Analysis of Rotavirus Nonstructural Protein NSP5 Phosphorylation

Abstract

The rotavirus nonstructural phosphoprotein NSP5 is encoded by a gene in RNA segment 11. Immunofluorescence analysis of fixed cells showed that NSP5 polypeptides remained confined to viroplasms even at a late stage when provirions migrated from these structures. When NSP5 was expressed in COS-7 cells in the absence of other viral proteins, it was uniformly distributed in the cytoplasm. Under these conditions, the 26-kDa polypeptide predominated. In the presence of the protein phosphatase inhibitor okadaic acid, the highly phosphorylated 28- and 32- to 35-kDa polypeptides were formed. Also, the fully phosphorylated protein had a homogeneous cytoplasmic distribution in transfected cells. In rotavirus SA11-infected cells, NSP5 synthesis was detectable at 2 h postinfection. However, the newly formed 26-kDa NSP5 was not converted to the 28- to 35-kDa forms until approximately 2 h later. Also, the protein kinase activity of isolated NSP5 was not detectable until the 28- and 30- to 35-kDa NSP5 forms had been formed. NSP5 immunoprecipitated from extracts of transfected COS-7 cells was active in autophosphorylation in vitro, demonstrating that other viral proteins were not required for this function. Treatment of NSP5-expressing cells with staurosporine, a broad-range protein kinase inhibitor, had only a limited negative effect on the phosphorylation of the viral polypeptide. Staurosporine did not inhibit autophosphorylation of NSP5 in vitro. Together, the data support the idea that NSP5 has an autophosphorylation activity that is positively regulated by addition of phosphate residues at some positions.

Rotavirus is the major etiologic agent of gastroenteritis in the young of human and animal species (3). The genome of rotavirus consists of 11 segments of double-stranded RNA. Six of them code for structural proteins, arranged in a core with VP1, VP2, and VP3, an inner shell made of VP6, and an outer shell composed of VP7 and VP4. The rest of the genome segments code for nonstructural proteins (NSP1 through NSP6) present in rotavirus-infected cells but not in virions (18).

Rotavirus NSP5 protein is encoded by the smallest genomic RNA segment. In the simian rotavirus SA11 strain, it is 198 amino acids long. The polypeptide contains 20% serine residues (23). The major intracellular polypeptides have sizes corresponding to 26, 28, and 30 to 35 kDa when resolved by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE). This heterogeneous mobility of the protein is due to phosphorylation at multiple sites (1, 5, 29, 32) and reportedly to addition of O-linked GlcNac (11).

NSP5 is phosphorylated at serine and, to a lesser extent, threonine residues. Tryptic mapping and two-dimensional PAGE analysis of the protein indicated that multiple amino acid residues are phosphorylated (1, 5). Recently, it was reported that treatment of the NSP5 polypeptides isolated from infected cells with protein phosphatase 2A (PP2A), calf intestinal phosphatase, or lambda protein phosphatase resulted in dephosphorylation of 28- and 30- to 35-kDa polypeptides and accumulation of the 26-kDa form (1, 5, 29). However, phosphatase treatment did not remove the phosphate groups from the 26-kDa band. Incubation of NSP5 isolated from infected cells with ATP leads to autophosphorylation. The phosphate residues are incorporated chiefly into 28- to 35-kDa polypeptides, just as in infected cells (1, 5, 29). Furthermore, in vitro protein kinase activity has been demonstrated with NSP5 produced in transient transfection with the segment 11 gene, by infection with gene 11 recombinant vaccinia virus or baculovirus and bacterially expressed polypeptide (5, 29). However, in this autophosphorylation reaction the 26-kDa material did not become fully phosphorylated, as judged by the absence of 28- to 35-kDa forms (5, 29). The necessity of a cofactor that could generate full kinase activity of NSP5 in infected cells was suggested (29).

NSP5, together with NSP2 and the structural proteins VP2 and VP6, accumulate in viroplasms, the sites where rotavirus RNA replication and assortment of segments into provirions occur (19, 27, 28). During RNA replication and virion morphogenesis, NSP2 and NSP5 can be isolated in association with replicative intermediate particles (26). However, the function of NSP5 and its protein kinase activity are unknown. In the study described herein, we have analyzed the kinetics of synthesis and phosphorylation of NSP5, both during rotavirus infection of MA104 cells and when expressed from cDNA in transfected COS-7 cells. To investigate whether cellular protein kinases participate in the phosphorylation of NSP5, we tested the effects of several inhibitors on both infected and transfected cells.

MATERIALS AND METHODS

Cells and virus.

Rhesus monkey kidney MA104 cells and African green monkey kidney COS-7 cells were grown in Dulbecco modified Eagle medium (DMEM) supplemented with 5% fetal calf serum (Gibco BRL). Simian rotavirus strain SA11 clone 3 was obtained from R. F. Ramig (Baylor College of Medicine). MA104 and COS-7 cells were infected with SA11 rotavirus by using a previously described procedure (17). Virus titers were determined by counting cells expressing viral capsid antigen after staining with the immune peroxidase method (2). Titers were expressed as focus-forming units per milliliter.

NSP5 expression in transfected cells.

To prepare a recombinant expression plasmid, a DNA fragment containing the NSP5 gene was excised from plasmid pGEX-NSP5 (6), using restriction endonucleases EcoRI and BamHI. The fragment was then inserted into the corresponding cleavage sites of the pcDNA3 vector (Invitrogen) downstream of the cytomegalovirus immediate-early promoter. Subconfluent monolayers of COS-7 cells in 60-mm-diameter petri dishes were transfected with pcDNA3 or pcDNA-NSP5 (2 μg of plasmid DNA per dish), using Lipofectamine (Gibco BRL) according to the supplier’s instructions. The cells were harvested after 48 h, proteins were separated by SDS-PAGE, and the NSP5 polypeptides were detected by immunoblotting.

Radioactive labeling and immunoprecipitation of NSP5.

To label protein with ³⁵S or ³²P, MA104 or COS-7 cells were grown in 60-mm-diameter petri dishes. Cultures of infected and of transfected cells were incubated for 1 and 3 h, respectively, in methionine-free DMEM supplemented with 50 μCi of [³⁵S]methionine per ml or in phosphate-free DMEM supplemented with 200 μCi of ³²P_i per ml. For pulse-chase experiments, the cells were first radioactively labeled, then rinsed with DMEM, and incubated for the same time period with regular DMEM. After labeling, the cell monolayers were rinsed with 1 ml of Tris-buffered saline (20 mM Tris-HCl [pH 7.5], 150 mM NaCl) containing 50 mM NaF, 0.10 mM Na₃VO₄, and 1.0 mM dithiothreitol (DTT). The cells were scraped off the surface, transferred to microcentrifuge tubes, and treated for 10 min at 20°C with 100 μl of TX lysis buffer (20 mM Tris-HCl [pH 7.5], 150 mM NaCl, 30 mM NaPP_i, 1.0 mM DTT, 10% glycerol, 1.0% Triton X-100, 50 mM NaF, 0.10 mM Na₃VO₄, 1.0 μg of aprotinin and 50 μg of phenylmethylsulfonyl fluoride per ml). Nucleic and cell debris were removed by centrifugation at 20,000 × g for 20 min, and the NSP5 polypeptides were precipitated from the supernatant with a specific NSP5 antiserum as previously described (5). These samples, or fractions of TX lysates not subjected to immunoprecipitation, were mixed with an equal volume of 2× SDS sample buffer, boiled, and resolved by SDS-PAGE as described by Laemmli (15). Okadaic acid (Gibco BRL) at 0.5 μM, staurosporine (BIOMOL) at 1.0 mM, and 5,6-dichloro-1-β-d-ribofuranosyl benzimidazole (DRB; BIOMOL) at 0.2 mM were included during the pulse-labeling or chase periods in some experiments.

Immunoblotting procedure.

Cytosolic proteins of rotavirus-infected MA104 cells or transfected COS-7 cells were extracted with TX lysis buffer for 10 min at 22°C and then clarified by centrifugation. Polypeptides were resolved by SDS-PAGE and transferred to nitrocellulose membranes. After blocking with 5% nonfat milk in phosphate-buffered saline (PBS), membranes were probed with mouse NSP5 antiserum (diluted 1:2,000 in PBS–0.1% milk) for 1 h at 37°C. The membrane with bound specific antibodies was incubated for 1 h at 37°C with peroxidase-conjugated secondary antibodies. Membranes were developed either with diaminobenzidine and photographed or with the Amersham enhanced chemiluminescence detection system. Chemiluminescence of NSP5-related material was quantified with a GS-250 molecular imager (Bio-Rad).

Immunofluorescence microscopy.

MA104 cells were grown to semiconfluency on glass coverslips, fixed with 3% paraformaldehyde in PBS at different time points after rotavirus infection, and then stained for NSP5- or VP6-specific immunofluorescence based on a protocol described by Loo et al. (16). Cells were permeabilized with 0.2% Triton X-100, dissolved in PBS, rinsed in PBS, and then incubated with mouse NSP5 antiserum (1:200 dilution in PBS) overnight at 4°C. After washing with PBS, the preparations were blocked with normal goat serum (1:50 dilution in PBS) and then incubated with fluorescein isothiocyanate-conjugated goat anti-mouse secondary antibodies (1:100 dilution in PBS) for 1 h at 4°C. Coverslips with the cells were wet mounted in glycerol containing 2% 1,4-diazabicyclo-(2,2,2)-octane. The slides were visualized in a Nikon Optophot-2 microscope using phase-contrast and fluorescence optics. Images were digitized by using a charge-coupled device camera (Sony Instruments) connected to a computerized image analysis system (software and hardware from Bergströms Instruments, Solna, Sweden). The digitized images were printed with a Sony UP-860/CE video printer.

Assay of protein kinase activity.

Phosphorylation in vitro was performed with NSP5 protein immunoprecipitated from infected MA104 cells or from transfected COS-7 cells. NSP5 protein immobilized on protein G-Sepharose beads (Pharmacia) was incubated for 10 min at 20°C in 25 μl of kinase buffer (20 mM HEPES [pH 7.5], 10 mM MnCl₂, 1.0 mM DTT, 5 μM ATP) containing 10 μCi of [γ-³²P]ATP. The reactions were stopped by adding 25 μl of 2× SDS sample buffer. The samples were incubated in a boiling water bath for 5 min before separation of polypeptides by SDS-PAGE.

RESULTS

Cellular localization of NSP5.

The cellular localization of NSP5 during the infection cycle was examined by immunofluorescence. MA104 cells were infected with simian rotavirus SA11, fixed with paraformaldehyde at different time points postinfection, and then stained for NSP5-specific immunofluorescence (Fig. 1A and B). NSP5 was visible in the cytoplasmic inclusions characteristic of rotavirus infection already 2 h after infection. As the infection proceeded, the viroplasms increased in size and released proviral particles. However, NSP5 remained in the viroplasmic inclusions even late in infection, suggesting that the polypeptide forms part of a structure involved in assortment and replication of rotavirus RNA.

FIG. 1.

Localization of NSP5 in rotavirus-infected and transfected cells. MA104 cells were infected with rotavirus SA11 and then analyzed at 2 (A) and 8 (B) h, respectively, postinfection. COS-7 cells were transfected with pcDNA-NSP5 DNA and incubated for 48 h (C and D). One set of transfected COS-7 cultures was treated for 2 h with 0.5 μM okadaic acid before fixation (D). Cells were fixed in paraformaldehyde and incubated with mouse NSP5 antiserum and then with fluorescein-conjugated goat anti-mouse immunoglobulin G. Phase-contrast and immunofluorescence (IF) micrographs are shown.

To examine its localization in the absence of other rotavirus proteins, NSP5 was expressed in monkey COS-7 cells in a transfection experiment. COS-7 and MA104 cells are both susceptible to rotavirus SA11 infection. Virus production proceeded with similar kinetics in the two types of infected cells, and there was no significant difference in the synthesis of NSP5 (see Fig. 7, lanes 7 and 9). In COS-7 cells transfected with pcDNA-NSP5, expression and localization of NSP5 polypeptides were analyzed by indirect immunofluorescence (Fig. 1C and D). NSP5 was uniformly distributed in the cytoplasm of the transfected cells. In agreement with previous observations (29), no inclusions with accumulated NSP5 were detected. To assess the effect of phosphorylation on NSP5 localization, transfected COS-7 cultures were treated with okadaic acid. However, extensive phosphorylation of NSP5 in the presence of this phosphatase inhibitor (see below) did not induce formation of viroplasm-like structures observable by immunofluorescence.

FIG. 7.

Effect of okadaic acid on NSP5 autophosphorylation activity. COS-7 cells were transfected with pcDNA3 or pcDNA-NSP5 DNA and incubated for 48 h. Okadaic acid at 0.5 μM (+) was then added during 3 h (pulse). One set of cultures was subsequently incubated for another 3 h (chase) in DMEM without okadaic acid (−). Cell extracts were prepared, and a portion were used for SDS-PAGE. NSP5 polypeptides were detected by immunoblotting (A). From the remaining cell extracts, NSP5 was immunoprecipitated and incubated with [γ-³²P]ATP in a protein kinase assay. Phosphorylated polypeptides were separated by SDS-PAGE and detected by autoradiography (B). Controls from rotavirus-infected MA104 (lanes 6 and 7) and COS-7 (lanes 8 and 9) cells were included in the experiment. The positions of NSP5 polypeptides (in kilodaltons) in relation to protein size markers are indicated on the right.

Kinetics of synthesis and phosphorylation of NSP5.

To quantify the accumulation of NSP5 polypeptide during rotavirus infection of MA104 cells, the polypeptide was analyzed by immunoblotting. In earlier experiments, we extracted protein from cells by using a buffer containing the nonionic detergent Triton X-100. To determine what proportion of NSP5 remained insoluble in cells lysed by this procedure, the extracts were centrifuged at high speed. Polypeptides in both the supernatant and sedimented fractions were then treated with SDS at 100°C under reducing conditions, resolved by SDS-PAGE, and transferred to a membrane. An antiserum raised against bacterially produced purified NSP5 was used to detect the polypeptide expressed in the rotavirus-infected MA104 cells. As previously described (1, 5, 29), multiple bands of NSP5 polypeptides were found at positions corresponding to 26 and 28 kDa and as a smear migrating at 28 to 35 kDa (Fig. 2). The two major bands at 26 and 28 kDa appeared in samples taken at 4 h postinfection. As the infection proceeded, NSP5 accumulated and material migrating at the 30- to 35-kDa position emerged. Late during infection NSP5 synthesis dropped, probably due to the cytopathic effect. The experiment also showed that most of NSP5 was soluble after extraction with Triton X-100 and that the insoluble fraction of the polypeptide did not change during the course of the infection.

FIG. 2.

Kinetics of NSP5 formation in infected cells analyzed by immunoblotting. MA104 cells were infected with rotavirus SA11 and harvested at the indicated time points postinfection. Cells were lysed with TX buffer, and the insoluble material was sedimented by centrifugation. Both insoluble (pellet) and soluble (supernatant) fractions were boiled in SDS sample buffer, and proteins were separated by SDS-PAGE. After electrophoresis, protein was transferred to a nitrocellulose membrane and incubated with NSP5 antiserum. The positions of NSP5 polypeptides are indicated on the right, and the positions of size markers (in kilodaltons) are indicated on the left.

Synthesis and phosphorylation of NSP5 during infection were further assessed by metabolic labeling with [³⁵S]methionine and ³²P_i for 1 h immediately before cell lysis. Mock- and rotavirus-infected cells were lysed with TX buffer, and soluble protein was incubated with NSP5 antiserum. Immune complexes were precipitated with protein G-Sepharose and subjected to SDS-PAGE. Again, bands with mobilities corresponding to 26, 28, and 30 to 35 kDa were observed (Fig. 3). Synthesis of NSP5 had commenced at 2 h and reached its maximum at 4 to 6 h after infection. At early times, the 26-kDa form predominated (Fig. 3, lanes 4 and 5). Later, the 26-kDa polypeptide seemed to be modified more rapidly. At 6 to 8 h postinfection, the relative amounts of heavier material increased, without a concomitant increase in the rate of synthesis (Fig. 3, lanes 6 and 7). At 8 to 10 h after infection, the synthesis of NSP5 dropped to low levels, together with the synthesis of other viral proteins (data not shown). This observation is in agreement with earlier work on the synthesis of rotavirus protein in infected cells (9, 20). Similar amounts of soluble NSP5 at different times postinfection were detected by immunoblotting (Fig. 2) and precipitated with the anti-NSP5 serum (Fig. 3) except at late times postinfection, when incorporation of radioactivity into newly synthesized polypeptide was very low. The data show that the NSP5 antiserum was efficient in immunoprecipitating all isoforms of the protein.

FIG. 3.

Synthesis and phosphorylation of NSP5 in infected cells, and protein kinase activity of isolated protein. MA104 cells were infected with rotavirus SA11 and harvested at the indicated time points postinfection. Protein was metabolically labeled with [³⁵S]methionine (lanes 1 to 8), or ³²P_i (lanes 9 to 14) for 1 h before extraction. TX buffer extracts of mock (−)- or rotavirus-infected MA104 cells were subjected to SDS-PAGE in a 12% gel, either directly (lanes 1 and 2) or after immunoprecipitation with NSP5 antiserum (lanes 3 to 20). To test protein kinase activity, immunoreactive material immobilized on protein G-Sepharose beads was incubated with [γ-³²P]ATP for 10 min at 22°C in kinase buffer (lanes 15 to 20). Radioactively labeled proteins were visualized by autoradiography. The positions of NSP5 polypeptides are indicated on the right, and the positions of molecular mass markers (in kilodaltons) are indicated on the left.

When ³²P_i was used as the radioactive precursor, we observed phosphorylation of NSP5 at 4 h postinfection, approximately 2 h after the start of its synthesis. Maximal phosphorylation occurred at 6 to 8 h postinfection (Fig. 3). The gap between incorporation of ³⁵S and ³²P suggests that NSP5 was not phosphorylated immediately following its synthesis. Moreover, only a minor part of NSP5 became highly phosphorylated, since the polypeptides recovered at the 30- to 35-kDa position, in relative terms, had incorporated much more phosphate than methionine (Fig. 3). Also immunoreactive material at 30 to 35 kDa represented a small fraction of the whole (Fig. 2).

To establish the precursor-product relationship of the NSP5 forms, a pulse-chase experiment was performed. After a 60-min pulse with [³⁵S]methionine or with ³²P_i, the infected cultures were incubated for 60 min in nonradioactive medium. Analysis of NSP5 showed that the amount of ³⁵S-labeled 26-kDa NSP5 decreased during the chase period. Besides this change, there were no significant differences in the amount or distribution of ³⁵S and ³²P label in NSP5 polypeptides (data not shown). Together the results suggest that 26-kDa NSP5 is a precursor and that the 28- and 30- to 35-kDa forms are modified products with different and relatively stable phosphorylation patterns.

To investigate whether the protein kinase activity of isolated NSP5 was related to its synthesis and phosphorylation in infected cells, extracts of infected cells were mixed with NSP5 antiserum and the immune complexes were isolated on protein G-Sepharose beads. They were then incubated for 10 min at room temperature with [γ-³²P]ATP in protein kinase buffer. The appearance of NSP5 with protein kinase activity in vitro coincided with the formation of highly phosphorylated forms of the polypeptide in cells (Fig. 3, lanes 18 to 20). However, the protein synthesized at 4 h postinfection had a very low autophosphorylation activity in the cell-free system. In addition, the in vitro protein kinase activity remained even at 10 h, when there was practically no synthesis or phosphorylation of NSP5 in vivo. The low activity early in infection suggests that efficient autophosphorylation of NSP5 requires a cofactor(s). Whether this cofactor consists of NSP5 modification or another polypeptide, we do not know. However, it is unlikely that a polypeptide factor would coprecipitate with NSP5 at late but not at early times after infection.

Effect of staurosporine and okadaic acid on NSP5 phosphorylation.

NSP5 synthesized early during rotavirus infection was phosphorylated in the infected cells but had very low autophosphorylation activity in vitro (Fig. 3). This result suggests that cellular protein kinases might be involved in the phosphorylation of NSP5. Analysis of the NSP5 amino acid sequence predicted the presence of several recognition sites of protein kinase C and casein kinase II. Therefore, we tested the effects of staurosporine, DRB, and okadaic acid on NSP5 phosphorylation. Staurosporine is a potent inhibitor of protein kinase C (31) and a wide variety of other enzymes of both the serine/threonine and tyrosine kinase families (21). DRB inhibits casein kinase II (34), and okadaic acid is a specific inhibitor of PP1 and PP2A (7). The latter two enzymes are major cytosolic protein phosphatases of mammalian cells with specificity for phosphoserine and phosphothreonine residues (4).

To investigate the effects of the three enzyme inhibitors on NSP5 phosphorylation, either compound was added to the infected cells at different times postinfection during the last 60 min before harvest. Concentrations of the inhibitors—1.0 mM staurosporine, 0.2 mM DRB, and 0.5 μM okadaic acid—that were just below the cytotoxic level were determined in initial experiments (data not shown). The NSP5 polypeptides extracted from treated cells were analyzed by SDS-PAGE followed by immunoblotting. To detect newly synthesized or newly phosphorylated NSP5, material extracted from cells labeled with [³⁵S]methionine or ³²P_i during 1 h immediately before harvest was immunoprecipitated and then resolved by SDS-PAGE. The amount of immunoreactive material and the radioactivity at the positions corresponding to 26, 28, and 30 to 35 kDa were determined.

The influence of treatment with staurosporine for 1 h on the accumulated amount and isoform distribution of NSP5 was small (Fig. 4), an expected result considering the stability of the protein. Staurosporine had a moderate effect on newly synthesized polypeptides and induced a strong decrease of their phosphorylation. The compound had similar effects at early and late times after infection, and the inhibition of ³²P_i incorporation was more obvious than the effect on electrophoretic mobility shift. Thus, we conclude that the effect of staurosporine was to reduce the number of phosphate groups per polypeptide chain to a somewhat lower value. At the same time, the inhibitor had a much stronger effect on cellular protein phosphorylation. Incorporation of ³²P_i into cellular protein decreased by 50 to 60% (data not shown). DRB had no effect on NSP5 phosphorylation in infected cells, and therefore data are not shown.

FIG. 4.

Effects of staurosporine and okadaic acid on NSP5 synthesis and phosphorylation. Cells were infected with rotavirus SA11 and at different time points after infection incubated with 1.0 mM staurosporine or 0.5 mM okadaic acid for 60 min. Cells were lysed, and the soluble fraction was subjected to SDS-PAGE. The resolved proteins from one set of cell extracts were analyzed by immunoblotting using a specific NSP5 antiserum. Chemiluminescence corresponding to NSP5 was then quantitated (top row). Two other sets of infected cell cultures were labeled, one set with [³⁵S]methionine and one with ³²P_i, during the 60-min period before harvest. NSP5 extracted from these cells was immunoprecipitated and resolved by SDS-PAGE. ³⁵S (middle row) and ³²P (bottom row) radioactivity corresponding to NSP5 was quantitated with a molecular imager. The areas indicate the quantitated material migrating at 26 kDa (black), 28 kDa (dark grey), and 30 to 35 kDa (light grey). PDUnits, photon decay units.

Treatment of rotavirus-infected cells with the protein phosphatase inhibitor okadaic acid had little effect on the accumulated amount of different isoforms of the protein. The analysis of the newly synthesized polypeptides showed a marginal decrease in the 26-kDa isoform and an increase of the 28-kDa isoform of NSP5 (Fig. 4). The 30- to 35-kDa bands in the gel also became more abundant compared to the amounts extracted from untreated cells. Particularly, the ³²P incorporation into 30- to 35-kDa forms of NSP5 was increased. Thus, the phosphorylation of NSP5 appeared to depend on a balance between addition and removal of phosphate residues.

Synthesis and phosphorylation of NSP5 in transfected cells.

To examine the synthesis and modifications of NSP5 in the absence of other viral protein, the polypeptide was transiently expressed in COS-7 cells after transfection with the recombinant plasmid pcDNA-NSP5. Transfected COS-7 cells were incubated for 48 h before protein was pulse-labeled with [³⁵S]methionine or ³²P_i for 3 h. Extracted NSP5 was immunoprecipitated and analyzed by SDS-PAGE. A 26-kDa polypeptide labeled with ³⁵S or ³²P was synthesized in cells transfected with pcDNA-NSP5 (Fig. 5, lane 3). However, although this polypeptide was stable during a 3-h chase period, more slowly migrating NSP5 polypeptides did not appear (Fig. 5, lane 4). When the transfected cells were incubated with okadaic acid, to inhibit protein phosphatases during the labeling period, NSP5 was apparently protected from dephosphorylation and shifted in size to 28 to 35 kDa (Fig. 5, lane 6). When protein was labeled in the absence of inhibitor and the label was chased in the presence of okadaic acid, the shift to the 30- to 35-kDa position was even more pronounced (Fig. 5, lane 5). Conversely, some of the 30- to 35-kDa material formed in the presence of okadaic acid disappeared when the pulse-chase was performed in the absence of the inhibitor (Fig. 6, lanes 6 to 8).

FIG. 5.

Effect of okadaic acid on the synthesis and phosphorylation of NSP5 in transfected cells. COS-7 cells were transfected with pcDNA3 or pcDNA-NSP5 DNA and incubated for 48 h. Protein was labeled in one set of transfected cultures with [³⁵S]methionine (top) and in a second set with ³²P_i (bottom) for 3 h before harvest (pulse). In the indicated cases, the pulse was followed by a 3-h chase period without the radioactive compounds. The pulse-chase was done in the presence (+) or absence (−) of 0.5 μM okadaic acid. NSP5 was immunoprecipitated from cytosolic extracts and resolved by SDS-PAGE. After electrophoresis, the gel was autoradiographed. The positions (in kilodaltons) of NSP5 forms in relation to size markers are indicated on the right.

FIG. 6.

Effects of protein kinase inhibitors on NSP5 phosphorylation in transfected cells. COS-7 cells were transfected with pcDNA3 or pcDNA-NSP5 DNA and incubated for 48 h. Okadaic acid at 0.5 μM was added to one set of cultures. Thirty minutes later, 1.0 mM staurosporine or 0.2 mM DRB was added to the same cultures, and incubation was continued for 3 h. During this final 3-h period before harvest, the cells were also labeled with ³²P_i. NSP5 was immunoprecipitated from cell extracts and analyzed by SDS-PAGE. The positions of NSP5 polypeptides, as visualized by autoradiography, are indicated on the right, and the positions of molecular mass markers (in kilodaltons) are indicated on the left.

The detection of NSP5 polypeptides with an authentic size distribution suggests that no other viral proteins were required for phosphorylation of the NSP5 polypeptide.

Effect of protein kinase inhibitors on NSP5 expression in transfected cells.

Protein kinase inhibitors were added during labeling with ³²P_i to investigate their effect on NSP5 phosphorylation in COS-7 cells transfected with pcDNA-NSP5. The cells were treated with 0.5 μM okadaic acid before and during the incubation with protein kinase inhibitors, to allow the formation of all NSP5 isoforms (Fig. 6, lane 3). NSP5 phosphorylation decreased significantly when the cells were treated for 3 h with 1.0 mM staurosporine (Fig. 6, lane 4). The inhibition of NSP5 phosphorylation appeared to be stronger in the transfected cells than in cells infected with rotavirus (Fig. 4).

The adenosine analog DRB did not inhibit NSP5 phosphorylation in the transfected cells (Fig. 6, lane 5), consistent with the absence of effect in infected cultures. These results indicate that casein kinase II is not involved in NSP5 phosphorylation. In contrast, DRB partially inhibited NSP5 autophosphorylation in vitro (see below).

Effect of NSP5 phosphorylation and protein kinase inhibitors on autophosphorylation in vitro.

Since the NSP5 protein expressed in COS-7 cells was phosphorylated independently of other viral proteins, we investigated whether it maintained this ability in vitro. To obtain highly phosphorylated NSP5 polypeptides, protein phosphatase activity was inhibited with okadaic acid as described above. Polypeptides from cell extracts resolved by SDS-PAGE were first analyzed by immunoblotting, to verify that NSP5 was expressed and modified in the COS-7 cells. As a control, NSP5 produced during a 6-h infection of MA104 and COS-7 cells by rotavirus SA11 was included in the experiment (Fig. 7A). To test protein kinase activity in vitro, immunoprecipitated NSP5 protein immobilized on protein G-Sepharose beads was incubated in kinase buffer containing 10 μCi of [γ-³²P]ATP. The reaction products were separated by SDS-PAGE and identified by autoradiography. Figure 7B shows that the 26-kDa NSP5 produced in transfected cells had a weak autophosphorylation activity. Still, there was a large amount of 26-kDa NSP5, as judged by the amount of protein detected in the immunoblot (Fig. 7A). When the same experiment was performed with the NSP5 polypeptides isolated from transfected cells treated with okadaic acid, or isolated from rotavirus-infected MA104 or COS-7 cells, all forms of NSP5 were modified (Fig. 7, lanes 4, 7, and 9). However, larger forms of the polypeptide contained most of the phosphate. When the transfected cells were incubated with okadaic acid for 3 h and then removed from phosphatase inhibition for another 3 h before harvest, the amount of highly phosphorylated NSP5 decreased, with a concomitant increase of 26- to 28-kDa forms (Fig. 5). The isolated NSP5 showed a parallel distribution of ³²P incorporation from autophosphorylation in vitro (Fig. 7, lane 5). Hence, this activity of NSP5 in cells and in vitro appeared to be related to the number of phosphate residues already incorporated.

To further explore the possibility that cellular protein kinases are involved in NSP5 phosphorylation, particularly early in infection, the effects of protein kinase inhibitors on isolated NSP5 were tested. Cells were extracted at different times after rotavirus infection, and NSP5 was isolated by immunoprecipitation. The immune complexes were incubated with 1.0 mM staurosporine or 0.2 mM DRB for 30 min. [γ-³²P]ATP and MnCl₂ were then added, and incubation was continued for 20 min. Analysis of ³²P incorporation into polypeptides separated by SDS-PAGE showed (Fig. 8) that staurosporine had little effect. In contrast, DRB decreased the autophosphorylation of NSP5 in this cell-free system by more than 50%. The effect of the casein kinase II inhibitor DRB was unexpected, since the compound did not inhibit NSP5 phosphorylation in cells.

FIG. 8.

Autophosphorylation of isolated NSP5 in the presence of protein kinase inhibitors. NSP5 isolated by immunoprecipitation from MA104 cells at different times after rotavirus infection was incubated for 30 min in 50 mM HEPES (pH 7.5)–1.0 mM DTT alone (□) or containing 1.0 mM staurosporine (⧫), or 0.2 mM DRB (○). Protein kinase activity was then assayed in a 20-min reaction by addition of 10 mM MnCl₂ and 5.0 μM ATP containing 10 μCi of [γ-³²P]ATP. Polypeptides were separated by SDS-PAGE. Radioactivity at the position of NSP5 polypeptides was determined with a molecular imager.

DISCUSSION

NSP5 accumulates together with VP6 in viroplasms (28). The work presented here shows that the viroplasms gradually increased in size during infection and that NSP5 remained in these structures (Fig. 1), suggesting that it forms part of a scaffold for the early steps of virion morphogenesis. However, NSP5 alone cannot be the viroplasm organizer, since expression of the polypeptide in the absence of the other rotavirus gene products did not lead to formation of viroplasm-like inclusions (Fig. 1). Instead, material reactive with NSP5 antibodies formed a diffuse staining in the cytoplasm, even when the posttranslational modification of the polypeptides apparently was the same as in infected cells (Fig. 5 and 7).

The present experiments provide further evidence that phosphorylation of NSP5 is not a determinant for the localization of the polypeptide. Labeling with [³⁵S]methionine and immunoblot analysis showed NSP5 formation at 2 h postinfection. The immunoreactive material was located in viroplasms (Fig. 1). Phosphorylation of NSP5 as determined by labeling with ³²P_i became detectable at 2 to 4 h after infection, approximately 2 h after synthesis of the polypeptide. Thus, phosphorylation of NSP5 did not appear to be necessary for its localization in viroplasms. Similarly, rotavirus NSP2 expressed in the absence of other viral proteins did not accumulate in cytoplasmic inclusions, as it does in infected cells (14). The establishment of the viroplasm structure may require interaction between several viral proteins (28).

All phosphorylated forms of NSP5 (26, 28, and 30 to 35 kDa) accumulating in viroplasms were observed only in cells infected with rotavirus. In COS-7 cells expressing NSP5 from a plasmid DNA vector, only the 26-kDa form appeared, and it had a diffuse cytoplasmic distribution. However, extensive phosphorylation of NSP5, leading to a series of phosphorylated polypeptides as in infected cells, was achieved without expression of the other rotavirus proteins. The critical factor appeared to be removal from NSP5 of newly added phosphate groups by phosphatases. In the presence of the phosphatase inhibitor okadaic acid, the phosphorylated NSP5 forms were stable for several hours.

Okadaic acid increased the phosphorylation of NSP5 also in rotavirus-infected cells. However, the relative effect was much less than with NSP5 expressed from a vector. This result indicates that the viroplasms are sites protected from the activities of cellular phosphatases and possibly other cellular enzymes.

In a previous study of NSP5 processing (32), protein was pulse-labeled with [³⁵S]methionine for 10 min and then chased for 40 min with unlabeled methionine. During the chase period, most of the labeled 26-kDa polypeptide was transferred to the 28-kDa position. We extended that study by analyzing the phosphorylation of NSP5 at different times postinfection. At all time points after infection, most of the 26-kDa polypeptide labeled with [³⁵S]methionine during 60 min was processed to the 28-kDa, and some to the 30- to 35-kDa, form during the following hour. In contrast, NSP5 that had incorporated ³²P label during a 60-min pulse remained essentially stable. Neither the amount of radioactivity nor the electrophoretic mobility of the labeled polypeptide changed during a 60-min chase period, indicating that a number of stable phosphorylated NSP5 forms exist. Apparently, 26-kDa material is processed to both 28- and 30- to 35-kDa forms. However, at least the majority of 28-kDa NSP5 is not an intermediate used for further phosphorylation.

Purified NSP5 isolated from rotavirus-infected cells, or other eukaryotic or bacterial cells expressing the protein, is capable of autophosphorylation (1, 5, 29). However, the activity of the purified protein isolated from bacteria has low activity and does not add more than one phosphate residue per polypeptide chain (5, 29). Thus, at present it is unclear whether cellular protein kinases also participate in NSP5 phosphorylation. Considering that the autophosphorylation activity of 26-kDa NSP5 is quite low and that there are several recognition sites of protein kinase C and casein kinase II in NSP5 (5), these two enzymes might participate in the phosphorylation of the polypeptide. Therefore, the effects of specific inhibitors were investigated. Staurosporine was first described as a potent inhibitor of protein kinase C, and inhibition by this compound has been considered diagnostic of the involvement of protein kinase C (31). However, the inhibitor was later shown to block a wide variety of protein kinases (8, 13, 21, 22, 25, 33), including several with tyrosine specificity (10, 21, 24, 30). There is also a group of protein kinases including casein kinase II (12, 21) that is relatively refractory to staurosporine inhibition.

Although the addition of staurosporine to infected cells produced a general decrease of the phosphorylation of all NSP5 forms, we did not observe a shift in the ratio of different forms. In rotavirus-infected cells, we observed a reduction of approximately 50% in the formation of 28- and 30- to 35-kDa forms, as measured by ³²P incorporation. Phosphorylation of NSP5 protein expressed in COS-7 cells was also partially inhibited by staurosporine, whereas the autophosphorylation of the purified polypeptide was not. Together, these results show that part of the NSP5 phosphorylation in cells is produced by cellular protein kinases.

The casein kinase II inhibitor DRB (34) did not influence the phosphorylation of NSP5 when it was added to cell cultures infected with rotavirus or transfected with pcDNA-NSP5. In contrast, DRB partially inhibited autophosphorylation of immunoprecipitated NSP5 in vitro. We do not take this result as an indication that casein kinase II and NSP5 are coimmunoprecipitated. It is more likely that DRB is a weak inhibitor of NSP5 activity, but did not reach an inhibitory intracellular concentration when the compound was added to cultures. Analysis of the casein kinase II and NSP5 amino acid sequences did not reveal any similarity indicating that the active sites are related.

In the early phase of rotavirus infection, NSP5 becomes phosphorylated approximately 2 h after its synthesis (Fig. 3). This delay was probably not caused by susceptibility to phosphatases before NSP5 was translocated to viroplasms. After addition of okadaic acid to infected cells, the phosphorylation of NSP5 was still delayed approximately 2 h, although ³²P incorporation into the 28- and 30- to 35-kDa forms was increased (Fig. 4). Moreover, the failure to induce full phosphorylation of the 26-kDa NSP5 precursor early in infection, when phosphatases were inhibited, probably reflects a low initial autophosphorylation activity of the polypeptide.

NSP5 immunoprecipitated from the infected cells did not show any autophosphorylation activity until 6 h postinfection, when highly phosphorylated forms of the protein had emerged in the infected cells (Fig. 3). Conversely, the in vitro protein kinase activity remained in NSP5 isolated at 10 h postinfection, when synthesis and phosphorylation of the polypeptide had ceased in the cells. These result suggests that the enzyme activity is regulated. At least part of the regulation seemed to be achieved by modification of the NSP5 polypeptide. Phosphorylation itself appears to be one factor of importance for the autophosphorylation activity in vitro. First, the 35-kDa form had the highest specific activity (Fig. 3), and second, NSP5 expressed from a vector had significant autophosphorylation activity only when it remained phosphorylated by inhibition of cellular phosphatases (Fig. 7).

The present study does not shed light on the function of NSP5 in rotavirus infection. No phosphorylation of other virus-encoded polypeptides has been identified. We are investigating whether cellular polypeptides are substrates of NSP5. So far, analysis by two-dimensional PAGE has not revealed any cellular phosphopeptides that appeared after expression of NSP5 in transfected COS-7 cells. In infected cells, NSP5 is produced in unusually large amounts to serve solely a catalytic function. Thus, it is possible that the highly phosphorylated polypeptide has a structural function in the organization of viroplasms.