Mutations in conserved domains IV and VI of the large (L) subunit of the sendai virus RNA polymerase give a spectrum of defective RNA synthesis phenotypes.

The Sendai virus RNA polymerase is a complex of two virus-encoded proteins, the phosphoprotein (P) and the large (L) protein. When aligned with amino acid sequences of L proteins from other negative-sense RNA viruses, the Sendai L protein contains six regions of good conservation, designated domains I-VI, which have been postulated to be important for the various enzymatic activities of the polymerase. To directly address the roles of domains IV and VI, 14 site-directed mutations were constructed either by changing clustered charged amino acids to ala or by substituting selected Sendai L amino acids with the corresponding sequence from measles virus L. Each mutant L protein was tested for its ability to transcribe and replicate the Sendai genome. The series of mutations created a spectrum of phenotypes, from those with significant, near wild-type, activity to those being completely defective for all RNA synthesis. The inactive L proteins, however, were still able to bind P protein and form a polymerase capable of binding the nucleocapsid template. The remainder of the mutations reduced, but did not abolish, enzymatic activity and included one mutant with a specific defect in the synthesis of the leader RNA compared with mRNA, and three mutants that replicated genome RNA much more efficiently in vivo than in vitro. Together, these data suggest that even within a domain, the function of the Sendai L protein is likely to be very complex. In addition, SS3 and SS10 L in domain IV and SS13 L in domain VI were shown to be temperature-sensitive. Both SS3 and SS10 gave significant, although not wild-type, activity at 32 degrees C; however, each was completely inactivated for all RNA synthesis at 37 and 39.6 degrees C. SS13 was completely inactive only when synthesized at the higher temperature. Each polymerase synthesized at 32 degrees C could only be partially heat inactivated in vitro at 39.6 degrees C, suggesting that inactivation involves both thermal lability of the protein and temperature sensitivity for its synthesis.

Mutations in conserved domain II of the large (L) subunit of the Sendai virus RNA polymerase abolish RNA synthesis.

The large (L) protein of Sendai virus complexes with the phosphoprotein (P) to form the active RNA-dependent RNA polymerase. The L protein is believed to be responsible for all of the catalytic activities of the polymerase associated with transcription and replication. Sequence alignment of the L proteins of negative-strand RNA viruses has revealed six conserved domains (I-VI) thought to be responsible for the enzymatic activities. Charged-to-alanine mutagenesis was carried out in a highly charged, conserved region (amino acids 533-569) within domain II to test the hypothesis of Müller et al. [J. Gen. Virol. 75, 1345-1352 (1994)] that this region may contribute to the template binding domain of the viral RNA polymerase. The mutant proteins were tested for expression and stability, the ability to synthesize viral RNA in vitro and in vivo, and protein-protein interactions. Five of the seven mutants were completely defective in all viral RNA synthesis, whereas two mutants showed significant levels of both mRNA and leader RNA synthesis. One of the transcriptionally active mutants also gave genome replication in vitro although not in vivo. The other mutant was defective in all the replication assays and thus the mutation uncoupled transcription and replication. Because the completely inactive L mutants can bind to the P protein to form the polymerase complex and the polymerases bind to the viral nucleocapsid template, these amino acids are essential for the activity of the L protein.

The Sendai virus C protein binds the L polymerase protein to inhibit viral RNA synthesis.

The Sendai virus nested set of C proteins which are expressed in an alternative open reading frame from the P mRNA has been shown to downregulate viral RNA synthesis. Utilizing a glutathione S-transferase (gst) C fusion protein (gstC), we have shown that C protein forms a complex with the L, but not the P, subunit of the viral RNA polymerase. When P, L, and gstC are coexpressed, an oligomer of P, through its interaction with L, is also bound to beads. Since binding of C to L in the P-L complex does not disrupt P binding, the C and P binding sites appear to be different. GstC binding to L occurs only when the proteins are coexpressed in the same cell. The gstC, but not gst, protein inhibits viral transcription in vitro, showing that the fusion protein retains biological function. Pulse-chase experiments of the various complexes show that L protein synthesized alone has a half-life of 1. 2 hr, which is increased 12.5-fold by binding P, but is not significantly increased by binding gstC. Analyses of complex formation with truncations of L protein show that the C-terminal 1333 amino acids of L are not required for binding C. The dose-response curves show that replication of the genomic DI-H RNA is more sensitive to inhibition by C protein than is the synthesis of DI leader RNA, suggesting that the downregulation of RNA synthesis may be more complex than just the inhibition of the initiation of RNA synthesis.

The Sendai virus V protein interacts with the NP protein to regulate viral genome RNA replication.

The interactions of Sendai virus proteins required for viral RNA synthesis have been characterized both by the yeast two-hybrid system and through the use of glutathione S-transferase (gst)-viral fusion proteins synthesized in mammalian cells. Using the two-hybrid system we have confirmed the previously identified P-L (RNA polymerase), NPo-P (encapsidation substrate), and P-P complexes and now demonstrate NP-NP and NPo-V protein interactions. Expression of gstP and P proteins and binding to glutathione-Sepharose beads as a measure of complex formation confirmed the P-P interaction. The P-gstP binding occurred only on expression of the proteins in the same cell and was mapped to amino acids 345-411. We also show that full-length and deletion gstV and gstW proteins bound NPo protein when these sets of proteins were coexpressed and have identified one required region from amino acids 78-316. Neither gstV nor gstW bound NP assembled into nucleocapsids. Furthermore, both V and W proteins lacking the N-terminal 77 amino acids inhibited DI-H genome replication in vitro, showing the biological relevance of the remaining region. We propose that the specific inhibition of genome replication by V and W proteins occurs through interference with either the formation or the use of the NPo-P encapsidation substrate.

Domains of the measles virus N protein required for binding to P protein and self-assembly.

The nucleocapsid protein (N, 525 amino acids) of measles virus plays a central role in the replication of the viral genomic RNA. Its functions require interactions with itself and with other viral components. The N protein encapsidates genomic RNA, a function reflected in its ability to self-assemble into nucleocapsid-like particles in the absence of other viral proteins. The substrate for the packaging of nascent RNA during RNA replication is a complex between the N and phosphoprotein (P). The domains on the N protein that promote binding to P protein and self-assembly have been identified utilizing a series of N protein deletions. Two noncontiguous regions, amino acids 4-188 and 304-373 of N protein, are required for the formation of the soluble N-P complex, while deletion of amino acids 189-239 did not affect N-P binding. Amino acids 240-303 appear to be necessary for the stability of the protein. The N-terminal 398 amino acids are all required for the formation of organized nucleocapsid-like particles, since deletion of the central region from amino acids 189-373 completely abolished N-N interaction, and deletion of amino acids 4-188 and 374-492 caused the formation of unstructured aggregates.

Mutations in conserved domain I of the Sendai virus L polymerase protein uncouple transcription and replication.

To begin to map functional domains of the Sendai P-L RNA polymerase complex we wanted to characterize the P binding site on the Sendai L protein. Analysis of in vitro and in vivo P-L polymerase complex formation with carboxyl-truncations of the L protein showed that the N-terminal half of the protein was required. Site-directed mutagenesis of the Sendai virus L gene was employed to change amino acids within a highly conserved region of the N-terminal domain I from amino acids (aa) 348-379 singly or in pairs from the Sendai to the corresponding measles L sequence or to alanine. The mutant L proteins coexpressed with the viral P and NP proteins in mammalian cells were assayed for their ability to form the P-L complex and to synthesize RNA in vitro and showed a variety of defective phenotypes. While most of the mutant L proteins still formed the P-L polymerase complex, a change from serine to arginine at aa 368 and a three-amino-acid insertion at aa 379 virtually abolished both complex formation and RNA synthesis. Changes of aas 370 and 376-377 in the L protein gave only small decreases in viral RNA synthesis. Substitutions at either aas 349-350 or aas 354-355 and a three-amino-acid insertion at aa 348 in the L protein yielded enzymes that catalyzed significant transcription, but were defective in DI RNA replication, thus differentially affecting the two processes. Since DI leader RNA, but not genome RNA, was still synthesized by this class of mutants, the defect in replication appears to be in the ability of the mutant enzyme to package newly synthesized nascent RNA. Single changes at aas 362, 363, and 366 in the L protein gave enzymes with severely decreased overall RNA synthesis, although some leader RNA was synthesized, suggesting that they cannot transcribe or replicate past the leader gene. These studies have identified a region in conserved domain I critical for multiple functions of the Sendai virus L protein.

Alternative amino acids at a single site in the Sendai virus L protein produce multiple defects in RNA synthesis in vitro.

Our long-term goal is to define the catalytic domains of the L protein subunit of the Sendai virus RNA polymerase. An aberrant polyadenylation phenotype in the vesicular stomatitis virus tsG16 L protein mutant has recently been identified as a phenylalanine to serine change at amino acid 1488 (Hunt and Hutchinson, Virology 193, 786-793, 1993). To test if functional domains are conserved in the L proteins of negative-strand RNA viruses, we attempted to create a similar polyadenylation defect in the Sendai virus L protein. Nine different amino acid substitutions at the analogous site in the Sendai L protein (cysteine at amino acid 1571) were constructed by site-directed mutagenesis of the gene. Each mutant L protein was synthesized and bound to the Sendai P protein to form the P-L polymerase complex. While none of these L mutants exhibited a change in polyadenylation, the single amino acid changes yielded a variety of activities in vitro. Mutants containing valine, leucine, or phenylalanine at amino acid 1571, amino acids found naturally in the L proteins of other paramyxoviruses, yielded polymerases that had biological activity equal to or better than the wild-type (WT) polymerase. Serine or threonine substitutions in the L protein at this position also resulted in polymerases with nearly WT synthetic activity. In contrast, a glycine substitution significantly decreased overall polymerase activity, whereas a tyrosine substitution gave decreased transcription, but virtually no DI genome replication in vitro. The tyrosine-substituted polymerase may be unable to carry out the packaging step of replication, since DI leader RNA synthesis was normal in this mutant. Mutant L proteins with basic arginine or histidine substitutions were inactive in all viral RNA synthesis in vitro, although the polymerase complexes could bind the nucleocapsid template.

Double-stranded RNA adenosine deaminase activity during measles virus infection.

It has been postulated that the cellular double-stranded (ds) RNA adenosine deaminase enzyme is responsible for biased hypermutation during persistent SSPE measles infections in humans. As a test of this hypothesis we studied the effect of negative-strand RNA virus infection on enzyme activity. The adenosine deaminase activity was found in nuclear extracts of both uninfected CV-1 and A549 cells and in cytoplasmic extracts of A549, but not CV-1, cells. During measles or Sendai virus infection of either CV-1 or A549 cells the adenosine deaminase activity in the nucleus remained fairly constant up to 24 h post infection, and there was no apparent re-partitioning of the enzyme between the nucleus and the cytoplasm. Transcription complexes of Sendai virus in vitro or measles virus in vivo did not serve as substrates for the enzyme. These data suggest that even though some portion of the adenosine deaminase enzyme may be present in the cytoplasm of at least some cells during virus infection, modification of the viral RNAs by this enzyme, if it occurs at all, must be at a very low level not directly detectable by biochemical analysis.

An amino-proximal domain of the L protein binds to the P protein in the measles virus RNA polymerase complex.

The RNA polymerase of measles virus consists of two virus-encoded subunits, the L and P proteins with 2183 and 507 amino acids, respectively. When these proteins were coexpressed from plasmids in a mammalian expression system, a complex was formed as detected by the coimmunoprecipitation of the L protein with the P protein by anti-P antibodies. Pulse-chase experiments showed that complex formation increased the stability of the L protein. We have used the coimmunoprecipitation assay in conjunction with a series of C-terminal truncations of the L protein to map the region of the L protein which is involved in complex formation with the P protein. Mutant L proteins consisting of the N-terminal 1139, 916, 511, and 408 amino acids all bound to the P protein. An L protein truncation consisting of only the N-terminal 292 amino acids, which deleted part of the conserved domain I, however, did not bind the P protein. The data show that the N-terminal 408 amino acids of the L protein contain the P binding domain and suggest that domain I within this region of the L proteins of (-) strand RNA viruses may be important for RNA polymerase complex formation.

Complexes of Sendai virus NP-P and P-L proteins are required for defective interfering particle genome replication in vitro.

We present evidence that the formation of NP-P and P-L protein complexes is essential for replication of the genome of Sendai defective interfering (DI-H) virus in vitro, using extracts of cells expressing these viral proteins from plasmids. Optimal replication of DI-H nucleocapsid RNA required extracts of cells transfected with critical amounts and ratios of each of the plasmids and was three- to fivefold better than replication with a control extract prepared from a natural virus infection. Extracts in which NP and P proteins were coexpressed supported replication of the genome of purified DI-H virus which contained endogenous polymerase proteins, but extracts in which NP and P were expressed separately and then mixed were inactive. Similarly, the P and L proteins must be coexpressed for biological activity. The replication data thus suggest that two protein complexes, NP-P and P-L, are required for nucleocapsid RNA replication and that these complexes must form during or soon after synthesis of the proteins. Biochemical evidence in support of the formation of each complex includes coimmunoprecipitation of both proteins of each complex with an antibody specific for one component and cosedimentation of the subunits of each complex. We propose that the P-L complex serves as the RNA polymerase and NP-P is required for encapsidation of newly synthesized RNA.

Synthesis of leader RNA and editing of the P mRNA during transcription by purified measles virus.

A transcription system with detergent-disrupted purified measles virus was developed. Synthesis of authentic, full-length measles virus N, P, M, and F mRNAs by purified virus occurred as identified by dot-blot hybridization analysis of individual measles virus clones and gel electrophoresis. The relative abundance of the first five viral mRNAs synthesized in vitro decreased significantly with their distance from the 3' end. The addition of the soluble protein fraction from uninfected A549 cells stimulated overall viral RNA synthesis but did not alter the relative abundance of each of the mRNAs. Measles virus synthesized in vitro a leader RNA of approximately 55 nucleotides in length, suggesting that like other negative-strand viruses, transcription initiated only at the 3' end of the genome RNA. Purified measles virus also catalyzed RNA editing during the synthesis of the P mRNA as shown by modified primer extension analysis of the mRNA products and by translation of the modified RNA into the V protein in rabbit reticulocyte lysates. These data suggested that the RNA editing activity was virus encoded.

Host cell proteins required for measles virus reproduction.

We have developed a cell-free system derived from measles virus-infected cells that supported the transcription and replication of measles virus RNA in vitro. The data suggest that tubulin may be required for these reactions, since an anti-beta-tubulin monoclonal antibody inhibited viral RNA synthesis and the addition of purified tubulin stimulated measles virus RNA synthesis in vitro. Tubulin may be a subunit of the viral RNA polymerase, since two different anti-tubulin antibodies, one specific for the beta- and another specific for the alpha-subunit of tubulin, coimmunoprecipitated the measles virus L protein as well as tubulin from extracts of measles virus-infected cells. Other experiments further implicated actin in the budding process during virus maturation, as there appeared to be a specific association of actin in vitro only with nucleocapsids that have terminated RNA synthesis, which is presumably a prerequisite to budding.

The vesicular stomatitis virus L protein possesses the mRNA methyltransferase activities.

We have previously shown that the vesicular stomatitis virus (VSV) host range mutant, hr 1, is completely defective for the mRNA methyltransferase activities, but can synthesize full-length, unmethylated mRNAs in vitro [S. M. Horikami and S. A. Moyer (1982). Proc. Natl. Acad. Sci. USA 79, 7694-7698] and in vivo [S. M. Horikami, F. De Ferra, and S. A. Moyer (1984). Virology 138, 1-15]. Here we have used the hr 1 mutant to identify the viral protein which possesses the methyltransferase activities. The wild-type VSV L and NS proteins, subunits of the viral RNA polymerase, were separately purified and added to high salt dissociated mutant hr 1 nucleocapsids for in vitro transcription reactions. The results show that the purified wild-type L protein, but not the NS protein, restores methylation and thus possesses the viral mRNA methyltransferase activities.

Characterization of the infections of permissive and nonpermissive cells by host range mutants of vesicular stomatitis virus defective in RNA methylation.

Two host range mutants of VSV, hr 1 and hr 8, which, unlike the wild-type virus, have a mRNA methylation defect and direct the in vitro synthesis of full-length capped but unmethylated viral mRNAs have been described previously (S.M. Horikami and S.A. Moyer, 1982, Proc. Natl, Acad. Sci. USA 79, 7694-7698). It is shown that the in vivo nonpermissive infection of HEp-2 cells by either of these two mutants is characterized by the reduced synthesis of full-length mRNAs at levels characteristic of primary transcription and the total lack of synthesis of genome-length RNA. The VSV mRNAs synthesized by either mutant in HEp-2 cells are not translated either in vivo or in vitro in mRNA-dependent rabbit reticulocyte lysates. Subsequent isolation and analysis of the mRNAs from infected HEp-2 cells has shown that the 5' termini of the messages contain a cap structure which is guanylylated, but unmethylated (GpppA), a finding that might account for the lack of translatability. Hence these mutants are unable to properly methylate mRNAs whether they are synthesized in vitro or in vivo within nonpermissively infected cells. It is also shown that unlike hr 1, the undermethylation of mRNA synthesized by hr 8 is partially reversible by the addition of high levels of AdoMet in vitro. It is interesting to note, therefore, that permissive baby hamster kidney (BHK) cells have a 10-fold higher level of endogenous AdoMet than the nonpermissive HEp-2 cells. Unlike singly infected cells, the coinfection of HEp-2 cells with either hr mutant and a poxvirus yields a permissive infection for these two host range mutants. Analysis of the VSV mRNAs produced in vivo under the conditions of rescue reveals the presence of fully methylated caps (7mGppp(m)Am), suggesting that poxvirus may rescue the mutants by converting the VSV mRNAs to a translationally active form due to methylation by the cytoplasmic poxvirus mRNA methyltransferase enzymes. Both mutants are, however, able to grow normally in permissive BHK cells. An analysis of the translationally active mRNAs from infected permissive cells shows the presence primarily of a 5'-monomethylated cap, 7mGpppA. Finally, we have examined the nonpermissive infections of two other host range mutants of VSV (hr 5 and hr 7). Unlike mutants hr 1 and hr 8 described above, these two mutants synthesize mRNA in HEp-2 cells which is translated both in vivo and in vitro.

Host range mutants of vesicular stomatitis virus defective in in vitro RNA methylation.

The viral RNA polymerase of detergent-treated vesicular stomatitis virus normally synthesizes viral mRNAs in vitro that are both guanylylated and methylated to give 5'-terminal 7mGpppAm caps. We have characterized a virus host range mutant, hr 1, that is totally defective in vitro in the methylation of mRNA, although full-length polyadenylylated mRNAs with 5' termini of the form GpppA are synthesized in normal yields. A second mutant, hr 8, is partially defective in methylation and synthesizes mRNAs in vitro with primarily GpppA and some GpppAm 5' termini. When used for in vitro translation, the unmethylated hr 1 mutant mRNA shows, as expected, reduced synthesis of viral proteins. These data provide direct evidence that the vesicular stomatitis virus-associated methyltransferase activities are virus encoded.