Role of the Sonchus Yellow Net Virus N Protein in Formation of Nuclear Viroplasms

Abstract

Sonchus yellow net virus is a plant nucleorhabdovirus whose nucleocapsid (N), phosphoprotein (P), and polymerase (L) proteins form large viroplasms in the nuclei of infected plants (C. R. F. Martins, J. A. Johnson, D. M. Lawrence, T. J. Choi, A. Pisi, S. L. Tobin, D. Lapidus, J. D. O. Wagner, S. Ruzin, K. McDonald, and A. O. Jackson, J. Virol. 72:5669-5679, 1998). When expressed alone, the N protein localizes to the nuclei of plant and yeast (Saccharomyces cerevisiae) cells and the P protein is distributed throughout the cells, but coexpression of N and P results in formation of subnuclear viroplasm-like foci (M. M. Goodin, J. Austin, R. Tobias, M. Fujita, C. Morales, and A. O. Jackson, J. Virol. 75:9393-9406, 2001; M. M. Goodin, R. G. Dietzgen, D. Schichnes, S. Ruzin, and A. O. Jackson, Plant J. 31:375-383, 2002). We now show that the N protein and various fluorescent derivatives form similar subnuclear foci in plant cells and that homologous interactions mediated by a helix-loop-helix region near the amino terminus are required for formation of the foci. Mutations within the helix-loop-helix region also interfere with N- and P-protein interactions that are required for N and P colocalization in the subnuclear foci. Affinity purification of N proteins harboring single mutations within the motif revealed that Tyr40 is critical for N-N and N-P interactions. Additional in vitro binding assays also indicated that the N protein binds to yeast and plant importin α homologues, whereas mutations in the carboxy-terminal nuclear localization signal abrogate importin α binding. The P protein did not bind to the importin α homologues, suggesting that the N and P proteins use different pathways for nuclear entry. Our results in toto support a model suggesting that during infection, the N and P proteins enter the nucleus independently, that viroplasm formation requires homologous N-protein interactions, and that P protein targeting to the viroplasm requires N-P protein interactions that occur after N and P protein import into the nucleus.

Plant rhabdoviruses are classified into the genera Cytorhabdovirus and Nucleorhabdovirus on the basis of their sites of replication, morphogenesis, and maturation (for a review, see reference 17). Sonchus yellow net nucleorhabdovirus (SYNV) replicates in the nucleus and is the most extensively characterized among the plant rhabdoviruses. SYNV encodes six genes in a negative-sense orientation; these six genes encode a nucleocapsid protein (N), a phosphoprotein (P), a putative movement protein (sc4), a matrix protein (M), a glycoprotein (G), and a large polymerase protein (L). The N, P, and L proteins are components of an infectious nucleocapsid core (15) with RNA-dependent RNA polymerase activity that can be purified from the nuclei of virus-infected cells (34, 35). These core components form viroplasm-like structures within the nucleus that are thought to be the sites of viral replication (22). The N protein contains a carboxy (C)-terminal bipartite nuclear localization signal (NLS) that is required for nuclear import (7). The P protein when expressed alone localizes both inside and outside of the nucleus, but coexpression of the N and P proteins in plant and yeast (Saccharomyces cerevisiae) cells results in formation of compact subnuclear foci that are reminiscent of viroplasms (7). Sedimentation and immunological analyses have shown that in vivo associations of the N, P, and L proteins are required for RNA-dependent RNA polymerase activity (35). Yeast two-hybrid analyses and affinity chromatography experiments have also verified homologous (N-N) and heterologous (N-P) binding that is mediated by a region near the amino (N) terminus of the N protein (7). These results indicate that complex interactions of the N, P, and L proteins are required for subnuclear viroplasm formation and that the nucleocapsid cores within the viroplasms function in replication of genomic and antigenomic RNAs and in mRNA transcription (17).

In the current study, we have conducted experiments to define the contributions of the N protein to the formation of viroplasms. These experiments include refined mapping to identify amino acids in an N-terminal helix-loop-helix motif of the N protein that result in subnuclear localization and N-N and N-P protein binding. We have also shown that mutations within the helix-loop-helix motif that disrupt N-N and N-P interactions interfere with the formation of subnuclear foci. The C-terminal NLS that is required for nuclear import of the N protein (7) mediates binding to importin α homologues from yeast (ySrp1) and plant Arabidopsis thaliana (AtSrp1), but the P protein fails to bind to the importin α homologues. These results suggest that the N and P proteins differ in nuclear import pathways and provide a model for N-N and N-P associations required for formation of subnuclear viroplasms.

MATERIALS AND METHODS

General.

Healthy Nicotiana benthamiana plants were maintained as described by Jackson and Wagner (16) or in a growth room at 23°C under 1,000 lumens with a 16-hour daylight regimen. Agrobacterium tumefaciens strain EHA105 was grown on LB agar containing 50 μg/ml of rifampin as described by Guo and Ding (11). The Saccharomyces cerevisiae strain PJ69-4a used for yeast two-hybrid analyses was maintained in yeast extract-peptone-dextrose (or glucose) medium as described by James et al. (18). SYNV was propagated by serial mechanical passages in N. benthamiana under ambient greenhouse conditions (14, 16).

All plasmids were maintained in Escherichia coli strain DH5α or TOP10. Plasmid DNA was extracted using an alkaline lysis procedure (28), and DNA used for sequencing was given an additional polyethylene glycol precipitation step (26). DNA for subcloning was recovered from gel slices using a QIAEX II gel extraction kit (QIAGEN, Valencia, CA). Restriction enzymes were obtained from New England Biolabs (Beverly, MA), and chemicals were purchased from Sigma Chemical (St. Louis, MO), or Fisher (Springfield, N.J.). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed on 10% gels using the Mini-Protean II system according to the manufacturer's instructions (Bio-Rad, Richmond, CA).

Cloning of N- and P-protein derivatives.

All oligonucleotides used for PCRs were purchased from Operon Technologies (Alameda, CA). Primer names indicate the specific SYNV nucleocapsid (N) or phosphoprotein (P) genes followed by numerals corresponding to the N or P sequences terminating the primers (Table 1). The f or r designation in the primer names denotes whether the primer is a forward (5′) or reverse (3′) primer, respectively. PCR amplifications were performed using Pfu DNA polymerase (Stratagene, La Jolla, CA) as described by the manufacturer. PCR amplification was set to have an initial 3-min denaturation at 94°C, followed by 30 cycles of 30 seconds at 93°C, 30 seconds at 58°C, and 1 to 2 min at 68°C, with a final extension at 68°C for 8 min. After the final extension step, one unit of Taq polymerase (Stratagene, La Jolla, CA) was added, and the mixture was maintained at 72°C for another 20 min to add an extra A residue at the 3′ termini. Deletions and site-specific mutations were introduced into the N-protein open reading frame (ORF) through overlapping mutagenesis (28), using the primers shown in Table 1. The PCR products were purified by gel electrophoresis and cloned using the TOPO TA cloning kit (Invitrogen, Carlsbad, CA). Clones derived from the PCR products were confirmed by sequencing at the University of California-Berkeley DNA Sequencing Facility (Berkeley, CA).

TABLE 1.

Synthetic oligonucleotide primers used for mutagenesis

Primera	Sequenceb	Description
N1f-Hind	CTCAAGCTTCGATGAGCACTACACCAACAATCACT	Add HindIII site at 5′ of N ATG
N475r-Bam	GAGGGATCCTTAAAAGTCCGGTATGTTTGGTAG	Add BamHI site at 3′ of N
NmutYYLF-f	AACCCATCAGGCCAATGCACCGCTCGTGAGGCTGCAGCCAGTGATGCAGTCAAGTATCCT	Mutate Y37A, Y40A, L41A, and F42A
NmutYYLF-r	AGGATACTTGACTGCATCACTGGCTGCAGCCTCACGAGCGGTGCATTGGCCTGATGGGTT	Reverse for NmutYYLF-f
NmutYIY-f	CTGTTCAGTGATGCAGTCAAGGCTCCTGCAGCCAAACGTGCCACAATGACAAATGAGGAG	Mutate Y48A, I50A, and Y51A
NmutYIY-r	CTCCTCATTTGTCATTGTGGCACGTTTGGCTGCAGGAGCCTTGACTGCATCACTGAACAG	Reverse for NmutYIY-f
NmutIFFI-f	GCCACAATGACAAATGAGGAGGCTGTTACCGCCGCTGGGAAGGCCACCTCCGACAAGCACACTCAT	Mutate I61A, F64A, F65A, and I68A
NmutIFFI-r	ATGAGTGTGCTTGTCGGAGGTGGCCTTCCCAGCGGCGGTAACAGCCTCCTCATTTGTCATTGTGGC	Reverse for NmutIFFI-f
N30(Y37A)-f	AACCCATCAGGCCAATGCACCGCCCGTGAGTATCTGTTCAGTGAT	Mutate Y37A
N30(Y37A)-r	ATCACTGAACAGATACTCACGGGCGGTGCATTGGCCTGATGGGTT	Reverse for N30(Y37A)-f
N33(Y40A)-f	GGCCAATGCACCTACCGTGAGGCTCTGTTCAGTGATGCAGTCAAG	Mutate Y40A
N33(Y40A)-r	CTTGACTGCATCACTGAACAGAGCCTCACGGTAGGTGCATTGGCC	Reverse for N33(Y40A)-f
N34(L41A)-f	CAATGCACCTACCGTGAGTATGCGTTCAGTGATGCAGTCAAGTAT	Mutate L41A
N34(L41A)-r	ATACTTGACTGCATCACTGAACGCATACTCACGGTAGGTGCATTG	Reverse for N34(L41A)-f
N35(F42A)-f	TGCACCTACCGTGAGTATCTGGCAAGTGATGCAGTCAAGTATCCT	Mutate F42A
N35(F42A)-r	AGGATACTTGACTGCATCACTTGCCAGATACTCACGGTAGGTGCA	Reverse for N35(F42A)-f
N41(Y48A)-f	CTGTTCAGTGATGCAGTCAAGGCTCCTATATACAAACGTGCCACA	Mutate Y48A
N41(Y48A)-r	TGTGGCACGTTTGTATATAGGAGCCTTGACTGCATCACTGAACAG	Reverse for N41(Y48A)-f
N43(I50A)-f	AGTGATGCAGTCAAGTATCCTGCATACAAACGTGCCACAATGACA	Mutate I50A
N43(I50A)-r	TGTCATTGTGGCACGTTTGTATGCAGGATACTTGACTGCATCACT	Reverse for N43(I50A)-f
N44(Y51A)-f	GATGCAGTCAAGTATCCTATAGCCAAACGTGCCACAATGACAAAT	Mutate Y51A
N44(Y51A)-r	ATTTGTCATTGTGGCACGTTTGGCTATAGGATACTTGACTGCATC	Reverse for N44(Y51A)-f
N54(I61A)-f	GCCACAATGACAAATGAGGAGGCTGTTACCTTCTTTGGGAAGATC	Mutate I61A
N54(I61A)-r	GATCTTCCCAAAGAAGGTAACAGCCTCCTCATTTGTCATTGTGGC	Reverse for N54(I61A)-f
N57(F64A)-f	ACAAATGAGGAGATTGTTACCGCCTTTGGGAAGATCACCTCCGAC	Mutate F64A
N57(F64A)-r	GTCGGAGGTGATCTTCCCAAAGGCGGTAACAATCTCCTCATTTGT	Reverse for N57(F64A)-f
N58(F65A)-f	AATGAGGAGATTGTTACCTTCGCTGGGAAGATCACCTCCGACAAG	Mutate F65A
N58(F65A)-r	CTTGTCGGAGGTGATCTTCCCAGCGAAGGTAACAATCTCCTCATT	Reverse for N58(F65A)-f
N61(I68A)-f	ATTGTTACCTTCTTTGGGAAGGCCACCTCCGACAAGCACACTCAT	Mutate I68A
N61(I68A)-r	ATGAGTGTGCTTGTCGGAGGTGGCCTTCCCAAAGAAGGTAACAAT	Reverse for N61(I68A)-f
P1f-apgp-Sal	GTCGACGCACCAGGACCAATGGAAATCGATCCAAATTACGTTAAC	Add SalI site at the 5′ of P
P308r-Bam	GGATCCTCACGCCTTCTTTGGGTCAAT	Add BamHI site at the 3′ of P
5eGFP-Bam	GGATCCATGGTGAGCAAGGGCGAG	Add BamHI site at the 3′ of eG
3eGFP-Bgl	AGATCTCTTGTACAGCTCGTCCATGCC	Add BglII site at the 3′ of eG
5GST-Bam	GGATCCATGTCCCCTATACTAGGTTAT	Add BamHI site at the 5′ of GST
5GST-Bgl	AGATCTTTTTGGAGGATGGTCGCCACC	Add BglII site at the 5′ of GST
3GST-TEVpro-Bgl	CAAGATCTTGGTCCCTGGAAGTACAGATTCTCTTTTGGAGGATGGTCGCCACCACCAAACGT	Add TEV NIa protease cleavage site and BglII site at the 3′ of GST

^aPrimer names indicate the specific SYNV nucleocapsid (N) or phosphoprotein (P) genes followed by numerals corresponding to the N or P sequences terminating the primers. The f or r designation in the primer names denotes whether the primer is a forward (5′) or reverse (3′) primer, respectively.

^bRestriction sites embedded in the primers are noted in bold type.

Sequence analyses during the initial cloning of the SYNV N protein indicated that the mRNA contains a 1,425-nucleotide (nt) ORF encoding a 475-amino-acid (aa) polypeptide (37). The sequences of eight full-length cDNA clones isolated more recently have revealed three silent consensus variants (a C-to-U substitution at nt 264 of the ORF, a U-to-C substitution at nt 402, and a C-to-A substitution at nt 960) that differ from the published sequence. In addition, all eight clones contained a G-to-A substitution at position 409 in the N-protein ORF that introduces a V137I substitution in the N protein, and nucleotide variants at positions 428 to 431 (AGUA to GUAC) resulted in K143S and Y144T changes. Seven additional single-nucleotide substitutions were also found in individual clones; these substitutions appear to be sporadic mutations (Table 2). We have selected an N-protein cDNA clone (N_WT) that contains the consensus polymorphisms and have used this derivative in all experiments described in this study. The sequence of this clone has been submitted to the National Center for Biotechnology Information (GenBank accession no. NC_001615).

TABLE 2.

Mutations in the cDNA clones of the SYNV N protein

Mutation	Position(s)	Nucleotide change	Amino acid change(s)
Consensus	264	C to U	None
	402	U to C	None
	409	G to A	V137I
	428-431	AGUA to GUAC	K143S and Y144T
	960	C to A	None
Sporadic	348	A to G	None
	356	A to G	Q119R
	535	A to G	None
	1033	A to G	None
	1045	U to G	None
	1195	G to U	None
	1423	C to U	None

Construction of protein expression cassettes.

To ensure that fluorescent protein fusions were of sufficient size to prevent diffusion into plant nuclei (9), we developed expression vectors containing fusions to a single enhanced green fluorescent protein (eGFP), two eGFP proteins, and a larger derivative consisting of two eGFP proteins fused to glutathione S-transferase (GST). For these fusions, the eGFP sequence was amplified from pEGFP-N1 (Clontech, Palo Alto, CA) using the primers 5eGFP-Bam and 3eGFP-Bgl (Table 1) and cloned into the TOPO vector. The resulting BamHI-BglII fragment was then inserted into the BglII site downstream of the 35S promoter in the pGD vector (8) to generate pGDeG. Another eGFP BamHI-BglII fragment was cloned into BglII-treated pGDeG to construct the plasmid pGDeGII which expresses the eG-eG fusion protein. An additional BamHI-BglII GST fragment was cloned into BglII-treated pGDeGII to produce a plasmid (pGDeGIIGST) which expresses the eG-eG-GST fusion protein. A monomeric red fluorescent protein (mRFP) fusion protein expression vector (pGDmR) was constructed by amplifying the region encoding monomeric Discosoma sp. fluorescent protein DsRed (29) from pRSET-B mRFP (a gift of Roger Tsien), and the fragment was cloned into NcoI-BglII-digested pGDR (8). The red-shifted GFP (RSGFP) vector pGDG (8) and yellow fluorescent protein (YFP) and cyan fluorescent protein (CFP) expression vectors (32) were used as previously described. For affinity chromatography, a GST fusion derivative containing the tobacco etch virus (TEV) NIa protease cleavage site was engineered into the C terminus of the GST protein by amplifying the GST fragment from pGEX-2T (Amersham, Piscataway, NJ) using the primers 5GST-Bam and 3GST-TEVpro-Bgl (Table 1), and the fragment was cloned into the pGD vector to generate pGDGSTpro. The five resulting vectors (pGDeG, pGDeGII, pGDeGIIGST, pGDmR, and pGDGSTpro) have the same cloning sites as the pGD serial vectors (8, 32). The full-length N gene and various mutants were cloned into the pGD expression vectors in frame with the respective fluorescent proteins at the HindIII and BamHI sites. The full-length P gene was cloned in frame into these vectors at the SalI and BamHI sites.

Transient expression in planta.

The pGD plasmids were transformed into A. tumefaciens strain EHA105 by a freeze-thaw method (1) and selected on LB agar plates containing rifampin (50 μg/ml) and kanamycin (100 μg/ml). To prepare for agroinfiltration, 1- to 2-day-old EHA105 cells were scraped from LB agar plates and resuspended in morpholineethanesulfonic acid (MES) buffer containing 10 mM MgCl₂ and 10 mM MES, pH 5.6, without acetylsyringone (Shou-Wei Ding, personal communication). After 2 to 4 h of induction at room temperature, the cell suspensions were adjusted to an optical density at 600 nm of 0.8 and infiltrated into leaves of 2-month-old N. benthamiana plants (8). For coexpression of different fusions, equal volumes of culture suspensions from different transformants were mixed, and infiltrated plants were maintained under a 16-h light regimen at room temperature for 2 days.

Laser-scanning confocal microscopy.

Confocal microscopy was performed at 36 to 48 h postinfiltration. To stain the nuclei of the leaf epidermal cells, 0.1 μg/ml of 4′-6-diamidino-2-phenylindole dihydrochloride (DAPI) was pressure infiltrated into leaves with a syringe. The water-soaked area was immediately sliced into 2-mm by 4-mm pieces and submerged in the DAPI solution for 10 to 20 min prior to mounting onto slides.

Microscopy was performed using a Zeiss LSM 510 META confocal laser-scanning microscope using the multitrack scan configuration. GFP was excited at 488 nm with an argon laser, and emission light was captured through a band-pass emission filter (BP505-530). DAPI-stained nuclei were excited by a 364-nm laser, and the emission was collected through a long pass 385-nm emission filter. DsRed was excited at 543 nm using a 543-nm helium neon laser, and the emission was selected by the META detector set for 570 to 600 nm. Leaf tissue was examined with a Plan-Neofluar objective with the pinhole set to yield a 1.5-μm optical slice. Images were captured using Zeiss LSM 510 META software, converted to TIFF files for export, and processed in Adobe Photoshop 7 (Adobe Systems, Inc., San Jose, CA).

Immunofluorescence assays.

Immunostaining procedures were adapted from Martins et al. (22), with minor modifications to accommodate staining of leaf tissue instead of protoplasts. At 2 days postinfiltration, uninfected N. benthamiana leaves or leaves expressing viral proteins were detached and fixed at room temperature by pressure infiltration of 3.7% formaldehyde and 10% mannitol, followed by soaking in the solution for 30 min. The fixed leaf tissue was then cut into 2-mm² pieces and permeabilized by soaking in phosphate-buffered saline (PBS) (150 mM NaCl, 15 mM NaHPO₄, pH 7.2) buffer containing 3% bovine serum albumin (BSA) and 0.05% Tween 20 for 5 min. After two additional washing and soaking steps in PBS containing 3% BSA (buffer A), the leaf tissue was incubated overnight at 4°C in N-protein antisera (37) diluted in buffer A. The tissue was then given five additional 30-min washes in buffer A, mounted on slides, and incubated for 1 h in buffer A containing 0.1 μg/ml of DAPI and Alexa Fluor 488-conjugated anti-rabbit secondary antibody (Invitrogen, Carlsbad, CA). The slides were finally washed three times in buffer A and examined by confocal microscopy.

Yeast two-hybrid assays.

The yeast two-hybrid vectors described by James et al. (18) were designed to produce fusions of either the Gal4 activation domain (AD) or the Gal4 binding domain (BD) to the N terminus of the proteins to be tested for interactions. Fusions of the wild-type N (N_WT) and P sequences to the AD and BD were described previously (8). PCR primers (Table 1) were designed to introduce either a 5′ EcoRI site and a 3′ BamHI site or a 5′ SalI site and a 3′ BamHI site to facilitate ligation into the EcoRI and BamHI or SalI and BamHI sites of the pGAD and pGBDU vectors, respectively.

Two-hybrid interaction experiments were performed using the yeast strain PJ69-4A (18). PJ69-4A constructs containing LEU2-selected activation domains (AD) and URA3-selected binding domains (BD) were transformed by the method of Becker and Guarente (2). To select yeast containing both the AD and BD plasmids, transformants were plated on SD-glucose medium (0.67% Bacto yeast nitrogen base lacking amino acids and containing 2% glucose) and supplemented with 30 μg/ml of Lys and 20 μg/ml of Ade, His, Met, and Trp, and growth was evaluated after 3 days at 28°C. Interactions were tested at 28°C or at room temperature (∼20°C) by assessing colony growth for 3 to 5 days after streaking on SD medium containing Lys (30 μg/ml) and Met and Trp (20 μg/ml).

Affinity chromatography.

For GST pull-down experiments, two proteins (with one fused to GST) were coexpressed in N. benthamiana plants through agroinfiltration as described above. At 2 or 3 days postinfiltration, 1.5 gram of leaf tissue was macerated with a mortar and pestle containing 4 ml of STE buffer (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1 mM EDTA) containing 2% bovine serum albumin, 10% glycerol, 1.0 mM phenylmethanesulfonyl fluoride, and 0.1 μg/ml of antipain, leupeptin, pepstatin, with or without 10 mM β-mercaptoethanol (β-ME). The brei was filtered through four layers of cheesecloth and then centrifuged at 10,000 × g at 4°C for 40 min in a Sorvall SS-34 or SA600 rotor. The supernatants were then transferred to 15-ml Falcon tubes, adjusted to 2% Triton X-100, and incubated with 100 μl of 50% glutathione-Sepharose 4B slurry as described by the manufacturer (Amersham, Piscataway, NJ). After 1 to 3 hours of incubation at 4°C, the beads were washed three times with 12 ml of PBS buffer by centrifugation at 5,000 rpm in a Sorvall SS-34 rotor and decanting the buffer. The Sepharose slurry was transferred to 1.5-ml Eppendorf tubes, and the proteins bound to the beads were eluted by adding 80 μl of 2× SMASH buffer (140 mM Tris-HCl, pH 6.8, 20% glycerol, 5% SDS, 400 mM β-ME, and 0.01% bromophenol blue) and boiled for 5 min. Alternatively, proteins were eluted three times in a total volume of 50 μl elution buffer containing 50 mM Tris-HCl (pH 8.0) and 20 mM reduced glutathione. Protein recoveries were assessed by loading 10 μl of each eluted sample into a gel lane, followed by separation in 10% SDS-polyacrylamide gels.

In vitro binding assays.

The BamHI-BglII fragments of yeast importin α (ySrp1 [NCBI accession no. NP 014210]), Arabidopsis thaliana importin α (AtSrp1 [NCBI accession no. AY125529]), and human importin α (hSrp1, from pKW230 [36]) were inserted individually into BamHI-cleaved pGEX-2T plasmids. The resulting expression vectors were transformed into E. coli strain BL21 harboring the plasmid pREP-4. One colony from each construct was grown in 5 ml of LB medium containing 200 μg/ml of ampicillin and 100 μg/ml of kanamycin at 37°C for 5 h. The cultures were then diluted 1:20 in the same medium and grown at 37°C to an optical density at 600 nm of 0.8. Protein expression was induced by adding β-d-thiogalactopyranoside to 0.1 mM, followed by shaking for 5 h at room temperature. Cells were harvested by centrifugation at 5,000 rpm in a Sorvall SS-34 rotor, resuspended in 3 ml of STE buffer as described above, and disrupted on ice with a sonic dismembrator (Fisher Scientific, Hampton, NH) by five sonication cycles consisting of 10-second bursts interspersed with 2-second cooling periods. The SYNV N, N_KKRR, and P proteins (7) used for binding were transiently expressed in N. benthamiana and extracted as described above. For affinity chromatography, 0.5 ml of GST-importin α extracted from BL21 cells and 1 ml of the N- or P-protein derivatives from plants were mixed with 100 μl of glutathione-Sepharose beads at 4°C for 2 hours to allow in vitro binding. Proteins bound to the beads were washed, boiled in 100 μl of 2× SMASH buffer, and analyzed on gels by SDS-PAGE as described above. For N-protein detection, 15-μl samples were loaded on gels, and 2-μl samples were loaded for detection of GST fusion proteins.

RESULTS

The N protein localizes to subnuclear foci.

We have previously shown that the SYNV N protein localizes completely inside the nuclei of yeast and plant cells when fused with either RSGFP or the DsRed fluorescent proteins (7, 8). The subnuclear foci can be more easily visualized in plant nuclei than in yeast nuclei because of the larger size of the plant nuclei. Moreover, agroinfiltration results in transformation of a large number of leaf cells (>95%), each of which can transiently express several different proteins from mixtures of Agrobacterium harboring different expression vectors (8). Therefore, the N-protein fusions to eGFP, RSGFP, DsRed, mRFP, YFP, and CFP were evaluated in leaves after Agrobacterium infiltration to determine whether localization patterns might be affected by the reporter genes. The images presented in Fig. 1 show cells representing the localization patterns of N protein fused to eGFP (Fig. 1A and B), RSGFP (Fig. 1C and D), DsRed (Fig. 1E and F), mRFP (Fig. 1G and H), YFP (Fig. 1I and J), and CFP (Fig. 1K and L [false-color green in Fig. 1K[) fluorescent proteins. The N-protein fusions each localized completely inside the nuclei as indicated by DAPI staining, and as previously shown in viroplasms of SYNV-infected plants (22), DAPI staining was reduced in the subnuclear regions exhibiting N-protein fluorescence (see white arrows in Fig. 1B, H, J, L, N, and P). With the exception of RSGFP-N, the fusion proteins normally formed one or more intense fluorescent subnuclear foci. However, the individual foci varied in size and shape, and these differences are thought to represent variations among different cells and the planes of the confocal sections, rather than differences among the different fusion proteins. In contrast to the eGFP, DsRed, mRFP, YFP, and CFP fusions, the RSGFP-N protein fluorescence was generally distributed throughout the nuclei and appeared not to form discrete foci (Fig. 1C and D). N-protein fusions with two eGFP reporter genes (eG-eG-N) or two eGFP reporter genes and a GST reporter gene (eG-eG-GST-N) also localized completely inside of the nucleus and formed subnuclear foci similar to those of the eG-N fusion (data not shown). These results indicate that the sizes of the fluorescent reporter genes do not have appreciable affects on the localization patterns.

FIG. 1.

Subcellular localization patterns of the N and P proteins. N-protein fusions with eGFP (A and B), RSGFP (C and D), DsRed (E and F), mRFP (G and H), YFP (I and J), or CFP (K and L). Panels M to P show immunostaining of the N protein expressed alone (M and N) or with the unfused P protein (O and P). Panels Q, R, S, and T show the P protein expressed as a fusion with eGFP (eG-P). Panels Q and R show a single cell, and images S and T show the enlarged nucleus of this cell. Panels U, V, W, and X show eG-P fusion coexpressed with the N protein. Panels U and V were photographed at a low magnification to show complete nuclear localization of the P protein, and panels W and X show an enlarged nucleus. The images in panels B, D, F, H, J, L, N, P, R, T, V, and X show DAPI staining of the nuclei in blue. White arrows indicate the area inside the nuclei where DAPI staining is excluded. Bars, 10 μm. All images were visualized by confocal microscopy and processed with Adobe Photoshop as outlined in Materials and Methods.

To determine whether N-protein localization at subnuclear sites is characteristic of the native N protein or is affected by the fluorescent protein fusions, immunofluorescence experiments were conducted with fluorescent dye-conjugated antibodies (22). As shown in Fig. 1M and N, the native N protein also formed subnuclear foci in regions with reduced DAPI staining. These observations are similar to those reported for SYNV-infected protoplasts (22) and plant cell bombardment experiments (7). In additional control experiments similar to those reported previously (22), background fluorescence was not observed in the nuclei of uninfiltrated leaves (data not shown). Thus, our results indicate that with the exception of RSGFP-N, the reporter gene fusions to the N protein form subnuclear patterns similar to those of the unfused wild-type N protein and also suggest that the N protein contains signals conditioning formation of the subnuclear foci.

In contrast to the N fluorescent fusion proteins, the eG-P fusion protein fluoresced inside the nucleus and throughout the cytoplasm (Fig. 1Q and R). When the same cells were examined at a higher magnification to highlight the nuclei, fluorescence was distributed throughout the nuclei, and subnuclear foci were not observed (Fig. 1S and T). The eG-P-expressing cells had more intense fluorescence in the nuclei than in the cytoplasm; this also resulted in reduced DAPI staining throughout the nucleus. Similar incidences of higher P-protein nuclear fluorescence than cytoplasmic fluorescence were also obtained when the P protein was fused with a double eGFP (data not shown).

Additional immunostaining of infiltrated leaf tissue expressing both the N and P proteins suggests that the P protein does not have obvious effects on the subcellular localization patterns of the N protein. In these experiments, immunostained N protein localized as intense foci completely inside the nuclei in the presence of the P protein (Fig. 1O and P) to produce patterns similar to those observed when the N-protein fusions were expressed alone (Fig. 1M). In addition, coexpression of the N protein with eG-P resulted in subnuclear fluorescence patterns that appeared similar to the foci formed when the eG-N protein was expressed alone (Fig. 1U to X). These results show that coexpression of the P protein does not affect the subnuclear patterns of the N fusion proteins and suggest that signals in the N protein result in formation of the subnuclear foci formed during coexpression of N and P.

Overlapping N-terminal residues of the N protein are required for N-N and N-P binding.

To evaluate the physical interactions of the N and P proteins, we modified our Agrobacterium-mediated transient protein expression plasmids (8) to develop a GST affinity chromatography system using proteins expressed in planta. For this purpose, the GST ORF was inserted into the pGD vector to construct the pGDGSTpro vector, which permits fusion of viral proteins in frame with GST, bridged by a TEV NIa protease cleavage site. The unfused N protein was then coexpressed with GST-N or GST-P in N. benthamiana leaves for 2 or 3 days, and GST pull-down experiments were performed in the presence or absence of 10 mM β-ME. The native N protein eluted from the glutathione matrix when coexpressed with GST-N (Fig. 2A, lanes 1 and 2) or GST-P (Fig. 2B, lanes 2 and 3) but did not bind to the matrix when coexpressed with GST alone (Fig. 2A, lanes 5 and 6, and B, lane 1). These results demonstrate formation of homologous N-N and heterologous N-P binding and confirm the N-N and N-P interactions previously revealed by yeast two-hybrid experiments (8).

FIG. 2.

Binding of the N protein to GST-N or GST-P affinity columns. The SYNV N protein was coexpressed in N. benthamiana leaves with either GST-N (A) or GST-P (B) (+) and purified over glutathione-Sepharose 4B affinity columns in the presence (+) or absence (−) of 10 mM β-mercaptoethanol (β-ME). Designations below each lane in panel A indicate the expression of GST, GST-N, and the unfused N protein. Designations in panel B show GST, GST-P, and the unfused N protein. Proteins eluted from the matrix were separated by SDS-PAGE and detected with anti-N antiserum. Arrows indicate the monomer, dimer, and aggregate (agg) forms of the N or GST-N proteins.

In the absence of β-ME, the N protein was observed at positions corresponding to monomers, dimers, and large aggregates that migrated near the top of the gels (Fig. 2A, lanes 2 and 4, and B, lane 2). However, when β-ME was added to the extraction buffer to a final concentration of 10 mM, the native N protein was observed in the monomeric form (Fig. 2A, lanes 1 and 3, and B, lane 3). The GST-N fusion protein also formed large aggregates when β-ME was omitted from the extraction buffer (Fig. 2A, lane 4). These data indicate that reducing conditions are necessary to prevent the formation of N-protein aggregates during extraction. Therefore, for all subsequent GST pull-down experiments, β-ME was added to the extraction buffer in order to prevent aggregation of the N and P proteins.

Our previous results using yeast two-hybrid experiments indicated that the first 73 aa of the N protein are critical for homologous N-N and heterologous N-P binding and that deletion mutants containing residues 1 to 195 are able to form both N-N and N-P interactions (7). More-detailed yeast two-hybrid analyses revealed that a mutant lacking the N-terminal 27 residues (N_28-475) was able to form N-N and N-P interactions but that deletion of the N-terminal 53 residues (N_53-475) interfered with both the N- and P-protein interactions (Table 3). Subsequently, we focused on aa 28 through 73, a region that contains a putative helix-loop-helix structure that was identified using the PROFsec Profile Network Prediction HeiDelberg (27). To extend these experiments, alanine residues were substituted for hydrophobic residues within the helix-loop-helix to create three full-length N mutants that were used for yeast two-hybrid and GST affinity chromatography experiments (Fig. 3A). Mutant N_YYLF contains the substitutions Y37A, Y40A, L41A, and F42A that reside in the first helix, mutant N_YIY contains Y48A, I50A, and Y51A substitutions within the loop area, and mutant N_IFFI contains four substitutions (I61A, F64A, F65A,and I68A) within the second helix. Immunoblot analyses of crude protein extracts prior to the addition of the glutathione beads indicate that the wild-type N protein and mutant N-protein derivatives were expressed at similar levels (Fig. 3B, bottom blot). As shown in the eluted samples (Fig. 3B, top blot), the N_WT protein and GST-N_WT bound at nearly a 1:1 ratio (Fig. 3B, lane 1), but the N_YYLF and N_IFFI mutants failed to bind GST-N_WT (Fig. 3B, lanes 2 and 4). The N_YIY mutant retained the ability to bind to GST-N_WT (Fig. 3B, lane 3), albeit less efficiently than the N_WT protein. These mutants were also subjected to yeast two-hybrid analyses. In these experiments, the N_YYLF and N_IFFI mutants incorporated into the full-length N protein binding domain failed to elicit yeast growth, whereas growth was observed with the N_YIY mutant (Table 3). Taken together, these results indicate that the N-terminal helix-loop-helix region of N (aa 37 to 68) contributes substantially to homologous interactions of the N protein and that the first and second helices have greater contributions to the interactions than the loop region, which may serve primarily as a linker to constrain positioning of the two helices.

TABLE 3.

Yeast two-hybrid analyses

ADa	BDb	Interactionc
N	N	+
N	N28-475	+
N	N53-475	−
N	NYYLF	−
N	NYIY	+
N	NIFFI	−
P	N	+
P	N28-475	+
P	N53-475	−
P	NYYLF	−
P	NYIY	+
P	NIFFI	−

^aAD, activation domain fusion.

^bBD, binding domain fusion. The numbers designate the amino acid residues in the N protein.

^c+, interaction; −, no interaction.

FIG. 3.

Involvement of N-protein N-terminal amino acid residues in homologous N-N and heterologous N-P interactions. (A) Alanine mutations introduced into the helix-loop-helix region. Mutated amino acid residues of the N protein are indicated by bold italic type; numerals refer to the amino acid residues. (B) Wild-type (WT) and mutant N proteins coexpressed with GST-N. (C) Wild-type and mutant N proteins coexpressed with GST-P. Crude protein extracts recovered after agroinfiltration of N. benthamiana leaves were purified over glutathione-Sepharose 4B affinity columns and analyzed by Western blotting as described in Materials and Methods.

The effects of the helix-loop-helix mutations on N-P associations were also evaluated in GST pull-down experiments with the GST-P fusion protein (Fig. 3C). In these experiments, the mutant N derivatives and N_WT protein were also expressed at similar levels (Fig. 3C, bottom blot). The N_WT protein binds to GST-P, but the N_YYLF and N_IFFI mutants failed to interact with P (Fig. 3C, top blot), and the N_YIY mutant retained a lower affinity for GST-P than the N_WT did (Fig. 3C, top blot). As was the case with the N-N interactions, yeast containing activation domain fusions to the P protein exhibited growth in pairings with the N_YIY binding domain mutant, but not with the N_YYLF and N_IFFI mutants. Thus, both the yeast two-hybrid experiments and the affinity chromatography experiments provide persuasive evidence that the N-terminal region is important for binding of both the N and the P proteins. All of the available evidence indicates that the N-N interactions result from direct binding to amino acids within the helix-loop-helix region. The results also suggest that N-P binding occurs in this region, and an attractive hypothesis is that the P protein may increase N-protein solubility during infection by blocking N-N interactions. However, an alternative hypothesis is that P-protein binding requires N-N dimerization or multimerization interactions and that the N-N interactions elicit conformational changes that expose distal N-protein residues needed for N-P binding. If so, P-protein binding must occur within the N-terminal 195 residues of the N protein, because in our previous experiments an N-protein deletion mutant consisting of this region was able to bind to the P protein in yeast two-hybrid analyses (7).

To pinpoint individual amino acid residues in the helix-loop-helix region that affect N-N and N-P binding, an additional panel of 11 full-length N mutants (N_Y37A, N_Y40A, N_L41A, N_F42A, N_Y48A, N_I50A, N_Y51A, N_I61A, N_F64A, N_F65A, and N_I68A) that contained only single-amino-acid substitutions were constructed (Fig. 4A) and shown by Western blot analyses to be expressed at similar levels when expressed with the GST-N_WT or GST-P proteins (data not shown). The affinity chromatography results indicated that most of the single-amino-acid mutants retained the full capacity to bind to GST-N_WT but that binding of the N_Y40A mutant was drastically impaired (Fig. 4B). When the N mutants were tested for binding to GST-P, the N_Y40A, N_F42A, and N_Y51A mutants had reduced binding, but these mutants did not totally eliminate N-P associations (Fig. 4C). However, the N_Y40A and N_F42A mutants were compromised to a greater extent than the N_Y51A mutant was. These experiments indicate that the Tyr40 residue has an important role in both N-N and N-P interactions.

FIG. 4.

Binding of native wild-type and mutant SYNV N proteins to N- and P-GST fusion proteins. (A) Alanine substitutions were incorporated into the N-terminal region of the N protein at the positions shown in bold. (B) Affinity chromatography of GST-N protein expressed with wild-type N protein or one of the alanine mutants shown in panel A. (C) Binding of GST-P to the WT or mutant N proteins. (D) Binding of GST-P to the WT N protein or alanine mutants harboring two amino acid changes (KR, RE, EE, or YF; see double underlined residues in panel A). Negative controls correspond to single expression of GST-N or GST-P, or coexpression of GST with N. Asterisks indicate residues that are of key importance to the interactions. Note that all of the mutant proteins were expressed at levels similar to those of the N_WT protein (not shown).

To further assess the N-N and N-P interactions, double-amino-acid substitutions were engineered to produce three double mutations in basic or acidic amino acids (N_KR, N_RE, and N_EE) and a double aromatic amino acid mutant (N_YF). Mutant N_RE has R38A and E39A substitutions, mutant N_KR has K52A and R53A substitutions, and mutant N_EE has E59A and E60A substitutions. Mutant N_YF has the Y40A and F42A substitutions that individually diminish the N-N and N-P interactions. As shown in Fig. 4D, amino acid substitutions of the charged amino acids had no effect on either the N-N or N-P interactions, whereas mutagenesis of the tyrosine and phenylalanine residues (N_YF) completely disrupted both the N-N and N-P interactions.

N-terminal interactions of the N protein are required for formation of subnuclear foci.

The effects of the N-terminal mutations on nuclear localization and formation of subnuclear foci were assessed by incorporating the N_YYLF, N_YIY, and N_IFFI mutants into the full-length eGFP-N protein for agroinfiltration into N. benthamiana leaves (Fig. 5A). Two days after infiltration, the transient-expression results revealed that each of the three eGFP-N protein mutants localized completely inside the nucleus but that none of the mutants formed subnuclear inclusions corresponding to those observed for the N_WT protein (Fig. 5B). These results indicate that homologous N-protein interactions are required for formation of subnuclear foci but that disruption of the N terminus does not affect the C-terminus-mediated nuclear import of the N protein. We also assessed single N-terminal mutations (N_Y40A, N_F42A, and N_Y51A) in the eG-N protein for their effects on the formation of subnuclear foci. After expression, all three mutants localized completely inside of the nuclei as expected, but after an extensive search, subnuclear foci were not observed in any of the fluorescing cells (data not shown). However, the fluorescence of these derivatives was low, and to increase their expression levels, the tomato bushy stunt virus p19 silencing suppressor was included in the infiltration experiments (33). Under these conditions, subnuclear foci were not observed in leaves expressing mutant N_F42A, but we were able to discern subnuclear foci in 5% and 2% of the cells expressing higher levels of the N_Y40A or N_Y51A mutant, respectively (data not shown). Thus, the three aromatic residues each appear to contribute to the formation of subnuclear foci by the N protein.

FIG. 5.

Localization and subnuclear foci formed by wild-type or mutant N proteins in the absence or presence of the P protein. (A) Schematic diagram of the N protein highlighting the sequence of the N-terminal region and the positions of mutations introduced into the helix-loop-helix region. Mutated amino acid residues are highlighted in red, and the numerals refer to the positions of the amino acid residues in the N protein. (B) Localization of WT and mutant N proteins expressed as fusions with eGFP (eG-N). (C) Localization of the P protein fused to a double eGFP protein (eG-eG-P) that was coexpressed with N-protein derivatives. Confocal images of N. benthamiana leaf epidermal cells were taken 2 days after infiltration with Agrobacterium vectors. One whole cell is shown in each picture, and the insets show the enlarged nuclei. White arrows indicate the subnuclear foci. Bars, 10 μm.

To determine whether mutations in the N-terminal region of the N protein influence the nuclear import of the P protein, an eG-eG-P fusion protein was constructed and used in coexpression experiments with unfused N-protein derivatives. In agreement with previous results (7, 8), coinfiltration with the wild-type N protein and the eG-eG-P derivative resulted in characteristic subnuclear foci (Fig. 5C). In contrast, widespread cytoplasmic and nuclear fluorescence was observed in cells expressing the eG-eG-P protein reporter gene in the presence of the unfused N_YYLF or N_IFFI mutant, and subnuclear foci were not seen (Fig. 5C). Therefore, N-P binding mediated by the helix regions is essential for complete nuclear import of the P protein and for colocalization with the N protein in subnuclear foci. Interestingly, discrete subnuclear foci were noted in leaves expressing the eG-eG-P reporter protein and the unfused N_YIY mutant, even though the eG-eG-P protein continued to exhibit considerable levels of cytoplasmic fluorescence. Therefore, these results indicate that the N protein has the predominant role in the formation of subnuclear foci, albeit some activities of the P protein, perhaps weak interactions that could increase solubility of the N_YIY mutant, may contribute to formation of the foci during coexpression.

We also assessed the formation of subnuclear foci during coexpression of the P fusion protein with the eG-N_Y40A, eG-N_F42A, and eG-N_Y51A single-substitution mutants. When expressed alone, the three mutants formed only occasional subnuclear foci (data not shown). However, the mutant eG-N_F42A formed subnuclear foci in 95% of the cells when it was coexpressed with the P protein; subnuclear foci failed to form during coexpression of P with eG-N_Y40A and eG-N_Y51A (data not shown). These results indicate that Phe42 has a less pronounced contribution to the formation of subnuclear foci in the presence of the P protein than the Tyr residues do.

N_KKRR NLS mutant interactions affect nuclear localization of the P protein.

Previous molecular genetic experiments indicated that the C terminus of the N protein has a bipartite nuclear localization signal that is essential for nuclear import (7). The signal that resides between aa 445 and 461 consists of an arginine-rich motif (PSRKRR), a seven-residue spacer, and a lysine-rich motif (KPKK). The mutant N_KKRR contains alanine substitutions for two of the arginine residues and two of the lysine residues in the NLS motifs (indicated by the underlined residues above), and this derivative localized mostly outside of the yeast nuclei (7). To provide more definitive information in plant cells, localization was analyzed in N. benthamiana leaf epidermal cells after agroinfiltration of strains expressing GFP-P and either DsRed-N_WT or the mutant DsRed-N_KKRR. Confocal microscopy experiments clearly indicated that the mutant N_KKRR localized predominately in the cytoplasm, with little detectable red fluorescence inside the nucleus (Fig. 6A). When the DsRed-N_WT and eG-eG-P fusion proteins were coexpressed, both fusion proteins colocalized completely inside the nucleus where they formed one to several bright spots (Fig. 6B and C). However, when the DsRed-N_KKRR mutant was coexpressed with the eG-eG-P fusion protein, both fusion proteins localized throughout the cell (Fig. 6D and E). Moreover, the P-protein GFP fluorescence was much less intense in the nuclei than when it was expressed alone (compare with Fig. 1Q to T), and there were no visible subnuclear foci typical of those present during coexpression with N_WT. We also constructed a double mutant of the N protein (N_YYLF/KKRR) that contains both the N-terminal YYLF and C-terminal KKRR mutations and assessed their effects on localization of the P protein. When this derivative was coexpressed with the eG-eG-P fusion protein, the proportions of cells showing both cytoplasmic and nuclear fluorescence were indistinguishable from those observed when the P protein was expressed alone (data not shown). These results suggest that the N and P proteins can interact in the cytoplasm but that the signals responsible for P nuclear localization are not sufficient to overcome a dysfunctional N-protein NLS. Additional GST-N pull-down experiments indicated that the binding affinities of the N_WT and helix-loop-helix mutants are not affected by mutations in the C-terminal NLS (data not shown). These results taken together suggest that the C-terminal NLS mutations do not have spurious effects on the N-N or N-P interactions mediated by the N-terminal residues.

FIG. 6.

Localization of wild-type and mutant N proteins expressed singly or coexpressed with the P protein. (A) Cytoplasmic expression of the N_KKRR mutant fused with DsRed. (B to E) Coexpression of DsRed-P protein with N_WT and N_KKRR mutant protein derivatives fused to eGFP. (B) Low-magnification view showing cells expressing DsRed-P and eG-N_WT. (C) Higher-magnification view to illustrate the nuclei of a cell shown in panel B. (D) Low-magnification view to show cells expressing DsRed-P and eG-N_KKRR. (E) Higher-magnification view showing the nuclei of a cell shown in panel D. Bars, 10 μm. Confocal images of N. benthamiana leaf epidermal cells were taken 2 days after infiltration with the Agrobacterium vectors.

N protein, but not P protein, binds to importin α homologues.

We have investigated host components that might mediate nuclear transport to extend our understanding of N- and P-protein nuclear localization. In eukaryotic cells, cargo proteins that contain a bipartite NLS are recognized in the cytoplasm by components of the importin α complex (19). The resulting complexes dock at the nuclear pore complex and are transported through the nuclear pore complex, where the cargo proteins are released into the nucleoplasm and the importins are exported back into the cytoplasm. To assess the binding of the N and P proteins with importin α, we expressed ySrp1, plant AtSrp1, and hSrp1 as GST fusions in E. coli cells and used these fusions for affinity assays to N-protein derivatives. The GST-importin α homologues were extracted from the bacteria, incubated in vitro with SYNV N- or P-protein derivatives recovered from agroinfiltrated N. benthamiana leaves, and subjected to glutathione affinity chromatography. The in vitro binding experiments included N_WT, the N_KKRR mutant, an N-protein mutant (N_delNLS) containing a 32-aa C-terminal deletion that encompassed the bipartite NLS, and the P protein. The results revealed that the N_WT protein binds to the ySrp1 and AtSrp1 importin α homologues (Fig. 7A and B) but failed to bind specifically to hSrp1 (data not shown). As expected, the N_KKRR and N_delNLS mutants failed to bind to any of the importin α derivatives (Fig. 7A and B; also data not shown). The P protein also failed to interact directly with ySrp1 or AtSrp1 (Fig. 7C), suggesting that the two proteins have independent mechanisms for nuclear import. Thus, the yeast and Arabidopsis importin α proteins used in these experiments are each capable of forming direct associations with the N protein that require positively charged residues within the bipartite NLS of the N protein.

FIG. 7.

Associations of SYNV N-protein derivatives and the P protein with importin α homologues. (A) Binding to yeast importin α (ySrp1); (B) binding to Arabidopsis thaliana importin α (AtSrp1); (C) failure of the P protein to bind to the yeast and A. thaliana importin α proteins. The importin α homologues were expressed (+) as fusion proteins with GST in E. coli strain BL21 (pREP-4). The SYNV N_WT, N_KKRR, N_delNLS, and P proteins were expressed in N. benthamiana leaves via agroinfiltration. Crude protein extracts from BL21 cells and plants were mixed in vitro to allow binding and separated by glutathione-Sepharose affinity chromatography. Note that the unfused GST in the bottom blots of panels A, B, and C is thought to have arisen either during expression in E. coli or during bacterial purification.

DISCUSSION

The N proteins of negative-strand RNA viruses carry out a number of functions required for formation of the viral nucleocapsid core, which represents the minimal infectious unit of this class of viruses (3, 25). In the two most extensively characterized animal rhabdoviruses, vesicular stomatitis virus (VSV) and rabies virus (RV), a number of studies provide a general model whereby the N protein binds viral RNA at its phosphodiester backbone in a complex with the P and L proteins (10, 13) so that the bases are accessible for transcription and replication (5, 6). SYNV fits within this model, and the results of our previous studies show that the nucleocapsid of SYNV can be purified from infected plants as a transcriptionally active complex consisting of the viral RNA and the N, P, and L proteins (34, 35). Despite these important similarities, there are substantial differences between the replication strategies of SYNV and those of the mammalian rhabdoviruses that relate to properties of the N protein. For example, SYNV and other plant nucleorhabdoviruses differ from VSV, RV, and the plant cytorhabdoviruses by replicating in the nucleus (17) and by forming large subnuclear viroplasms that contain the viral RNA and the N, P, and L proteins (22). Both the N and P proteins of SYNV are able to enter the nucleus independently when expressed alone, and the N protein contains a bipartite nuclear localization signal that functions in nuclear entry (7). The N and P proteins also form heterologous interactions, and when coexpressed in plant or yeast cells, the two proteins colocalize in subnuclear foci that are reminiscent of the viroplasms observed in SYNV-infected plant cells (7, 8). Although the nuclear sites of replication of SYNV and other plant nucleorhabdoviruses differ from those of all known mammalian rhabdoviruses, Borna disease virus (20), another member of the Mononegavirales order, and the segmented orthomyxoviruses (4) also replicate in the nucleus. Therefore, information obtained about the features of SYNV contributes to our general understanding of the complexity and evolution of diverse negative-strand viruses that share a common property of replication in the nucleus.

The present study provides insight into the homologous interactions of the SYNV N protein and its heterologous interactions with the P protein. We have mapped the N-N and N-P interaction domains of the N protein to a helix-loop-helix region within the N-terminal 37 to 68 residues. The results of yeast two-hybrid and GST pull-down experiments provide persuasive evidence that aromatic amino acid residues within both helices are required for N-N and N-P binding. When individual alanine substitutions were evaluated, Tyr40 appeared to be critical for N-N binding, and substitution of alanine for Tyr40, Phe42, or Tyr51 reduced the N-P binding levels. Double alanine substitutions for Tyr40 and Phe42 completely disrupted the N-N and N-P complexes. The N-P interactions are important for the solubility of the N protein and decrease N-protein aggregates (7) and likely are essential for formation of nucleocapsid complexes, as is also the case with VSV (24) and RV (12).

The requirements for SYNV N- and P-protein interactions differ substantially from those of the vertebrate rhabdoviruses. Mutations within the N-protein helix-loop-helix motif abrogate both N-N and N-P binding. Moreover, yeast two-hybrid interactions reveal that the first 195 residues of the N protein function in homologous and heterologous interactions (7). Therefore, if N-protein conformational changes resulting from N-N interactions are required for N-P interactions, these must reside within N-terminal half of the protein. In contrast, with VSV and RV, C-terminal amino acids are required for N-P associations (30, 31). Although we have not evaluated SYNV RNA binding, the P protein of VSV is required for specific binding of the N protein to viral RNA in vitro (23, 24), and encapsidation of the leader RNA requires the five C-terminal residues of the N protein (6). In the case of RV, N-P binding requires the C-terminal 74 amino acids of the N protein because trypsin treatment releases the P protein from nucleocapsid complexes formed in insect cells (12, 21). Unfortunately, more direct evaluations of VSV and RV N-N and N-P interactions have not been determined because recombinant N protein isolated from cells binds tenaciously to cellular RNAs and these complexes are difficult to dissociate (12). However, in vitro RNA binding assays with N protein released from RV nucleocapsids by detergent denaturation have shown that the N-terminal fragment (aa 1 to 376) produced by trypsin treatment is able to bind viral leader RNA and that residues 298 to 352 are required for RNA binding activity (21). Thus, the available data indicate that SYNV has substantial differences from both VSV and RV in the mechanisms of N-P interactions.

Our cytological analyses have shown that the N and P proteins colocalize in subnuclear viroplasms in SYNV-infected plant cells (14, 17, 22). The confocal imaging results presented above provide further evidence that the subnuclear foci formed during expression of N-protein reporter genes are very similar to those formed when the N protein is expressed alone or in the context of an SYNV infection. Our mutagenesis experiments also demonstrate that disruption of the helix-loop-helix structure by the N_YYLF, N_YIY and N_IFFI mutants interferes with the formation of subnuclear foci in plant cells. Taken together, these results imply that signals within the N protein mediate formation of subnuclear foci and demonstrate that homologous N-protein interactions are essential for formation of the foci. Our data also confirm earlier results showing that the P protein can enter the nucleus independently but that complete nuclear localization in the presence of the N protein is driven by N-P interactions (7, 8). The P protein accumulates in both the nucleus and the cytoplasm when expressed alone, and little change in localization occurs when the P protein is coexpressed with two of the three N mutants (N_YYLF and N_IFFI) that affect N-P interactions. However, subnuclear foci do form when GFP-P is coexpressed with the N_YIY mutant, which has a weak affinity for the P protein, although a substantial portion of the fluorescence remains in the cytoplasm. This result suggests that weak N-P interactions are sufficient for formation of subnuclear foci even though the N protein is not able to direct homologous binding. We believe that the formation of the foci may be a consequence of P-mediated increases in N-N mutant interactions or increased solubility of the N protein resulting from a chaperone-like function of the P protein, or possibly that conformational changes during N-P interactions may provide access to signals that have as yet unidentified roles that can contribute to subnuclear localization.

In eukaryotic cells, nuclear import of cargo proteins, such as the N protein, normally contain nuclear import signals that are recognized in the cytoplasm by importin α and importin β proteins (19). The resulting complexes dock at the nuclear pore complex and are transported through the pore to release their cargo proteins into the nucleoplasm. In this regard, it is of particular interest to ascertain the karyophilic activities of viruses that replicate in the nucleus because such activities are critical to the infection cycle and may provide valuable information about the mechanisms of nuclear import in host cells. The C-terminal region of the SYNV N protein contains a bipartite NLS between aa 445 and 461 that consists of an arginine-rich motif (PSRKRR), a seven-residue spacer, and a lysine-rich motif (KPKK). Mutagenesis has previously revealed that disruption of either of the two motifs results in GFP derivatives that are unable to enter yeast nuclei and that the NLS alone is sufficient to facilitate movement of a GFP-GST fusion protein into the nuclei (7). To extend these studies, we have conducted in vitro binding experiments with GST fusions to yeast, plant, and human importin α homologues. The results show that the N protein binds to the yeast and Arabidopsis homologues, but not to human importin α. These results, coupled with the absence of canonical importin α NLS signals and the failure of the P protein to interact with any of the importin α derivatives, indicate that the N and P proteins use separate mechanisms for nuclear import.

We previously have shown that mutations in the N-protein NLS result in nearly complete sequestering of the GFP-P protein in the cytoplasm (7). These observations suggest that nuclear import of the N protein by importin α is very rapid and occurs before substantial P protein can associate with the N protein. Furthermore, the observations suggest that P-protein nuclear import signals are either masked by N protein binding or that they are not sufficiently strong to mediate nuclear import of the cytoplasmic complexes. We have complemented these experiments to show that abrogation of the N-P interactions by alanine substitutions for Tyr40 and Phe42 in the N protein results in a nuclear/cytoplasmic pattern identical to that occurring when the GFP-P protein is expressed alone.

Our results in toto can be used to derive a model for the formation of viroplasms and the early stages of replication. This hypothesis assumes that immediately after translation, the N protein interacts with importin α and is transported rapidly into the nucleus. After nuclear import, homologous interactions occur, and these interactions culminate in N-N transport to as yet unidentified subnuclear locales. At some point in this sequence of events, the P protein enters the nucleus via a different mechanism from the N protein-importin α pathway, and along with the L polymerase protein, the P protein associates with the N-N complexes to initiate viroplasm formation. This sets the stage for encapsidation of nascent genomic and antigenomic RNAs that are replicated from the primary infecting nucleocapsids. The progeny nucleocapsids then participate in secondary transcription and replication cycles to ultimately produce viroplasm foci that vary in number and shape and eventually occupy a considerable volume of the enlarged nuclei. Information needed to complete this model will require identification of the P-protein signals required for nuclear localization and interactions with the N protein and an understanding of the mechanisms whereby the L protein enters the nucleus and interacts with the nucleocapsid to form an active transcription complex. We are currently investigating these areas, and these investigations will be the subjects of subsequent communications.