Previous Article | Next Article ![]()
Journal of Virology, September 2002, p. 9545-9550, Vol. 76, No. 18
0022-538X/02/$04.00+0 DOI: 10.1128/JVI.76.18.9545-9550.2002
Copyright © 2002, American Society for Microbiology. All Rights Reserved.
Veterinary Molecular Biology, Montana State University, Bozeman, Montana 59717
Received 11 March 2002/ Accepted 12 June 2002
| ABSTRACT |
|---|
|
|
|---|
| TEXT |
|---|
|
|
|---|
The rotavirus segmented double-stranded RNA genome encodes six structural proteins (VP) and six nonstructural proteins (NSP) (reviewed in reference 17). The structural proteins VP1, VP2, VP3, VP4, VP6, and VP7 are well characterized in terms of their antigenic, structural, and biochemical properties. The functions of the rotavirus nonstructural proteins NSP1 to NSP6 are less well defined with regard to the roles that these proteins play in the rotavirus replication cycle. Intriguing functions have recently been described for some. NSP3 binds the 3' consensus sequence of viral mRNAs (37) and acts as a functional analog of poly(A) binding protein through its interaction with eIF4GI (36). NSP4 is both an intracellular glycoprotein receptor for maturating rotavirus particles that bud through the endoplasmic reticulum (3, 32) and a viral enterotoxin that induces diarrhea in mice in the absence of any other viral protein (5). Functions of the remaining nonstructural proteins, NSP1, NSP2, NSP5, and NSP6, have been proposed based predominately on biochemical properties and activities of recombinant proteins (reviewed in reference 17).
NSP1 displays several interesting properties that warrant investigation. NSP1 has a calculated molecular weight of approximately 54,000 and is the least conserved protein encoded by the rotavirus genome when NSP1s of different strains are compared (23, 34). The N terminus contains a conserved zinc finger motif that binds zinc and viral mRNA in vitro (10, 22). Immunofluorescent staining showed NSP1 to be localized throughout the cytoplasm, in contrast to most other rotavirus proteins, which concentrate in viroplasms (24). NSP1 is also found associated with the cytoskeleton when analyzed by subcellular fractionation (24). NSP1 apparently is not required for rotavirus replication because strains with rearrangements in gene 5 that result in the synthesis of truncated NSP1 have been isolated from animals and from both immune-deficient and immune-competent children (1, 7, 15, 25, 35, 44, 45). Each of the described strains replicates in cell culture to titers close to those of their wild-type counterparts, but they yield small- to minute-plaque phenotypes. These observations suggest that NSP1 plays a role in regulating the efficiency of viral gene expression or in modulating host cell responses. We addressed the second possibility by constructing an MA104 cell cDNA library into the activation domain vector of the Matchmaker 3 yeast two-hybrid interaction trap (Clontech). We screened the library with NSP1 as bait to identify candidate partners of NSP1 that would provide clues to its function in rotavirus-infected cells.
Polyadenylated mRNA was isolated from MA104 cells in exponential growth phase with TriZol reagent (Life Technologies) and purified by column chromatography with the mRNA purification system from Amersham Pharmacia. cDNA was synthesized according to the protocol supplied with reagents from Clontech, with modifications. Briefly, 5 µg of mRNA was reverse transcribed with Moloney murine leukemia virus reverse transcriptase and an oligo(dT)-XhoI linker primer. Double-stranded cDNA was generated with DNA polymerase I, and the ends were blunted with Pfu DNA polymerase (Stratagene). Double-stranded cDNA fragments ligated to EcoRI adapters were size fractionated with ChromaSpin 400 columns (Clontech) and ligated into the Matchmaker pGADT7 vector prepared by EcoRI-XhoI digestion. Transformations that were sufficient to obtain 2.5 x 106 CFU were performed. The colonies were collected into Luria broth, and DNA was purified by two sequential centrifugations through CsCl density gradients. The library was quantified by measuring the absorbance at 260 nm, aliquoted, and stored at -80°C.
Gene 5 encoding NSP1 of bovine rotavirus strain B641 was cloned by reverse transcriptase PCR (unpublished data). The nucleotide sequence of B641 gene 5 is 98% identical to that of gene 5 of bovine strain RF (9). A plasmid containing gene 5 of murine strain EW was a generous gift of Harry Greenberg, Stanford University School of Medicine. Gene 5 cDNAs from B641 and EW were cloned into pGBKT7 by standard cloning techniques to generate pGBK-bNSP1 (bovine) and pGBK-mNSP1 (murine), respectively. All yeast transformations followed the lithium acetate-polyethylene glycol (PEG) method described by Gietz and Woods (19). The DNA binding domain vector pGBKT7 carries the TRP1 nutritional selection marker, and colonies containing bait plasmids were isolated by culture on synthetic complete medium lacking tryptophan (SC-W medium). The Matchmaker activation domain vector pGADT7 carries the LEU2 nutritional marker for selection of yeast.
Positive two-hybrid interactions were scored by determining colony growth on SC medium without leucine and tryptophan (SC-L-W medium) and activation of the reporter genes HIS3, ADE2, and MEL1. Reporter gene activation was indicated by growth in the absence of histidine and adenine and by the ability to metabolize the chromogenic substrate X-
-Gal(
-5-bromo-4-chloro-3-indolyl-ß-D-galactopyranoside; ICN).pGBK-bNSP1 yeast cells (6 x 109) cultured in SC-W liquid medium were transformed with 120 µg of the cDNA library by the lithium acetate-PEG procedure. Transformants were plated on SC medium without leucine, tryptophan, histidine, and adenine (SC-L-W-H-A medium) and cultured for 2 to 4 days at 30°C. Colonies measuring 2 to 4 mm in diameter were restreaked on SC-L-W-H-A plus X-
-Gal medium and again incubated for 2 to 4 days at 30°C. Plasmid DNA was isolated by electroporation into DH10B cells (Life Technologies) and cultured on antibiotic selection media.
NSP1 interacts with IRF-3.
We performed a two-hybrid interaction screen with bovine NSP1 as bait to identify candidate cellular partners of NSP1 that would shed light on the function of this protein in rotavirus-infected cells. We identified several colonies that grew on defined media that selected for the presence of both transcription activation and DNA binding domain plasmids and for activation of reporter gene expression. We isolated six library clones by transforming them into bacteria and sequenced the activation domain plasmid inserts. A BLASTX search with the nucleotide sequence of each of these clones yielded matches with human interferon (IFN) regulatory factor 3 (IRF-3). The cDNA clones ranged from 0.9 to 1.2 kb. The longest cDNA corresponded to a sequence from nucleotide 110 to the 3' polyadenylate tail, according to the published sequence for human IRF-3 (4) (GenBank accession no. NM 001571). IRF-3 is a transcription factor that responds to virus infection by inducing the expression of the genes for IFN-
and IFN-ß (4, 21, 26, 31). Subsequent IFN-stimulated gene products drive a cell to an antiviral state. Thus, identification of IRF-3 as a partner of NSP1 was relevant and warranted further study.
We performed a number of control experiments to ensure that the defined interaction was not an artifact or a false-positive interaction (Fig. 1). We retransformed pGAD-IRF-3 into yeast containing pGBK-bNSP1 and plated transformants on SC-L-W medium. Colonies capable of growth on this medium were then streaked onto SC-L-W-H-A medium containing X-
-Gal. Only yeast cells containing both plasmids were capable of growth on selective medium (Fig. 1A). pGAD-IRF-3 was negative for autoactivation, as was evidenced by its failure to activate reporter gene expression in the absence of pGBK-bNSP1. We also tested potential interactions between (i) bovine NSP1 and an irrelevant protein (simian virus 40 T antigen) in the activation domain plasmid, (ii) IRF-3 and an irrelevant protein in the binding domain plasmid (v15), and (iii) IRF-3 with an empty bait plasmid (Fig. 1B and data not shown). These tests were also negative.
|
NSP1 synthesized in rotavirus-infected cells binds IRF-3. The ability of NSP1 synthesized in infected cells to interact with IRF-3 was examined by a glutathione S-transferase (GST) pull-down assay as previously described (29). MA104 cells were infected with trypsin-activated B641 at a multiplicity of infection of 10 in the presence of 5 µg of actinomycin D per ml. Fifty microcuries of 35S-trans label (>1,000 Ci/mmol; ICN) per ml was added 2 h postinfection, and the infection was allowed to proceed for an additional 4 h. Cells were harvested in lysis buffer (50 mM Tris-Cl [pH 8.0], 15 mM NaCl, 140 mM KCl, and 2% NP-40). Radiolabeled lysates from infected or mock-infected cells were then incubated with glutathione-Sepharose 4B beads bound to either GST or GST-IRF-3. Interacting proteins were eluted with 10 mM reduced glutathione, resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and visualized by autoradiography. Figure 2 (left panel) shows that NSP1 bound GST-IRF-3 but not GST. The identity of the band migrating at the predicted molecular weight of NSP1 in the GST-IRF-3 eluate was confirmed by Western blotting with an anti-NSP1 monoclonal antibody (right panel). Rotavirus proteins VP7 and NSP3 were also retained by GST-IRF-3 but not GST. Whether these proteins are bound by IRF-3 or by NSP1 complexed to IRF-3 is not known. NSP1 binds the 5' end of viral mRNA, and NSP3 binds the 3' end (22, 37). NSP3 interacts with translation initiation factor eIF4GI, resulting in circularization of the viral mRNA in polysomes (36). A direct interaction between NSP1 and NSP3 was shown in a yeast two-hybrid assay (20) but thus far has not been reported to occur in infected cells. It is possible that NSP3 is brought down in this assay through interaction with NSP1. Alternatively, NSP3 may be pulled down through a viral mRNA bound to NSP1. The detection of VP7 in the GST-IRF-3 eluate is more difficult to interpret. It may be that this protein interacts with NSP1 in infected cells, but these interactions have not been investigated or described. Numerous attempts to coimmunoprecipitate NSP1 and IRF-3 were unsuccessful, indicating that the interaction may be transient or not sufficiently stable to survive the conditions of the immunoprecipitation. The data presented in Fig. 2 show that NSP1 synthesized in a native rotavirus infection interacted with IRF-3, and thus NSP1 and IRF-3 are not binding partners unique to the yeast two-hybrid system.
|
Schematic diagrams of the bovine NSP1 mutants used in the following analyses are shown in Fig. 3A. NSP1 deletion mutants were constructed by PCR and by standard cloning techniques. Deletion mutant
62 is missing the first 62 amino acids. The fragment was amplified by PCR with oligonucleotides 5'-CGCGGATCCAGATCAATAACCATTACAAAGTGTGG-3'(positive) and 5'-CCGCTCGAGGGTTCAACATCTGAAAGTTC-3' (negative) and the full-length B641 gene 5 clone as the template. BamHI and XhoI sites for cloning are underlined. Deletion mutant
326 is missing the N-terminal 326 amino acids. This mutant was constructed by PCR amplification with the negative-sense primer described above and with the positive-sense primer 5'-CGCGGATCCAGATCAATAACCATTACAAAGTGTGG-3'. Deletion mutant
290 contains only the N-terminal 200 amino acids and was constructed by PCR amplification with primers 5'-CGCGGATCCTGATGGCGACTTTTAAGGAC-3' (positive) and 5'-CCGCTCGAGTATTACTGGTAAGTGTGACAT-3' (negative). PCR fragments were cloned into the pGBKT7 vector. All constructs were sequenced prior to transformation into yeast to ensure that the correct reading frame had been maintained. Each mutant NSP1 bait construct was transformed into AH109 cells alone or into AH109 cells containing pGAD-IRF-3. Transformants were then scored for interaction based on the ability to grow on nutrient-deficient medium.
|
62 was constructed to ascertain the contribution of the N-terminal zinc finger to the interaction between NSP1 and IRF-3. Figure 3C shows that deletion of the N-terminal 62 amino acids abolished the interaction with IRF-3, as evidenced by the inability of yeast cotransformed with
62 and IRF-3 to grow on selective medium. A second construct (
326) contained only the C-terminal 164 amino acids. This mutant scored positive for interaction with IRF-3, though the colonies consistently grew more slowly than those with full-length NSP1 (Fig. 3C). A construct containing the analogous C-terminal residues of EW NSP1 (
327) also interacted with IRF-3 (data not shown). A final deletion mutant,
290, consisted of the N-terminal 200 amino acids and contained the zinc finger. Surprisingly, this mutant failed to interact with IRF-3. The results of the deletion analyses were not straightforward, and they suggested that structure played a significant role in the IRF-3 interaction. Therefore, we performed oligonucleotide-directed site mutagenesis (QuikChange; Stratagene) to disrupt the N-terminal zinc finger so that the overall structure of the protein would be maintained as much as possible. In one mutant the cysteine at position 54 was replaced with alanine (C54A mutant). In a second mutant, the cysteines at both position 54 and position 57 were replaced with alanines (C54A/57A mutant) (Fig. 3B). Neither of these mutants interacted with IRF-3 (Fig. 3D).
Taken together, the mutagenesis data suggest that proper folding and the structure of NSP1 are important for interaction with IRF-3. The inability of mutant
62, the C54A mutant, and the C54A/57A mutant to interact with IRF-3 and the reduced growth rate of
326 suggested that the zinc finger was important for binding. However, this domain was not sufficient for binding because the
290 mutant, which contained the entire zinc finger, did not interact with IRF-3. The interaction of the C-terminal 164 amino acids of NSP1 with IRF-3 (
326 and EW
327) and the lack of interaction between
290 and IRF-3 suggest that the interaction is not an artifact mediated simply by a zinc finger domain that may be present in a wide variety of proteins.
The data lead us to propose that at least one point of contact with IRF-3 resides in the C-terminal domain of NSP1. The inability of mutant
62 or the Zn2+ finger point mutants to interact with IRF-3 in yeast might be explained by a structural inhibition of C-terminal binding in the absence of the Zn2+ finger. Alternatively, proper folding of full-length NSP1 may require formation of the zinc finger, and thus the authentic conformation may be disrupted even by the point mutations. Computer predictions of the secondary structures of the C-terminal
164 amino acids of B641 and EW NSP1 show conserved regions of
-helices connected by random coils. Consistent with the mutagenesis data, it is likely that this structure is critical, and the interaction between IRF-3 and NSP1 may be driven primarily by the
-helical content of the C terminus of NSP1. Secondary-structure analysis of three viral proteins known to interact with or modulate the function of IRF-3 (see below), influenza virus NS1 (43), adenovirus E1A (26), and human papillomavirus E6 (39) also predicts
40%
-helices and 40% random coils, suggesting that these structures may be important. Further mutational analyses that disrupt discrete structures in the C terminus are required to refine domains important in the interaction.
Role of NSP1-IRF-3 interaction in rotavirus-infected cells.
The role of IFN in rotavirus infection is unclear, as the data are conflicting and dependent on the species studied. Both humans and animals infected with rotavirus have serum IFN-
and excrete IFN-
in stool (11, 14, 27). Calves treated with recombinant human IFN-
2 were protected from diarrhea induced by tissue culture-adapted bovine rotavirus (42), whereas piglets were not (28). The colon adenocarcinoma cell line HT-29 shows weak or no IFN-ß response following rotavirus infection, although evidence for IFN-induced gene expression was apparent (38). HT-29 and CaCo-2 cells become resistant to rotavirus infection when they are pretreated with IFN-
(6). Despite evidence for a protective effect of IFN, studies in mice which were administered exogenous IFN-
/ß or mice lacking the IFN receptor (IFN-
/ß receptor-/-) showed no differences in disease severity or duration compared to controls (2). These data suggested that the IFN response was not an important nonimmune mediator of protection from rotavirus disease in mice. Collective interpretation of these data is difficult because of the wide range of rotavirus strains, cell lines, species of virus, means of cell isolation, methods of virus adaptation, and animal models used. Given that the cellular IFN response is promiscuous in reaction to virus infection or exposure to extracellular pathogens, it is difficult to accept the idea that the IFN response is not important in rotavirus infection. In fact, a recent analysis of global gene expression in rotavirus-infected CaCo-2 cells showed up-regulation of IFN-stimulated genes 16 h postinfection (12). It is therefore very likely that viral interference with the cellular IFN response is not absolute and that it primarily serves to diminish the magnitude or effectiveness of the response. There are no reports of mechanisms that rotaviruses may have evolved to ameliorate the IFN response to promote infection and spread. The data presented here suggest that NSP1 may fulfill such a function in the rotavirus replication cycle. In support of this hypothesis, the small-plaque phenotype displayed by rotavirus strains that do not encode NSP1 may be explained if the IFN response in cells infected with these viruses is not down-regulated.
We identified IRF-3 as a cellular partner of NSP1. IRF-3 is a 427-amino-acid transcription factor that is constitutively expressed in all cell types (4). IRF-3 resides latently in the cytoplasm, and upon infection by a virus or exposure of cells to double-stranded RNA, IRF-3 is phosphorylated, dimerizes, and translocates to the nucleus, where it complexes with transcription coactivators such as CBP/p300 (8, 30, 41). We predict that the consequence of the NSP1-IRF-3 interaction in rotavirus-infected cells is the functional inhibition of IRF-3 and down-regulation of the innate cellular IFN response. Precedent for this prediction is provided by several viral systems. The adenovirus E1A protein down-regulates IRF-3-induced transcription by competing with IRF-3 for binding to CBP/p300 (13, 26). A second mechanism of IRF-3 inhibition is displayed by human papillomavirus type 16. The E6 protein of human papillomavirus type 16 binds IRF-3 directly and down-regulates induction of IFN-ß (39). Finally, the NS1 protein of influenza virus inhibits nuclear translocation of IRF-3 in infected cells (43). Clearly, IRF-3-mediated gene induction is an important cellular response to infection, and viruses apparently have evolved mechanisms to thwart such a defensive response. Ongoing studies are directed toward identification of mechanisms by which NSP1 may inhibit the action of IRF-3.
| ACKNOWLEDGMENTS |
|---|
We extend appreciation to Michael White and Jay Radke for help and advice in cDNA library construction and to Jean Cohen, INRA, Jouy-en-Josas, France, for the gift of the anti-NSP1 monoclonal antibody.
| FOOTNOTES |
|---|
This article has been assigned journal number 2002-11 by the MAES. ![]()
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| J. Bacteriol. | Mol. Cell. Biol. | Microbiol. Mol. Biol. Rev. |
|---|
| Clin. Vaccine Immunol. | ALL ASM JOURNALS |
|---|