Elimination of Phosphorylation Sites of Semliki Forest Virus Replicase Protein nsP3
2001; Elsevier BV; Volume: 276; Issue: 8 Linguagem: Inglês
10.1074/jbc.m006077200
ISSN1083-351X
AutoresHelena Vihinen, Tero Ahola, Minna Tuittila, Andres Merits, Leevi Kääriäinen,
Tópico(s)Viral Infections and Outbreaks Research
ResumonsP3 is one of the four RNA replicase subunits encoded by alphaviruses. The specific essential functions of nsP3 remain unknown, but it is known to be phosphorylated on serine and threonine residues. Here we have completed mapping of the individual phosphorylation sites on Semliki Forest virus nsP3 (482 amino acids) by point mutational analysis of threonine residues. This showed that threonines 344 and 345 represented the major threonine phosphorylation sites in nsP3. Experiments with deletion variants suggested that nsP3 itself had no kinase activity; instead, it was likely to be phosphorylated by multiple cellular kinases. Phosphorylation was not necessary for the peripheral membrane association of nsP3, which was mediated by the N-terminal region preceding the phosphorylation sites. Two deletion variants of nsP3 with either reduced or undetectable phosphorylation were studied in the context of virus infection. Cells infected with mutant viruses produced close to wild type levels of infectious virions; however, the rate of viral RNA synthesis was significantly reduced in the mutants. A virus totally defective in nsP3 phosphorylation and exhibiting a decreased rate of RNA synthesis also exhibited greatly reduced pathogenicity in mice. nsP3 is one of the four RNA replicase subunits encoded by alphaviruses. The specific essential functions of nsP3 remain unknown, but it is known to be phosphorylated on serine and threonine residues. Here we have completed mapping of the individual phosphorylation sites on Semliki Forest virus nsP3 (482 amino acids) by point mutational analysis of threonine residues. This showed that threonines 344 and 345 represented the major threonine phosphorylation sites in nsP3. Experiments with deletion variants suggested that nsP3 itself had no kinase activity; instead, it was likely to be phosphorylated by multiple cellular kinases. Phosphorylation was not necessary for the peripheral membrane association of nsP3, which was mediated by the N-terminal region preceding the phosphorylation sites. Two deletion variants of nsP3 with either reduced or undetectable phosphorylation were studied in the context of virus infection. Cells infected with mutant viruses produced close to wild type levels of infectious virions; however, the rate of viral RNA synthesis was significantly reduced in the mutants. A virus totally defective in nsP3 phosphorylation and exhibiting a decreased rate of RNA synthesis also exhibited greatly reduced pathogenicity in mice. Sindbis virus Semliki Forest virus nonstructural protein baby hamster kidney plaque-forming unit phosphate-buffered saline postinfection casein kinase II polyacrylamide gel electrophoresis amino acid(s) 4-morpholineethanesulfonic acid Phosphorylation and dephosphorylation have been recognized as major processes by which protein function is regulated. A wide range of proteins display phosphorylation state-dependent activity, including proteins involved in signal transduction, transcription, and the cell cycle. In the field of RNA virus replication, the phosphoprotein P of negative-strand RNA viruses with unsegmented genomes such as rhabdoviruses (order Mononegavirales) has a central function in regulating viral mRNA transcription and possibly RNA replication (1Hwang L.N. Englund N. Das T. Banerjee A.K. Pattnaik A.K. J. Virol. 1999; 73: 5613-5620Crossref PubMed Google Scholar).The alphaviruses are a globally distributed group of enveloped positive-strand RNA animal viruses capable of causing fatal encephalitis; representative members include Sindbis virus (SIN)1 and Semliki Forest virus (SFV). The RNA synthesis of alphaviruses occurs in the cytoplasm, where the 5′ two-thirds of the genomic RNA (total length, ∼11.5 kilobases) is translated into a large polyprotein of ∼2500 amino acids (aa). This polyprotein, termed P1234, is autoproteolytically cleaved to yield the four subunits of the viral RNA-dependent RNA polymerase, the nonstructural proteins nsP1–nsP4 (reviewed in Refs. 2Kääriäinen L. Takkinen K. Keränen S. Söderlund H. J. Cell Sci. Suppl. 1987; 7: 231-250Crossref PubMed Google Scholar and 3Strauss J.H. Strauss E.G. Microbiol. Rev. 1994; 58: 491-562Crossref PubMed Google Scholar). Polyprotein processing intermediates have distinct essential functions during the early phase of RNA replication, the synthesis of negative strands (4Lemm J.A. Rümenapf T. Strauss E.G. Strauss J.H. Rice C.M. EMBO J. 1994; 13: 2925-2934Crossref PubMed Scopus (231) Google Scholar, 5Shirako Y. Strauss J.H. J. Virol. 1994; 68: 1874-1885Crossref PubMed Google Scholar, 6Wang Y.-F. Sawicki S.G. Sawicki D.L. J. Virol. 1994; 68: 6466-6475Crossref PubMed Google Scholar). Later in infection, the negative strands are used as stable templates for synthesis of progeny-positive strands, and for synthesis of subgenomic mRNAs coding for the structural proteins of the virus.nsP1 is an enzyme responsible for methylation and capping of viral mRNAs (7Laakkonen P. Hyvönen M. Peränen J. Kääriäinen L. J. Virol. 1994; 68: 7418-7425Crossref PubMed Google Scholar, 8Ahola T. Kääriäinen L. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 507-511Crossref PubMed Scopus (228) Google Scholar). In addition, it mediates membrane association of the RNA replication complex (9Ahola T. Lampio A. Auvinen P. Kääriäinen L. EMBO J. 1999; 18: 3164-3172Crossref PubMed Scopus (137) Google Scholar) and its targeting onto the cytoplasmic surface of endosomes and lysosomes (10Peränen J. Laakkonen P. Hyvönen M. Kääriäinen L. Virology. 1995; 208: 610-620Crossref PubMed Scopus (84) Google Scholar). nsP2 is an RNA helicase (11Gomez de Cedrón M. Ehsani N. Mikkola M.L. Garcı́a J.A. Kääriäinen L. FEBS Lett. 1999; 448: 19-22Crossref PubMed Scopus (132) Google Scholar), RNA triphosphatase (12Vasiljeva L. Merits A. Auvinen P. Kääriäinen L. J. Biol. Chem. 2000; 275: 17281-17287Abstract Full Text Full Text PDF PubMed Scopus (150) Google Scholar), and an autoprotease responsible for the cleavage of the nonstructural polyprotein (3Strauss J.H. Strauss E.G. Microbiol. Rev. 1994; 58: 491-562Crossref PubMed Google Scholar). nsP4 is the catalytic subunit of this RNA-dependent RNA polymerase (13Barton D.J. Sawicki S.G. Sawicki D.L. J. Virol. 1988; 62: 3597-3602Crossref PubMed Google Scholar). Each of these three polypeptides has conserved amino acid sequence motifs, when compared with cellular and viral proteins of similar function: nsP4 with RNA-dependent polymerases, nsP2 with nucleic acid helicases and papain-like cysteine proteases, and nsP1 with methyltransferases (14Koonin E.V. Dolja V.V. Crit. Rev. Biochem. Mol. Biol. 1993; 28: 375-430Crossref PubMed Scopus (907) Google Scholar, 15Ahola T. Laakkonen P. Vihinen H. Kääriäinen L. J. Virol. 1997; 71: 392-397Crossref PubMed Google Scholar).In contrast, the functions of nsP3 are not well defined, although the protein is essential for RNA replication (16Hahn Y.S. Strauss E.G. Strauss J.H. J. Virol. 1989; 63: 3142-3150Crossref PubMed Google Scholar). Studies of temperature-sensitive and linker insertion mutants of SIN nsP3 have implicated it in negative-strand RNA synthesis, and possibly also in the synthesis of subgenomic mRNA (6Wang Y.-F. Sawicki S.G. Sawicki D.L. J. Virol. 1994; 68: 6466-6475Crossref PubMed Google Scholar, 17LaStarza M.W. Lemm J.A. Rice C.M. J. Virol. 1994; 68: 5781-5791Crossref PubMed Google Scholar). The amino acid sequence of SFV nsP3 (482 aa) can be divided into three almost equal sections. The first third forms a small domain conserved in alphaviruses, rubella virus, hepatitis E virus, and coronaviruses (14Koonin E.V. Dolja V.V. Crit. Rev. Biochem. Mol. Biol. 1993; 28: 375-430Crossref PubMed Scopus (907) Google Scholar). Recently, through genome sequencing, it has become apparent that this domain is widely, although not universally distributed in bacteria, archae, and eukaryotes. It usually exists on its own as a small open reading frame, but a more divergent version can be found attached to unusual histone variants, the macrohistones H2A (18Pehrson J.R. Fuji R.N. Nucleic Acids Res. 1998; 26: 2837-2842Crossref PubMed Scopus (121) Google Scholar). For the moment, this unanticipated conservation imparts little insight, since none of the nsP3-related proteins have been functionally characterized. The middle third of nsP3 is only conserved between alphaviruses, whereas the last third (after Tyr324; Fig. 1) is hypervariable, showing no discernible conservation. Even the size of this C-terminal "tail" varies in different alphaviruses: SFV nsP3 has 158 aa and SIN nsP3 232 aa (3Strauss J.H. Strauss E.G. Microbiol. Rev. 1994; 58: 491-562Crossref PubMed Google Scholar). The tail is rich in acidic residues, as well as in serine, threonine, and proline, and devoid of predicted secondary structure.nsP3 is the only alphavirus nonstructural protein modified by phosphorylation (19Peränen J. Takkinen K. Kalkkinen N. Kääriäinen L. J. Gen. Virol. 1988; 69: 2165-2178Crossref PubMed Scopus (76) Google Scholar, 20Li G. LaStarza M.W. Hardy W.R. Strauss J.H. Rice C.M. Virology. 1990; 179: 416-427Crossref PubMed Scopus (84) Google Scholar). Phosphoamino acid analysis of SFV nsP3 showed that serine and threonine residues are phosphorylated, approximately in 2:1 ratio, whereas no phosphotyrosine could be found (19Peränen J. Takkinen K. Kalkkinen N. Kääriäinen L. J. Gen. Virol. 1988; 69: 2165-2178Crossref PubMed Scopus (76) Google Scholar). The major phosphorylation sites of SFV nsP3 were determined by mass spectrometric analysis in conjugation with on-target alkaline phosphatase digestion, as well as two-dimensional peptide mapping and Edman sequencing (21Vihinen H. Saarinen J. J. Biol. Chem. 2000; 275: 27775-27783Abstract Full Text Full Text PDF PubMed Scopus (55) Google Scholar). In SFV nsP3 the serines 320, 327, 332, and 335 and from 7 to 12 residues in peptide Gly338–Lys415 can be phosphorylated. SIN nsP3 is even more heavily phosphorylated on serine and threonine, leading to formation of several species of different electrophoretic mobility (20Li G. LaStarza M.W. Hardy W.R. Strauss J.H. Rice C.M. Virology. 1990; 179: 416-427Crossref PubMed Scopus (84) Google Scholar). Deletions in the nonconserved tail region of SIN nsP3 considerably reduce its phosphorylation, suggesting that the modification may primarily occur in this part of the protein (22LaStarza M.W. Grakoui A. Rice C.M. Virology. 1994; 202: 224-232Crossref PubMed Scopus (71) Google Scholar). nsP3 is capable of associating with cellular membranes when expressed alone (23Peränen J. Kääriäinen L. J. Virol. 1991; 65: 1623-1627Crossref PubMed Google Scholar).Here we have constructed truncated and point-mutated derivatives of SFV nsP3 and studied their phosphorylation. This enabled the introduction of phosphorylation-defective nsP3 variants to the SFV genome, to better understand the role of phosphorylation of nsP3 in the alphavirus life cycle. The resultant mutant viruses were characterized in cell culture with respect to growth and RNA synthesis. Finally, a virus encoding nonphosphorylated nsP3 was used in studies of neurovirulence in mice.DISCUSSIONIn this study, we have completed the identification of the main phosphorylation sites of SFV nsP3 by showing that the major threonines phosphorylated are Thr344 and/or Thr345 (Fig.2). Several serine residues between amino acids 320 and 368 were also phosphorylated, possibly in a heterogeneous manner. Thus, all the identified phosphorylation sites were concentrated in a small, highly phosphorylated region (Fig. 1 B). Mutation of one or more serines/threonines affected the phosphorylation of other residues, since the effects of point mutations were not additive (Fig. 4). Specifically, mutation of either Thr344 and Thr345 or Ser320 reduced the overall phosphorylation level of nsP3 by 40–50%, but combination of these two mutations gave no additional reduction. The phosphorylated region was located in the beginning of the nonconserved C-terminal "tail" region of nsP3 (Fig. 1), extending slightly into the conserved domain, as Ser320 could be phosphorylated. However, this serine is not conserved among alphaviruses, and thus there appear to be no conserved phosphorylation sites shared by nsP3s from different alphaviruses. Nevertheless, the general feature of extensive phosphorylation of the tail seems to be shared with SIN nsP3 (22LaStarza M.W. Grakoui A. Rice C.M. Virology. 1994; 202: 224-232Crossref PubMed Scopus (71) Google Scholar).Among alphaviruses, the nsP3 variable region differs in length and in composition (3Strauss J.H. Strauss E.G. Microbiol. Rev. 1994; 58: 491-562Crossref PubMed Google Scholar). General similarities in the nonconserved regions of alphavirus nsP3s include the presence of proline-rich and acidic regions, which, however, show variation in length and distribution. In the SFV nsP3 nonconserved region, there are two acidic domains: aa 340–406 (15 Asp+Glu in 67 aa, 22%), and aa 453–473 (8 Asp+Glu in 21 aa, 38%) separated by a proline-rich stretch between aa 408–439 (11 prolines in 32 aa, 34%). Most of the phosphorylated sites of SFV nsP3 were in the first acidic region (Fig. 1 B), making it even more negatively charged. It remains to be determined whether phosphorylation is concentrated in a small subregion of nsP3 also in other alphaviruses. Both acidic and proline-rich domains have been suggested to be consensus motifs for transcriptional activation domains (30Ma J. Ptashne M. Cell. 1987; 48: 847-853Abstract Full Text PDF PubMed Scopus (602) Google Scholar, 31Mermod N. O'Neill E.A. Kelly T.J. Tjian R. Cell. 1989; 58: 741-753Abstract Full Text PDF PubMed Scopus (538) Google Scholar).The C-terminal region of nsP3 (C-nsP3; Fig. 1 A) was capable of becoming phosphorylated when expressed alone, although at reduced level as compared with the full-length protein (Fig. 3). This reduced phosphorylation may reflect a different conformation leading to differential kinase recognition. Coexpression with the wild type nsP3 did not increase the phosphorylation of the C-terminal peptide, indicating that the N-terminal portion is unlikely to contribute any kinase activity itself. The N-terminal domain is related to a large protein family with unknown function, but unrelated to known kinases (18Pehrson J.R. Fuji R.N. Nucleic Acids Res. 1998; 26: 2837-2842Crossref PubMed Scopus (121) Google Scholar). An alternative explanation for the reduced phosphorylation of C-nsP3 is offered by the fact that it was localized in the cytoplasm, whereas the wild type nsP3 was associated with vesicular structures. Thus, C-nsP3 might not be equally accessible to membrane-associated kinases, which may play a role in phosphorylation of the wild type nsP3 since membrane-associated nsP3 is more heavily phosphorylated than soluble nsP3 (19Peränen J. Takkinen K. Kalkkinen N. Kääriäinen L. J. Gen. Virol. 1988; 69: 2165-2178Crossref PubMed Scopus (76) Google Scholar). The phosphorylation of nsP3 was, however, not required for its peripheral membrane association.Casein kinase II (CK II) has been suggested to be the host enzyme phosphorylating SIN nsP3 (20Li G. LaStarza M.W. Hardy W.R. Strauss J.H. Rice C.M. Virology. 1990; 179: 416-427Crossref PubMed Scopus (84) Google Scholar), based on an experiment in which nsP3 was immunoprecipitated from SIN-infected cells and phosphorylated in vitro by a kinase present in the precipitate. The properties of this kinase resembled those of CK II, as assessed by heparin inhibition, polyamine activation, and use of both ATP and GTP as substrate (20Li G. LaStarza M.W. Hardy W.R. Strauss J.H. Rice C.M. Virology. 1990; 179: 416-427Crossref PubMed Scopus (84) Google Scholar). However, the SIN nsP3 studied in this experiment was most likely already phosphorylated within the cells. This may have favored the binding of, and phosphorylation by CK II, which recognizes serines/threonines followed by acidic residues, including phosphorylated residues (32Kennelly P.J. Krebs E.G. J. Biol. Chem. 1991; 266: 15555-15558Abstract Full Text PDF PubMed Google Scholar). Thus, CK II is one of the kinases involved in nsP3 phosphorylation, but there are likely to be others. In SFV nsP3, phosphorylation sites at Ser327 and Ser367 are the best potential sites for CKII (preferred recognition sequence (S/T)-X-X-(D/E)), but phosphorylation of other sites may become possible, if other sites are already phosphorylated by additional kinases (see sequence in Fig.1 B). Among SFV nsP3 phosphorylation sites, Ser320, Ser332, and Ser335 are potential protein kinase C sites (recognition sequence (S/T)-X-(R/K); Ref. 32Kennelly P.J. Krebs E.G. J. Biol. Chem. 1991; 266: 15555-15558Abstract Full Text PDF PubMed Google Scholar). SFV nsP3 Ser320 is a consensus site also for p34cdc2 kinase. Among negative-strand RNA viruses, the phosphoprotein P may also be a substrate for multiple kinases; examples include phosphorylation of canine distemper and measles virus P proteins by CK II and protein kinase C (33Das T. Schuster A. Schneider-Schaulies S. Banerjee A.K. Virology. 1995; 211: 218-226Crossref PubMed Scopus (63) Google Scholar, 34Liu Z. Huntley C.C. De B.P. Das T. Banerjee A.K. Oglesbee M.J. Virology. 1997; 232: 198-206Crossref PubMed Scopus (25) Google Scholar), rabies virus P protein by a unique cellular protein kinase and protein kinase C (35Gupta A.K. Blondel D. Choudhary S. Banerjee A.K. J. Virol. 2000; 74: 91-98Crossref PubMed Scopus (85) Google Scholar), vesicular stomatis virus P protein by CKII and an L protein-associated kinase (36Barik S. Banerjee A.K. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 6570-6574Crossref PubMed Scopus (126) Google Scholar, 37Barik S. Banerjee A.K. J. Virol. 1992; 66: 1109-1118Crossref PubMed Google Scholar), and Sendai virus P protein by protein kinase C and a proline-directed protein kinase (38Huntley C.C. De B.P. Banerjee A.K. J. Biol. Chem. 1997; 272: 16578-16584Abstract Full Text Full Text PDF PubMed Scopus (32) Google Scholar,39Byrappa S. Pan Y.-B. Gupta K.C. Virology. 1996; 216: 228-234Crossref PubMed Scopus (30) Google Scholar).Mutant viruses coding for truncated forms of SFV nsP3 replicated to high titers in BHK, Vero, and NIH cells (Fig. 7 A). nsP3 mutant Δ26 had 10% level of phosphorylation remaining as compared with wild type, whereas mutant Δ50 was not detectably phosphorylated. These results indicate that the mutated nsP3 proteins were able to carry out all their essential functions needed for RNA replication and for high rates of virus production. Therefore, these internal deletions in a region that was predicted to be devoid of secondary structure did not cause gross misfolding of the protein.When used to infect mice, the SFV nsP3Δ50 mutant virus was avirulent even with a dose of 107 pfu/mouse. The rate of viral RNA synthesis of this mutant was significantly lower than that of the wild type SFV resulting the decreased virulence in mice. However, it cannot be simply stated that the decreased RNA synthesis rate is caused solely by the lack of phosphorylation of nsP3, since deletion of amino acids around the phosphorylation sites may have made a contribution to this phenomenon. Recent study with an avirulent SFV mutant A7(74) (40Atkins G.J. Sheahan B.J. Liljeström P. J. Gen. Virol. 1999; 80: 2287-2297Crossref PubMed Scopus (78) Google Scholar) has revealed the importance of nsP3 in the neuropathogenicity of SFV (28Tuittila M.T. Santagati M.G. Röyttä M. Määttä J.A. Hinkkanen A.E. J. Virol. 2000; 74: 4579-4589Crossref PubMed Scopus (65) Google Scholar). Phosphorylation of nsP3, as well as the entire heterogenous C-terminal tail, which has been subject to rapid alteration during alphavirus evolution, may serve to fine-tune the replication of alphaviruses in different cell types. Further studies should reveal the role of the phosphorylated C-terminal region of nsP3 in specific interactions with components of the neuronal cells leading to neurovirulence of Semliki Forest virus. Phosphorylation and dephosphorylation have been recognized as major processes by which protein function is regulated. A wide range of proteins display phosphorylation state-dependent activity, including proteins involved in signal transduction, transcription, and the cell cycle. In the field of RNA virus replication, the phosphoprotein P of negative-strand RNA viruses with unsegmented genomes such as rhabdoviruses (order Mononegavirales) has a central function in regulating viral mRNA transcription and possibly RNA replication (1Hwang L.N. Englund N. Das T. Banerjee A.K. Pattnaik A.K. J. Virol. 1999; 73: 5613-5620Crossref PubMed Google Scholar). The alphaviruses are a globally distributed group of enveloped positive-strand RNA animal viruses capable of causing fatal encephalitis; representative members include Sindbis virus (SIN)1 and Semliki Forest virus (SFV). The RNA synthesis of alphaviruses occurs in the cytoplasm, where the 5′ two-thirds of the genomic RNA (total length, ∼11.5 kilobases) is translated into a large polyprotein of ∼2500 amino acids (aa). This polyprotein, termed P1234, is autoproteolytically cleaved to yield the four subunits of the viral RNA-dependent RNA polymerase, the nonstructural proteins nsP1–nsP4 (reviewed in Refs. 2Kääriäinen L. Takkinen K. Keränen S. Söderlund H. J. Cell Sci. Suppl. 1987; 7: 231-250Crossref PubMed Google Scholar and 3Strauss J.H. Strauss E.G. Microbiol. Rev. 1994; 58: 491-562Crossref PubMed Google Scholar). Polyprotein processing intermediates have distinct essential functions during the early phase of RNA replication, the synthesis of negative strands (4Lemm J.A. Rümenapf T. Strauss E.G. Strauss J.H. Rice C.M. EMBO J. 1994; 13: 2925-2934Crossref PubMed Scopus (231) Google Scholar, 5Shirako Y. Strauss J.H. J. Virol. 1994; 68: 1874-1885Crossref PubMed Google Scholar, 6Wang Y.-F. Sawicki S.G. Sawicki D.L. J. Virol. 1994; 68: 6466-6475Crossref PubMed Google Scholar). Later in infection, the negative strands are used as stable templates for synthesis of progeny-positive strands, and for synthesis of subgenomic mRNAs coding for the structural proteins of the virus. nsP1 is an enzyme responsible for methylation and capping of viral mRNAs (7Laakkonen P. Hyvönen M. Peränen J. Kääriäinen L. J. Virol. 1994; 68: 7418-7425Crossref PubMed Google Scholar, 8Ahola T. Kääriäinen L. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 507-511Crossref PubMed Scopus (228) Google Scholar). In addition, it mediates membrane association of the RNA replication complex (9Ahola T. Lampio A. Auvinen P. Kääriäinen L. EMBO J. 1999; 18: 3164-3172Crossref PubMed Scopus (137) Google Scholar) and its targeting onto the cytoplasmic surface of endosomes and lysosomes (10Peränen J. Laakkonen P. Hyvönen M. Kääriäinen L. Virology. 1995; 208: 610-620Crossref PubMed Scopus (84) Google Scholar). nsP2 is an RNA helicase (11Gomez de Cedrón M. Ehsani N. Mikkola M.L. Garcı́a J.A. Kääriäinen L. FEBS Lett. 1999; 448: 19-22Crossref PubMed Scopus (132) Google Scholar), RNA triphosphatase (12Vasiljeva L. Merits A. Auvinen P. Kääriäinen L. J. Biol. Chem. 2000; 275: 17281-17287Abstract Full Text Full Text PDF PubMed Scopus (150) Google Scholar), and an autoprotease responsible for the cleavage of the nonstructural polyprotein (3Strauss J.H. Strauss E.G. Microbiol. Rev. 1994; 58: 491-562Crossref PubMed Google Scholar). nsP4 is the catalytic subunit of this RNA-dependent RNA polymerase (13Barton D.J. Sawicki S.G. Sawicki D.L. J. Virol. 1988; 62: 3597-3602Crossref PubMed Google Scholar). Each of these three polypeptides has conserved amino acid sequence motifs, when compared with cellular and viral proteins of similar function: nsP4 with RNA-dependent polymerases, nsP2 with nucleic acid helicases and papain-like cysteine proteases, and nsP1 with methyltransferases (14Koonin E.V. Dolja V.V. Crit. Rev. Biochem. Mol. Biol. 1993; 28: 375-430Crossref PubMed Scopus (907) Google Scholar, 15Ahola T. Laakkonen P. Vihinen H. Kääriäinen L. J. Virol. 1997; 71: 392-397Crossref PubMed Google Scholar). In contrast, the functions of nsP3 are not well defined, although the protein is essential for RNA replication (16Hahn Y.S. Strauss E.G. Strauss J.H. J. Virol. 1989; 63: 3142-3150Crossref PubMed Google Scholar). Studies of temperature-sensitive and linker insertion mutants of SIN nsP3 have implicated it in negative-strand RNA synthesis, and possibly also in the synthesis of subgenomic mRNA (6Wang Y.-F. Sawicki S.G. Sawicki D.L. J. Virol. 1994; 68: 6466-6475Crossref PubMed Google Scholar, 17LaStarza M.W. Lemm J.A. Rice C.M. J. Virol. 1994; 68: 5781-5791Crossref PubMed Google Scholar). The amino acid sequence of SFV nsP3 (482 aa) can be divided into three almost equal sections. The first third forms a small domain conserved in alphaviruses, rubella virus, hepatitis E virus, and coronaviruses (14Koonin E.V. Dolja V.V. Crit. Rev. Biochem. Mol. Biol. 1993; 28: 375-430Crossref PubMed Scopus (907) Google Scholar). Recently, through genome sequencing, it has become apparent that this domain is widely, although not universally distributed in bacteria, archae, and eukaryotes. It usually exists on its own as a small open reading frame, but a more divergent version can be found attached to unusual histone variants, the macrohistones H2A (18Pehrson J.R. Fuji R.N. Nucleic Acids Res. 1998; 26: 2837-2842Crossref PubMed Scopus (121) Google Scholar). For the moment, this unanticipated conservation imparts little insight, since none of the nsP3-related proteins have been functionally characterized. The middle third of nsP3 is only conserved between alphaviruses, whereas the last third (after Tyr324; Fig. 1) is hypervariable, showing no discernible conservation. Even the size of this C-terminal "tail" varies in different alphaviruses: SFV nsP3 has 158 aa and SIN nsP3 232 aa (3Strauss J.H. Strauss E.G. Microbiol. Rev. 1994; 58: 491-562Crossref PubMed Google Scholar). The tail is rich in acidic residues, as well as in serine, threonine, and proline, and devoid of predicted secondary structure. nsP3 is the only alphavirus nonstructural protein modified by phosphorylation (19Peränen J. Takkinen K. Kalkkinen N. Kääriäinen L. J. Gen. Virol. 1988; 69: 2165-2178Crossref PubMed Scopus (76) Google Scholar, 20Li G. LaStarza M.W. Hardy W.R. Strauss J.H. Rice C.M. Virology. 1990; 179: 416-427Crossref PubMed Scopus (84) Google Scholar). Phosphoamino acid analysis of SFV nsP3 showed that serine and threonine residues are phosphorylated, approximately in 2:1 ratio, whereas no phosphotyrosine could be found (19Peränen J. Takkinen K. Kalkkinen N. Kääriäinen L. J. Gen. Virol. 1988; 69: 2165-2178Crossref PubMed Scopus (76) Google Scholar). The major phosphorylation sites of SFV nsP3 were determined by mass spectrometric analysis in conjugation with on-target alkaline phosphatase digestion, as well as two-dimensional peptide mapping and Edman sequencing (21Vihinen H. Saarinen J. J. Biol. Chem. 2000; 275: 27775-27783Abstract Full Text Full Text PDF PubMed Scopus (55) Google Scholar). In SFV nsP3 the serines 320, 327, 332, and 335 and from 7 to 12 residues in peptide Gly338–Lys415 can be phosphorylated. SIN nsP3 is even more heavily phosphorylated on serine and threonine, leading to formation of several species of different electrophoretic mobility (20Li G. LaStarza M.W. Hardy W.R. Strauss J.H. Rice C.M. Virology. 1990; 179: 416-427Crossref PubMed Scopus (84) Google Scholar). Deletions in the nonconserved tail region of SIN nsP3 considerably reduce its phosphorylation, suggesting that the modification may primarily occur in this part of the protein (22LaStarza M.W. Grakoui A. Rice C.M. Virology. 1994; 202: 224-232Crossref PubMed Scopus (71) Google Scholar). nsP3 is capable of associating with cellular membranes when expressed alone (23Peränen J. Kääriäinen L. J. Virol. 1991; 65: 1623-1627Crossref PubMed Google Scholar). Here we have constructed truncated and point-mutated derivatives of SFV nsP3 and studied their phosphorylation. This enabled the introduction of phosphorylation-defective nsP3 variants to the SFV genome, to better understand the role of phosphorylation of nsP3 in the alphavirus life cycle. The resultant mutant viruses were characterized in cell culture with respect to growth and RNA synthesis. Finally, a virus encoding nonphosphorylated nsP3 was used in studies of neurovirulence in mice. DISCUSSIONIn this study, we have completed the identification of the main phosphorylation sites of SFV nsP3 by showing that the major threonines phosphorylated are Thr344 and/or Thr345 (Fig.2). Several serine residues between amino acids 320 and 368 were also phosphorylated, possibly in a heterogeneous manner. Thus, all the identified phosphorylation sites were concentrated in a small, highly phosphorylated region (Fig. 1 B). Mutation of one or more serines/threonines affected the phosphorylation of other residues, since the effects of point mutations were not additive (Fig. 4). Specifically, mutation of either Thr344 and Thr345 or Ser320 reduced the overall phosphorylation level of nsP3 by 40–50%, but combination of these two mutations gave no additional reduction. The phosphorylated region was located in the beginning of the nonconserved C-terminal "tail" region of nsP3 (Fig. 1), extending slightly into the conserved domain, as Ser320 could be phosphorylated. However, this serine is not conserved among alphaviruses, and thus there appear to be no conserved phosphorylation sites shared by nsP3s from different alphaviruses. Nevertheless, the general feature of extensive phosphorylation of the tail seems to be shared with SIN nsP3 (22LaStarza M.W. Grakoui A. Rice C.M. Virology. 1994; 202: 224-232Crossref PubMed Scopus (71) Google Scholar).Among alphaviruses, the nsP3 variable region differs in length and in composition (3Strauss J.H. Strauss E.G. Microbiol. Rev. 1994; 58: 491-562Crossref PubMed Google Scholar). General similarities in the nonconserved regions of alphavirus nsP3s include the presence of proline-rich and acidic regions, which, however, show variation in length and distribution. In the SFV nsP3 nonconserved region, there are two acidic domains: aa 340–406 (15 Asp+Glu in 67 aa, 22%), and aa 453–473 (8 Asp+Glu in 21 aa, 38%) separated by a proline-rich stretch between aa 408–439 (11 prolines in 32 aa, 34%). Most of the phosphorylated sites of SFV nsP3 were in the first acidic region (Fig. 1 B), making it even more negatively charged. It remains to be determined whether phosphorylation is concentrated in a small subregion of nsP3 also in other alphaviruses. Both acidic and proline-rich domains have been suggested to be consensus motifs for transcriptional activation domains (30Ma J. Ptashne M. Cell. 1987; 48: 847-853Abstract Full Text PDF PubMed Scopus (602) Google Scholar, 31Mermod N. O'Neill E.A. Kelly T.J. Tjian R. Cell. 1989; 58: 741-753Abstract Full Text PDF PubMed Scopus (538) Google Scholar).The C-terminal region of nsP3 (C-nsP3; Fig. 1 A) was capable of becoming phosphorylated when expressed alone, although at reduced level as compared with the full-length protein (Fig. 3). This reduced phosphorylation may reflect a different conformation leading to differential kinase recognition. Coexpression with the wild type nsP3 did not increase the phosphorylation of the C-terminal peptide, indicating that the N-terminal portion is unlikely to contribute any kinase activity itself. The N-terminal domain is related to a large protein family with unknown function, but unrelated to known kinases (18Pehrson J.R. Fuji R.N. Nucleic Acids Res. 1998; 26: 2837-2842Crossref PubMed Scopus (121) Google Scholar). An alternative explanation for the reduced phosphorylation of C-nsP3 is offered by the fact that it was localized in the cytoplasm, whereas the wild type nsP3 was associated with vesicular structures. Thus, C-nsP3 might not be equally accessible to membrane-associated kinases, which may play a role in phosphorylation of the wild type nsP3 since membrane-associated nsP3 is more heavily phosphorylated than soluble nsP3 (19Peränen J. Takkinen K. Kalkkinen N. Kääriäinen L. J. Gen. Virol. 1988; 69: 2165-2178Crossref PubMed Scopus (76) Google Scholar). The phosphorylation of nsP3 was, however, not required for its peripheral membrane association.Casein kinase II (CK II) has been suggested to be the host enzyme phosphorylating SIN nsP3 (20Li G. LaStarza M.W. Hardy W.R. Strauss J.H. Rice C.M. Virology. 1990; 179: 416-427Crossref PubMed Scopus (84) Google Scholar), based on an experiment in which nsP3 was immunoprecipitated from SIN-infected cells and phosphorylated in vitro by a kinase present in the precipitate. The properties of this kinase resembled those of CK II, as assessed by heparin inhibition, polyamine activation, and use of both ATP and GTP as substrate (20Li G. LaStarza M.W. Hardy W.R. Strauss J.H. Rice C.M. Virology. 1990; 179: 416-427Crossref PubMed Scopus (84) Google Scholar). However, the SIN nsP3 studied in this experiment was most likely already phosphorylated within the cells. This may have favored the binding of, and phosphorylation by CK II, which recognizes serines/threonines followed by acidic residues, including phosphorylated residues (32Kennelly P.J. Krebs E.G. J. Biol. Chem. 1991; 266: 15555-15558Abstract Full Text PDF PubMed Google Scholar). Thus, CK II is one of the kinases involved in nsP3 phosphorylation, but there are likely to be others. In SFV nsP3, phosphorylation sites at Ser327 and Ser367 are the best potential sites for CKII (preferred recognition sequence (S/T)-X-X-(D/E)), but phosphorylation of other sites may become possible, if other sites are already phosphorylated by additional kinases (see sequence in Fig.1 B). Among SFV nsP3 phosphorylation sites, Ser320, Ser332, and Ser335 are potential protein kinase C sites (recognition sequence (S/T)-X-(R/K); Ref. 32Kennelly P.J. Krebs E.G. J. Biol. Chem. 1991; 266: 15555-15558Abstract Full Text PDF PubMed Google Scholar). SFV nsP3 Ser320 is a consensus site also for p34cdc2 kinase. Among negative-strand RNA viruses, the phosphoprotein P may also be a substrate for multiple kinases; examples include phosphorylation of canine distemper and measles virus P proteins by CK II and protein kinase C (33Das T. Schuster A. Schneider-Schaulies S. Banerjee A.K. Virology. 1995; 211: 218-226Crossref PubMed Scopus (63) Google Scholar, 34Liu Z. Huntley C.C. De B.P. Das T. Banerjee A.K. Oglesbee M.J. Virology. 1997; 232: 198-206Crossref PubMed Scopus (25) Google Scholar), rabies virus P protein by a unique cellular protein kinase and protein kinase C (35Gupta A.K. Blondel D. Choudhary S. Banerjee A.K. J. Virol. 2000; 74: 91-98Crossref PubMed Scopus (85) Google Scholar), vesicular stomatis virus P protein by CKII and an L protein-associated kinase (36Barik S. Banerjee A.K. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 6570-6574Crossref PubMed Scopus (126) Google Scholar, 37Barik S. Banerjee A.K. J. Virol. 1992; 66: 1109-1118Crossref PubMed Google Scholar), and Sendai virus P protein by protein kinase C and a proline-directed protein kinase (38Huntley C.C. De B.P. Banerjee A.K. J. Biol. Chem. 1997; 272: 16578-16584Abstract Full Text Full Text PDF PubMed Scopus (32) Google Scholar,39Byrappa S. Pan Y.-B. Gupta K.C. Virology. 1996; 216: 228-234Crossref PubMed Scopus (30) Google Scholar).Mutant viruses coding for truncated forms of SFV nsP3 replicated to high titers in BHK, Vero, and NIH cells (Fig. 7 A). nsP3 mutant Δ26 had 10% level of phosphorylation remaining as compared with wild type, whereas mutant Δ50 was not detectably phosphorylated. These results indicate that the mutated nsP3 proteins were able to carry out all their essential functions needed for RNA replication and for high rates of virus production. Therefore, these internal deletions in a region that was predicted to be devoid of secondary structure did not cause gross misfolding of the protein.When used to infect mice, the SFV nsP3Δ50 mutant virus was avirulent even with a dose of 107 pfu/mouse. The rate of viral RNA synthesis of this mutant was significantly lower than that of the wild type SFV resulting the decreased virulence in mice. However, it cannot be simply stated that the decreased RNA synthesis rate is caused solely by the lack of phosphorylation of nsP3, since deletion of amino acids around the phosphorylation sites may have made a contribution to this phenomenon. Recent study with an avirulent SFV mutant A7(74) (40Atkins G.J. Sheahan B.J. Liljeström P. J. Gen. Virol. 1999; 80: 2287-2297Crossref PubMed Scopus (78) Google Scholar) has revealed the importance of nsP3 in the neuropathogenicity of SFV (28Tuittila M.T. Santagati M.G. Röyttä M. Määttä J.A. Hinkkanen A.E. J. Virol. 2000; 74: 4579-4589Crossref PubMed Scopus (65) Google Scholar). Phosphorylation of nsP3, as well as the entire heterogenous C-terminal tail, which has been subject to rapid alteration during alphavirus evolution, may serve to fine-tune the replication of alphaviruses in different cell types. Further studies should reveal the role of the phosphorylated C-terminal region of nsP3 in specific interactions with components of the neuronal cells leading to neurovirulence of Semliki Forest virus. In this study, we have completed the identification of the main phosphorylation sites of SFV nsP3 by showing that the major threonines phosphorylated are Thr344 and/or Thr345 (Fig.2). Several serine residues between amino acids 320 and 368 were also phosphorylated, possibly in a heterogeneous manner. Thus, all the identified phosphorylation sites were concentrated in a small, highly phosphorylated region (Fig. 1 B). Mutation of one or more serines/threonines affected the phosphorylation of other residues, since the effects of point mutations were not additive (Fig. 4). Specifically, mutation of either Thr344 and Thr345 or Ser320 reduced the overall phosphorylation level of nsP3 by 40–50%, but combination of these two mutations gave no additional reduction. The phosphorylated region was located in the beginning of the nonconserved C-terminal "tail" region of nsP3 (Fig. 1), extending slightly into the conserved domain, as Ser320 could be phosphorylated. However, this serine is not conserved among alphaviruses, and thus there appear to be no conserved phosphorylation sites shared by nsP3s from different alphaviruses. Nevertheless, the general feature of extensive phosphorylation of the tail seems to be shared with SIN nsP3 (22LaStarza M.W. Grakoui A. Rice C.M. Virology. 1994; 202: 224-232Crossref PubMed Scopus (71) Google Scholar). Among alphaviruses, the nsP3 variable region differs in length and in composition (3Strauss J.H. Strauss E.G. Microbiol. Rev. 1994; 58: 491-562Crossref PubMed Google Scholar). General similarities in the nonconserved regions of alphavirus nsP3s include the presence of proline-rich and acidic regions, which, however, show variation in length and distribution. In the SFV nsP3 nonconserved region, there are two acidic domains: aa 340–406 (15 Asp+Glu in 67 aa, 22%), and aa 453–473 (8 Asp+Glu in 21 aa, 38%) separated by a proline-rich stretch between aa 408–439 (11 prolines in 32 aa, 34%). Most of the phosphorylated sites of SFV nsP3 were in the first acidic region (Fig. 1 B), making it even more negatively charged. It remains to be determined whether phosphorylation is concentrated in a small subregion of nsP3 also in other alphaviruses. Both acidic and proline-rich domains have been suggested to be consensus motifs for transcriptional activation domains (30Ma J. Ptashne M. Cell. 1987; 48: 847-853Abstract Full Text PDF PubMed Scopus (602) Google Scholar, 31Mermod N. O'Neill E.A. Kelly T.J. Tjian R. Cell. 1989; 58: 741-753Abstract Full Text PDF PubMed Scopus (538) Google Scholar). The C-terminal region of nsP3 (C-nsP3; Fig. 1 A) was capable of becoming phosphorylated when expressed alone, although at reduced level as compared with the full-length protein (Fig. 3). This reduced phosphorylation may reflect a different conformation leading to differential kinase recognition. Coexpression with the wild type nsP3 did not increase the phosphorylation of the C-terminal peptide, indicating that the N-terminal portion is unlikely to contribute any kinase activity itself. The N-terminal domain is related to a large protein family with unknown function, but unrelated to known kinases (18Pehrson J.R. Fuji R.N. Nucleic Acids Res. 1998; 26: 2837-2842Crossref PubMed Scopus (121) Google Scholar). An alternative explanation for the reduced phosphorylation of C-nsP3 is offered by the fact that it was localized in the cytoplasm, whereas the wild type nsP3 was associated with vesicular structures. Thus, C-nsP3 might not be equally accessible to membrane-associated kinases, which may play a role in phosphorylation of the wild type nsP3 since membrane-associated nsP3 is more heavily phosphorylated than soluble nsP3 (19Peränen J. Takkinen K. Kalkkinen N. Kääriäinen L. J. Gen. Virol. 1988; 69: 2165-2178Crossref PubMed Scopus (76) Google Scholar). The phosphorylation of nsP3 was, however, not required for its peripheral membrane association. Casein kinase II (CK II) has been suggested to be the host enzyme phosphorylating SIN nsP3 (20Li G. LaStarza M.W. Hardy W.R. Strauss J.H. Rice C.M. Virology. 1990; 179: 416-427Crossref PubMed Scopus (84) Google Scholar), based on an experiment in which nsP3 was immunoprecipitated from SIN-infected cells and phosphorylated in vitro by a kinase present in the precipitate. The properties of this kinase resembled those of CK II, as assessed by heparin inhibition, polyamine activation, and use of both ATP and GTP as substrate (20Li G. LaStarza M.W. Hardy W.R. Strauss J.H. Rice C.M. Virology. 1990; 179: 416-427Crossref PubMed Scopus (84) Google Scholar). However, the SIN nsP3 studied in this experiment was most likely already phosphorylated within the cells. This may have favored the binding of, and phosphorylation by CK II, which recognizes serines/threonines followed by acidic residues, including phosphorylated residues (32Kennelly P.J. Krebs E.G. J. Biol. Chem. 1991; 266: 15555-15558Abstract Full Text PDF PubMed Google Scholar). Thus, CK II is one of the kinases involved in nsP3 phosphorylation, but there are likely to be others. In SFV nsP3, phosphorylation sites at Ser327 and Ser367 are the best potential sites for CKII (preferred recognition sequence (S/T)-X-X-(D/E)), but phosphorylation of other sites may become possible, if other sites are already phosphorylated by additional kinases (see sequence in Fig.1 B). Among SFV nsP3 phosphorylation sites, Ser320, Ser332, and Ser335 are potential protein kinase C sites (recognition sequence (S/T)-X-(R/K); Ref. 32Kennelly P.J. Krebs E.G. J. Biol. Chem. 1991; 266: 15555-15558Abstract Full Text PDF PubMed Google Scholar). SFV nsP3 Ser320 is a consensus site also for p34cdc2 kinase. Among negative-strand RNA viruses, the phosphoprotein P may also be a substrate for multiple kinases; examples include phosphorylation of canine distemper and measles virus P proteins by CK II and protein kinase C (33Das T. Schuster A. Schneider-Schaulies S. Banerjee A.K. Virology. 1995; 211: 218-226Crossref PubMed Scopus (63) Google Scholar, 34Liu Z. Huntley C.C. De B.P. Das T. Banerjee A.K. Oglesbee M.J. Virology. 1997; 232: 198-206Crossref PubMed Scopus (25) Google Scholar), rabies virus P protein by a unique cellular protein kinase and protein kinase C (35Gupta A.K. Blondel D. Choudhary S. Banerjee A.K. J. Virol. 2000; 74: 91-98Crossref PubMed Scopus (85) Google Scholar), vesicular stomatis virus P protein by CKII and an L protein-associated kinase (36Barik S. Banerjee A.K. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 6570-6574Crossref PubMed Scopus (126) Google Scholar, 37Barik S. Banerjee A.K. J. Virol. 1992; 66: 1109-1118Crossref PubMed Google Scholar), and Sendai virus P protein by protein kinase C and a proline-directed protein kinase (38Huntley C.C. De B.P. Banerjee A.K. J. Biol. Chem. 1997; 272: 16578-16584Abstract Full Text Full Text PDF PubMed Scopus (32) Google Scholar,39Byrappa S. Pan Y.-B. Gupta K.C. Virology. 1996; 216: 228-234Crossref PubMed Scopus (30) Google Scholar). Mutant viruses coding for truncated forms of SFV nsP3 replicated to high titers in BHK, Vero, and NIH cells (Fig. 7 A). nsP3 mutant Δ26 had 10% level of phosphorylation remaining as compared with wild type, whereas mutant Δ50 was not detectably phosphorylated. These results indicate that the mutated nsP3 proteins were able to carry out all their essential functions needed for RNA replication and for high rates of virus production. Therefore, these internal deletions in a region that was predicted to be devoid of secondary structure did not cause gross misfolding of the protein. When used to infect mice, the SFV nsP3Δ50 mutant virus was avirulent even with a dose of 107 pfu/mouse. The rate of viral RNA synthesis of this mutant was significantly lower than that of the wild type SFV resulting the decreased virulence in mice. However, it cannot be simply stated that the decreased RNA synthesis rate is caused solely by the lack of phosphorylation of nsP3, since deletion of amino acids around the phosphorylation sites may have made a contribution to this phenomenon. Recent study with an avirulent SFV mutant A7(74) (40Atkins G.J. Sheahan B.J. Liljeström P. J. Gen. Virol. 1999; 80: 2287-2297Crossref PubMed Scopus (78) Google Scholar) has revealed the importance of nsP3 in the neuropathogenicity of SFV (28Tuittila M.T. Santagati M.G. Röyttä M. Määttä J.A. Hinkkanen A.E. J. Virol. 2000; 74: 4579-4589Crossref PubMed Scopus (65) Google Scholar). Phosphorylation of nsP3, as well as the entire heterogenous C-terminal tail, which has been subject to rapid alteration during alphavirus evolution, may serve to fine-tune the replication of alphaviruses in different cell types. Further studies should reveal the role of the phosphorylated C-terminal region of nsP3 in specific interactions with components of the neuronal cells leading to neurovirulence of Semliki Forest virus. We are grateful to Petra Nygårdas (Åbo Academi and Turku Immunology Center, Turku, Finland) for performing the immunohistochemical assays. We also thank Airi Sinkko and Tarja Välimäki for excellent technical assistance and Dr. Marja Makarow for critical reading of this manuscript.
Referência(s)