INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Influenza A virus contains eight different RNA segments of negative polarity in its genome, each encoding one or two unique viral proteins. The viral RNA polymerase is associated with each RNA segment as a viral component and in infected cells, responsible for both transcription and replication of the viral genome (for reviews, see references 7, 13, and 21). The pathway for the synthesis of viral mRNA involves multiple-step reactions, consisting of endonucleolytic cleavage of host cell mRNA, capped oligonucleotide-primed transcription of viral RNA (vRNA), and the addition of a poly(A) tail to the nascent mRNA. On the other hand, replication takes place in two steps, vRNA-directed synthesis of full-length complementary RNA (cRNA) without any modification at both the 5' and 3' termini and cRNA-directed reproduction of vRNA.

Starting from the viral ribonucleoprotein (RNP), we isolated an RNA-3P (PB1, PB2, and PA) protein complex without NP by equilibrium centrifugation in cesium sulfate or cesium chloride (15, 16). The RNA-3P complex is enzymatically active in the in vitro synthesis of short attenuated RNA chains (but NP is needed for RNA chain elongation [12]), indicating that the three P proteins participate in the catalytic function of RNA polymerization. The vRNA-free 3P complex, consisting of one molecule each of the three P proteins, was isolated from the RNA-3P complex by centrifugation in cesium chloride or cesium trifluoroacetate (11). The 3P complexes exhibited RNA synthesis only when a model vRNA template with 5'- and 3'-terminal conserved vRNA sequences was added (31).

The molecular composition of influenza virus RNA polymerase was confirmed after establishment of the in vitro reconstitution system using the three P proteins which were individually expressed in insect cells after infection with recombinant baculoviruses (20, 39), but the reconstitution of the functional RNA polymerase from the three overexpressed P proteins was not high, mainly due to insufficient refolding of the P proteins from inclusion bodies in expressed cell lysates. Recently, we have succeeded in expressing each P protein in Escherichia coli by changing the nucleotide sequence (without changing the amino acid sequence) near the translation initiation site so as to adjust the codon usage to the E. coli pattern (Y. Asano and A. Ishihama, unpublished data), but the E. coli system has the same drawbacks as the baculovirus expression system. On the other hand, an enzymatically active RNP was also reconstituted from a mixture of three P proteins and NP, which were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), electroblotted onto a membrane, recovered from the membrane, and refolded with the help of E. coli thioredoxin (43), but this system is not practical for large-scale preparation of the functional RNA polymerase. At present, the use of in vitro reconstitution systems for large-scale production of the RNA polymerase is costly, time-consuming, and technically inconvenient.

In order to meet the demand for a large amount of functional influenza virus RNA polymerase in template-free form, an alternative approach has since been employed, in which all three P proteins were expressed in the same cells transiently after infection of recombinant vaccinia viruses (25, 41, 48) or permanently after integration of all three P protein cDNAs into the chromosome (18, 29). The expression levels of P proteins in these coexpression systems were, however, not sufficient for large-scale purification of functional RNA polymerase. In order to improve the expression levels, we used the methylotrophic yeast Pichia pastoris system as well as the recombinant baculovirus system. Here we describe the first successful expression of negative-strand viral RNA polymerase in a yeast.

The idea of using a yeast as a host for the growth of animal viruses originated from the finding that the budding yeast Saccharomyces cerevisiae can be a host for the replication of the genome of a plant virus, brome mosaic virus (14). P. pastoris is able to utilize methanol as its sole carbon source and has been developed as a host for the expression of heterologous proteins. The major advantages of this expression system include (i) a strong, tightly regulated alcohol oxidase (AOX) promoter, 5'AOX1, is available; (ii) large-scale protein production can be achieved in a large-volume fermentor culture; (iii) a secretary pathway allows the product to be secreted into the medium, separating the foreign protein from most of the host proteins; (iv) the expression system can be easily set up; and (v) the cost is as low as that of the E. coli expression system (2, 4, 6, 23). Two membrane proteins of influenza virions, neuraminidase (NA) and hemagglutinin (HA), were successfully expressed in P. pastoris and produced as secreted forms in the culture medium (24, 37). These influenza virus proteins served as recombinant vaccines that elicit partial or fully protective antibodies in mice.

The main concern of this study is to express and purify template-free influenza virus RNA polymerase using P. pastoris as a host strain and subsequently to examine its catalytic properties. The results indicate that (i) an expression system producing reasonable amounts of the three P proteins in P. pastoris was established after a search for optimum induction times and the optimum concentration of methanol to give the maximum level of induction; (ii) by adding a histidine tag to the PB2 protein, the 3P protein complex was isolated from the cell lysates after Ni²⁺-nitrilotriacetic acid (NTA)-agarose affinity chromatography; (iii) by quantitative Western blotting, all three P proteins in the isolated 3P complex were detected in stoichiometric molar ratios; (iv) the catalytic activity of RNA synthesis was detected for the 3P complex only when model vRNA or cRNA templates with conserved terminal sequences were added; and (v) both synthetic dinucleotides and globin mRNA served as primers, but the 3P protein complex was virtually inactive in the absence of primers. Template recognition specificity was also examined by using various kinds of model vRNA or cRNA with and without the terminal conserved sequences. Taking all the results together, we conclude that functional influenza virus RNA polymerase can be produced in the methylotrophic yeast P. pastoris in amounts sufficient for detailed functional analysis and structural studies.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Construction of plasmids and transformation into P. pastoris. The cDNAs for the PB1, PB2, and PA proteins of influenza virus were prepared as described previously (20). The PB1 cDNA was inserted into plasmid pPic9 (Invitrogen) at the BamHI site between the promoter and the terminator of the AOX gene to construct pPic9-PB1 (Fig. 1). The PA cDNA was integrated into the pPic9 vector at the BamHI site to generate pPic9-PA, into which the HaeII fragment of pPicZA containing the Zeo^r coding sequence was inserted to generate pPic9Z-PA. The PB2 cDNA was first inserted into pEH31 between the SacI and XhoI sites to generate pEH-PB2, from which the PB2 cDNA was PCR amplified with a 5' primer containing a BamHI site and a histidine tag sequence and a 3' primer containing an XhoI site. The N-terminal leader sequence, MSHTHEHLHHHEL, of the Klebsiella nitrile hydratase  subunit fused to the factor Xa recognition sequence IEGR was used as the histidine tag sequence (27) and was added at the N terminus of the PB2 gene (this sequence is hereafter referred to as Hx in this paper). The PCR-amplified HxPB2 sequence was inserted into pPic9 between the BamHI and XhoI sites to construct pPic-HxPB2. A HaeII fragment of pPicZA containing the Zeo^r coding sequence was inserted into pPic-HxPB2 to produce pPicZ-HxPB2 (Fig. 1). All three P protein genes are under the independent control of the AOX promoter (Fig. 1). The construction of all expression plasmids was confirmed by DNA sequencing.

View larger version (22K): [in this window] [in a new window]   FIG. 1.   Construction of P. pastoris F3P carrying the 3P subunit genes for influenza virus RNA polymerase. Transformation of the cDNA for each of the three P proteins of influenza virus was carried out in two steps. In the first-step transformation, plasmid pPic9-PB1 carrying cDNA for the PB1 protein was linearized by treatment with AatI within the HIS4 gene and transformed into P. pastoris KM71 by the LiCl method, and the transformants were screened for His+. Since Southern blot analysis indicated integration of two copies of the PB1 cDNA (Fig. 2), the crossing-over must have taken place twice to yield P. pastoris FPB1. In the second-step transformation, plasmid pPicZ-Amp-HxPB2-PA carrying the cDNAs for both PB2 and PA was linearized by treatment with PvuI and transformed into P. pastoris FPB1, and cultures were screened for zeocin-resistant transformants. Southern blot analysis indicated the integration of two copies of the PB2 gene and one copy of the PA gene in the transformant P. pastoris F3P used for expression of the influenza virus RNA polymerase. The most probable pathway involves two consecutive cross-overs of the plasmid followed by deletion of the PA gene after the integration into the Pichia chromosome. An alternative mechanism involves the integration of two different sequences into the Pichia chromosome, one complete copy of the PB2-PA plasmid and the other incomplete copy of only the PA gene. The F3P genome, as estimated from the Southern blot patterns, contains two PB1, two HxPB2 (Hx tag conjugated at the N terminus of PB2), and one PA cDNAs. The promoter (5'AOX1) and terminator (3'AOXTT) of the alcohol oxidase gene are located upstream and downstream, respectively, of each P gene.

Expression of the 3P proteins in P. pastoris. P. pastoris F3P, which carries the PB1, PB2, and PA cDNAs integrated in the chromosome, was cultured overnight at 30°C in YPD medium supplemented with zeocin (final concentration, 0.1 mg/ml). The preculture was inoculated into 20 ml of fresh MG medium (1.34% yeast nitrogen base, 0.00004% biotin, and 1% [vol/vol] glycerol) containing zeocin (0.1 mg/ml). Cells were grown with vigorous shaking at 30°C until the cell density reached an A₆₀₀ of 1.5 or 2.2.

Purification of the 3P protein complex. The frozen cells were resuspended in breaking buffer, and an equal volume of acid-washed glass beads was added to the cell suspension. After cell breakage by vigorous vortexing at 4°C, the disrupted cell suspension was centrifuged at 12,000 rpm for 15 min. The remaining cell pellet was resuspended in an equal volume of breaking buffer and centrifuged. The combined supernatants were used as the cell lysate.

PAGE and Western blotting. Samples of either the crude cell extract or the Ni²⁺-agarose column chromatography eluates were separated by SDS-10% PAGE, and the proteins were visualized by staining with Coomassie brilliant blue. For Western blotting, the proteins in the gel were transferred to a polyvinylidine difluoride membrane (Pall Gelman Laboratories) using semidry apparatus in transfer buffer (39 mM glycine, 48 mM Tris base, 0.037% SDS [electrophoresis grade], 20% [vol/vol] methanol). After blocking with bovine serum albumin, the membrane was incubated with rabbit polyclonal antibodies against PB1, PB2, or PA (1:1,000 dilution). The membrane was washed and then reacted with the secondary antibody, goat anti-rabbit immunoglobulin G conjugated to horseradish peroxides (1:1,000 dilution) (Promega). Finally, the membrane was washed and developed with 0.05% diaminobenzidine tetrahydrochloride in the presence of H₂O₂.

Isolation of genomic DNA from P. pastoris F3P. A single colony of P. pastoris F3P was inoculated into 10 ml of YPD containing Zeocin (0.1 mg/ml) and cultured at 30°C overnight. A 0.05-ml aliquot of the preculture was inoculated into 10 ml of minimal medium MD (1.34% YNB, 0.00004% biotin, 2% glucose) containing zeocin (0.1 mg/ml) and grown for at least for 20 h at 30°C with shaking at 300 rpm. Cells were harvested by centrifugation, washed twice with distilled water, and resuspended in 2 ml of SCED buffer (1 M sorbitol, 10 mM sodium citrate [pH 7.5], 10 mM EDTA, 10 mM DTT). To the cell suspension, 0.3 mg of Zymolyase (Seikagaku Kogyo, Tokyo, Japan) was added, and the mixture was incubated at 37°C for 1 h with continuous shaking. For isolation of DNA, 2 ml of 1% SDS was added, and the cell suspension was set on ice for 5 min. After adding 1.5 ml of 5 M potassium acetate (pH 8.9), the cell lysate was centrifuged to remove the cell debris. To the supernatant, 90 µl of 10-mg/ml RNase A was added, and the mixture was incubated at 37°C for 1 h. DNA was precipitated by adding 2 volumes of ethanol. The DNA pellet was dissolved in 1 ml of distilled water and purified several times by treatment with phenol-chloroform (1:1, vol/vol).

Southern blot analysis. DNA probes for Southern blotting were generated by PCR using the genomic DNA from P. pastoris F3P as a template (Table 1). The PCR mixture contained, in 50 µl, 25 mM N-Tris(hydroxymethyl)methyl-3-aminopropanesulfonic acid (TAPS) buffer (pH 9.3); 50 mM KCl; 3.25 mM MgCl₂; 1 mM 2-mercaptoethanol; 20 ng of purified genomic DNA; 200 ng each of the forward and reverse primers; 0.5 mM each dATP, dGTP, and dTTP; 0.05 mM dCTP; 0.67 µM [-³²P]dCTP; and 2.5 U of Taq polymerase (Takara, Otsu, Japan). PCR products were purified by using the SUPREC-02 cartridge (Takara).

Preparation of RNA templates. cRNA and vRNA model templates for RNA synthesis were constructed as described previously (31, 46). Basically, plasmids containing the cDNA for vRNA or cRNA were digested with MboII and then transcribed by T7 RNA polymerase to produce RNA model templates. cDNAs for the synthesis of mutant model templates were generated by PCR. Standard PCR was carried out in a 50-µl reaction mixture containing 25 mM TAPS (pH 9.3); 50 mM KCl; 2 mM MgCl₂; 1 mM 2-mercaptoethanol; 20 pmol each of forward and reverse oligonucleotides; 0.25 mM each dATP, dGTP, dCTP, and dTTP; and 2.5 U of Taq polymerase (Takara). PCR products were subjected to electrophoresis on a 2% SeaKem GTG agarose gel, isolated from the gel, and then recovered with SUPREC-01 filter cartridges (Takara). Each purified DNA fragment was digested with PstI and HindIII, cloned into pUC19, and subsequently transformed into E. coli DH5. Following verification of the sequence of each mutant template, large quantities of plasmid DNA were obtained by using the Qiagen plasmid maxikit and used as templates for in vitro transcription by T7 RNA polymerase.

In vitro RNA synthesis. RNA synthesis in vitro was performed essentially as described previously (11, 46) with a slight modification. In brief, 20 µl of the reaction mixture contained 50 mM HEPES-KOH (pH 7.6); 5 mM magnesium acetate; 100 mM KCl; 2 mM DTT; 0.5 mM each ATP, CTP, and GTP; 50 µM UTP; 5 µCi of [-³²P]UTP (Amersham); 100 U of RNase inhibitor (Takara) per ml; 10 pmol of RNA template (c53, v53, c84, or v84); and 1 mM ApG (Sigma). For mutant templates vM3, cM9, and cM10, primers ApA, ApA, and CpA, respectively, were added to the reaction mixture instead of ApG. About 1 pmol of either the RNP core or the 3P complex, as estimated from the amount of P proteins, was added as the enzyme, and the reaction was carried out at 30°C for 2 h. For capped RNA-primed transcription assays, globin mRNA (Gibco-BRL) was used in place of dinucleotide primers. RNA products were extracted with an equal volume of a 1:1 (vol/vol) mixture of phenol and chloroform-isoamyl alcohol (24:1) and precipitated with ethanol. After centrifugation, samples were dissolved in gel loading buffer, heated at 90°C for 3 min, and subsequently analyzed by electrophoresis on an 8% denaturing polyacrylamide gel in the presence of 8 M urea. Gels were exposed to imaging plates, and the plates were analyzed with a PhosphoImage analyzer (BAS 2000; Fuji, Tokyo, Japan).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Construction of P. pastoris F3P carrying cDNAs for influenza virus RNA polymerase subunits. The construction of P. pastoris F3P carrying on its chromosome the cDNAs for all three P protein subunits of influenza virus RNA polymerase was performed in two transformation steps. In the first-step transformation, the cDNA for the PB1 subunit protein was inserted into P. pastoris expression plasmid pPic9 to construct pPic9-PB1 (Fig. 1). After treatment with AatI within the HIS gene, the linearized pPic9-PB1 DNA was transfected into P. pastoris KM71 mutS, and His⁺ transformants were selected. The resulting P. pastoris FPB1 containing the PB1 gene was then transformed with PvuI-treated pPicZ-Amp-HxPB2-PA (the PvuI site is located within the amp gene), which was constructed by inserting the cDNAs for the HxPB2 (Hx tag at the N terminus) and PA proteins, each under the independent control of the alcohol oxidase (AOX1) promoter, into plasmid pPicZ-Amp in tandem (Fig. 1). The selection of P. pastoris F3P containing the HxPB2-PA genes as well as the PB1 gene was carried out in the presence of zeocin. If one copy each of the two plasmids inserted into the P. pastoris chromosome at the expected sites (HIS for the first transformation and amp for the second transformation step), the order of gene integration should be HIS4-Amp-HxPB2-PA-Zeo-Amp-PB1-his, and all three P genes should be located under the independent control of the AOX1 promoter. The gene organization was then checked by Southern blot analysis.

Organization of the PB1, PB2, and PA genes in the P. pastoris F3P chromosome. The gene organization and copy numbers of PB1, PB2, and PA cDNAs in the chromosome of P. pastoris F3P were determined by Southern blotting analysis. For this purpose, genomic DNA was isolated from P. pastoris F3P and its parental strain KM71 and digested with NotI, SalI, XbaI, and HpaI (Fig. 1 [bottom] for the restriction enzyme sites). The digested DNA samples were separated by agarose gel electrophoresis, and the gels were subjected to Southern hybridization with ³²P-labeled 300-bp-long DNA probes with sequences complementary to the PB1, PB2, and PA genes (see Materials and Methods for probe preparation). The Southern hybridization patterns are shown in Fig. 2. All the probes hybridized to the selected fragments from the F3P genome, whereas none of the DNA fragments from KM71 hybridized with these probes, indicating that the F3P strain carries the sequences for all three P protein genes.

View larger version (58K): [in this window] [in a new window]   FIG. 2.   Southern blot analyses of F3P genome DNA. Genomic DNA was isolated from both P. pastoris F3P and its parental strain KM71 and digested with NotI, SalI, XbaI, or HpaI at 37°C overnight. The digested DNA samples were separated by electrophoresis on a 0.8% agarose gel and then transferred to Hybond-N+ membranes (Amersham). Southern hybridization was carried out at 60°C overnight with 32P-labeled 300-bp-long DNA probes with sequences complementary to the PB1, PB2, and PA genes. The membranes were washed several times, and the radioactive signals on the membranes were detected with a PhosphorImager analyzer. Lanes 1, DNA from parental strain KM71; lanes 2, DNA from P. pastoris F3P. (A) 32P-labeled PB1 probe; (B) 32P-labeled PB2 probe; (C) 32P-labeled PA probe. The migration positions (in base pairs) of the HindIII-digested  size markers are shown on the right.

Expression of 3P proteins in P. pastoris F3P. The expression of 3P proteins in P. pastoris F3P was examined by adding the inducer methanol to activate the alcohol oxidase promoter, which controls the P protein genes. The induced cell lysates were positive for P protein expression by Western blotting against specific anti-P protein antibodies, which were raised in rabbits against each of the three P proteins purified from E. coli. Since these antibodies showed little cross-reaction with each other (45), PB2 and PA, which apparently migrate to the same positions on SDS-PAGE gels, could be detected separately. In order to identify the optimum conditions for maximum expression of the three P proteins, induction was initiated at two different cell densities, 7 × 10⁷ and 1.2 × 10⁸ cells/ml, by transfer of the culture into induction medium containing various concentrations of methanol. At various times after induction, cell lysates were prepared and subjected to Western blotting against each of the anti-P protein antisera.

View larger version (29K): [in this window] [in a new window]   FIG. 3.   Western blotting analyses of the expression of 3P in P. pastoris. Cells were grown to a concentration of 7 × 107 or 1.2 × 108 cells/ml and then induced for P protein expression by adding 0.5% methanol. Methanol was added every 24 h to give a concentration of 0.5%, and cells were harvested at 24, 48, 72, and 96 h after methanol addition. The control uninduced cells were grown in a medium containing 0.5% glycerol in place of methanol. The cell lysates were analyzed by SDS-PAGE followed by Western blotting with anti-PB1 (A), anti-PB2 (B), and anti-PA (C) antibodies. Lanes: Mr, broad-range molecular weight markers (Bio-Rad); uninduced, a cell lysate at 72 h of culture in glycerol medium; cell lysates at 24, 48, 72, and 96 h after methanol addition; RNP, native RNP isolated from purified influenza virions. The migration positions of the three P subunits are indicated by arrowheads. The yields of each P protein and of the combination of all three P proteins were quantified from the intensities of Western blotting patterns using known amounts of P proteins as references. Symbols: ×, PB1 subunit; , PB2 subunit; , PA subunit; , combined three P proteins. The cell lysates were prepared from two different cell cultures with initial cell densities of 7 × 107/ml (D) and 1.2 × 108/ml (E).

Purification of the 3P protein complex. To purify the three P proteins from P. pastoris F3P, a 100-ml culture was induced for P protein expression, and the crude cell lysate was subjected to a two-step purification, ammonium sulfate fractionation and Ni²⁺-NTA-agarose column chromatography. Samples collected from each step of the purification were analyzed by SDS-PAGE and Western blotting. As shown in Fig. 4A, the crude cell lysate contained all three P proteins (lane 2), which were recovered at 60 to 77% saturation of the ammonium sulfate precipitates (lane 3). The ammonium sulfate precipitates were dissolved in Ni²⁺-NTA affinity binding buffer. The dialysate (Fig. 4A, lane 4) was applied to an Ni²⁺-NTA-agarose column, and the column was washed in two steps: a first wash with washing buffer containing 5 mM imidazole, and the second wash with elution buffer containing 60 mM imidazole. The initial wash fraction did not contain the 3P proteins (Fig. 4A, lane 6), but the elution fraction with 60 mM imidazole contained the 3P proteins (lane 7). Upon increasing the imidazole concentration in the elution buffer to 200 mM, little P protein was eluted (Fig. 4A, lane 8), indicating that the majority of the 3P protein complex was eluted at 60 mM imidazole. The affinity for the Ni²⁺-NTA column may not be high enough for the Hx tag sequence used in this study. We then inserted the second wash step with a wash buffer containing 25 mM imidazole prior to elution of the 3P protein complex with elution buffer containing 60 mM imidazole. For initial characterization of the enzymatic activities of the 3P protein complex, the 60 mM imidazole elution fractions were pooled, concentrated, and dialyzed against storage buffer containing 50% glycerol for storage.

View larger version (80K): [in this window] [in a new window]   FIG. 4.   Purification of the 3P complex. (A) The P proteins were precipitated from induced cell lysates with (NH4)2SO4, dissolved in 1× Ni2+-NTA binding buffer, dialyzed against the same buffer, and subjected to Ni2+-NTA-agarose affinity column chromatography. Fractions from each step were analyzed by Western blotting with anti-PB1, anti-PB2, and anti-PA antibodies. Stained bands were quantified by using a standard curve prepared from a known amount of RNP. Lane 1, low-range molecular weight markers (Bio-Rad); lane 2, cell lysate; lane 3, precipitates from (NH4)2SO4 fractionation; lane 4, dialysate in 1× binding buffer; lane 5, flowthrough fraction of the Ni2+ column; lane 6, wash fraction with 1× binding buffer; lane 7, fraction eluted with a buffer containing 60 mM imidazole; lane 8, fraction eluted with a buffer containing 200 mM imidazole; lane 9, precipitates from the second (NH4)2SO4 fractionation; lane 10, the 3P complex in storage buffer; lane 11, RNP; lane 12, broad-range molecular weight markers. See also Table 2 for yields of the three P proteins at each purification step. (B) The partially purified 3P complex was separated, in parallel with RNP, by SDS-PAGE, and the gel was stained with Coomassie brilliant blue. Lane Mr, size markers. NP, influenza virus nucleoprotein. (C) The partially purified 3P complex was separated, in parallel with RNP, by SDS-PAGE, and the gel was immunostained with specific antibodies against PB1, PB2, and PA proteins.

RNA synthesis in vitro by the 3P protein complex. Model RNA templates were developed for detection of in vitro RNA synthesis activity by such template-free RNA polymerases as the solubilized RNA polymerase from RNP (31) and the reconstituted RNA polymerase (20). The RNA synthesis activity of the partially purified 3P protein complex from P. pastoris F3P was examined using a pair of negative-sense model RNAs (v53 and v84) and a pair of positive-sense model RNAs (c53 and c84), each carrying 5'- and 3'-terminal conserved sequences of vRNA or cRNA (note that v and c represent the viral and complementary strand [or negative and positive strand], respectively). The 3P complex exhibited the catalytic activity of RNA synthesis only when a template was added, indicating that the 3P complex is free of any RNA with template activity. Figure 5 shows transcripts directed by the v84 (A) and c53 (B) model templates. The 3P complex catalyzed v84-directed ApG-primed synthesis of RNA (Fig. 5A, lane 7), which migrated on urea-PAGE to the same position as the template v84 (Fig. 5A, lane 1). The synthesis of this template-sized transcript was not detected in the absence of ApG primer (Fig. 5A, lane 8) or v84 template (Fig. 5A, lane 9). RNP added as a control produced several transcripts (Fig. 5A, lane 3), one of which migrated to the same position as the template v84 (Fig. 5A, lane 1). This v84-sized product was not synthesized without addition of the v84 template (Fig. 5A, lane 5). The products formed by RNP in the absence of v84 addition represent those initiated on the endogenous vRNA templates, but no RNP transcripts were detected in the absence of ApG primer (Fig. 5A, lane 4). From the definition of transcriptase (requirements for vRNA template and primers) (see references 7 and 13), the 3P complex and the RNP-associated RNA polymerase might be the transcriptase form.

View larger version (72K): [in this window] [in a new window]   FIG. 5.   In vitro RNA synthesis by the 3P complex. The 3P complex, prepared as described in the legend to Fig. 4, was subjected to RNA synthesis in vitro using both a negative-sense v84 template (A) and a positive-sense c53 template (B). The reaction mixture contained, in 20 µl, 10 pmol of each template, 1 mM ApG primer, and approximately 1 pmol of the RNP or the 3P complex, as estimated from the contents of 3P proteins. RNA synthesis was carried out at 30°C for 2 h. RNA products were ethanol precipitated and analyzed by electrophoresis on an 8% polyacrylamide sequencing gel. Lane 1, molecular weight marker RNA, v84 (A) and c53 (B).

Kinetic properties of the 3P RNA polymerase. Here we obtained for the first time a large amount of RNA-free functional RNA polymerase with which to measure the kinetic parameters of RNA synthesis. First, we investigated the effects of increasing concentrations of template (v84 and c84), ApG primer, and nucleoside 5'-triphosphate substrates on RNA synthesis (Fig. 6). The minimum amounts of these reaction components required to give the maximum catalytic activity of RNA synthesis were calculated after replotting the data to double-reciprocal graphs. The amounts of v84 and c84 templates that give maximum transcription are apparently different: the K_m value for the v84 template was 12.5 nM, while that for c84 was 36.5 nM (Table 3). The V_max was apparently similar for the two templates, indicating that the RNA polymerase has a greater affinity for the vRNA template than for the cRNA template, but once initiated, the rate of RNA synthesis is the same for both the vRNA and cRNA templates. The high affinity for vRNA is a reaction property expected for the transcriptase.

View larger version (19K): [in this window] [in a new window]   FIG. 6.   Effects of template, primer, and substrate concentrations on RNA synthesis by the 3P complex. The conditions for the in vitro RNA synthesis reaction were the same as those in Fig. 5 except that the concentration of one of the following reaction components was varied as indicated: (A) v84 template (pmol per 20-µl reaction mixture); (B) c84 template (pmol per 20-µl reaction mixture); (C) ApG primer; and (D) nucleotide substrates (ATP, CTP, GTP, and UTP). The results were replotted into double-reciprocal forms for estimation of the kinetic parameters, as summarized in Table 3.

Sequences on the templates recognized by the RNA polymerase. One type of experiment that can be done using RNA-free RNA polymerase is to examine the template recognition specificity of the viral RNA polymerase. As an initial attempt, we constructed various types of deletion mutant model templates, as illustrated in Fig. 7A, and tested whether these mutant templates can be used as templates for RNA synthesis by the 3P complex from P. pastoris F3P. All the mutant v53 templates were virtually inactive in directing RNA synthesis (Fig. 7B, lanes vM1 to vM4). Deletion of the first several nucleotides from either the 3' or the 5' terminus made the v53 template inactive, indicating that both 5'- and 3'-terminal conserved sequences are required for the vRNA to function as the template for the influenza virus RNA polymerase. The recognition specificity of the vRNA template by the 3P complex is essentially identical to those of virus-associated or recombinant RNA polymerases (30, 33, 38, 44).

View larger version (78K): [in this window] [in a new window]   FIG. 7.   Template activities of deletion mutant model RNAs at either 5' or 3' conserved sequences. A set of 5'- and 3'-terminal deletion mutants of both the v53 and c53 model templates were constructed by PCR and then cloned into the pUC19 vector between the PstI and HindIII sites. The sequences of these cDNA deletion mutants were verified by sequencing with DSQ-500L (Shimadzu, Kyoto, Japan). Model RNA templates were generated by transcribing the linearized v53 or c53 deletion mutant cDNA with T7 RNA polymerase. (A) Construction of v53 (vM1 through vM4) and c53 (cM5 through cM10) mutant templates. The sequences of the original v53 and c53 are shown, in which the 5' and 3' conserved sequences are boxed. The sequences remaining in the deletion mutant templates are shown by open boxes. (B) Dinucleotide-primed RNA synthesis in vitro was carried out using approximately 10 pmol of v53 and its deletion mutant templates and the 3P complex as the enzyme. (C) Dinucleotide-primed RNA synthesis by the 3P complex was carried out using 10 pmol each of c53 and its deletion mutants. Lane MW, size marker RNA (53 nucleotides in length).

Globin mRNA-directed transcription by the RNA polymerase. Finally, we examined whether the 3P complex from P. pastoris is able to utilize capped RNA as a primer for transcription. For this purpose, we used globin mRNA, which is known to be a good natural primer (1). In the presence of ApG primer and v84 template, a template-sized transcript was synthesized (Fig. 8, v84 template, ApG lane, open triangle). In the presence of globin mRNA instead of ApG primer, the 3P complex produced transcripts about 10 nucleotides longer than the ApG-primed transcript (Fig. 8, v84 template, mRNA lane, solid triangles). Thus, we concluded that the 3P complex is able to utilize capped RNA as a primer. The detection of capped RNA-primed transcription activity itself indicates the association of capped RNA endonuclease activity with the 3P complex. In addition to this globin mRNA-primed transcript of expected size, smaller transcripts migrating near the v84 size marker were detected (mRNA lanes, arrows). Since no transcript was detected in the absence of primer addition (Fig. 8, v84 template, None lane), it is unlikely that these represent products of unprimed RNA synthesis, but they may be either degradation products of globin mRNA-primed transcripts or internal-initiation transcripts.

View larger version (83K): [in this window] [in a new window]   FIG. 8.   Capped RNA-primed transcription by the 3P complex. RNA synthesis in vitro by the 3P complex was carried out in a 20-µl reaction mixture containing 15 pmol of either v84 or c84 template and either 20 nmol of ApG (1 mM) or 0.25 µg of globin mRNA as the primer. RNA products were fractionated, together with labeled v84 marker, by 8% PAGE in the presence of 8 M urea. Solid triangles point to globin mRNA-primed transcripts (for mRNA lanes), and open triangles indicate ApG-primed transcripts. The bands marked by arrows might represent degradation and/or internal-initiation products (mRNA lanes). The band produced in the absence of primer addition, marked with a solid star, represents a putative unprimed transcript (for the c84 template).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Several different approaches have been employed for the expression of influenza virus RNA polymerase proteins in various organisms. Animal cell lines transformed by recombinant viruses were effectively used for the expression of all three P proteins and for testing each P protein functions in vivo (22, 29), but the expression levels in all these cases were too low for purification and biochemical characterization of the RNA polymerase. In order to increase the expression level, the lytic infection system of recombinant viruses was also established using vaccinia virus (mammalian cell system) (25, 41, 48), simian virus 40 (mammalian cell system) (3), or baculovirus vectors (insect cell system) (20, 39). Recombinant virus infection system have the advantages that (i) all three individual P proteins can be expressed simultaneously or in various combinations, (ii) the three P proteins can be expressed at different ratios and at different times, and (iii) mutations can be introduced into each P protein gene on the recombinant virus vectors. Previously, we established a high-level expression system for individual P proteins using recombinant baculoviruses (20), but the reconstitution efficiency of functional RNA polymerase from isolated P proteins was not high enough for detailed biochemical analyses. However, after coinfection of all three kinds of recombinant baculovirus, each expressing one of the three P proteins, into the same cells, the 3P complex was formed, which showed the activity of model RNA-dependent RNA synthesis (A. Honda, A. Endo, and A. Ishihama, submitted for publication).

Attempts to express cDNAs for the PB1, PB2, and PA proteins cloned in various conventional E. coli expression vectors were all unsuccessful, but we have succeeded in expressing the P proteins in E. coli by changing the mRNA nucleotide sequences at 5'-terminal proximal regions so as to match the codon usage pattern to the E. coli type (Y. Asano and A. Ishihama, unpublished). Although the expression of all three P proteins increased to detectable levels, the highly expressed P proteins again formed inclusion bodies (10). So far we have failed to reconstitute the functional RNA polymerase using insoluble P proteins purified from E. coli.

In this study, we used the P. pastoris expression system for heterologous proteins for simultaneous expression of the influenza virus P proteins. Previously, expression of functional viral RNA polymerase in the budding yeast S. cerevisiae has been observed for the plant RNA virus brome mosaic virus (BMV) (14). The results herein described were quite remarkable, because three P proteins were expressed at significant levels in P. pastoris and furthermore the expressed P proteins were assembled into functional RNA polymerase. The presence and organization of all the genes for the three P proteins in the genomic DNA of P. pastoris F3P were confirmed by Southern blot analysis (Fig. 2). The gene copy numbers of PB1 and PB2 were twice that of PA (Fig. 1, bottom). A possible mechanism for the generation of such a construct includes four integration events, as shown in Fig. 1: two crossover events at the AatI site within the His4 gene of the plasmid carrying the PB1 clone, and two crossovers at the PvuI site within the amp gene of the plasmid carrying both the PB2 and PA genes. Since P. pastoris F3P contains only a single copy of the PA gene, one PA copy might be deleted after integration of two complete sets of the PB2 and PA genes. However, it is also possible, albeit not as likely, that only the PB2 gene was integrated into the chromosome at one of the two-step integration reactions in the second-step transformation.

The Western blot analysis clearly showed that all three P proteins, PB1, PB2, and PA, were expressed in P. pastoris F3P, and the expressed 3P proteins remained soluble in the cell lysate (Fig. 3). The expression yield (1.57 µg/ml) of 3P was not as high as those of several heterologous proteins, such as human tumor necrosis factor (42), the antigen pertactin (p69) of Bordetella pertussis (35), and glycosylated invertase (47), which have all been expressed in P. pastoris up to several grams per liter of culture. However, the expression levels of P proteins are as high as those of two membrane proteins of influenza virus, neuraminidase (NA) and hemagglutinin (HA), which ranged from 2.5 to 3 µg/ml (24) and 0.375 to 0.675 µg/ml (37), respectively. In the cases of NA and HA expression, the secreted products from P. pastoris were shown to be sufficient to serve as recombinant vaccines that elicit partial or fully protective antibodies in mice. Since the expression levels in yeast cells of proteins bearing a complete nuclear localization signal (NLS) are less than those of the corresponding proteins with deletions at the NLS (26, 36, 40), it appears that the NLS downregulates gene expression, presumably at the step of translation. Since all three P proteins carry the NLS sequence (13), their expression levels could be elevated by introducing mutations in the NLS sequences.

The three P proteins expressed in the P. pastoris F3P strain formed a complex(es) which binds to the Ni²⁺-NTA-agarose column via the Hx tag added at the N terminus of the PB2 protein, but the affinity of the 3P complex for the Ni²⁺ resin was not so great as to be used for a single-step purification by elution with a high concentration of imidazole. The affinity for Ni²⁺-NTA-agarose may be weaker for the Hx-tag sequence used (27) than for the widely used hexahistidine sequence. The subunit-subunit contact network within the influenza virus RNA polymerase is formed through two major contacts, the PB2 N-terminal domain to the PB1 C-terminal domain and the PB1 N-terminal domain to the PA C-terminal domain (45). Since the Hx tag was added at the N terminus of the PB2 protein, the PB2-PB1 contact may interfere with free access of the Hx tag to the Ni²⁺-NTA-agarose. PB2 carries the NLS (32) and two capped RNA-binding sites (10), which may interact with such cellular components as the nuclear transport machinery and the RNA cap-binding proteins, thereby preventing the Hx tag from binding tightly to the Ni²⁺ resin.

The catalytic activities of RNA synthesis in vitro were analyzed using the partially purified 3P complex. The 3P complex exhibited RNA synthesis activity only when exogenous model RNAs were added as templates (Fig. 5). The v84 model RNA-directed and ApG-primed activity of template-sized RNA by the 3P complex was about 20% of the RNP activity, as normalized to the content of P proteins (Fig. 4 and 5). The 3P complex has the same enzymatic characteristics as the purified or reconstituted influenza virus RNA: (i) the 3P complex is unable to synthesize RNA in the absence of primer (Fig. 5) (11, 20, 31, 43); (ii) the v-sense RNA is a better template than the c-sense RNA polymerase (Fig. 6 and Table 3) (11, 20, 28, 46); (iii) both 3' and 5' conserved sequences are required for transcription initiation (Fig. 7) (5, 31, 33, 38); and (iv) capped RNA such as globin mRNA can be used as primers for model template-dependent RNA synthesis (Fig. 8) (21). These reaction properties are characteristic of the transcriptase. Since the 5'-terminal triphosphate remains associated with the replication products (9), the replicase form of RNA polymerase must be able to initiate RNA synthesis de novo without primers. In addition, the replicase should use both vRNA and cRNA equally as templates. Thus, we concluded that the majority of the 3P complex formed in P. pastoris F3P is the transcriptase form of the viral RNA polymerase. The functional integrity in vivo of the 3P complex in P. pastoris cells is being tested.

The K_m for ATP in RNA synthesis by the 3P complex was much higher than the K_m values for other ribonucleoside triphosphates. Klumpp et al. (19) showed that the K_m for ATP is 10-fold higher than the K_m for other ribonucleotide triphosphates in transcription initiation. The high ATP concentration may be required for the functional conversion of RNA polymerase during the transition from transcription initiation to elongation (19). However, it has not yet been excluded that the requirement for a high ATP concentration is due to contamination with cellular ATPase in the 3P complex fraction.