12 (9) of RVFV was grown in Vero E6 cells in Dulbecco's medium containing 2% fetal calf serum and antibiotics. Vaccinia virus expressing T7 polymerase (vTF7-3) (kindly provided by B. Moss) was produced in HeLa cells as described previously (10).
Plasmids. The cDNAs coding for the RVFV strain MP12 nucleoprotein and its truncated forms or for TOSV nucleoprotein were obtained, respectively, by reverse transcription-PCR from MP12-infected cell RNA and by PCR from pGEM4Z-S-TOSV expressing the S segment of TOSV (a kind gift from C. Giorgi). The primer sequences used for amplification and containing the BglII restriction sites at the termini are available on request. The cDNA fragments were digested with BglII and cloned at the BamHI site of pTM1-GST (a generous gift from S. Moyer), pACT2, or pGBKT7 vector (Clontech). To generate pTM1-NRVF and pTM1-NTOS, the coding sequences of RVFV or TOSV nucleoprotein were amplified with specific primers containing the NcoI or BglII restriction sites at the 5' and 3' termini, respectively, and the cDNAs digested with NcoI and BglII were cloned at the NcoI and BamHI sites of the pTM1 vector (10).
The mutants pTM1-GST-NY4G, -NF11G, -ND17G, and -NW24G were constructed using PCR-directed mutagenesis from pTM1-GST-NRVF plasmid and with primers containing the mutated sequence. All the constructs described were sequenced with an ABIPRISM sequencer (Applied Biosystems).
Yeast two-hybrid assays. Saccharomyces cerevisiae strain SFY526 (MATa trp1 leu2 ura3 his3 lys2 ade2 gal4 gal80 URA3::GAL1UAS-GAL1TATA-lacZ) (14) was cotransformed with pGBKT7 and pACT2 plasmids by the lithium acetate method (12) and selected for tryptophan and leucine prototrophy on appropriate minimal media. Liquid ?-galactosidase assays were performed as described previously (7, 8).
RVFV RNP purification. Vero cells were infected by RVFV strain MP12 at a multiplicity of infection of 5 PFU per cell. At 24 h postinfection, infected cells were lysed in TNE buffer (Tris, 25 mM [pH 7.5]; NaCl, 50 mM; EDTA, 2 mM) containing 0.6% NP-40 and protease inhibitors (Roche Molecular) for 5 min at 4°C. After centrifugation, cytoplasmic extracts were treated with 6 mM EDTA. RNPs were then purified by ultracentrifugation through a 20 to 40% (wt/vol) cesium chloride gradient run for 16 h at 32,000 rpm at 12°C in a Beckman SW41 rotor. The opalescent band corresponding to a density of 1.32 was collected and dialyzed against phosphate-buffered saline through nitrocellulose membranes (Millipore).
Cross-linking of purified RNPs. Purified RNPs untreated or treated with RNase (250 μg per ml), 500 mM NaCl, or both for 30 min at room temperature were incubated with a solution of glutaraldehyde (25%; Merck) diluted to a final concentration of 0.05% for 30 min at room temperature or with suberic acid bis(N-hydroxy-succinimide ester) (SAB) (Sigma) at concentrations ranging from 1.0 μM to 333 μM for 16 h at 20°C. The samples were then denatured with an equal volume of 2x dissociation buffer (Tris-HCl [pH 6.8], 10 mM; glycerol, 25%; ?-mercaptoethanol, 10%; sodium dodecyl sulfate [SDS], 10%; bromophenol blue, 0.02%), and after 5 min of heating at 95°C, the proteins were separated in SDS-10% polyacrylamide gels. Immunoblotting was carried out with RVFV N-specific murine polyclonal antibodies. The N protein was detected using the enhanced chemiluminescence system (Pierce).
GST pull down. To assess protein-protein interactions in cells, subconfluent HeLa cells were infected with vT7-3 at a multiplicity of infection of 5 PFU per cell and at 1 h postinfection were transfected with pTM1-GST or pTM1-GST-N (1.5 μg) and pTM1-N constructs (1 μg), using DOTAP (Roche Molecular) and OptiMEM medium (GIBCO). The cytoplasmic extract was prepared at 24 h posttransfection in lysis buffer (Tris-HCl, pH 7.5; NaCl, 140 mM; EDTA, 1 mM; NP-40, 0.5%; protease inhibitor [Roche Molecular]), and membranes and cell debris were pelleted at 13,000 rpm for 5 min at 4°C. The GST complexes were selected on Sepharose beads coupled to glutathione (Pharmacia). After extensive washing with phosphate-buffered saline, the complex retained on the beads was treated with the dissociation buffer. The proteins, denatured for 5 min at 95°C, were analyzed by polyacrylamide gel electrophoresis (PAGE) and immunoblotting with RVFV or TOSV N-specific murine polyclonal antibodies.
RESULTS
RVFV nucleoprotein forms dimers in purified RVFV RNPs. To determine the oligomerization state of the RVFV N protein, RNPs from infected Vero cells were purified to equilibrium in a cesium chloride gradient and treated with two chemical cross-linking reagents: (i) glutaraldehyde, which reacts with Lys, His, Trp, and Tyr residues with a spacer arm length of 6 angstroms, and (ii) SAB, which is a bifunctional reagent of fixed size (13.1 angstroms) reacting only with lysine residues. When the RVFV RNPs were treated with glutaraldehyde (0.05%) prior to protein analysis by SDS-PAGE and immunoblotting with RVFV N-specific murine polyclonal antibodies (Fig. 1A), the N protein migrated as a band of 55 kDa, a size consistent with the migration of a dimeric form, compared to the monomeric form of 27 kDa observed in the absence of cross-linker (Fig. 1A, lane 1). Bands corresponding to higher molecular mass, including multimers which did not penetrate in the gel, were also detected (Fig. 1A, lanes 2 to 5). When the treatment was carried out with SAB, dimers were almost the only form of oligomers observed (Fig. 1B). The proportion of oligomers increased with the concentration of SAB. Of note, after treatment with both agents, the relative proportion of dimers compared to other oligomers suggests that the dimer represents the major oligomeric form of N in RNPs.
To eliminate the possibility that what we identified as dimers/oligomers might be an adduct between N and the viral RNA, purified RNPs were treated with RNase in the presence or absence of 500 mM NaCl prior to glutaraldehyde treatment. After this treatment, which would disrupt RNA-protein structures and some protein-protein interactions, dimers and multimers were still detected (Fig. 1A, lanes 3 and 5), suggesting that in RNPs, two N molecules are kept in close vicinity, allowing an efficient cross-linking, and that the interaction is relatively stringent.
The N-terminal 71 residues are involved in the N-N interaction. To ascertain the existence of N-N interactions, we used the yeast two-hybrid system. When the yeast strain SFY526 was transformed with pGBKT7-N and pACT2-N, expressing, respectively, the Gal4 DNA binding domain (DBD) and activating domain (AD) fused to the RVFV N protein, the homotypic interaction of RVFV N protein was clearly detected as monitored by histidine prototrophy (not shown) and ?-galactosidase activity (Fig. 2A and B). As controls, we checked by Western blotting that the fusion proteins DBD-N and AD-N containing the entire RVFV N sequence were expressed with the proper size (data not shown). We verified also that expression of DBD-N and AD-N was not toxic for the yeast and that these fusion proteins did not bind or transactivate nonspecifically the reporter gene when expressed alone (Fig. 2B). To localize the region involved in dimerization, N- and C-terminally deleted mutants were constructed in the pACT2-N plasmid (Fig. 2A), and the mutated AD-N proteins were tested for their ability to interact with wild-type DBD-N. Except for pACT2-N71-246, which transactivated the ?-galactosidase gene when transformed alone and was not assayed further in this system, the others had no transactivating activity (not shown) and could be tested for their interaction with DBD-N. Two C-terminally truncated forms, AD-N1-133 and AD-N1-152 (encoding residues 1 to 133 and 1 to 152, respectively), coexpressed with the wild-type DBD-N activated the expression of ?-galactosidase almost as efficiently as the wild-type NRVF (Fig. 2B). In contrast, the N-terminally truncated AD-N133-246 did not activate the reporter gene expression (Fig. 2B). These results indicate that the N-terminal moiety is essential for the N-N interaction.
To better delineate the domain of interaction, we coexpressed the native NRVF protein with GST-NRVF, the glutathione S-transferase fused to wild-type or mutated N, and assessed the interaction by the ability of nucleoprotein to copurify with GST-NRVF on glutathione-Sepharose beads. Expression was carried out via the T7-based plasmids pTM1-NRVF and pTM1-GST-NRVF after transfection in HeLa cells infected with vT7-3, a vaccinia virus expressing the T7 RNA polymerase (Fig. 3). When the wild-type NRVF and GST-NRVF were coexpressed, the two proteins were found to copurify as visualized in the immunoblot by the bands of 27 and 54 kDa, which represent NRVF and GST-NRVF, respectively (Fig. 3B, lane 1). As a negative control, GST and NRVF were coexpressed, and, as expected, the nucleoprotein was not retained on the glutathione-Sepharose beads (data not shown). These data confirmed the N-N interaction and the specificity of the assay. We then analyzed the interaction of the native NRVF with GST-N1-133, GST-N1-152, GST-N133-246, and GST-N71-246, in which the nucleoprotein is truncated from the C- or N-terminal region (Fig. 3A). When NRVF was coexpressed with GST-N1-133 or GST-N1-152, it copurified with the GST fusion protein (Fig. 3B, lanes 2 and 3), whereas it did not when it was coexpressed with GST-N133-246 or GST-N71-246 (Fig. 3B, lanes 4 and 5). These results clearly indicated that the N-terminal first 71 amino acids contain residues crucial for the N-N interaction. Of note, since some mutations reduced significantly the level of expression of the native form (not shown) but not that of the GST-NRVF mutants (Fig. 4B), the mutational analysis was carried out with the sequence of the GST-NRVF protein but not with the N native form.
Tyrosine 4 and phenylalanine 11 are crucial for the N-N interaction. Comparison of the nucleoprotein sequences of several phleboviruses, i.e., RVFV, TOSV, Punta Toro virus, and Sicilian fever sandfly virus, using the Clustal W algorithm indicated the presence of several conserved sequences (Fig. 4A) within the 71 first N-terminal amino acids. Four residues are conserved and aligned, i.e., Tyr 4, Phe 11, Asp 17, and Trp 24 (numbers refer to the RVFV sequence) (Fig. 4A), as well as two motifs (residues 28 to 34 and 46 to 68). The latter two are also conserved in the most divergent Uukuniemi virus sequence. Mutational analysis was performed to investigate the putative role of the conserved amino acids in the N-N interaction. A series of Ala substitutions were made within the two conserved motifs of GST-NRVF; neither the single mutations F28A, Y30A, D34A, R64A, N66A, K67A, and P68A nor the double mutations F28A Y30A, Q31A D34A, R64A N66A, and K67A P68A disrupted the interaction (not shown). Regarding the conserved amino acids at positions 4, 11, 17, and 24, each of these amino acid residues was mutated into a Gly residue in the N sequence of pTM1-GST-NRVF, generating plasmids pTM1-GST-NY4G, -GST-NF11G, -GST-ND17G, and -GST-NW24G, and the ability of the mutants to interact with NRVF was tested using the GST pull-down assay. When GST-ND17G or GST-NW24G was coexpressed with the wild-type NRVF protein, NRVF copurified with GST-ND17G or GST-NW24G (Fig. 4B, lanes 5 and 6), but coexpression of NRVF and GST-NY4G or GST-NF11G drastically impaired the complex formation (Fig. 4B, lanes 3 and 4), indicating that substitutions of Tyr 4 and Phe 11 impeded the interaction with nucleoprotein.
The crucial role of phenylalanine 11 in N-N interaction is conserved in TOSV. Because of the amino acid conservation between phlebovirus nucleoproteins, we investigated whether assembly of N into dimers is a general property among phleboviruses and tested the ability of the nucleoprotein from another phlebovirus, Toscana virus, to interact with itself. Thus, pTM1-NTos and pTM1-GST-NTos plasmids were generated. After coexpression of the proteins in HeLa cells infected with vT7-3 and analysis by GST pull-down assay, NTOS was found to interact with the GST-NTOS protein, indicating that NTOS interacts with itself (Fig. 5, lane 2).
When coexpressed with GST-NRVF, NTOS copurified with the GST fusion protein (Fig. 5, lane 3), showing that RVFV and TOSV nucleoproteins can form heterodimers. To estimate the role of the conserved residues in the heterotypic interaction, we assayed the interaction between NTOS and the substitution mutants in GST-NRVF. NTOS copurified with GST-NY4G and GST-ND17G (Fig. 5, lanes 4 and 6), but the N-N interaction was abolished when NTOS was coexpressed with GST-NF11G or GST-NW24G (Fig. 5, lanes 5 and 7). These results show that in both homotypic and heterotypic interactions, Phe 11 in RVFV nucleoprotein plays an important role. In contrast with the homotypic interaction, Tyr 4 in NRVF is not essential for the heterotypic interaction with NTOS, but instead Trp 24, another conserved residue, becomes crucial in the interaction with NTOS.
DISCUSSION
In this paper, we demonstrate that the RVFV N protein interacts with itself and forms homodimers. These structures were observed in purified RNPs treated with glutaraldehyde and SAB (Fig. 1). The interaction was not affected after treatments that disrupt the association between viral RNA and N proteins. Similar dimeric forms were detected with cross-linked recombinant nucleoprotein (data not shown). Although we cannot ruled out the possibility that the nucleoprotein is also associated with cellular proteins contaminating the RNPs, these complexes must be in minor amounts, since only trace amounts of contaminating proteins were detected in the purified RNP fraction after Coomassie blue staining (data not shown). The N-N interaction was analyzed using the yeast two-hybrid system and GST pull-down method from proteins expressed in HeLa cells (Fig. 2B and 3B). It must be noted that the two-hybrid system could not be utilized extensively to delineate the interaction domain, because some of the deletion mutants (e.g., AD-N71-246) transactivated expression of the reporter gene. Using the GST pull-down assay, we tested the N-N interaction with the wild-type protein and the wild-type or mutated NRVF fused to GST. The mutations were introduced into the GST fusion protein and in the native NRVF, but the latter ones were not used because N-terminal truncations led to drastically reduced expression levels, probably due to the encephalomyocarditis virus internal ribosome entry site present in pTM1 and the sequence downstream of the initiating ATG (16). Using this system, we demonstrated that the C-terminal region did not seem to be essential for the N-N interaction but that the N terminus, and Tyr 4 and Phe 11 in particular, plays an important role. (Fig. 3 and 4). Except for Uukuniemi virus N protein, these two amino acid residues are conserved within phlebovirus N proteins (Fig. 4A). In fact, it could well be that homodimer formation is a property of the phlebotomine and mosquito-borne phlebovirus nucleoproteins, as illustrated by the data showing that the N protein of TOSV is able to interact with itself and with NRVF (Fig. 5). However, it seems that this does not apply to N protein of the tick-borne Uukuniemi virus, in which both N and C termini are necessary for dimerization (A. Plyusnin, personal communication). For the formation of RVFV homodimers and heterodimers with TOSV, Phe 11 in the RVFV sequence appeared to be crucial for both interactions. A second amino acid, Tyr 4 or Trp 24, is also important in, respectively, homo- or heterodimer formation. The N-terminal region was predicted to contain alpha-helices. Thus, being separated by five or six residues, the conserved amino acid residues at the N terminus (Tyr 4, Phe 11, Asp 17, or Trp 24) are likely localized on the same face of the alpha-helix. Thus, except for Asp 17, for which no role was found, changes at one of these positions might disrupt the folding of the N-terminal region and impair the interaction. At the present time, the three-dimensional structure of bunyavirus nucleoproteins is still unknown. This information would be of value to better understand dimer formation.
To our knowledge, no data are available for the N protein of nairoviruses, but, interestingly, the N proteins of tospoviruses, hantaviruses, and probably orthobunyaviruses (31) have been shown to assemble into oligomers. However, the N-N interactions appear to differ from those in the present situation. In the case of the tomato spotted wilt tospovirus, the nature of the oligomers in the RNPs is not known, but using two-hybrid and plasmon resonance analyses, Uhrig et al. showed that both the N and C termini are crucial for the N-N interaction, which implies a head-to-tail interaction forming a multimeric chain (37). In the case of hantaviruses, the N protein present in viral RNPs or expressed as a recombinant protein in Escherichia coli assembles into dimers and trimers, the latter being the major form (1, 18). Our results suggest that the structure of RVFV (and probably phlebovirus) nucleoprotein oligomers is different from that of hantavirus nucleoprotein oligomers. These differences are evidenced by several observations: (i) among the different bunyavirus genera, nucleoproteins are poorly conserved in size and sequence, with the hantavirus N being more complex, with a size twice that of the phlebovirus N; (ii) gel analysis of cross-linked N oligomers clearly showed that trimers are the major forms in hantavirus RNPs (1, 18), whereas N dimers are the main oligomers in RVFV RNPs (this paper); and (iii) the domains important for the hantavirus N trimer assembly are conserved only among hantaviruses, and not phleboviruses or other bunyaviruses, and vice versa. For hantaviruses, the N-terminal residues assemble into trimeric coiled coils. A prediction for such a motif in the N terminus of RVFV N protein, using the algorithm of Lupas et al. (24), has a rather low score of 0.2. Notwithstanding, we explored this possibility and changed Ala-Ala at positions 12 and 13 to Gly-Gly to disrupt this putative domain. However, when tested in the GST pull-down assay, the double mutant was still able to interact with N (data not shown), indicating that if a coiled-coil motif exists in the N terminus, it does not contribute significantly to the interaction. Together these results point out a major difference in the structures of RVFV and hantavirus nucleoproteins, leading, respectively, to dimers and trimers. In RVFV nucleoprotein, only the N-terminal region is involved in N-N interaction whereas in the hantavirus protein, the N and C termini both contribute to the interaction, with the primary domain being located at amino acids 393 to 398 in Tula virus N and the domain from amino acid 1 to 43 representing a secondary interaction (19, 20). The mono- and dimeric forms may exist transiently, while the trimeric form is highly stable (27).
Finally, the question which remains is whether oligomers have specific functions in the viral cycle. The N proteins of bunyaviruses must have at least two intrinsic functions, one for binding to the viral RNA and the other for the formation of a complex active in transcription and replication, since only encapsidated RNA can serve as template for transcription and replication. Investigations of the N proteins of viruses in the Orthobunyavirus, Hantavirus, and Tospovirus genera have shown that N binds to RNA essentially in a nonspecific manner, but it seems to preferentially bind double-stranded RNA or sequence present at the 5' end of the genomic RNA (31, 39). Recently, Mir and Panganiban (26, 27) reported that the mono- and dimeric forms of Sin Nombre hantavirus N bind RNA, forming a complex that is semispecific and salt sensitive. In contrast, purified N trimers are able to discriminate between viral and nonviral RNA molecules and recognize and bind with high affinity the panhandle structure composed of the 5' and 3' ends of the viral RNA. The discrimination of N trimers for RNA binding between viral and nonviral RNA molecules could be very important during the early steps of encapsidation by acting as assembly intermediates during RNP formation, and therefore oligomerization may be involved in replication. In the case of RVFV, further work will be required to determine if dimers play a similar role in RNP formation.
ACKNOWLEDGMENTS
N.L.M. is a recipient of fellowships from Ministère de la Recherche, Association de la Recherche Contre le Cancer (ARC), and Fondation de France. N.G. is a recipient of a "Contrat Jeune Chercheur" from DGA and a fellowship from Académie de Médecine. This work was supported in part by grant MIC0310 from INSERM.
REFERENCES
Alfadhli, A., Z. Love, B. Arvidson, J. Seeds, J. Willey, and E. Barklis. 2001. Hantavirus nucleocapsid protein oligomerization. J. Virol. 75:2019-2023.
Anonymous. 2000. Outbreak of Rift Valley fever, Yemen, August-October 2000. Wkly. Epidemiol. Rec. 75:392-395.
Anonymous. 2000. Rift Valley fever, Saudi Arabia, August-October 2000. Wkly. Epidemiol. Rec. 75:370-371.
Becker, S., C. Rinne, U. Hofsass, H. D. Klenk, and E. Muhlberger. 1998. Interactions of Marburg virus nucleocapsid proteins. Virology 249:406-417.
Bouloy, M., N. Pardigon, P. Vialat, S. Gerbaud, and M. Girard. 1990. Characterization of the 5' and 3' ends of viral messenger RNAs isolated from BHK21 cells infected with Germiston virus (Bunyavirus). Virology 175:50-58.
Bouloy, M., S. J. Plotch, and R. M. Krug. 1978. Globin mRNAs are primers for the transcription of influenza viral RNA in vitro. Proc. Natl. Acad. Sci. USA 75:4886-4890.
Campbell, K. S., A. Buder, and U. Deuschle. 1995. Interaction of p56lck with CD4 in the yeast two-hybrid system. Ann. N. Y. Acad. Sci. 766:89-92.
Campbell, K. S., A. Buder, and U. Deuschle. 1995. Interactions between the amino-terminal domain of p56lck and cytoplasmic domains of CD4 and CD8 alpha in yeast. Eur. J. Immunol. 25:2408-2412.
Caplen, H., C. J. Peters, and D. H. Bishop. 1985. Mutagen-directed attenuation of Rift Valley fever virus as a method for vaccine development. J. Gen. Virol. 66:2271-2277.
Fuerst, T. R., E. G. Niles, F. W. Studier, and B. Moss. 1986. Eukaryotic transient-expression system based on recombinant vaccinia virus that synthesizes bacteriophage T7 RNA polymerase. Proc. Natl. Acad. Sci. USA 83:8122-8126.
Garcin, D., M. Lezzi, M. Dobbs, R. M. Elliott, C. Schmaljohn, C. Y. Kang, and D. Kolakofsky. 1995. The 5' ends of Hantaan virus (Bunyaviridae) RNAs suggest a prime-and-realign mechanism for the initiation of RNA synthesis. J. Virol. 69:5754-5762.
Gietz, D., A. St Jean, R. A. Woods, and R. H. Schiestl. 1992. Improved method for high efficiency transformation of intact yeast cells. Nucleic Acids Res. 20:1425.
Giorgi, C., L. Accardi, L. Nicoletti, M. C. Gro, K. Takehara, C. Hilditch, S. Morikawa, and D. H. Bishop. 1991. Sequences and coding strategies of the S RNAs of Toscana and Rift Valley fever viruses compared to those of Punta Toro, Sicilian sandfly fever, and Uukuniemi viruses. Virology 180:738-753.
Harper, J. W., G. R. Adami, N. Wei, K. Keyomarsi, and S. J. Elledge. 1993. The p21 Cdk-interacting protein Cip1 is a potent inhibitor of G1 cyclin-dependent kinases. Cell 75:805-816.
Hewlett, M. J., R. F. Pettersson, and D. Baltimore. 1977. Circular forms of Uukuniemi virion RNA: an electron microscopic study. J. Virol. 21:1085-1093.
Hunt, S. L., A. Kaminski, and R. J. Jackson. 1993. The influence of viral coding sequences on the efficiency of internal initiation of translation of cardiovirus RNAs. Virology 197:801-807.
Jin, H., and R. M. Elliott. 1993. Characterization of Bunyamwera virus S RNA that is transcribed and replicated by the L protein expressed from recombinant vaccinia virus. J. Virol. 67:1396-1404.
Kaukinen, P., V. Koistinen, O. Vapalahti, A. Vaheri, and A. Plyusnin. 2001. Interaction between molecules of hantavirus nucleocapsid protein. J. Gen. Virol. 82:1845-1853.
Kaukinen, P., V. Kumar, K. Tulimaki, P. Engelhardt, A. Vaheri, and A. Plyusnin. 2004. Oligomerization of Hantavirus N protein: C-terminal alpha-helices interact to form a shared hydrophobic space. J. Virol. 78:13669-13677.
Kaukinen, P., A. Vaheri, and A. Plyusnin. 2003. Mapping of the regions involved in homotypic interactions of Tula hantavirus N protein. J. Virol. 77:10910-10916.
Krug, R. M., M. Bouloy, and S. J. Plotch. 1980. RNA primers and the role of host nuclear RNA polymerase II in influenza viral RNA transcription. Philos. Trans. R. Soc. London B 288:359-370.
Linthicum, K. J., A. Anyamba, C. J. Tucker, P. W. Kelley, M. F. Myers, and C. J. Peters. 1999. Climate and satellite indicators to forecast Rift Valley fever epidemics in Kenya. Science 285:397-400.
Luban, J., K. B. Alin, K. L. Bossolt, T. Humaran, and S. P. Goff. 1992. Genetic assay for multimerization of retroviral Gag polyproteins. J. Virol. 66:5157-5160.
Lupas, A., M. Van Dyke, and J. Stock. 1991. Predicting coiled coils from protein sequences. Science 252:1162-1164.
Matsumoto, M., S. B. Hwang, K. S. Jeng, N. Zhu, and M. M. Lai. 1996. Homotypic interaction and multimerization of hepatitis C virus core protein. Virology 218:43-51.
Mir, M. A., and A. T. Panganiban. 2005. The hantavirus nucleocapsid protein recognizes specific features of the viral RNA panhandle and is altered in conformation upon RNA binding. J. Virol. 79:1824-1835.
Mir, M. A., and A. T. Panganiban. 2004. Trimeric hantavirus nucleocapsid protein binds specifically to the viral RNA panhandle. J. Virol. 78:8281-8288.
Morrill, J. C., and D. J. McClain. 1996. Epidemiology and pathogenesis of the Rift Valley fever and other phleboviruses, p. 281-293. In H. Fraenkel-Conrat and R. R. Wagner (ed.), The viruses. Plenum Press, New York, N.Y.
Myers, T. M., A. Pieters, and S. A. Moyer. 1997. A highly conserved region of the Sendai virus nucleocapsid protein contributes to the NP-NP binding domain. Virology 229:322-335.
Nolandt, O., V. Kern, H. Muller, E. Pfaff, L. Theilmann, R. Welker, and H. G. Krausslich. 1997. Analysis of hepatitis C virus core protein interaction domains. J Gen. Virol. 78:1331-1340.
Osborne, J. C., and R. M. Elliott. 2000. RNA binding properties of bunyamwera virus nucleocapsid protein and selective binding to an element in the 5' terminus of the negative-sense S segment. J. Virol. 74:9946-9952.
Patterson, J. L., and D. Kolakofsky. 1984. Characterization of La Crosse virus small-genome transcripts. J. Virol. 49:680-685.
Pettersson, R. F., and C. H. von Bonsdorff. 1975. Ribonucleoproteins of Uukuniemi virus are circular. J. Virol. 15:386-392.
Samso, A., M. Bouloy, and C. Hannoun. 1975. Circular ribonucleoproteins in the virus Lumbo (Bunyavirus). C. R. Acad. Sci. D 280:779-782.
Samso, A., M. Bouloy, and C. Hannoun. 1976. Demonstration of circular ribonucleic acid in the Lumbo virus (Bunyavirus). C. R. Acad. Sci. D 282:1653-1655.
Schmaljohn, C., and J. W. Hooper. 2001. Bunyaviridae: the viruses and their replication, 4th ed. Lippincott Williams & Wilkins, Philadelphia, Pa.
Uhrig, J. F., T. R. Soellick, C. J. Minke, C. Philipp, J. W. Kellmann, and P. H. Schreier. 1999. Homotypic interaction and multimerization of nucleocapsid protein of tomato spotted wilt tospovirus: identification and characterization of two interacting domains. Proc. Natl. Acad. Sci. USA 96:55-60.
Vialat, P., and M. Bouloy. 1992. Germiston virus transcriptase requires active 40S ribosomal subunits and utilizes capped cellular RNAs. J. Virol. 66:685-693.
Xu, X., W. Severson, N. Villegas, C. S. Schmaljohn, and C. B. Jonsson. 2002. The RNA binding domain of the hantaan virus N protein maps to a central, conserved region. J. Virol. 76:3301-3308.(Nicolas Le May, Nicolas G)