key: cord-0687278-l6wh0gud
authors: Galinski, Mark S.; Troy, Roberta M.; Banerjee, Amiya K.
title: RNA editing in the phosphoprotein gene of the human parainfluenza virus type 3
date: 1992-02-29
journal: Virology
DOI: 10.1016/0042-6822(92)90020-p
sha: e17116c3e2697c586f3e36296479f931f3f6c48e
doc_id: 687278
cord_uid: l6wh0gud

Abstract RNA editing of the human parainfluenza virus type 3 (HPIV3) phosphoprotein (P) gene was found to occur for the accession of an alternate discontinuous cistron. Editing occurred within a purine-rich sequence (AAUUAAAAAAGGGGG) found at the mRNA nucleotides 791–805. This sequence resembles an HPIV3 consensus transcription termination sequence and is located at the 5′-end of the putative D protein coding sequences. Editing at an alternate site (AAUUGGAAAGGAAAGG), mRNA nucleotides 1121–1136, for accession of a conserved V cistron, which is present in a number of paramyxovirus P genes, was not found to occur in HPIV3. In contrast with many other paramyxoviruses, editing was indiscriminate with the insertion of 1–12 additional G residues not present in the gene template. RNA editing was found to occur in both in vivo (HPIV3 infected cells) and in vitro (purified nucleocapsid complexes) synthesized mRNAs. Further, the in vitro prepared mRNA was edited regardless of whether the nucleocapsid complexes were transcribed in the presence or absence of uninfected human lung carcinoma (HLC) cell lysates. These results support the notion that RNA editing appears to be exclusively a function of viral proteins.

The phosphoprotein genes of members of the Paramyxovirus and Morbillivirus genera of the Paramyxoviridae family are unique among the nonsegmented negative-strand RNA viruses in their capacity to encode multiple proteins (Kingsbury 1990 (Kingsbury , 1991 . Interestingly, the protein coding information is contained within a variety of continuous, discontinuous, and overlapping cistrons. Accession of this information is dependent upon transcriptional editing of the primary mRNA transcripts (discontinuous cistrons) as well as translation initiation at alternate codons (overlapping and continuous cistrons).

(BPIV3; Pelet eta/., 1991) ; and Sendai virus (Vidal eta/., 1990) . While SV5, HPIV2, HPIV4A, HPIV4B, and mumps viruses require the insertion of nontemplated nucleotides into their respective P mRNAs for the synthesis of complete P proteins, measles virus (Bellini et al., 1985) HPIVl (Matsuoka et al., 1991) , HPIV3 (Galinski et a/., 1986b; Luk et a/., 1986; Spriggs and Collins, 1986b) , BPIV3 (Sakai et al., 1987; Pelet et al., 1991) and Sendai virus (Giorgi eta/., 1983) do not. This latter group of viruses encode their P proteins in single continuous cistrons; however, in addition to the continuous P cistrons, these genes contain overlapping (continuous) C protein cistrons which are accessed via independent translation initiation, Molecular cloning and sequence analysis of the sim-A notable feature of the 5'-proximal (mRNA) cistron in ian virus type 5 (SV5) phosphoprotein (P) gene/mRNA SV5 (designated as protein V in SV5), mumps, HPIV4A, revealed that insertion of nontemplated nucleotides and HPIV4B is the presence of a conserved amino acid during transcription of the P mRNA was necessary for sequence (approximately 55 amino acids) at the C-terthe fusion of two discontinuous (and overlapping) cismini of the encoded proteins. A prominent feature trons in order to produce a template capable of encodamong the residues in this domain is the presence of ing a complete P protein (Thomas et al., 1988) . The six highly conserved cysteine residues. The similarity addition of nontemplated nucleotides in the P mRNA of these sequences to motifs found in nonrepetitive occurred within a guanosine track just downstream zinc fingers has been noted (Thomas et a/., 1988) . The from a purine-rich sequence which closely resembles existence of similar amino acid sequences in "cryptic" an SV5 transcription terminator. Subsequently, similar cistrons, present in the P genes of other paramyxoviobservations for RNA editing were made for measles ruses, suggests some functional or biological signifivirus (Cattaneo et al., 1989) ; mumps virus (Elliott et al., cance for the maintenance of these sequences within 1990; Paterson and Lamb, 1990) ; human parain-this group of viruses. These observations, together fluenza viruses types 2 (Ohgimoto et al., 1990; South- with the known association of the P protein with the ern er a/., 1990) 4A, and 4B (HPIV4A and HPIV4B; nucleocapsid complex (Chinchar and Portner, 1981; Kondo et al., 1990) ; bovine parainfluenza virus type 3 Deshpande and Por-tner, 1985; Hamaguchi et al., 1983; Ryan and Kingsbury, 1988; Ryan et al., 1990; Ryan and Portner, 1990) would suggest a transcriptional regulatory function for this domain. The HPIV3 P gene/mRNA nucleotide sequence has been detailed in a number of independent reports (Galinski et al., 198613; Luk et al., 1986; Spriggs and Collins, 198613) . The molecular organization of the P gene allows for the encoding of continuous cistrons for the P, C, and putative D proteins. And, as described above, a cryptic V cistron, which shares 33% amino acid identity with the C-terminus of the SV5 V protein, is also present. A sequence, similar to the SV5 RNAediting site, is adjacent to the cryptic V cistron and an additional sequence, similar to an HPIV3 transcription terminator, is present upstream of the D cistron. In both instances, purine-rich tracts are followed by a string of guanosine residues. Interestingly, the number of G nucleotides reported from the various laboratories in the mRNA sequences proximal to the D cistron differ, with 5, 8, or 11 Gs being reported.

In this report we characterize the transcriptional editing that occurs in HPIV3. The results presented demonstrate that RNA editing occurs proximal to the D cistron and not at the V cistron, and that the number of nucleotides inserted at the D editing site is indiscriminate. In addition to mapping the location of transcriptional editing, we provide evidence that this process is solely a function of viral proteins and does not appear to require any cellular factor(s).

The preparation and maintenance of the CV-1 cell line and viral stock of HPIV3 (HA-l ; NIH 47885; catalog number V323-000-020) was as previously described (Galinski et a/., 1986a) . Radiolabeling of uninfected and HPIV3-infected CV-1 cells was as previously described (Wechsler et a/., 1985) using [35S] methionine.

The various RNA species, including total cytoplasmic RNA and nucleocapsid-associated RNA, were prepared from uninfected and HPIV3-infected CV-1 cells as previously -described (Galinski et al., 1986a) . The recombinant clane, 33-9, containing a full-length copy of the P mRNA in the transcription vector pSPl9, was used to synthesize RNA with a "fixed" sequence as previously described (Galinski et a/., 1986b) .

Nucleocapsid-associated genomic RNA (vRNA) was obtained from HPIV3-infected CV-1 cells by banding nucleocapsid cores in cesium chloride as previously described (Udem and Cook, 1984; Galinski et a/., 1986a) . Free cytoplasmic RNA, which pelleted through the cesium chloride, consisted of total cytoplasmic RNA depleted of any viral genomic/antigenomic RNA.

In vitro synthesized mRNA, a generous gift from Bishnu De (Cleveland Clinic Foundation) , was prepared as previously described (De et al., 1990 (De et al., , 1991 using purified nucleocapsid complexes in the presence or absence of cell lysates. Transcribed mRNA was removed from the genomic vRNA following oligo dT-cellulose chromatography.

cDNA amplification and recombinant library construction cDNA was prepared using various RNA templates and synthetic oligodeoxyribonucleotides complementary to the genomic viral RNA (vRNA) nucleotides 701-720 (5'-TGCTGCGACACCAGATGATG; P + 701) or complementary to the P mRNA nucleotides 1221-1240 (5'-TTGTCTTGATATAAATCTAG; P-l 240) as previously described (Galinski et al., 1986a) . Briefly, the RNA templates were mixed with 200 ng of each primer in water, incubated at 95" for 1 min, and allowed to cool on ice. The reactions were adjusted with 10X stocks to a final concentration of 50 mM Tris-HCI, pH 8.0, 50 mA# KCI, 10 mhll dithiothreitol, 5 mM magnesium chloride, 30 U/30 ~1 RNAsin, 20 U/30 ~1 AMV reverse transcriptase, and incubated at 37" for 60 min. Four (30 ~1) cDNA reactions were performed using the following RNA templates (and 200 ng primer): (1) total cytoplasmic RNA from uninfected CV-1 cells (P-1240) (2) total cytoplasmic RNA (depleted of nucleocapsid vRNA) from HPIV3 infected CV-1 cells (P-l 240),

(3) nucleocapsid vRNA (P + 701) and (4) SP6 in vitro transcribed P specific mRNA (P-1240) as previously described (Galinski et a/., 1986a) . Following first strand cDNA synthesis, the entire reaction was transferred to a 500 ~1 eppendorf tube and adjusted with 1 OX stock solution to make a final concentration of 10 mM Tris-HCI, pH 8.0, 50 mM potassium chloride, 1.5 mA# magnesium chloride, 0.250 mM each deoxyribonucleotide triphosphate, 200 ng/lOO ~1 P + 701/P-1240 primer, and 2U/lOO ~1 Taq DNA polymerase. The reaction was overlaid with mineral oil and amplified by polymerase chain reaction (PCR) in a Perkin-Elmer Cetus DNA Thermal Cycler using the following parameters: initial denaturation, 94", 7 min; cycle denaturation, 94", 3 min; cycle annealing, 37", 2 min; cycle extension, 72", 3 min; final extension, and 72", 3 min. The PCR reaction was run for 35 cycles and then held at 4". Following the amplification of the cDNA, the ends of the amplified DNA were repaired by the addition of 1 U of Klenow fragment to the PCR reaction and further incubated at 37" for 30 min. The amplified cDNA was puri-fied following electrophoresis in low melting 1% agarose gels as previously described (Galinski et a/., 1983) .

The PCR amplified DNA was cloned into the Smal site of the vector pGEM-3. Following ligation with T4 DNA ligase, the samples were treated with Smal to reduce the background intramolecular religations of the vector. Plasmids were transformed into the bacterial host JM83 and selected with ampicillin as previously described (Galinski et al., 1986a) . Recombinant clones were identified using P gene-specific probes which were prepared by end-labeling of restriction endonuclease fragments. The fragments were separated on 1% low melting agarose gels and then used to probe recombinant colonies by in situ hybridization as previously described (Galinski et a/., 1986a) . Mini plasmid preparations of positive colonies were prepared and sequenced by dideoxy sequence analysis (U.S. Biochemical Sequenase Kit, following the manufacturer's protocol) using the same oligonucleotides utilized above for cDNA amplification.

Since the PCR-generated clones were not full-length copies of the P gene/mRNA, reassembly of transcription vectors containing defined sequences was performed. Selected PCR generated clones, with either six or seven Gs were used to reconstruct full-length P gene sequences. A Fokl to C/al restriction endonuclease fragment spanning the editing/chatter site was reassembled into the plasmid 33-9 by religation of a BarnHI to Fokl fragment from the clone 33-9 (N-terminus P) and a C/al to BarnHI fragment from the clone 33-9 (C-terminus of P and pSPl9 vector). Recombinant clones were identified by in situ hybridization and sequenced across the fused fragments domains as previously described (Galinski et al., 1986b) .

The recombinant pSPl9 plasmids 33-9 (containing a full-length P protein gene) and pPG6 and pPG7 (containing six and seven Gs, respectively, at the D editing site) were used to produce RNA transcripts, following restriction endonuclease digestion with Pvull, as described by Konarska et al. (1984) . The RNA transcripts, recovered following DNAse I treatment, phenol extraction, and successive ethanol precipitations, were then used to charge a translation system.

Translation of in vivo and in vitro produced RNA transcripts was performed in a rabbit reticulocyte lysate as directed by the manufacturer (Promega Biotec) using [?S]methionine as the radiolabel. The translation products, total or immune precipitated, were resolved by polyacrylamide gel electrophoresis on 10% polyacrylamide gels (Maizel, 1971 ) followed by fluorography (Bonner and Laskey, 1974) .

Analysis of nucleotide and protein sequences was performed using algorithms built by Queen and Korn (1984; MicroGenie) and Devereux et al. (1984; Genetics Computer Group) .

The utilization of discontinuous cistrons for encoding a paramyxoviral P protein was first described for SV5 (Thomas et a/., 1988) . Figure 1 illustrates the cistron usage map of the SV5 P gene as compared with that for HPIV3. HPIV3 follows the more general pattern of a single continuous P cistron with an alternate overlapping C cistron similar to that first described for Sendai virus (Giorgi era/., 1983; Shioda et al., 1983) . Cloning and sequence analysis of the HPIV3 mRNA/genes by a number of labs (Galinski et a/., 1986b; Luk et a/., 1986; Spriggs and Collins, 1986b; Galinski, unpublished observations) revealed a heterogeneity in the number of G nucleotides between position 801 and 805, which is proximal to the 5'-end of the D cistron ( Fig. 1, labeled D) . Interestingly, the sequence just upstream of this G tract, AATTAAAAAA, resembles the consensus transcription termination sequence, AA-TAACWAAA, found in the gene-end boundaries of HPIVS (Spriggs and Collins, 1986a) . Together with the observation that transcriptional editing occurs in a number of paramyxoviruses at similar sequences suggests that editing is occurring at this site for accession of the D cistron. Since the cryptic V cistron is also present in the HPIV3 P gene, the question of whether transcriptional chattering occurs in the accession of the D and/or V cistrons needed to be resolved.

Initially, restriction fragment length polymorphisms present in PCR amplified cDNA was examined. The results of restriction site analysis indicated that the PCR amplified cDNA was homogeneous (data not shown). Thus it was necessary to analyze individual cloned molecules to determine whether any sequence heterogeneity was present. The amplified cDNA was cloned into the Smal site of pGEM-3 and recombinant clones were identified by in situ hybridization using radiolabeled P gene-specific sequences as a probe. Individual clones were then sequenced across the putative editing sites.

Approximately 50 clones, derived from either vRNA or mRNA templates, were examined. All the clones from the vRNA template had invariant sequences spanning both putative editing sites (the sequence was identical to that determined by dideoxy sequence analysis of the vRNA). In contrast, clones derived from the mRNA template had variable numbers of G residues at the D editing site and an invariant sequence at the V editing site. The results for the mRNA derived clones are illustrated in Fig. 2 . This figure is a composite of information (dideoxy G or dideoxy C sequencing reactions only) from several representative mRNA clones aligned with the precise sequence spanning the two editing sites derived from a genomic RNA clone. Panel D shows the sequence across the D editing site while panel V shows the sequence across the V editing site. This figure serves to illustrate the indiscriminate insertion of additional G residues into the mRNA template at the D editing site beyond the 5 Gs which are present in the genomic template and, further, the absence of any additional G insertions at the V editing site. The number of Gs inserted into the mRNA template varied from 5-14 residues. These results demonstrate that RNA editing occurs exclusively at the D editing site.

Inspection of Fig. 1 shows that downstream of the D editing site there are a number of stop codons in the same reading frame (+2) as the V cistron. Thus any editing which occurs with the insertion of 1,4, 7, or 10, additional G residues would result in the generation of a truncated P cistron following fusion to the termination codons.

The role of host cell factors in optimal HPIVS transcription has recently been described (De et al., 1990 (De et al., , 1991 . Extension of these observations to address the question of whether host cell factors participate in RNA editing, or if this process is dependent solely upon viral components was pursued. Nucleocapsid complexes were isolated from purified HPIV3 virions and used in an in v&o transcription system. Transcription was performed either in the presence or absence of uninfected cell lysates. The mRNA was separated from the genomic RNA by oligo dT-cellulose chromatography. The in vitro synthesized mRNA was then used as a template for cDNA synthesis, PCR amplification, and molecular cloning and sequence analysis as described above for the in viva synthesized RNAs. Individual clones were identified and then sequenced across the two putative editing sites. As was observed with the in viva synthesized mRNA, sequence heterogeneity was observed only at the D editing site, and not at the V editing site. Fig. 3A summarizes the results of the in viva and in vitro produced RNAs. Since the putative editing site for the accession of the V cistron was invariant in all clones, only the results for the D editing site are shown in this figure.

The effects on cistron usage following the insertion of nontemplated nucleotides in the mRNA are shown in Fig. 4 . In this figure, the three potential reading frames (+l, +2, +3) in the faithfully transcribed P gene and following the addition of one or two nontemplated nucleotides in the edited transcripts are indicated in (A). In all instances following editing, the V cistron is never fused to any other cistron and thus would not be expected to be functional. In addition, since the D editing site is downstream of the C cistron termination codon, this cistron is never affected by the addition of nontemplated nucleotides.

The addition of a single nucleotide at the D editing site results in the introduction of a stop codon and the subsequent truncation of the P cistron. This truncated protein consists of 242 N-terminal residues of P and six additional residues (KRERLV) followed by the stop codon. Further, the C-terminal 361 amino acids of the P cistron are fused to nine of the N-terminal residues of the D cistron including an initiating methionine (MTKELKKGG). Displacement of the remaining portion of the D cistron does not provide any methionine codon for internal translation initiation.

The most noticeable feature following the addition of two nontemplated nucleotides at the D editing site is the fusion of the main portion of the D cistron (residues 1 O-l 40) onto the N-terminal 241 residues of the P cistron. Displacement of the C-terminal 361 residues of the P cistron does not provide for an initiating methionine within 1300 nucleotides of the 5'-end of the mRNA, and thus, this portion of the P cistron would not be expected to be functional.

Since the insertion of three (or multiples of three) nucleotides into the mRNA will add a new codon, multi- The cistron usage map in all three reading frames of the faithfully transcribed HPIV3 P gene mRNA and the mRNAs produced followrng the insertion of one or two additional G residues at the D editing site (denoted with a vertical line). The primary mRNA transcript (faithful copies of the P gene) encodes intact P and C proteins. The principal products following the introduction of a single G are truncated P and C proteins, while the introduction of two G residues leads to fusion of the D cistron and the amino terminus of the P cistron. Recombinant transcription vectors containing genes with fixed number of G residues at the D editing site were assembled as described under Materials And Methods. Each of these clones was transcribed and used to charge a rabbit reticulo- ple "codon" insertions in any reading frame will result in the production of nearly identical proteins except for the addition of a glycine residue. Thus, the insertion of G residues can be viewed as the accession or fusion of three cistrons, the intact P proteins, truncated P (P,) proteins, and a fused P/D protein. Fig. 3B summarizes these details by charting the percentage of recombinant clones encoding P, P,, and P/D. In order to investigate the expression of the various cistrons, P, C, P,, and P/D, we reassembled recombinant genes containing a fixed number of Gs at the D editing site to duplicate the various mRNA species illustrated in Fig. 4A . Three clones, containing either six, seven, or eight G residues in the editing site, were individually transcribed with SP6 RNA polymerase following restriction endonuclease digestion with Pvull at a site downstream of the P genes. The RNAs were then recovered and used to charge a rabbit reticulocyte lysate system for the production of various proteins. Proteins were resolved by SDS-PAGE using 10% polyacrylamide resolving gels. The results of this analysis are shown in Fig. 4B , which compares the proteins produced by each of the fixed mRNA sequences (Lanes 5, 6, and 7) to the proteins synthesized from HPIV3-infected and uninfected CV-1 cell mRNA (lanes 3 and 4, respectively). In addition to the in vitro synthesized proteins, in viva synthesized proteins from HPIV3-infected and uninfected cells are shown (lanes 1 and 2, respectively). The results show that the encoded proteins translated from the recombinant P mRNAs are also present in the translation products of the infected cell mRNAs and that proteins with identical electrophoretic mobilities are present in HPIV3-infected cell lysates.

Fusion of discontinuous cistrons through RNA editing for the expression of an intact P protein appears to be an obligatory process for SV5, mumps, HPIV2, HPIV4A, and HPIV4B (PII group; indicating two discontinuous cistrons). In contrast, measles, Sendai, BPIV3, HPlVl , HPIV3, NDV, and CDV encode their respective P proteins in a single cistron (PI group). For many of these latter viruses, RNA editing and fusion of discontinuous cistrons have been shown to be a feature for the accession of alternate V or D cistrons. The distinction in the molecular organization and expression of the P genes between these two groups is congruent with the amino acid similarities shared among the P proteins from each group of viruses. Thus, the P pro teins of the PII group appear to be more closely related to one another than they are to the PI group. Con-versely, the PI group appears to be more closely related to one another then they are to the PII group.

The distinction between these two groups is also apparent when analysis is extended to the nucleotide sequence found at the RNA editing site. Alignment of the viral RNA sequences accommodates two related consensussequences: GUAA ;CGU uuuuuuccc PI :UAA""C"CCC PII t AG

In the PI sequences the stretch of U residues appears to vary with either five for Sendai, NDV, CDV, and MV, or six for BPIV3 and HPIV3 (indicated with an arrow). These sequences, which are pyrimidine-rich, resemble transcription termination sequences found at the gene-end boundaries of these viruses . This resemblance suggests that the insertion of nontemplated nucleotides in the P mRNA may occur in a fashion similar to that for polyadenylation of the mRNAs. However, unlike polyadenylation, a more limited number of residues is added to the 3'-end and transcription termination does not occur.

In paramyxoviruses, the gene-end boundaries are composed of three domains, transcription termination, intergenic, and transcription initiation sequences. If the addition of nontemplated nucleotides is occurring through a mechanism analogous to aberrant transcription termination, then a role for the intergenic and transcription initiation sequences in proper transcription termination is suggested since these sequences are not present in the editing site. In further support of this notion is the observation that the viral RNA template, which has an invariant sequence, would need to suppress transcriptional editing as the polymerase complex crossed the RNA editing region to make viral complementary RNA. The utilization of a common control mechanism for these processes would be a parsimonious means for the virus to control transcription (both polyadenylation and polyguanylation) and replication (replicative and progeny RNAs).

The results of the in vitro synthesized mRNA suggest that RNA editing is a function solely of virion-associated proteins. That is, the addition of host cell lysate did not appreciably alter the pattern of clones containing additional nontemplated nucleotides. While these data are not conclusive in that host cell factors packaged in the virions (i.e., actin, tubulin, or other as yet uncharacterized proteins) might participate in this process, it is certainly suggestive that viral proteins contain all the requisite functions. Until an in vitro transcription system using sufficiently defined products (i.e., reconstituted nucleocapsid complexes from recombinant proteins and RNA) is available, the exclusive role of viral proteins in this process will remain circumstantial. Finally, these results are also in agreement with the observations that RNA editing is a function of (primary) transcription as discussed above.

The results presented here indicate that the cryptic V cistron in HPIVS is not expressed through transcriptional editing. However the D cistron is accessed in a similar fashion as the V cistron is in other paramyxoviruses. There are no significant amino acid sequence similarities between the D and V cistrons. In contradistinction, the D cistron contains a single cysteine and is extremely rich in glutamine (170/o), properties which are divergent with those observed for the V cistron. Clearly, accession of the D cistron was not the result of selection for a protein with similar properties as V.

Frameshifting by eucaryotic ribosomes for the accession of discontinuous cistrons has been shown to occur in both retroviruses (Craigen and Caskey, 1987) and coronaviruses (Brierly et a/., 1989) . The formation of a pseudoknot downstream of a "slippery" RNA sequence (UUUAAAC) appears to be essential for effective shifting of the reading frame and translational fusion of an alternate cistron. Comparison of the RNA sequence downstream of the D editing site did not disclose any sequence homologies with the frameshift sequences. This result supports the notion that accession of the V cistron does not occur via ribosomal frameshifting.

Recently the HPIVl P gene/mRNA nucleotide sequence has been determined. In contrast with other paramyxoviruses, HPIVl does not contain an intact V cistron (Matsuoka et al., 1991) . Thus, although remnants of the amino acid sequence of V can be found within the P gene, there are several stop codons spanning this region. In addition, there is no RNA editing sequence proximal to the 5'-end of the cistron. These results, together with the results reported here for HPIV3, indicate that V is nonessential for replication in these two viruses.

Interestingly, there is another cistron, just downstream of the C cistron in HPIVl , remarkably similar in position to that observed for the D cistrons in HPIV3 and BPIV3. However, the encoded amino acids show no significant homology with the D proteins, and there is no RNA editing site present for the accession of this cistron.

The BPIV3 virus was recently shown to access both the D and V cistrons by utilizing an RNA editing site proximal to the BPIV3 D cistron . In this virus, the number of G residues inserted into the mRNA template was found to be indiscriminate as was reported here for HPIV3. However, there are no stop codons in any of the three potential reading frames between the RNA editing site and either the D or V cistrons. This results in the fusion of either the D cistron or V cistron to the amino terminus of the P cistron. Thus, BPIV3 has selected an alternate editing site for the accession of two different discontinuous cistrons.

It would appear that there is a spectrum of "types" of P genes in the paramyxoviruses including: (1) those which obligately require RNA editing for the expression of an intact P protein and normally access the V cistron; (2) those which edit their RNA, presumably for access of the V cistron; (3) BPIV3 which accesses both the V and D cistrons; (4) HPIV3 which accesses only the D cistron; and finally (5) HPIVl , which does not edit its P mRNA and expresses the intact P (and C) protein only.

A search of the Swiss, GenPept, and National Biomedical Research Foundation data banks for proteins with similar amino acids to the P/D proteins did not reveal any significant homologies with nonparamyxoviral proteins. In addition, a search for similarity to 18 proteins with defined structural motifs did not reveal any resemblance with these domains. The P proteins of Paramyxoviridae are extremely divergent at the amino acid level (Galinski and Wechsler, 1991) . In fact, the P proteins are the least conserved of all of the viral proteins. If one seeks to define functionally important domains within these proteins by sequence alignment then this wide divergence rapidly obscures any salient features. Thus, it is not unexpected that sequence similarity searches of the protein data banks did not reveal any significant similarities. These results again underscore the uniqueness of the P proteins.

Measles virus P gene codes for two proteins. 1. Viral

A film detection method for tritium-labeled proteins and nucleic acids in polyacrylamide gels

Characterization of an efficient coronavirus ribosomal frameshifting signal: Requirement for an RNA pseudoknot

Measles virus editing provides an additional cysteine-rich protein

Functions of Sendai virus nucleocapsid polypeptides: Enzymatic activities in nucleocapsids following cleavage of polypeptide P by Staphylococcus aureus protease V8

Translational frameshifting: Where will it stop?

Characterization of an in vitro system for the synthesis of mRNA from human parainfluenza virus type 3.1. Viral

Human parainfluenza virus type 3 transcription in vitro: role of cellular actin in mRNA synthesis

Monoclonal antibodies to the P protein of Sendai virus define its structure and role in transcription

A comprehensive set of sequence analysis programs for the VAX

Strain variable editing during transcription of the P gene of mumps virus may lead to the generation of non-structural proteins NSl (V) and NS2. 1

Paramyxoviridae: Transcription and replication. ln

Molecular cloning and sequence analysis of the human parainfluenza 3 virus RNA encoding the nucleocapsid protein

Molecular cloning and sequence analysis of the human parainfluenza 3 virus mRNA encoding the P and C proteins

The molecular biology of the Paramyxovirus genus

Comparison of sequence diversity in several cytoplasmic polyhedrosis viruses

Sendai virus contains overlapping genes expressed from a single mRNA

Transcriptive complex of Newcastle disease virus. I, Both Land P proteins are required to constitute an active complex

Paramyxoviridae and their replication. ln "Virology

The Paramyxoviruses

Recognition of cap structure in splicing in vitro of mRNA

Sequence analysis of the phosphoprotein (P) genes of human parainfluenza type 4A and 48 viruses and RNA editing at transcript of the P genes: The number of G residues added is imprecise

Messenger RNA encoding the phosphoprotein (P) of human parainfluenza virus 3 is bicistronic

Polyacrylamide gel electrophorests of viral proteins. ln

The P gene of human parainfluenza virus type 1 encodes P and C proteins but not a cysteine-rich V protein

Sequence analysis of P gene of human parainfluenza type 2 virus: P and cysteine-rich proteins are translated by two mRNAs that differ by two nontemplated G residues

RNA editing by G-nucleotide insertion in mumps virus P-gene mRNA transcripts

The P gene of bovine parainfluenza virus 3 expresses all three reading frames from a single mRNA editing site

A comprehensive sequence analysis program for the IBM personal computer

Carboxyl-terminal region of Sendai virus P protein is required for binding to viral nucleocapsids

Localization of P protein binding sites on the Sendai virus nucleocapsid

Separate domains of Sendai virus P protein are required for binding to viral nucleocapsids

Nucleotide sequence of the bovine parainfluenza 3 virus genome: its 3' end and the genes of NP, P, C and M proteins

Sequence of 3,687 nucleotides from the 3'end of Sendai virus genome RNA and the predicted amino acid sequences of viral NP, P, and C proteins

Two nontemplated nucleotide additions are required to generate the P mRNA of parainfluenza virus type 2 since the RNA genome encodes protein V

Human parainfluenza virus type 3: messenger RNAs, polypeptide coding assignments, intergenic sequences, and genetic map

Sequence analysis of the P and C protein genes of human parainfluenza virus type 3: Patterns of amino acid sequence homology among paramyxovirus proteins

Two mRNAs that differ by two nontemplated nucleotides encode the amino coterminal proteins P and V of the paramyxovirus SV5

Isolation and characterization of measles virus intracellular nucleocapsid RNA

Editing of the Sen-da1 virus P/C mRNA by G insertion occurs during mRNA synthesis via a virus-encoded activity

Human parainfluenza virus 3: Purification and characterization of subviral components, viral proteins and viral RNA

The authors thank Bishnu De for providing the HPlV3 in vitro transcribed mRNA for analysis, and Beverly R. Heminway for technical support. This work was supported in part by Public Health Service Grant Al-26256.