Involvement of a Cellular Glycolytic Enzyme, Phosphoglycerate Kinase, in Sendai Virus Transcription*

In vitro mRNA synthesis of Sendai virus is almost entirely dependent on the addition of cellular proteins (host factors). Previous studies indicated that the host factor activity from bovine brain was resolved into at least two complementary fractions, one of which may be tubulin. In this study, the host factor activity that stimulates the transcription in the presence of tubulin was further purified from bovine brain. This fraction was found to contain at least two complementary factors, and one of them was purified to a single polypeptide chain with an apparentM r of 46,000 (p46). From the amino acid sequence, biochemical, and immunological analyses, p46 was identified as a glycolytic enzyme, phosphoglycerate kinase (PGK). Purified native PGK from rabbit and yeast, and a recombinant human PGK substituted for p46. Although, as previously suggested, tubulin was involved in the transcription initiation complex formation by being integrated into the complex, p46 and its complementary factor had little effect on the complex formation. On the other hand, when p46 and the complementary factor were added to the RNA chain elongation reaction from the isolated initiation complex formed with tubulin, mRNA synthesis was dramatically stimulated. The enzymatic activity per se of PGK did not seem to be required for its activity. West-Western blot analysis showed that PGK could directly interact with tubulin. These data suggest that PGK stimulates Sendai virus mRNA synthesis at the elongation step, probably through its interaction with tubulin in the initiation complex.

genome of 15.3 kilobases and three viral proteins: NP, P, and L. The RNA genome encapsidated with the NP proteins serves as the template to synthesize a positive strand leader RNA and mRNAs as well as a full-length positive strand anti-genomic RNA, which serves as the template for the negative strand RNA genome (1). The genetic information of the RNA genome is expressed through at least six monocistronic mRNA species. The P gene mRNA products which encode the C and V protein families, in addition to the P protein, use multiple translation initiation sites as well as an open reading frame newly generated by RNA editing (1). According to a current model (1), the viral RNA synthesis starts with a viral RNA-dependent RNA polymerase entering at the 3Ј end of the genome RNA, where the viral RNA polymerase initiates the sequential synthesis of a leader RNA and mRNAs or synthesizes a full-length antigenome RNA. However, little is known about the mechanism of mode switching of the RNA polymerase from transcription to replication, although it has been shown that three viral proteins, the NP, P, and L proteins are required for transcription and replication of the genome RNA (4 -6) and that the L protein interacts with the P protein to form the polymerase complex for replication and transcription (4,7). It has also been proposed that, during replication, P protein acts as a chaperon for the NP to encapsidate newly synthesized genome or antigenome RNAs (8).
Several studies using in vitro transcription systems of paramyxoviruses, e.g. human parainfluenza virus type 3 (HPIV3) (9,10), mumps virus (11), measles virus (12), respiratory syncytial virus (13), rinderpest virus (14), and canine distemper virus (15), suggest the involvement of cellular proteins in the viral RNA synthesis in addition to viral proteins (reviewed in Refs. 16 and 17). We have established an accurate and efficient in vitro mRNA synthesizing system using purified SeV particles or viral RNP complexes, in which viral mRNA synthesis is almost completely dependent on the presence of host cell proteins (host factors) (18). We have demonstrated that cellular tubulin, which is essential for in vitro transcription of SeV, is integrated into the transcription initiation complex and activates viral mRNA synthesis (18,19). The involvement of tubulin in SeV RNA synthesis was also supported by the inhibition of in vitro transcription of SeV by a monoclonal anti-tubulin antibody (20). On the other hand, De et al. (21,22) showed that the polymeric form of actin binds viral RNP of human parainfluenza virus type 3 (HPIV3) and activates in vitro transcription. They also found that the HPIV3 RNPs are specifically localized on the actin microfilaments in the virusinfected cells (23). The participation of tubulin or actin in transcription has also been observed in measles virus (12) and respiratory syncytial virus (24,25), respectively. Furthermore, cellular protein kinases such as casein kinase II and protein * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. kinase C isoform have been shown to phosphorylate the P protein (26 -30). However, the precise mode of action of these host factors in activating and regulating the RNA polymerase of paramyxoviruses remains to be studied. In addition, there seem to be other factors required for the transcription as well as replication of paramyxovirus genome that have not been identified yet (18,19). Therefore, the purification and the functional analysis of each host factor may, of course, provide an important step toward understanding the molecular mechanism of transcription and replication of the paramyxovirus genome.
In this study, we purified the host factor activity complementary to tubulin, and showed it has at least two components, and we identified one of them as phosphoglycerate kinase (PGK), a glycolytic enzyme. This is the first example showing the involvement of a glycolytic enzyme in the transcription of paramyxoviruses. We have shown that PGK stimulates viral mRNA synthesis at the elongation step, probably through its interaction with tubulin integrated into the initiation complex. Tubulin stimulated both mRNA synthesis and leader RNA synthesis of SeV, while PGK failed to stimulate leader RNA synthesis, suggesting that mRNA and leader RNA synthesis may be regulated by different sets of host proteins.

Materials
ATP, GTP, UTP, and CTP were purchased from Yamasa Shoyu, Japan. Hypatyte C was obtained from Clarkson Chemical Co.

Viruses
Sendai virus strain Z, propagated in the allantoic cavity of 10-day-old chicken eggs, was purified as described by Shibuta et al. (31). Vesicular stomatitis virus strain New Jersey was kindly supplied by Dr. H. Shibuta (Institute of Medical Science, University of Tokyo, Tokyo, Japan).

Preparation of Bovine Brain Extract (S100 Fraction)
The 100,000 ϫ g supernatant fraction (S100) from bovine brain was prepared as described by Takagi et al. (19).

Purification of Cellular Tubulin
Microtubular proteins were prepared form bovine brain through three assembly-disassembly cycles as described by Shelanski et al. (32). Purified tubulin free from microtubule-associated protein was obtained as a heterodimer by phosphocellulose column chromatography of microtubular proteins as reported by Weingarten et al. (33).

In Vitro Transcription of SeV and Vesicular Stomatitis Virus (VSV)
In vitro mRNA synthesis of Sendai virus or vesicular stomatitis virus was carried out under standard conditions as described by Mizumoto et al. (18), with 50 M [␣-32 P]UTP (3,000 -5,000 cpm/pmol) and 6 g of purified Sendai virus particles, or with 50 M [␣-32 P]UTP (500 -1,000 cpm/pmol) and 1 g of purified vesicular stomatitis virus particles. After incubation for 120 min at 30°C, the reaction mixtures were treated with proteinase K. The transcripts were extracted with phenolchloroform and precipitated with ethanol, and then electrophoresed in an 1.2% agarose gel after denaturation with glyoxal. In vitro plus leader RNA synthesis of Sendai virus was carried out with 10 M [␣-32 P]GTP (40,000 -50,000 cpm/pmol) and 6 g of virus particles under the same conditions for mRNA synthesis except that UTP was omitted, as described elsewhere. 2 After incubation for 120 min at 30°C, the RNA products were denatured with formamide, and analyzed by electrophoresis in a 20% polyacrylamide gel containing 8 M urea.

Isolation of Transcription Initiation Complexes
The transcription initiation complex was isolated as described by Takagi et al. (19). Purified Sendai virus particles (30 g) were incubated for 30 min at 30°C with the host factor in a 125-l reaction mixture containing 40 mM HEPES-KOH (pH 7.9 at 20°C), 30 mM NaCl, 50 mM KCl, 6 mM MgCl 2 , 2 mM DTT, and 0.1% Nonidet P-40, and the reaction mixture was centrifuged in a 0.8-ml ultraclear tube (Beckman), which contained two cushions, A and B (200 l each), at 100,000 ϫ g for 120 min at 4°C in a Hitachi P55ST-2 rotor with adapters. The viral RNP pellets were resuspended in cushion B and were subjected to the elongation reaction without additional host factors.

Purification of Host Factors
All operations were performed at 4°C.
Step 2: Blue-Sepharose Column Chromatography-A portion (43 mg, 86 ml) of the HA75 fraction was loaded onto a Blue-Sepharose column (inner diameter, 1.6 ϫ 18 cm) pre-equilibrated with TEMG/20 mM KCl, and proteins were step-eluted with TEMG containing 20, 150, and 300 mM KCl at a flow rate of 18 ml/h. The 20, 150, and 300 mM KCl eluates were referred to as BS20, BS150, and BS300, respectively. This column gave two active, complementary fractions, BS20 and BS150. The BS150 fractions from two batches were combined and dialyzed against TEMG/ 150 mM KCl. The dialysate of BS150 (50 ml, 0.26 mg/ml) was subjected to Heparin-Sepharose column chromatography. The BS20 fraction was kept at Ϫ80°C.
Step 3: Heparin-Sepharose Column Chromatography-A batch (6.5 mg, 25 ml) of the BS150 fraction was loaded onto a Heparin-Sepharose column (inner diameter, 0.8 ϫ 9.6 cm) pre-equilibrated with TEMG/150 mM KCl. After washing the column with 10 ml of the same buffer, proteins were eluted with a 50-ml linear gradient of 150 -350 mM KCl in TEMG at a flow rate of 6 ml/h and fractions of 0.54 ml were collected. The active fractions (220 mM KCl eluate, referred to HS220) containing a single polypeptide with M r of 46,000 (p46) (0.035 mg/ml, 19.5 ml from two batches) were pooled and stored at Ϫ80°C.

Amino Acid Sequence Analysis
Highly purified p46 (90 g, ϳ2 nmol) was digested with TPCKtreated trypsin (2 g) at 1:45 ratio (enzyme/substrate, w/w) in 180 l of a trypsin-digestion buffer (50 mM Tris-HCl (pH 8.0), 10 mM CaCl 2 ) at 30°C overnight. The tryptic peptides were fractionated by reversephase high performance liquid chromatography. The individual peptides were sequenced by automated Edman degradation using an Applied Biosystems model 477A sequencer. The obtained amino acid sequences were compared with sequences in the protein data base (PIR) by using the BLAST program.

Assay for Phosphoglycerate Kinase Activity
PGK activity was assayed in a coupled reaction with GAPDH reaction as described by Lee (34). The assay was performed at room temperature (ϳ25°C) in a total volume of 0.5 ml containing 0.1 M Tris-HCl (pH 7.9), 10 mM MgCl 2 , 0.15 mM NADH, 2 mM ATP, 6 mM 3-phosphoglycerate, 0.1 mg/ml BSA, 50 g of GAPDH, and 5-30 ng of PGK fraction. The consumption of NADH is monitored at 340 nm. One unit of PGK activity was defined as the amount of enzyme that catalyzes the formation of 1 mol of 1,3-bisphosphoglycerate (ϭ oxidation of 1 mol of NADH) per minute, using a molar extinction coefficient for NADH of 6.3 ϫ 10 3 M Ϫ1 cm Ϫ1 .

Preparation of Antibody against Rabbit Muscle PGK
Rabbit muscle PGK (300 g, obtained from Sigma) was mixed with an equal volume of Freund's complete adjuvant and injected subcutaneously into a New Zealand White rabbit (2 kg) four times at about 3-week intervals. After 11 weeks from the first immunization, polyclonal rabbit muscle PGK antibodies were affinity-purified using a rabbit muscle PGK-immobilized affinity column that had been prepared by coupling of PGK to 2-fluoro-1-methylpyridinium toluene-4-sulfonateactivated Cellulofine according to the manufacture's protocol.

Bacterial Expression of Human PGK-1
The coding sequence of human PGK-1 (GenBank accession nos. L00159 and L00160) was amplified from human leukocyte RNA (35) by RT-PCR using a sense primer, 5Ј-CCA GGA TCC ATG TCG CTT TCT AAC AAG C-3Ј (BamHI site underlined) and an antisense primer, 5Ј-GGT AGA TCT AAT ATT GCT GAG AGC ATC CAC-3Ј (BglII site underlined). The amplified PCR fragments of 1269 base pairs were digested with BamHI and BglII, inserted into BamHI and BglII sites of the pQE-16 plasmid (Qiagen), and sequenced to confirm their sequences and orientations. The resultant plasmid, pQE-hPGK, was transformed into Escherichia coli strain XL1-Blue. E. coli XL1-Blue harboring the plasmid pQE-hPGK were grown in LB medium containing ampicillin (100 g/ml) at 37°C until the A 600 reached 0.6. Expression of the recombinant human PGK with the hexahistidine tag at the carboxyl terminus was induced by adding isopropyl-1-thio-␤-D-galactopyranoside to a final concentration of 1 mM and cells grown for another 2 h before harvesting. The His-tagged human PGK was purified using a nickel-nitrilotriacetic acid-agarose column according to the manufacturer's protocol (Qiagen).
Immunoblotting SDS-polyacrylamide gel electrophoresis was performed essentially as described by Laemmli (36). For Western blotting, proteins were resolved by electrophoresis in a 10% polyacrylamide gel containing SDS, and electroblotted onto a PVDF membrane. The membrane was probed with monoclonal ␤-tubulin antibody (1:500 dilution) and HRP-conjugated goat anti-mouse IgG polyclonal antibody for the detection of tubulin, or with affinity-purified polyclonal PGK antibody (1:100 dilution) and HRP-linked protein A for the detection of PGK. The immunocomplex was detected by ECL detection system. To detect the protein band(s) that interacts with tubulin by West-Western blotting, the blotted membrane was first blocked with 3% BSA in a binding buffer (20 mM HEPES-KOH (pH 7.9), 100 mM KCl, 5 mM MgCl 2 , 5 mM 2-mercaptoethanol) at 4°C overnight, and then incubated with or without highly purified tubulin (50 g/ml) in a binding buffer containing 0.1% BSA at room temperature for 3 h. After extensive washing of the membrane, tubulin on the membrane was detected by using anti-␤-tubulin monoclonal antibody and 4-chloro-1-naphthol detection system.

Protein Concentration
Protein concentrations were determined by the methods of Bradford (37) using BSA as the standard.

RESULTS
Purification of the Host Factors-Cellular proteins (host factors) have been considered to have essential roles in the in vitro mRNA synthesis of SeV. We have previously shown that the host factor activity in the bovine brain extract is separated into two complementary fractions by hydroxylapatite column chromatography (18). These two fractions synergistically stimulated the in vitro mRNA synthesis of SeV, and one could be replaced by highly purified tubulin (18). In order to identify and characterize the host factor(s) that acts complementally to tubulin, we attempted to purify the active components from the bovine brain extract. In this study, hydroxylapatite column chromatography was performed as the first step essentially as described by Mizumoto et al. (18) with some modifications. Bovine brain extract (S-100 fraction) was loaded onto a hydroxylapatite (Hypatite C) column, and proteins were stepeluted with 75, 150, 300, and 600 mM potassium phosphate buffer, as illustrated in Fig. 1A. In these conditions, the host FIG. 1. Purification of host factor activity for SeV mRNA synthesis in the HA75 by Blue-Sepharose column chromatography. A, purification scheme for the host factors, required for the mRNA synthesis of SeV, from bovine brain extract (S-100). The host factors were purified by successive column chromatography using Hypatite C and Blue-Sepharose, as described under "Experimental Procedures". The 75, 150, 300, and 600 mM potassium phosphate eluates from Hypatite C column are referred to as HA75, HA150, HA300, and HA600, respectively. The 20, 150, and 300 mM KCl eluates from Blue-Sepharose column are referred to as BS20, BS150, and BS300, respectively. B, Hypatite C column 75 mM potassium phosphate eluate (HA75, 43 mg), containing an activity that stimulates SeV mRNA synthesis in the presence of HA300, was loaded onto a Blue-Sepharose column (inner diameter, 1.6 ϫ 18 cm) pre-equilibrated with TEMG (20 mM Tris-KCl (pH 7.9 at 20°C), 0.5 mM EDTA, 5 mM 2-mercaptoethanol, 20% glycerol) containing 20 mM KCl, and proteins were step-eluted with TEMG containing 20, 150, and 300 mM KCl. Proteins in each column fraction were concentrated and subjected to in vitro mRNA synthesis reactions with SeV particles (6 g) in the absence (lanes 1-7) or presence (lanes 8 -14) of HA300 (10 g) as indicated on the top of the figure. 32 P-Labeled transcripts were resolved by electrophoresis in an 1.2% agarose gel and detected by autoradiography. Lane 1 indicates transcription products with virus particles alone (SeV alone). Lanes 2 and 9 showed products from reactions with sample before column chromatography (HA75, 10 g) without or with HA300, respectively. Proteins used in these assays were 8.0, 1.5, and 0.5 g for the 20 (BS20), 150 (BS150), and 300 (BS300) mM KCl eluates, respectively. factor activity was separated as observed before (18) into two complementary fractions, the 75 and 300 mM potassium phosphate eluates (referred to HA75 and HA300, respectively). Western blot analysis with monoclonal anti-tubulin antibody revealed that tubulin was mainly detected in HA300, but almost not in HA75 (data not shown). Although the addition of HA75 alone (Fig. 1B, lane 2) or HA300 alone (lane 8) to the in vitro transcription mixture containing SeV particles (lane 1) gave weak transcription-stimulatory activities, the combinatorial addition of HA75 and HA300 synergistically stimulated the reaction to give 18 S size transcript (lane 9). To purify the host factor(s) contained in HA75, the fraction was loaded onto a Blue-Sepharose column, and proteins were step-eluted with 20, 150, and 300 mM KCl (Fig. 1A). The 20, 150, and 300 mM KCl eluates are referred to as BS20, BS150, and BS300, respectively. When BS fractions were added to the transcription reaction mixture without HA300, each BS fraction alone (Fig.  1B, lanes [3][4][5] or in combination (lanes 6 and 7) supported little, if any, stimulatory activity. However, the simultaneous addition of BS20 and BS150 (lane 13) or all Blue-Sepharose column fractions (lane 14) in the presence of a saturating amount of HA300 restored the transcription-stimulatory activity to levels comparable to that obtained with the combination of HA75 and HA300 (lane 9), while each Blue-Sepharose column fraction alone failed to stimulate transcription (lanes 10 -12), even in the presence of HA300. These results indicate that the host factor activity in HA75 can be further separated into the BS20 and BS150 fractions.
One of the two complementary fractions from Blue-Sepharose column, BS150, was further purified through a Heparin-Sepharose column. As shown in Fig. 2A, proteins were eluted with a linear gradient of 150 -350 mM KCl, and the transcription-stimulatory activity of each fraction was assayed in the presence of HA300 and BS20. The activity was eluted as a single peak at 220 mM KCl. As seen in Fig. 2B, the active fractions (fractions 110 -126) contained a single polypeptide with an apparent M r of 46,000 (p46), which was coeluted with the transcription-stimulatory activity (lanes 5-9). To investigate the complementarity of the host factors for the transcription stimulation in a more purified system, p46 and BS20 were subjected to an in vitro transcription system reconstituted with highly purified tubulin instead of HA300 using viral particles or viral RNP (Fig. 3). In the case of transcription with virions ( Fig. 3B), the addition of BS20 alone (lane 2), p46 alone (lane 3), or their combination (lane 4) caused a weak stimulation activity. In the presence of a saturating amount of highly purified tubulin, which by itself had some stimulatory activity (lane 5), neither BS20 (lane 6) nor p46 (lane 7) altered the reaction. However, the combination of three factors, BS20, p46, and tubulin, resulted in significant stimulation of mRNA synthesis (lane 8). In order to characterize the host factors using a more purified viral transcription machinery, the viral RNP complex was isolated from detergent-disrupted viral particles as described under "Experimental Procedures" (Fig. 3A, lane 2) in which most of the envelope glycoproteins were removed, and used for transcription in place of viral particles (Fig. 3C). Transcription was similarly and effectively stimulated depending on the addition of three factors: tubulin, p46, and BS20 (lane 8). These data suggest that the simultaneous presence of p46, tubulin, and unknown factor(s) in BS20 is required for the maximal transcription of the SeV genome, and that these factors act directly on the viral RNP to stimulate transcription.
Identification of p46 as PGK-To identify p46, we performed amino acid sequence analysis of tryptic peptides of p46 and then searched for similar sequences in the protein data base (PIR) using the BLAST program. Surprisingly, amino acid sequences of three peptides were highly homologous to those of a eukaryotic glycolytic enzyme, PGK (ATP:3-phospho-D-glycerate 1-phosphotransferase, EC 2.7.2.3) (Fig. 4A). These sequences almost completely matched the sequences of mammalian PGK, such as mouse PGK-1 (38) and human PGK-1 (39). The molecular weights of various PGKs were reported to be 46,000 Ϯ 2,000 by SDS-polyacrylamide gel electrophoresis (40), which also corresponds to the M r of p46 (also see Fig. 5A).
PGK catalyzes the reversible interconversion of ADP ϩ 1,3bisphosphoglycerate to ATP ϩ 3-phosphoglycerate. To assess p46 as bovine PGK, we examined the enzymatic activity of PGK at various steps during purification of p46. As shown in Table  I, a majority of PGK activity was copurified with transcriptionstimulatory activity of p46. The specific activity as the enzyme of the final column fraction, HS220 (481 units/mg protein) was comparable to that of rabbit muscle PGK (431 units/mg protein) or of yeast (S. cerevisiae) PGK (646 units/mg protein). An effective purification of PGK (p46) was achieved by three steps of column chromatography with an increase of 125-fold in the specific activity and with 59% recovery. Furthermore, PGK activity was coeluted with the transcription-stimulatory activity from a Heparin-Sepharose column (Fig. 4B). Since p46 was not separated from the other complementary factor(s) until Blue-Sepharose column chromatography, the specific activities of transcription in the presence of BS20 and tubulin were measured for the last two steps of purification. The final step of purification led to a 15-fold increase in the specific activity of transcription-stimulatory activity, which was nearly comparable to a 13-fold increase in PGK activity. As shown in Fig. 4C, affinity-purified polyclonal anti-rabbit muscle PGK antibodies, which reacted with rabbit muscle PGK and yeast one (lane 9), recognized p46 in BS150 (lane 1) as well as in HS fractions (lanes 3-7).
To investigate whether the transcription-stimulatory activity of p46 can be replaced by PGK from other sources, highly purified PGKs from rabbit muscle or Saccharomyces cerevisiae instead of p46 were added to the transcription reaction in the presence of complementary factors, tubulin and BS20 (Fig. 5). Each preparation from rabbit muscle (lane 2) or yeast (lane 3) contained a single polypeptide with approximate M r values of 43,000 or 48,000, respectively (Fig. 5A). As shown in Fig. 5B, PGK either from rabbit muscle or from yeast was also capable of stimulating SeV transcription in the presence of tubulin and BS20 in a dose-dependent manner as was p46, with a similar specific activity. By contrast, the addition of the rabbit muscle type GAPDH with tubulin and BS20 had no effect on the SeV transcription. To further provide evidence in support of a direct involvement of PGK in the SeV transcription, we tested whether a recombinant PGK can substitute for p46. We constructed human PGK-1 expression vector to produce Histagged recombinant protein and purified the protein to almost homogeneity, which had an apparent M r of 46,000 (Fig. 5A,  lane 4) corresponding to the calculated M r (46,112) of the tagged protein. The purified recombinant PGK exhibited the PGK activity (363 units/mg protein). As seen in Fig. 5B, the purified recombinant human PGK also stimulated the SeV transcription reaction in the presence of tubulin and BS20 with a specific activity similar to those obtained with native PGKs. From these biochemical and immunological data, we conclude that p46 is the bovine PGK.
Functional Properties of the Host Factors-It is interesting to see whether the enzymatic activity of PGK is directly involved in the transcription-stimulatory activity or not. We tested the effects of 3-phosphoglycerate, a substrate, and DL-glycerol-3phosphate, a competitive inhibitor against 3-phosphoglycerate,  4. Identification of p46 as phosphoglycerate kinase. A, the tryptic peptides of p46 were fractionated by reverse-phase high performance liquid chromatography. Three (peptides 1-3) of the obtained peptides were subjected to the amino acid sequence analysis, followed by the homology search from the protein data base using the BLAST program. Partial amino acid sequences of the three peptides were shown with sequences of phosphoglycerate kinases from mouse, human, and yeast (S. cerevisiae). Numbers indicate the amino acid positions of each PGK. Identical residues are marked by vertical lines. An amino acid that could not be identified is indicated by a question mark. B, the aliquots (0.3 l) of the Heparin-Sepharose column fractions ( Fig. 2A) was assayed for PGK activity (Ⅺ) by measuring the rate of the formation of 1,3-bisphosphoglycerate (nmol/min) and in vitro mRNA synthesis (q) as described under "Experimental Procedures." C, Western blot analysis for PGK. BS150 (lane 1, 3.0 g) and Heparin-Sepharose column fractions (lanes 2-8, each 6.0 l) were immunoblotted with an affinity-purified anti-rabbit muscle PGK polyclonal antibody. In lane 9, a mixture of PGKs (each 0.1 g) from rabbit muscle and yeast was used as the positive control.
on the transcription-stimulatory activity of p46 in the range of 1-3 mM (Table II), since the K m value for 3-phosphoglycerate (34,40) and the K i value for DL-glycerol-3-phosphate (41) were reported to be 1 and 0.7 mM, respectively. Increasing the concentrations of both reagents had little effect on the transcription-stimulatory activity of p46. Thus, the enzymatic activity per se of PGK does not seem to be required for its activity.
To elucidate the functions of the host factors in SeV mRNA synthesis, their effects on transcription initiation and elongation reactions were examined. First, viral particles were incubated with the host factors in various combinations without four NTPs, followed by ultracentrifugation to obtain viral RNPs (initiation complexes) as the pellets. Four NTPs were then added to the isolated viral RNPs to determine their mRNA synthesizing activity (Fig. 6A). We have previously shown that transcription initiation complex formed with bovine brain extract contains tubulin, an essential factor required for mRNA synthesis (19). When virus particles were preincubated with highly purified tubulin instead of bovine brain extract, the isolated viral RNP could also synthesize viral mRNAs (lane 5). These data further confirmed the notion that tubulin plays an important role in the initiation complex formation (19). In contrast, when virus particles were preincubated with BS20 (lane 2), p46 (lane 3), or their combination (lane 4), neither RNP pellets obtained by centrifugation supported mRNA synthesis. BS20 and p46 had little effect on the initiation complex formation with tubulin (lanes 6 -10). We analyzed these RNP pellets by electrophoresis in a SDS-polyacrylamide gel followed by silver staining (Fig. 6B) and Western blotting using antitubulin antibodies as the probe (Fig. 6C). All the active initiation complexes formed in the presence of tubulin contained nearly identical amounts of tubulin (Fig. 6B, lanes 7-10), which was confirmed by Western blotting (Fig. 6C, lanes 7-10), indicating that neither BS20 and p46 affected integration of tubulin into the initiation complexes. Thus, it seems that p46 and the unknown factor(s) in the BS20 fraction participate in transcription after the initiation complex formation.
We analyzed the effect of host factors on the mRNA chain elongation reaction from the isolated initiation complexes formed in the presence of tubulin (Fig. 7). The initiation complexes were prepared by incubating virus particles with or without highly purified tubulin, followed by ultracentrifugation, and then subjected to transcription reactions in the presence or absence of the host factors. As mentioned above, the initiation complex formed with tubulin showed a transcription activity, although at a low level, without addition of the other host factors (Fig. 7, lane 6), and further addition of tubulin to the complex had no effect on transcription activity (lane 10), indicating that a saturating amount of tubulin was integrated into the initiation complex. Either BS20 alone (lane 7) or p46 alone (lane 8) also failed to stimulate the transcription from the isolated initiation complex. However, when p46 (PGK) and BS20 were simultaneously added to the RNA chain elongation reaction from the isolated initiation complex, mRNA synthesis was dramatically stimulated (lane 9). On the other hand, addition of p46 (PGK) and BS20 to RNA chain elongation reaction from the RNP, which had not been preincubated with tubulin, showed only a little mRNA synthesizing activity (lane 4). From these data, it seems that p46 and BS20 act cooperatively on the tubulin-containing active initiation complex at the level of RNA chain elongation.
To investigate whether these factors can interact with tubulin, West-Western blot analysis was performed using highly purified tubulin as the probe (Fig. 8). Fig. 8A shows the patterns of protein transfer of BS20, p46, tubulin, and virus particles from the gel to a PVDF membrane monitored by Amido Black staining. For West-Western blotting, proteins were sim-  3, 4, 7, and 8), and 40 g of BS20 (lanes 2, 4, 6, and 8) in various combinations, as shown above each lane. After preincubation, the viral RNPs were isolated as pellets by ultracentrifugation, and then subjected to RNA chain elongation by adding nucleoside triphosphates, as described under "Experimental Procedures." Lane 1 indicates transcription products from the reaction with a pellet obtained after incubation of virus particles alone. B, the viral RNP pellets (lanes 3-10) corresponding to lanes 1-8 in panel A were analyzed by electrophoresis in a 10% SDS-polyacrylamide gel. Lanes 1 and 2 indicate 0.05 g of tubulin and 0.4 g of virus particles (SeV), respectively. After electrophoresis, proteins were detected by silver staining. Closed arrowhead indicates the migration position of tubulin incorporated into the complexes. Positions of viral proteins (NP, P, M, and L) and marker proteins are shown on the right and left, respectively. C, samples shown in panel B were immunoblotted with an anti-chicken brain ␤-tubulin monoclonal antibody. The migration position of ␤-tubulin is shown by a closed arrowhead.

TABLE I
Purification of p46 (PGK) as a host factor for mRNA synthesis of Sendai virus from bovine brain extract (S-100) PGK activity was assayed by measuring the formation of 1,3-bisphosphoglycerate using 3-phosphoglycerate and ATP as substrates as described under "Experimental Procedures." One unit of PGK activity was defined as the formation of 1 mol of 1,3-bisphosphoglycerate/min. Transcriptionstimulatory activity was measured in in vitro mRNA synthesis reactions with SeV particles (6 g) in the presence of tubulin (3.0 g) and BS20 (8.0 g) as described under "Experimental Procedures." The specific activity was expressed as the amounts of [ 32 P]UMP (nmol) incorporated into mRNA (18 S region in the agarose gel) per 1.0 mg of protein. a From 20 g of bovine brain. b The total activity of the BS150 was assumed to be 100%, because p46 was not separated from other complementary factor(s) until the step of Blue-Sepharose column chromatography.
ilarly blotted on a PVDF membrane, then the membranes were pretreated with (lanes 5-8) or without (lanes 1-4) highly purified tubulin. After extensive washing of the membrane, tubulin remaining on the membrane was detected with an anti-␤-tubulin monoclonal antibody. In the case of the pretreatment of the membrane without tubulin, no positive band was observed. By contrast, pretreatment of the membrane with tubulin allowed visualization of tubulin-binding proteins, p46 (lane 6) and two other proteins with apparent M r values of 52,000 (p52) and 27,000 (p27) in the BS20 fraction (lane 5). These data suggest that p46 (PGK) and factor(s) in BS20 stimulate SeV mRNA synthesis at the elongation step, possibly through their interaction with tubulin that has been integrated into the active initiation complex. However, it is not clear whether tubulin-binding proteins in BS20 are directly involved in transcription-stimulating activity.
Host Factors Act Specifically on mRNA Synthesis of Sendai Virus-Finally, to confirm a specific roles of host factors in SeV mRNA synthesis, we examined the effects of host factors on the in vitro plus leader RNA synthesis of SeV (Fig. 9) and in vitro mRNA synthesis of VSV, which belongs to the Rhabdovirus family in the order Mononegavirales (Fig. 10). The plus-sense leader RNA was transcribed from the 3Ј-end of a SeV genome. We have developed an in vitro plus leader RNA synthesis system using virus particles as described elsewhere. 2 In this system, artificially prematurely terminated plus leader RNA was detected to confirm that transcription took place from the genome leader sequence. Using this system, we have shown that plus leader RNA synthesis also requires host cell proteins. We tested whether the host factors for mRNA synthesis isolated in this study can also stimulate the leader RNA synthesis or not. As shown in Fig. 9, tubulin stimulated not only mRNA synthesis of SeV but also plus leader RNA synthesis (lane 5). In

TABLE II
Effects of a substrate and an inhibitor of PGK on mRNA synthesis of Sendai virus Transcription-stimulatory activity of p46 was measured in in vitro mRNA synthesis reactions with SeV particles (6 g), tubulin, and BS20 in the presence or absence of 3-phosphoglycerate or DL-glycerol-3-phosphate at the indicated concentrations. Proteins used were 0.1, 3.0, and 8.0 g for p46, tubulin, and BS20, respectively. The transcriptionstimulatory activity was expressed as the amounts of [ 32 P]UMP incorporated into the 18 S mRNA band. Effect of p46 and BS20 on the chain elongation step. The initiation complex were prepared by incubating SeV particles (300 g) without (lanes 1-5) or with (lanes 6 -10) tubulin (150 g), followed by ultracentrifugation as described under "Experimental Procedures." Aliquots of the RNP pellets (corresponding to 6 g of staining with SeV particles) were subjected to the chain elongation reaction in the presence of indicated factors. Factors used were tubulin (3.0 g), p46 (0.1 g), and BS20 (8.0 g). After elongation reaction, RNAs were analyzed by agarose gel electrophoresis and autoradiographed. contrast, BS20 potently repressed tubulin-mediated leader RNA synthesis (lane 6), while p46 (PGK) failed to stimulate tubulin-mediated leader RNA synthesis with (lane 8) or without (lane 7) BS20. In addition, no significant effect of these factors on the mRNA synthesis of VSV was observed (Fig. 10). These results suggest that p46 (PGK) and factor(s) present in BS20 may be specifically involved in the mRNA synthesis of SeV.

DISCUSSION
In our in vitro mRNA synthesizing system using purified SeV particles, viral mRNA synthesis was almost completely dependent on the presence of host cell proteins (host factors), one of which was suggested to be tubulin, a cytoskeletal protein (18). In this study, we attempted to purify the host factor(s) which acts complementarily to tubulin to stimulate viral mRNA synthesis and found the activity can be separated into two complementary fractions (BS20 and BS150) by Blue-Sepharose column chromatography (Fig. 1). These results suggest that the host factor is composed of at least three factors including tubulin. One of them was further purified to a single polypeptide with an apparent M r of 46,000 (p46) by Heparin-Sepharose column chromatography of BS150 (Fig. 2), and was identified as phosphoglycerate kinase (PGK), a glycolytic enzyme by biochemical and immunological analyses ( Fig. 4 and  5). The fact that the bacterially expressed PGK could substitute for p46 also supported the direct involvement of PGK in SeV transcription. To our knowledge, this is the first instance of functional identification of PGK as a host factor for mononegavirus transcription.
Since the molecules of PGK (0.1 g or 2.2 pmol per assay), required for SeV transcription in the presence of sufficient amounts of tubulin and BS20, is lower than that of tubulin (3 g or 30 pmol as heterodimers per assay) (Fig. 3), and is roughly the same order as that of L protein present in an assay mixture, it seems that PGK acts catalytically rather than stoichiometrically. PGK catalyzes the transfer of a phosphoryl group from the acyl phosphate of 1,3-bisphosphoglycerate to ADP, thereby forming ATP and 3-phosphoglycerate. However, the enzymatic activity of PGK does not seem to be involved in the transcription-stimulatory activity of PGK, because the substrates, 1,3-bisphosphoglycerate and 3-phosphoglycerate were not present in the transcription reaction mixture. Additionally, the addition of 3-phosphoglycerate, a substrate, or DL-glycerol-3-phosphate, a substrate analogue inhibitor, did not affect transcription (Table II). Transcription of SeV was stimulated by both mammalian PGKs and yeast PGK, although the identity of overall primary amino acid sequences between mammalian PGK and yeast PGK is about 64% (38,42).
Moyer et al. (20) reported that SeV transcription is inhibited by a monoclonal antibody against ␤-tubulin in an in vitro transcription system with SeV-infected cell lysate. They also mentioned that when the cell extract from uninfected BHK cells was added to detergent-disrupted SeV particles, several classes of viral mRNA were synthesized. By contrast, although addition of purified tubulin alone to SeV particles led to the synthesis of leader-like RNA and mRNA for the 5Ј-proximal NP gene, downstream mRNAs were not transcribed. Similar results were reported for measles virus (12,16). Thus it seems that an additional factor(s) may be required for complete transcription of these viral genomes. These observations are in agreement with our results presented before and in this work. At present, however, it is not clear whether the complementary factors PGK (p46) and BS20 preferentially stimulate transcription of downstream mRNAs rather than that of mRNA for the 5Ј-proximal NP gene.
We have previously shown that a transcription initiation complex formed with bovine brain extract contains tubulin molecules (about 250 copies/genome), and that the amounts of tubulin integrated into the complexes correlate with their transcriptional activity (19). In the present work, we showed that highly purified tubulin is also integrated into the viral RNP complex at the transcription initiation step and activates transcription from the complex (Fig. 6), confirming that tubulin is directly involved in transcription initiation complex formation. By contrast, p46 (PGK) and BS20, when added together, dramatically stimulated chain elongation from the initiation complex formed with purified tubulin, without affecting the initiation complex formation (Fig. 7). Moreover, direct interactions between p46 (PGK) and tubulin was shown by West-Western blot analysis. These observations suggest that p46 (PGK) stimulates SeV mRNA synthesis at the elongation step, probably through the interaction of p46 (PGK) with tubulin that has been integrated into the initiation complex. Alternative possibility is that p46 (PGK) and BS20 may function in the reutilization of the viral RNA polymerase.
De et al. (21) purified a host factor for in vitro transcription of human parainfluenza virus type 3 (HPIV3), a paramyxovirus closely related to SeV, from the cytoplasmic extract of uninfected cells, and identified it as actin, a cytoskeletal protein. Furthermore, they found that both polymeric and monomeric forms of actin are integrated into the viral RNP complex, but only the polymeric form can activate HPIV3 transcription, and HPIV3 transcription is not stimulated by tubulin (22). In our SeV mRNA synthesizing system, highly purified actin also stimulated mRNA synthesis, but specific activity of tubulin dimer was 7-fold higher than that of the actin monomer (data not shown). On the other hand, they reported that GAPDH, a glycolytic enzyme, interacts with the 3Ј-terminal leader sequence of the genome as well as the plus leader RNA (43). It is interesting to note that GAPDH catalyzes formation of 1,3bisphosphoglycerate, which is in turn the substrate for its downstream enzyme, PGK, in the sequential glycolytic reaction. It was reported that GAPDH is able to interact with not only PGK (44) but also tubulin (45). However, significant effect on the transcription with SeV particles was not observed by the addition of rabbit muscle type GAPDH in the various combinations with the host factors (tubulin, PGK, and BS20) for SeV transcription (Fig. 5, data not shown). We could not detect FIG. 10. Effects of tubulin, p46, and BS20 on mRNA synthesis of VSV. In vitro mRNA synthesis of VSV was carried out under the same conditions for SeV using 1 g of VSV particles instead of SeV particles in the presence of various factors as indicated. The following amounts of factors were added to the reaction: tubulin (3.0 g), p46 (0.1 g), and BS20 (8.0 g). After reaction, 32 P-labeled transcripts were resolved by electrophoresis in an 1.2% agarose gel and detected by autoradiography. Lane 1 indicates transcription products with virus particles alone (VSV alone). GAPDH in SeV particles either by Western blotting with anti-GAPDH antibody or by the assay for its enzymatic activity (data not shown).
Most of the glycolytic enzymes have been shown to interact with filamentous actin (46 -49) as well as microtubules (50 -52), suggesting that it is a common feature of these enzymes to bind to cytoskeletal proteins. We also demonstrated that PGK interacts with tubulin (Fig. 8). However, physiological significance of binding of the glycolytic enzymes to cytoskeletal proteins are still unknown. Many glycolytic enzymes were suggested to interact with the carboxyl-terminal tails of the tubulin subunits (53), which is important sites for regulation of the assembly of microtubules through the binding of microtubule-associated proteins (reviewed in Ref. 54). Previous studies suggested that the carboxyl-terminal tails of the tubulin subunits, which are highly negatively charged and exposed on the surface of the tubules, may be involved in SeV transcription (16,18). It is important to investigate the significance of the acidic tails of tubulin in the transcription-stimulatory activity and the recruitment of other host factors such as PGK. Tubulin may have multiple roles in SeV mRNA synthesis, e.g. tubulin 1) dissociates M protein, which acts as the negative regulator for RNA synthesis from the viral RNP; 2) stabilizes the RNP structure to maintain the template as a transcriptionally active state; and 3) recruits other host factors to the viral RNP. It remains to be seen at which precise steps of viral transcription PGK and BS20 act, and to which viral or cellular proteins they interact.
Our studies indicate that tubulin functions as a common host factor for both mRNA synthesis and plus leader RNA synthesis of SeV, while PGK acts specifically for mRNA synthesis (Fig. 9). Moyer et al. (20) also reported that purified tubulin stimulates the synthesis of leader-like RNA from the detergent-disrupted SeV particles. Interestingly, crude BS20 fraction, which exhibits the mRNA synthesis-stimulatory activity, was found to strongly repress leader RNA synthesis. Preliminary purification of BS20 revealed that both activities are separated into different fractions (data not shown). It is interesting to note that the negative factor which inhibits leader RNA synthesis, was found to be distinct from an RNase or a proteinase, and that it may act specifically to inhibit plus leader RNA synthesis, because this factor did not inhibit in vitro mRNA synthesis of SeV or of VSV. In addition, we previously reported an inhibitory factor that potently inhibits SeV mRNA synthesis from rat liver extract (55). Thus, mRNA synthesis as well as plus leader RNA synthesis may be regulated by both positive and negative factors.
The specific action of the host factors, tubulin, PGK, and BS20, in SeV mRNA synthesis also comes from the following observations. These factors did not show any effect on the in vitro mRNA synthesis with VSV particles (Fig. 10). Although either purified VSV particles or the viral RNP from virions are capable of transcribing mRNA in vitro, soluble cell extracts increase the level of mRNA synthesis (56). We also observed that in vitro transcription with VSV particles was stimulated 2-3-fold either by the addition of bovine brain extract or HeLa cell extract (data not shown), suggesting that at least tubulin, PGK, and BS20 are not responsible for the stimulation VSV transcription by crude cell extracts in our system. Hill et al. (56,57) reported that in vitro transcription and replication of VSV was stimulated by microtubule-associated proteins derived from HeLa cells or bovine brain, but not by tubulin. By contrast, Moyer et al. (20) reported that in vitro transcription of VSV was stimulated by purified tubulin and was inhibited by a monoclonal antibody against ␤-tubulin. At present, we do not know the reason for this discrepancy in the requirement of the host factors for in vitro mRNA synthesis of VSV. Recently, a reconstituted VSV transcription system was created using genome RNA encapsidated with nucleocapsid protein (N) and recombinant L and P proteins (58). In those studies, it was revealed that phosphorylation of P protein with cellular casein kinase II (59) and the association of L protein with protein synthesis elongation factor EF-1 (60) are essential for mRNA synthesis of VSV, and these host factors are packaged within the viral particles. Therefore, to elucidate precise functions of the host factors for transcription of SeV, it is important to develop a transcription system reconstituted with highly purified viral proteins and host factors.
Identification of a host factor as PGK will be helpful for understanding not only the regulatory mechanisms of SeV RNA synthesis but also the precise replication site of SeV in the cell. Since PGK is abundant and ubiquitous in cytoplasm of eukaryotic cells, its usage as a transcription factor seems to be feasible for the cytoplasmic replication of SeV. It is important to note that PGK was also identified as a subunit of primer recognition proteins, which are cofactors of DNA polymerase ␣ and may have a role in lagging strand DNA replication in nuclei (61,62). These observations including the involvement of PGK in cellular DNA replication and in viral transcription have suggested that PGK may contribute to regulation of multiple cellular as well as viral processes in addition to glycolysis. In this regard, a detailed study along the precise function of PGK and other host factors for SeV multiplication should lead to better understanding of host-virus interaction processes. Bacterially expressed recombinant PGK will make it possible to map the domain(s) of PGK required for the SeV transcription through site-directed mutational analysis. Experiments are in progress to address the significance of these interactions by delineating the role of PGK and BS20 in SeV transcription using a reconstituted RNA synthesizing system.