Role of a Highly Conserved NH₂-Terminal Domain of the Human Parainfluenza Virus Type 3 RNA Polymerase

Abstract

The RNA polymerase complex of human parainfluenza virus type 3 (HPIV 3), a member of the family Paramyxoviridae, is composed of two virally encoded polypeptides: a multifunctional large protein (L, 255 kDa) and a phosphoprotein (P, 90 kDa). From extensive deduced amino acid sequence analyses of the cDNA clones of a number of L proteins of nonsegmented negative-strand RNA viruses, a cluster of high-homology sequence segments have been identified within the body of the L proteins. Here, we have focused on the NH₂-terminal domain of HPIV 3 L protein that is also highly conserved. Following mutational analyses within this domain, we examined the ability of the mutant L proteins to (i) transcribe an HPIV 3 minireplicon, (ii) transcribe the viral RNA in vitro using the HPIV 3 nucleocapsid RNA template, and (iii) interact with HPIV 3 P protein. Our results demonstrate that the first 15 amino acids of the NH₂-terminal domain spanning a highly conserved motif is directly involved in transcription of the genome RNA and in forming a functional complex with the P protein. Substitution of eight nonconserved amino acids within this domain by the corresponding Sendai virus L protein residues yielded mutants with variable transcriptional activities. However, one mutant in which all eight amino acids were replaced with the corresponding residues of Sendai virus L protein failed to both transcribe the minireplicon and interact with HPIV 3 P and the Sendai virus P protein. The possible functional significance of the NH₂-terminal domain of paramyxovirus L protein is discussed.

The human parainfluenza virus type 3 (HPIV 3) is an enveloped, nonsegmented, single-stranded negative-sense RNA virus with a genome of 15,462 nucleotides and is the causative agent of lower respiratory tract infections in infants and children leading to bronchiolitis, pneumonia, and croup (5, 20). The virus encodes six major proteins, and three of them, namely the fusion (F), hemagglutinin-neuraminidase (HN), and matrix (M), are found within or closely associated with the viral lipid envelope. The nucleocapsid protein (N) encapsidates the viral genome to form a helical nucleocapsid, i.e., the N-RNA template (2). Two other structural polypeptides, the phosphoprotein (P) and the large protein (L), comprise the functional RNA polymerase holoenzyme complex that binds to the N-RNA template and initiates RNA synthesis to transcribe and eventually replicate the viral genome (2, 13). The L protein is the least abundant virion protein and possesses the RNA-dependent RNA polymerase activity and also encodes posttranscriptional modification activities, such as capping, methylation, and polyadenylation of mRNAs by analogy to the L protein of vesicular stomatitis virus (1, 13). Moreover, the L and P proteins must be coexpressed in the same cell to form a functional L-P complex because the L protein seems to be unstable in the absence of P protein (16, 18, 29). Recent studies have shown that the P protein of Sendai virus, a member of the same family, is a tetramer and is probably associated with a monomer of L protein to form a functional holoenzyme (31, 32).

Sequence alignment studies of all the L proteins of nonsegmented negative-strand RNA viruses have revealed the presence of six highly conserved domains, I to VI, which are joined by variable nonconserved regions (23, 25, 27). Although these domains share extensive sequence similarity with the corresponding domains in other L proteins, the precise roles of these conserved domains in the L protein function have not been determined. So far, available evidences suggest that domain II possesses a putative RNA binding motif (25), while domain III, with four highly conserved motifs (A, B, C, and D), is postulated to comprise the “polymerase module” of the L protein (21, 23, 25). The invariant pentapeptide QGDNQ within motif C of domain III is definitely the active site of the RNA polymerase, as mutations within this motif completely eliminate RNA synthesis (19, 22, 26, 28). Domain VI possesses a putative purine nucleotide binding/ATPase site necessary for RNA polymerization (3). Recently, mutational analyses within the various conserved domains have been carried out with the Sendai virus L protein by replacing charged amino acids with alanine or the corresponding residues found in the L protein of other viruses of the same family. In both cases, the mutant L proteins exhibited a spectrum of activity ranging from partial to complete defect in transcription and/or replication and a temperature-sensitive phenotype (4, 6, 11, 12, 30). Based on these results, it is suggested that the L protein may possess modular function, with each domain contributing to certain steps in RNA synthesis (6).

The purpose of this study was to probe the structure and function relationship of the L protein of HPIV 3, including another highly conserved domain present at the NH₂-terminal end of paramyxoviruses between amino acid residues 13 to 25. A sequence [IL(Y/L)PE(C/V)HL(N/D)SPIV] within the stretch of amino acids is highly conserved among seven paramyxoviruses (Fig. 1). Using site-directed mutagenesis and deletion analyses, we have identified the amino acid residues within this domain that are critically important in transcription and in the formation of the L-P complex. We have employed a previously described HPIV 3 minireplicon system (15) and in vitro assay (10) to measure the transcription of the HPIV 3 genome by the mutant L proteins. Our results demonstrate that an NH₂-terminal region of the HPIV 3 L protein comprising 15 amino acids which encompass a highly conserved motif is directly involved in the transcription of genome RNA and in forming a functional complex with the P protein. In addition, our studies on the effect of substitution of the eight nonconserved residues located within the NH₂-terminal domain of the HPIV 3 L protein with the corresponding residues of the Sendai virus L protein demonstrate that mutant HPIV 3 L proteins exhibit variable transcriptional activities.

FIG. 1.

Schematic representation of paramyxovirus L proteins. The arbitrary locations of the conserved domains (I to VI) are indicated by boxes. The amino-terminal sequences of paramyxovirus L proteins were aligned by a CLUSTAL sequence program. Dashes represents gaps introduced to optimize alignment. Stretches of strictly or conservatively maintained amino acids are shaded and are shown in boxes. PIV3, human parainfluenza virus type 3; SV, Sendai virus; MV, measles virus; CDV, canine distemper virus; SV5, simian virus 5; PIV 2, human parainfluenza virus type 2; MU, mumps virus; NDV, New Castle disease virus.

MATERIALS AND METHODS

Sequence analysis of L protein of negative-stranded RNA virus.

The Mac Vector 6.5.3 sequence analysis software was used for multiple-sequence alignment of the L proteins.

Cells and viruses.

HPIV 3 (HA-1, NIH 47885) was propagated in CV-1 cells as described previously (8, 9). Recombinant vaccinia viruses MVA and vTF7-3 expressing T7 RNA polymerase were grown in HeLa cells.

Site-directed mutagenesis.

Oligonucleotides harboring point mutations and/or deletions in the L protein residues were employed in PCR using the Expand polymerase (Roche Biochemicals, Indianapolis, Ind.). The plasmid pGEM4-L_FLAG, with a FLAG epitope (DYKDDDDK) fused to the 3′ terminus of the HPIV-3 L gene, was used as a template for the PCR (22). The amplified fragments possessing the deletions and mutations were cloned into pGEM4-L_FLAG, and dideoxy sequencing was done to ensure the sequence integrity.

In vivo minigenome replication and luciferase assay.

In vivo minigenome replication assay was performed following the procedure described by Hoffman and Banerjee (15). Briefly, HeLa cell monolayers in a 12-well plate, grown to 90% confluency, were infected with recombinant vaccinia virus MVA, which expresses T7 RNA polymerase, at a multiplicity of infection (MOI) of 3. After 1 h at 37°C, the HPIV 3 minigenome plasmid pHPIV 3-MG(-) (200 ng) and support plasmids pHPIV 3-N (pN) (600 ng), pHPIV 3-P (pP) (750 ng), and wild-type (WT) HPIV 3-L_WT (pL_WT) or HPIV 3 mutant L proteins (pL_MUT) (100 ng) were transfected using Lipofectin (Life Technologies Inc., Rockville, Md.) according to the manufacturer's instructions. After 4 h, the transfection medium was replaced with 1 ml of Dulbecco's modified Eagle's medium (DMEM)-10% fetal calf serum. At 28 h posttransfection, the monolayers were lysed in 150 μl of lysis buffer, from which 1.5 μl of lysate (equivalent to 2.3 × 10³ cells) was then used to determine luciferase activity in a Dynatech ML2250 luminometer according to the manufacturer's specifications (Luciferase Assay Kit; Roche Biochemicals).

Primer extension analysis.

Primer extension analysis was done as described by Hoffman and Banerjee (15). Briefly, HeLa cells in six-well plates were infected with MVA at an MOI of 6. After 1 h at 37°C, the HPIV 3 minigenome plasmid pHPIV 3-MG(-) (400 ng) and support plasmids pN (1.2 μg), pP (1.5 μg), and pL_WT or pL_MUT (200 ng) were transfected using Lipofectin (Life Technologies Inc.) according to the manufacturer's instructions. For the extraction of total RNA, the cells were lysed and RNA was purified with RNA STAT-60 (Tel Test B, Friendswood, Tex.) according to the manufacturer's instructions. The RNA products of the positive-sense transcription and replication were obtained using a negative-sense oligomer end labeled with 500 μCi of [γ-³²P]ATP (6,000 Ci/mmol; NEN, Boston, Mass.), which primes at nucleotide 2 of the luciferase gene. This primer was used with 50% of the total RNA extract in a standard reverse transcription (RT) reaction with Moloney murine leukemia virus (MMLV) reverse transcriptase (Roche Biochemicals) at 42°C. The extension products were separated on a 6% acrylamide-urea gel and analyzed by autoradiography.

HPIV 3 N-RNA transcription in vitro.

HeLa cells in a six-well plate were infected with vTF7-3 at an MOI of 3 and 1 h postinfection were transfected with pP alone (2 μg) or in the presence of pL_WT or pL_MUT plasmids (200 ng) following the procedure described by De et al. (10). At 5 h postinfection, the transfection medium was replaced with DMEM containing 10% fetal calf serum. At 24 h posttransfection, the cell monolayers were washed with ice-cold phosphate-buffered saline (PBS) and harvested by scraping in PBS. The cells were pelleted by centrifugation at 800 × g for 10 min. Cytoplasmic extracts were prepared by lysing cells in three cycles of freezing and thawing in 40 μl of hypotonic buffer containing 10 mM Tris-HCl (pH 8.0), 10 mM NaCl, and 1 mM dithiothreitol (DTT). Nuclei and cell debris were removed by centrifugation for 5 min in an Eppendorf centrifuge at 4°C. The soluble cytoplasmic extract was collected for use in the transcription reaction mixture. The protein concentration was estimated as 5 mg/ml.

N-RNA templates were prepared from HPIV 3-infected CV-1 cells (10⁸) by following the procedure of Curran et al. (7) with some modifications. Briefly, the cells were harvested in PBS and resuspended in 5 ml of buffer containing 50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 0.6% NP-40, 1% Triton X-100, and 1 mM DTT. The cells were lysed by vortexing, and nuclei and cell membranes were removed by centrifugation at 10,000 × g for 5 min. The concentration of cell extract was made 6 mM in EDTA and layered onto a 20 to 40% CsCl (wt/wt) gradient and centrifuged at 38,000 × g for 2 h at 12°C in an SW41 rotor. The visible N-RNA band was collected and repurified using the CsCl gradient. Finally, the N-RNA was sedimented through 40% glycerol in 50 mM HEPES-KOH (pH 8.0), 50 mM NaCl, 0.2% NP-40, and 1 mM DTT onto a 100-μl cushion of 100% glycerol and stored in liquid nitrogen.

The in vitro transcription reaction was performed in a 50-μl total volume. The reaction mixture contained 100 mM HEPES-KOH (pH 8.0), 100 mM KCl, 5 mM MgCl₂, 1 mM DTT, 1 mM (each) ATP, GTP, and CTP, 10 μM UTP, 20 μCi of [α-³²P]UTP, 25 U of human placental RNase inhibitor, 5 μg of actinomycin D per ml, 2 μg of N-RNA, and 10 μg of the cell extract coexpressing L and P proteins. The mRNA products were analyzed in a 5% polyacrylamide-urea gel.

Coexpression and immunoprecipitation of proteins.

HeLa cells in a 12-well plate were infected with vTF7-3 at an MOI of 1. At 1 h postinfection, the cells were transfected with the plasmid pP (1 μg) or Sendai virus P plasmid (SV-P) (1 μg) and pL_WT or pL_MUT (200 ng). At 12 h postinfection, the medium was replaced with 2 ml of methionine-free DMEM, and incubation was continued at 37°C. At 14 h postinfection, the cells were labeled with 50 μCi of [³⁵S]methionine (1,175 Ci/mmol; NEN, Boston, Mass.) in 1 ml of methionine-free DMEM for 6 h. Cells were washed with PBS and lysed in 150 μl of luciferase lysis buffer. An aliquot of the lysate was used for sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blot analysis using polyclonal anti-RNP antibody. The rest of the lysate was immunoprecipitated with anti-FLAG antibody (Sigma, St. Louis, Mo.) conjugated to agarose beads by following the manufacturer's protocol. For immunoprecipitation of SV-P protein, labeled cell extracts were incubated with the monoclonal antibody (MAb) M56, which was previously conjugated to protein A-Sepharose beads (Amersham Pharmacia, Piscataway, N.J.). The immune complexes were washed with buffer containing 25 mM Tris-HCl (pH 7.5), 50 mM NaCl, and 2 mM EDTA, and the immunoprecipitated proteins were boiled in SDS-polyacrylamide gel sample buffer followed by SDS-PAGE and fluorography.

RESULTS

Sequence alignment of paramyxovirus L proteins.

The CLUSTAL sequence alignment of the HPIV 3 L protein with seven other L proteins of paramyxoviruses was carried out using the Mac Vector 6.5.3 sequence analysis software. For the alignment, the minimum percent similarity score was 40% over a 60-amino-acid window, using an open gap penalty of 6.0. As predicted earlier (25, 27), the alignment demonstrated sequence relatedness between all the L proteins, displaying a colinear amino acid sequence similarity over nearly the entire L protein, with six domains linked by nonconserved variable regions as schematically depicted in Fig. 1. However, the alignment also revealed the presence of a highly conserved region in close proximity to the NH₂ terminus between amino acid residues 13 to 25, with a unique sequence [IL(Y/L)PE(C/V)HL(N/D)SPIV] being highly conserved among the paramyxovirus L proteins. It is also noteworthy that there are only eight amino acids (residues 3 to 10) in HPIV 3 L that are different from the corresponding sequence of the Sendai virus L protein within all of NH₂-terminal residues 1 to 25. In order to ascertain the role of this domain, if any, deletion of the L protein from the NH₂ terminus as well as point mutations within this domain were obtained. The activities of the mutant L proteins were determined by the previously described HPIV 3 minireplicon system (15) and in vitro transcription assay (10), and complex formation with P protein was studied by immunprecipitation of coexpressed L and P proteins.

Transcription of luciferase gene in HPIV 3 minireplicon by mutant L proteins.

The HPIV 3 minireplicon system was previously shown to serve as an efficient assay to measure the transcription efficiency of HPIV 3 genome in vivo by quantitative luciferase reporter gene expression in the presence of exogenous support plasmids encoding N, P, and L (15). Following infection with the recombinant vaccinia virus and transfection with the HPIV 3 MG(-) minireplicon fused to the reporter luciferase gene and with the support plasmids pN, pP, and pL_wt or pL_mut, bearing a T7 promoter, an encapsidated negative-sense minireplicon is formed. The minireplicon is subsequently transcribed by support plasmid-encoded proteins to generate the luciferase mRNA that are translated into the luciferase protein. Cytoplasmic extracts prepared from such infected and/or transfected HeLa cells were used to determine the luciferase activity as described in Materials and Methods. The assay was performed in duplicate, and the results from three individual experiments were calculated as percent activity relative to that of the wild type. The L plasmid was modified by the introduction of a nucleotide sequence encoding the FLAG protein (DYKDDDDK) at the 3′ terminus. FLAG-tagged L was confirmed to be as active as untagged L in the minireplicon system (data not shown) and was subsequently used in all experiments. The luciferase gene expression was amplified approximately 200-fold over the background levels (lacking L plasmid) and represented as 100% expression using the wild-type L protein. As shown in Fig. 2A, the mutant NΔ2-5 retained 30% luciferase activity compared to that of the wild type while NΔ2-15 was virtually inactive, with activity very close to the background level (without L plasmid; not shown) as a control. Interestingly, deletion of the 23 amino acids from the COOH terminus of the full-length L protein (CΔ2211-2233), used as a control, displayed 80% activity compared to that of the wild type (Fig. 2A), strongly suggesting that 23 COOH-terminal nonconserved amino acids (25, 27) are dispensable for the transcriptive function of the L protein. Taken together, our results demonstrated that a deletion of 4 and 14 amino acids from the NH₂ terminus of the HPIV 3 L protein results in an inhibition of luciferase activity by 70 and 100%, respectively.

FIG. 2.

(A) Activity of mutant L proteins in minigenome replication in vivo. HeLa cells were infected with MVA and transfected with pHPIV3-MG(-) and the support plasmids encoding N, P, and L_wt or L_mut as detailed in Materials and Methods. At 24 h postinfection, cell extracts were prepared and luciferase activity was determined. The data are the averages of three independent experiments ± standard deviation. (B) Primer extension analysis of positive-sense RNA products using total RNA extract. Primer extension was performed as described in Materials and Methods from total RNA extracted from HeLa cells transfected with plasmids as described for panel A. The dideoxy sequencing ladder was made using the same primer and pHPIV3-MG(-) as the template. Relevant regions of the gel with bands corresponding to the transcription (trans) and replication (rep) products are shown. Mock, vaccinia virus-infected and mock-transfected cell extract. (C) In vitro transcription of HPIV 3 L mutants. HeLa cells were infected with vTF7-3 and transfected with L_wt or L_mut along with P plasmid. Cell extracts containing the coexpressed L and P proteins were used in transcription reaction mixtures containing N-RNA template in the presence of [³²P]UTP, and the mRNA products were analyzed in a 5% polyacrylamide-urea gel. Mock, vaccinia virus-infected and mock-transfected cell extract. (D) Analysis of L-P interaction by coimmunoprecipitation assay in vitro. ³⁵S-labeled cell extracts prepared after transfection as described for panel C were immunoprecipitated with anti-FLAG antibody and analyzed by SDS-PAGE followed by fluorography as described in Materials and Methods (top). Western blot analysis of the cell extracts using anti-RNP antibody (bottom). The positions of L and P proteins are indicated. Mock, vaccinia virus-infected and mock-transfected cell extract.

A more precise and direct approach to study the role of mutant L proteins on the synthesis of luciferase mRNA was accomplished by primer extension analysis. Total RNA was isolated from HeLa cells infected with vaccinia virus and transfected with the minireplicon plasmid pHPIV 3 MG(-) and the support plasmids pN, pP, and pL_wt or pL_mut. Primer extension analysis was carried out as described by Hoffman and Banerjee (15), using a negative-sense primer that annealed within the luciferase coding region priming synthesis near the initiation codon AUG. This primer synthesizes a cDNA that is complementary to the luciferase mRNA extending up to the end of the mRNA with products of 68 and 123 nucleotides, corresponding to transcription and replication, respectively. The primer extension reaction was carried out with MMLV reverse transcriptase for 1 h at 42°C, and DNA products were analyzed by 6% acrylamide-urea gel. As expected, bands corresponding to the predicted transcription and replication products for the wild type were readily discernible (Fig. 2B). On the other hand, the mutant NΔ2-5 exhibited a 70% reduction in RNA synthesis compared to the wild type, while NΔ2-15 was virtually inactive in synthesizing the primer-directed product. In contrast, the mutant CΔ2211-2233 retained 70% activity compared to the wild type (Fig. 2B, bottom). These results correlated well with the luciferase assay described above. Similarly, the replication products (Fig. 2B, top) were also decreased proportionately for the mutants NΔ2-5 and NΔ2-15.

Transcription in vitro by L mutants.

The in vitro transcription assay utilizing the HPIV 3 N-RNA devoid of endogenous L and P proteins was previously shown to be an invaluable system for studying the role of the L and P proteins in transcription (7, 10). HeLa cells were infected with vaccinia virus and transfected with either L_wt or L_mut plasmids in the presence of P plasmid. At 24 h postinfection, cytoplasmic extracts were prepared and used directly in an in vitro transcription reaction mixture containing purified HPIV 3 N-RNA template in the presence of α-[³²P]UTP as described and the RNA products were analyzed in a 5% urea-acrylamide gel followed by autoradiography. As shown in Fig. 2C, the N-RNA template alone (mock) had virtually no RNA-synthesizing activity while addition of an extract containing coexpressed wild-type L and P proteins resulted in an efficient transcription of the genomic N-mRNA. Cytoplasmic extract of HeLa cells coexpressing P and the mutant NΔ2-15 was totally inactive in RNA synthesis and was therefore incapable of synthesizing RNA from the N-RNA template, while the mutants NΔ2-5 and CΔ2211-2233 exhibited activities similar to that of the wild type. Western blot analysis of the cytoplasmic extracts containing coexpressed L and P proteins using HPIV 3 antibody confirmed a uniform level of expression of both L and P proteins (data not shown). It is interesting to note that the mutant NΔ2-5 appeared to be fully active in the in vitro assay, suggesting that in the minireplicon assay described above some cellular protein(s) in vivo may have adversely affected transcription by the mutant L. In contrast, NΔ2-15 was equally defective in transcription both in vitro and in vivo. Thus, deletion of 14 amino acids from the NH₂ terminus of the L protein resulted in a complete loss of L activity.

Binding of P protein with L mutants.

Next, we wanted to study the complex formation between the mutant L proteins and P by coimmunoprecipitation. The unavailability of a suitable antibody to the HPIV 3 L protein was circumvented by tagging with FLAG epitope at the COOH terminus of the L protein. HeLa cells were infected with vaccinia virus and transfected with either L_wt or L_mut plasmids in the presence of P plasmid. Cells were labeled with [³⁵S]methionine and the FLAG-tagged L proteins were immunoprecipitated with anti-FLAG antibody as described in Materials and Methods, and the precipitate was analyzed by SDS-PAGE and autoradiography. The P percent binding to L protein was normalized by PhosphorImager analysis with respect to the L protein and used in all subsequent experiments. As expected, NΔ2-15 failed to immunoprecipitate P since no band corresponding to the P protein was detected (Fig. 2D, top), while complex formation between NΔ2-5 and the P protein was 80% compared to that of the wild type. Similarly, the mutant CΔ2211-2233, which retained the NH₂-terminal residues, formed a normal L-P complex and immunoprecipitated P protein with an efficiency of 80% compared to that of the wild type (Fig. 2D, top). Western blot analysis of cell extracts with anti-RNP antibody revealed similar expression levels for the P protein (Fig. 2D, bottom). Thus, the data obtained from the in vivo and in vitro analyses together with the results from the immunoprecipitation experiments strongly suggest that an NH₂-terminal domain of the HPIV 3 L protein comprising the first 15 amino acids encompassing a highly conserved motif is required for both the polymerase activity and binding to the P protein to form the functional polymerase complex.

Mutational analyses within ILYPE residues.

Sequence alignment of seven paramyxovirus L proteins (Fig. 1) revealed that a highly conserved motif comprised of ILYPECHLNSPIV residues is present within the NH₂-terminal region which, incidentally, overlaps with the first 15 amino acids from the NH₂ terminus. To study the role of these conserved residues in transcription and P protein binding, sequential point mutations within this motif were carried out by replacement with alanine. HeLa cells were infected with vaccinia virus and transfected with the HPIV 3 MG(-) minireplicon and support plasmids pN, pP, and pL_wt or pL_mut, and cell lysates were analyzed for luciferase expression as described earlier. As seen in Fig. 3A, the luciferase activities of two mutants, I13A and L14A, were significantly reduced by 60 and 90%, respectively, compared to that of the wild type. Moreover, the Y15A and P16A mutants, as well as NΔ13-16, a mutant devoid of ILYP residues, and mutant E17G (data not shown) were totally inactive. Primer extension analysis failed to reveal any transcription products for the three mutants Y15A, P16A and NΔ13-16, and E17G (data not shown); however, bands corresponding to the products of transcription were discernible for mutants I13A and L14A, albeit at a very low level (10 and 4%, respectively) (Fig. 3B). These results indicate that the ILYPE sequence is critical for the polymerase function, with Y, P, and E being the most important residues. Similar results were noted for the replication products of these mutants (data not shown).

FIG. 3.

(A) Functional analysis of mutant L proteins by using luciferase assay. Cell lysates prepared 24 h posttransfection were assayed for luciferase activity as detailed in Materials and Methods and as described for Fig. 2A. (B) Analysis of positive-sense RNA products by primer extension. Total RNA extracted after infection and transfection with plasmids was used in primer extension as described in Materials and Methods and for Fig. 2B. Relevant regions of the gel with the bands corresponding to the transcription (trans) products are shown. Mock, vaccinia virus-infected and mock-transfected cell extract. (C) L-P complex formation with HPIV 3 L mutants. [³⁵S]methionine-labeled cell extracts were immunoprecipitated with anti-FLAG antibody and analyzed by SDS-PAGE followed by fluorography (top) as described in Materials and Methods and for Fig. 2D. Shown is a Western blot analysis of the cell extracts using anti-RNP antibody (bottom). The positions of L and P proteins are indicated. Mock, vaccinia virus-infected and mock-transfected cell extract.

Binding of P with ILYPE mutants.

The observation that Y15A and P16A mutants were defective in transcription prompted us to investigate the ability of these mutants to interact with the P protein. Accordingly, the mutant L plasmids were transfected along with the P plasmid into HeLa cells previously infected with vaccinia virus and cell lysates were prepared after labeling cells with [³⁵S]methionine and immunoprecipitated with anti-FLAG antibody. As shown in Fig. 3C, P protein binding in both the mutants Y15A and P16A was significantly reduced by 76 and 73%, respectively. Interestingly, the mutants I13A, L14A, and NΔ13-16 retained 70, 90, and 47%, respectively, P binding compared to that of the wild type (Fig. 3C, top). The expression levels of the P proteins were monitored by Western blotting with anti-RNP antibody and found to be virtually similar (Fig. 3C, bottom). It is possible that these mutant L proteins bind to the P protein, but the structural integrity of the mutant (L-P) complex is sufficiently compromised to adversely affect transcription.

Role of nonconserved residues at NH₂ terminus.

It is quite apparent from Fig. 1 that within the nonconserved amino acids at the NH₂ terminus there is only an eight-amino-acid stretch (residues 3 to 10) that is different between HPIV3 L and Sendai virus L. To probe into this nonconserved domain, we decided to study the effect of substitution of each nonconserved residue of HPIV 3 L protein with the corresponding residue of the Sendai virus L protein. We employed two different approaches to study the effect of substitution within these eight amino acids. First, each HPIV 3 L amino acid was replaced individually with the corresponding Sendai virus residue. Eight such HPIV 3 L mutants (Fig. 4A) obtained in this manner were analyzed for transcription using the HPIV 3 minireplicon system and mRNA synthesis by the primer extension analysis as well as interaction with the P protein by ³⁵S labeling and coimmunoprecipitation analysis. As shown in Fig. 4B, mutants that possessed a single-residue change displayed a differential effect on the luciferase activity and thus the transcription activity. An enhancement in the activity ranging from 3 to 25% greater than the wild type was seen for the mutants S5E, N6S, N7S, G8Q, and T9N. On the other hand, three mutants, namely T3G, E4Q, and V10P, were 90, 70, and 80% active, respectively, compared to the wild type. The extent of transcription and replication measured by primer extension analysis further demonstrated that all eight mutants could efficiently transcribe and replicate luciferase mRNA like the wild type (Fig. 4C). Moreover, all eight mutants were capable of immunoprecipitating P protein as efficiently as the wild-type L protein, thus forming normal L-P complexes (Fig. 4D, top). In addition, the expression levels of the P protein remained unaffected, as determined by Western blot analysis using anti-RNP antibody (Fig. 4D, bottom). These results led to the conclusion that each residue within the eight HPIV 3 L amino acids substituted individually with the corresponding Sendai virus residue had essentially no adverse effect on the activity of the L protein.

FIG. 4.

(A) Sequence alignment of the amino termini of HPIV 3 (top) and Sendai virus L (bottom) proteins. The nonconserved amino acid residues are shown in boxes. (B) Luciferase activity from cell extracts was measured for mutant L proteins as described in Materials and Methods and for Fig. 2A. (C) Primer extension was performed using total RNA extract transfected with plasmids as detailed in Materials and Methods and for Fig. 2B. Relevant regions of the gel with the bands corresponding to the transcription (trans) and replication (rep) products are shown. (D) Coimmunoprecipitation of L and P proteins in vitro. ³⁵S-labeled cell extracts were immunoprecipitated as described for Fig. 3C (top). Shown is a Western blot analysis of the cell extracts using anti-RNP antibody (bottom). The positions of L and P proteins are indicated. Mock, vaccinia virus-infected and mock-transfected cell extract.

However, interesting results were obtained when the HPIV 3 sequence was sequentially replaced by Sendai virus amino acids (Fig. 5A). Eight such HPIV 3 L mutants obtained in this manner were analyzed using the HPIV 3 minireplicon system as well as interaction with the HPIV 3 P protein by ³⁵S labeling and immunoprecipitation analysis as described above. Except for mutants Sen3 (TES to GQE) and Sen8 (all replaced), all other mutants showed transcription ability similar to the wild type (Fig. 5B). Moreover, the ³⁵S-labeled extracts immunoprecipitated with anti-FLAG antibody demonstrated that all the mutants were capable of coprecipitating P protein as efficiently as the wild type, with the exception of Sen3 and Sen8 (Fig. 5C, top). Western blot analysis of the cell extracts using anti-RNP antibody exhibited similar expression levels of P protein (Fig. 5C, bottom). In Sen3 and Sen8, an important change in amino acid occurred, e.g., E (acidic) and P (helix breaker), respectively. Clearly, these two amino acids must have a profound effect on the structural integrity of the L protein to down regulate the transcriptive property of the L protein. However, when the ES sequence was subsequently restored in Sen4, the transcriptional activity was recovered similar to the wild type. These series of mutational studies clearly showed that the nonconserved eight amino acids may impart specificity to the HPIV 3 L protein, thereby allowing an optimal interaction with the P protein to form the functional holoenzyme.

FIG. 5.

(A) Sequences of mutant HPIV 3 L proteins. The amino-terminal residues of HPIV 3 and Sendai virus L proteins shown. The amino acid residue(s) within the boxes represent the corresponding residue(s) from Sendai virus L protein that were substituted within the HPIV 3 L protein. (B) Analysis of mutant L proteins using luciferase assay. The luciferase activity was determined for mutant L proteins as described in Materials and Methods and for Fig. 2A. (C) L-P complex formation between HPIV 3 L mutants and P in vitro. [³⁵S]methionine-labeled cell extracts were immunoprecipitated and analyzed by SDS-PAGE followed by fluorography as described earlier in Materials and Methods and for Fig. 2D (top). Shown is a Western blot analysis of the cell extracts using anti-RNP antibody (bottom). Mock, vaccinia virus-infected and mock-transfected cell extract.

Finally, immunoprecipitation experiments were performed using Sendai virus P protein to determine if the mutant Sen8 resembling the NH₂-terminal 25 amino acids of Sendai virus L protein may constitute the minimal binding domain for the interaction of Sendai virus P protein. Accordingly, the Sendai virus P plasmid was cotransfected with the mutant L plasmids, particularly Sen6, Sen7 and Sen8, and the ³⁵S-labeled extracts were immunoprecipitated with anti-FLAG antibody or by MAb M56 that specifically recognizes the Sendai virus P protein. As seen in Fig. 6, the anti-FLAG antibody efficiently immunprecipitated all the mutant L proteins; however, no band corresponding to the predicted size of the Sendai virus P protein was present in any of the lanes with the mutant L proteins. As a control, the anti-FLAG antibody efficiently immunoprecipitated the wild-type L protein and the HPIV 3 P protein bound to it. Conversely, the anti-Sendai virus P protein antibody, MAb M56, efficiently immunoprecipitated the Sendai virus P protein from the extracts containing the coexpressed mutant L proteins but failed to reveal any band corresponding to the expected size of the mutant L proteins (Fig. 6). These results underscore the point that the 25 amino acids spanning the NH₂ terminus of the HPIV 3 L protein in the mutant Sen8 resembling the corresponding domain in Sendai virus L protein do not confer the specificity for Sendai virus P to bind. Clearly, additional domains or structures in the Sendai virus L protein must play a role in determining specific binding with the homologous P protein at the NH₂ terminus if indeed Sendai virus P binds at the same domain (as in HPIV 3 L) of its cognate L protein.

FIG. 6.

Interaction of HPIV 3 L mutants with HPIV 3 P and Sendai virus P protein (SV-P). HeLa cells were infected with vTF7-3 and transfected with L_wt or L_mut plasmids along with HPIV 3 P plasmid or Sendai virus P plasmid as indicated. [³⁵S]methionine-labeled cell extracts were immunoprecipitated with anti-FLAG antibody conjugated to agarose beads as described in Materials and Methods and for Fig. 2D or anti-SV-P antibody (MAb M56) previously conjugated to Sepharose A and analyzed by SDS-PAGE followed by fluorography. Mock, vaccinia virus-infected and mock-transfected cell extract.

DISCUSSION

The functional RNA polymerase complex of the negative-strand RNA viruses responsible for the transcription and replication of the viral genome is comprised of the L and P proteins. Although the L protein is multifunctional and possesses RNA-dependent RNA polymerase activity, it can manifest its activity only in the presence of the P protein, an indispensable cofactor (20). Current studies were initiated based on the earlier sequence alignment studies of the paramyxovirus L proteins revealing the presence of six conserved motifs (25, 27). In this report, we have focused on yet another conserved domain present at the NH₂ terminus of all the paramyxovirus L proteins to gain insight into the structure and function of the HPIV 3 L protein (Fig. 1). Within the eight paramyxovirus L protein sequences aligned in Fig. 1, it is quite striking that the NH₂-terminal sequence IL(Y/L)PE(C/V)HL(N/D)SPIV is highly conserved (the residues in parentheses varied among some viruses). We systematically altered amino acid residues or deleted portions within the conserved domain and observed that this NH₂-terminal domain plays an important role in L protein function and P binding. It became apparent from the initial deletion experiments that removal of four nonconserved amino acid residues from the NH₂ terminus did not significantly alter the L function both in vitro and in vivo as well as its ability to bind P protein. However, deletion of an additional 10 residues (which overlapped with the conserved domain) had a drastic negative effect on L protein's ability to transcribe. Most importantly, the P protein failed to bind to the mutant L protein (NΔ2-15). Presumably, the NH₂-terminal 14 amino acids of L appear to constitute the P binding domain. Although such a small P binding domain has not been uncovered for other paramyxoviruses, it is important to note that a large domain spanning the NH₂ terminus has been implicated for Sendai virus, SV 5, and measles viruses by deletion analyses (4, 17, 24). However, it is possible that the NH₂-terminal 14 amino acids are critical in providing optimal conformation to the L protein, including a highly ordered structure required for efficient transcription and P binding. In that scenario, the precise location of P binding may reside within the NH₂-terminal region of the L protein such that mutations within the 14-amino-acid stretch will result in a misfolding of the native L structure resulting in a loss of P binding.

Since the NH₂-terminal 14 amino acids overlap with the two invariant amino acids, I and L, mutational analyses confirmed that ILYPE sequence is also critical to conferring transcription function to the L protein as well as binding to P protein (Fig. 3). Alteration of Y to A, P to A, and E to G completely eliminated L function and binding to P protein. In addition, deletion of the ILYP sequence had a similar drastic effect on the L function (Fig. 3). Although extensive mutational analyses have not been carried out within the rest of the highly conserved sequence of L, it is tempting to speculate that the 25 NH₂-terminal amino acids of other paramyxoviruses may also play a critical role in the activity of the corresponding L protein. So far, similar studies have been carried with three other paramyxoviruses, namely measles (17), Sendai virus (4), and SV 5 (24). In all cases, the NH₂-terminal domain has been implicated in binding to the cognate P protein by using deletion mutants of the L protein made from the COOH-terminal end. Using NH₂-terminal deletion mutants of HPIV 3 L protein, our data establish a small domain (possibly 25 amino acids) at this end that is possibly involved in P protein binding and the formation of a functionally competent polymerase complex.

Our subsequent mutational analyses on the nonconserved eight amino acids (residues 3 to 10) revealed that critical amino acid residues within this domain may alter the transcriptive phenotype of the L protein. We systematically altered residues 3 to 10 of HPIV 3 L protein to those of the corresponding Sendai virus L protein sequence (Fig. 1) to evaluate at what stage of alteration HPIV 3 L protein fails to bind with homologous P protein and may bind to the Sendai virus P protein instead. Substitution of all eight amino acids from the Sendai virus sequence would essentially convert the NH₂-terminal 25-amino-acid stretch identical to the Sendai virus L protein. However, the mutational analyses (Fig. 4 and 5) showed some interesting results. Alteration of HPIV 3 L residues individually to corresponding Sendai virus L residues did not have any effect on HPIV 3 L protein to transcribe or bind to the homologous P protein, suggesting that the nonconserved sequence has the ability to withstand alteration following substitution of E (acidic) to G or V to P (helix breaker). In direct contrast, swapping the entire first 25 amino acids of Sendai virus L protein within HPIV 3 L had a profound effect. First, altering E to S (Fig. 5) reduced the L function significantly, but it was fully recovered when N was altered to S to restore the ES sequence. Secondly, replacing seven residues of the Sendai virus sequence (Sen7) had no effect in L function, but the alteration of critical V to P totally eliminated the HPIV 3 L protein's function to transcribe and bind to the P protein. It is important to note that a similar V to P change in the HPIV 3 L sequence (Fig. 4) had no effect on the function of L protein. Thus, it seems that within the Sendai virus amino acid sequence context, the alteration of V to P must have a drastic effect on the secondary structure of the NH₂-terminal domain specifically with respect to binding of HPIV 3 P protein. However, the chimeric Sen8 L protein failed to bind to the Sendai virus P protein (Fig. 6), suggesting that other structural features and/or domains of the Sendai virus L protein is required to bind to its cognate L protein.

In summary, we have identified a highly conserved domain within the NH₂ terminus of the HPIV 3 L protein that is directly involved in viral transcription and plays a critical role in complex formation with the P protein. Furthermore, using mutational analyses, we have identified residues within this domain that govern the activity of the L protein, indicating the possible role of the nonconserved eight amino acids. We have previously reported the successful generation of an infectious clone of HPIV 3 (14), and efforts are underway to introduce these mutations in the infectious clone by using a reverse genetics system to study the role of the NH₂-terminal domain of the L protein in more detail.