De novo RNA synthesis by a recombinant classical swine fever virus RNA-dependent RNA polymerase Guang-Hui Yi, Chu-Yu Zhang, Sheng Cao, Hai-Xiang Wu and Yi Wang Institute of Virology, College of Life Sciences, Wuhan University, Wuhan, Hubei, China Classical swine fever virus nonstructural protein 5B (NS5B) encodes an RNA-dependent RNA polymerase, a key enzyme of the viral replication complex. To better under- stand the initiation of viral RNA synthesis and to establish an in vitro replication system, a recombinant NS5B protein, lacking the C-terminal 24-amino acid hydrophobic domain, was expressed in Escherichia coli. The truncated fusion protein (NS5BD24) was purified on a Ni-chelating HisTrap affinity column and demonstrated to initiate either plus- or minus-strand viral RNA synthesis de novo in a primer- independent manner but not by terminal nucleotidyle transferase activity. De novo RNA synthesis represented the preferred mechanism for initiation of classical swine fever virus RNA synthesis by RNA-dependent RNA polymerase in vitro.BothMg 2+ and Mn 2+ supported de novo initiation, however, RNA synthesis was more efficient in the presence of Mn 2+ than in the presence of Mg 2+ . De novo initiation of RNA synthesis was stimulated by preincubation with 0.5 m M GTP, and a 3¢-terminal cytidylate on the viral RNA template was preferred for de novo initiation. Furthermore, the purified protein was also shown, by North-Western blot analysis, to specifically interact with the 3¢-end of both plus- and minus-strand viral RNA templates. Keywords: classical swine fever virus; RNA-dependent RNA polymerase; nonstructural protein 5B; de novo RNA synthesis; RNA-binding activity. Classical swine fever virus (CSFV) is a small enveloped positive-strand RNA virus classified in the genus of Pestivirus, which also comprises bovine viral diarrhea virus (BVDV) and border disease virus (BDV). Together with the genera Flavivirus and Hepacivirus, they form the family Flaviviridae [1,2]. The genomic RNA,  12.3 kb in length, contains a single long ORF encoding a polyprotein of  3898 amino acids which is flanked by 5¢-and3¢-UTRs [3]. The 5¢-UTR contains an internal ribosomal entry site (IRES) for cap-independent translation of the viral poly- protein [4,5], whereas the 3¢-UTR may contain replication signals involved in minus-strand RNA synthesis, as in BVDV [6]. The polyprotein is processed, co- and post- translationally, into 12 polypeptides by viral and cellular proteases. The order of polypeptides is NH 2 -N pro -C-E rns - E1-E2-p7-NS2-NS3-NS4A-NS4B-NSA-NS5B-COOH [7]. N pro , a nonstructural autoprotease, can release itself from its precursor, but is not necessary for viral replication in cell culture [8]. Nuclecapsid protein C, and glycoproteins E rns , E1 and E2, represent four structural proteins, and form the capsid and envelope of the virion, respectively. The others are nonstructural proteins (NS) [3]. Most of the NS are speculated to be components in the viral replication cycle. Among them, NS3 is a multifunctional enzyme and responsible for functions associated with the replication and biotype of cytopathogenicity in cell culture [9]. The last viral protein (NS5B), at the C terminus of the polyprotein, is a key component responsible for the replication of viral RNA genome. It also contains motifs shared by RNA- dependent RNA polymerases (RdRps), such as the Gly– Asp–Asp (GDD) motif, which is highly conserved among RdRps [10] and has been demonstrated to possess RdRp activity in insect cells [11,12] and porcine kidney cells (PK- 15) [13]. It is believed that certain enzymatic functions of NS3 and NS5B may play an important role during the replication of viral RNA. The replication of the CSFV genome is generally thought to be similar to other positive-strand RNA viruses: synthesis of complementary minus-strand RNA with the plus-strand genomic RNA as template, and subsequent synthesis of the progeny RNA with the minus-strand RNA as template. Thus, the 3¢-end of both plus- and minus-strand RNAs may contain the cis-acting elements, such as promoter or enhancer, involved in the initiation of viral RNA synthesis by RdRp. Although several infectious cDNA clones of CSFV have been developed [14–18], facilitating the research of cytopathogenicity, replication and function of viral proteins by reverse genetic approach in the cell culture system [8,9,19], another method to study CSFV replication would be to work towards identification of the cis-acting elements at the 3¢-end of both plus- and minus-strand RNAs and possibly the viral or cellular proteins that interact with it to form a replication complex for initiating viral RNA synthesis in vitro. At present, the molecular mechanism of Correspondence to C Y. Zhang, Institute of Virology, College of Life Sciences, Wuhan University, Wuhan 430072, China. Fax: + 86 27 87883833, Tel.: + 86 27 87682833, E-mail: avlab@whu.edu.cn Abbreviations: ALP, alkaline phosphatase; BVDV, bovine viral diarrhea virus; CSFV, classical swine fever virus; DIG, digoxin; RdRp, RNA-dependent RNA polymerase; IRES, internal ribosomal entry site; NS5B, nonstructural protein 5B; TNTase, terminal nucleotidyle transferase. (Received 23 September 2003, accepted 24 October 2003) Eur. J. Biochem. 270, 4952–4961 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03897.x initiating CSFV RNA synthesis is not well understood. Previous reports have shown that the recombinant CSFV NS5B (expressed in insect cells) catalyzed RNA synthesis was strictly primer-dependent and that intramolecular priming copy-back synthesis represented the preferred mechanism for initiation of RNA synthesis [11]. However, the activity of the cellular terminal nucleotidyle transferase (TNTase) was also demonstrated to be present in the cytoplasmic extracts of insect cells [20,21]. The cellular TNTase could add extra nucleotides to the 3¢-terminus of the RNA template and might serve as primer for template- primed copy-back synthesis. Moreover, evidence was obtained that the TNTase activity associated with the hepatitis C virus NS5B might be the result of a contamin- ating cellular protein present in minute amounts in the enzyme preparation [21]. To better understand the initiation of CSFV RNA replication, we expressed and purified a recombinant NS5BD24 fusion protein from Escherichia coli BL21 (DE3). The fusion protein was demonstrated to have the ability to initiate de novo either plus- or minus-strand viral RNA synthesis in a primer-independent manner and to specifically interact with viral RNA templates. This in vitro RdRp assay will be useful for using to study the sequences and proteins required for the initiation of CSFV RNA synthesis. Materials and methods Plasmid constructs The plasmid pGEM5b, containing full-length CSFV (Shi- men strain) NS5B, was constructed as described previously [13]. The NS5B fragment lacking the C terminal 24 amino acids (NS5BD24) was PCR amplified with the following primer pair – NS5BFor and NS5BRev – from pGEM5b (Table 1). A polyhistidine tag (GSHHHHHH) was intro- duced at the C terminus to facilitate purification of the NS5BD24 protein. After purification, the PCR products were digested with NcoIandBglII and then inserted into the NcoI/BamHI sites of vector pET-28a (Novagen). The resulting expression vector, pET-NS5BD24, which was driven by the T7 RNA polymerase promoter, was trans- formed into E. coli DH5a. Site-directed mutagenesis of GDD to GAA, containing the double substitution of both Asp448 and Asp449 to alanine, was carried out by overlapping PCR. The N terminal 1370 bp fragment was amplified by NS5BFor and NS5Bm2, and the C terminal 780 bp fragment was amplified by NS5Bm1 and NS5BRev. Then, the two fragments were purified and combined to generate the mutant NS5BD24GAA using the outer primers NS5BFor and NS5BRev. After modification with NcoI and BglII, the mutant fragment was subcloned into vector pET-28a. Transformants were analyzed by restriction enzyme mapping and confirmed by the dideoxy sequencing method. Expression and purification of recombinant CSFV NS5BD24 E. coli BL21 (DE3), transformed with either pET-NS5BD24 or pET-NS5BD24GAA, was grown in Luria–Bertani (LB) medium, at 37 °C, to an attenuance (D) at 600 nm of  0.6– 0.8. Then, the temperature was lowed to 18 °C, and protein expression was induced for 20 h by the addition of 0.4 m M isopropyl thio-b- D -galactoside (IPTG). The cell pellet obtained from 500 mL of culture was resuspended in 30 mL of binding buffer containing 20 m M sodium phos- phate, pH 8.0, 500 m M NaCl, 20 m M imidazole, 10 m M b-mercaptoethanol, 20% glycerol, 1% Triton-X-100, 1 m M phenylmethanesulfonyl fluoride, and 10 lgÆmL )1 lysozyme. Then 20 lL of DNase I (Takara) was added to the suspension for 30 min at room temperature. The lysates were sonicated on ice to reduce viscosity, and any insoluble materials were removed by centrifugation at 13 000 g for 15 min. The clear supernatant was applied to a 1-mL Ni-chelating HisTrap affinity column (Amersham) pre- equilibrated with the binding buffer containing 50 m M imidazole. The bound protein was then eluted stepwise with elution buffer (20 m M sodium phosphate, pH 8.0, 500 m M NaCl, 5 m M b-mercaptoethanol, 20% glycerol, 0.2% Triton-X-100) containing a gradient imidazole con- centration from 150 to 450 m M .Thefractionswere monitored by SDS/PAGE and staining with Coomassie Brilliant Blue R250. Then, the His-tagged protein peaks were collected and dialyzed against buffer (20 m M Tris/HCl pH 8.0, 500 m M NaCl and 20% glycerol), followed by storage at )40 °C in small aliquots. The concentration of purified protein was determined by the Bradford method using BSA as a standard. SDS/PAGE and Western blot Protein fractions from the HisTrap affinity column were separated by 12% SDS/PAGE and electrotransferred to a nitrocellulose membrane. The membrane was blocked with Table 1. Sequence of the primers used in this study. The T7 polymerase promoter sequence is shown in italics. The mutated nucleotides are shown in bold and the additional polyhistidine amino acid sequences are underlined. Primers Sequence (5¢)3¢) NS5BFor CATGCCATGGGCAGTAATTGGGTGATGCA NS5BRev GAAG ATCTTAATGATGATGATGATGATG GCTGCCATTGTACCTGTCTGCCCCTT NS5Bm1 ATCAGGAGACCAGCAGCCCCGCACACAT NS5Bm2 ATGTGTGCGGGGCTGCTGGTCTCCTGAT 5¢-UTRfor GTATACGAGGTTAGTTCATTC 5¢-UTRfor1 CTATACGAGGTTAGTTCATTC 5¢-UTRfor2 TTATACGAGGTTAGTTCATTC 5¢-UTRfor3 ATATACGAGGTTAGTTCATTC 5¢-UTRrev TAATACGACTCACTATAGGGTGCCATGAA CAG 5¢-UTRrev2 GTGCCATGAACAGCAGAGATTTTTATAC P1 TAGCCATGCCCATAGTAGG P2 ATCAGGTCGTACTCCCATCAC 3¢-UTRfor TAATACGACTCACTATAGCGCGGGTAAC 3¢-UTRfor2 GCGCGGGTAACCCGGGATCTGAA 3¢-UTRrev GGGCCGTTAGGAAATTACCTTAGTC 3¢-UTRrev1 CGGCCGTTAGGAAATTACCTTAGTC 3¢-UTRrev2 TGGCCGTTAGGAAATTACCTTAGTC 3¢-UTRrev3 AGGCCGTTAGGAAATTACCTTAGTC Ó FEBS 2003 De novo RNA synthesis by CSFV RdRp (Eur. J. Biochem. 270) 4953 3% BSA in NaCl/P i and treated with rabbit anti-swine serum infected with CSFV. Alkaline phosphatase (ALP)- conjugated goat anti-(rabbit IgG) was used as the secondary antibody. After washing three times with NaCl/P i contain- ing 0.1% Tween-20, membrane-bound antibodies were detected with Nitro Blue tetrazolium/5-bromo-4-chloro- indol-2-yl phosphate. Preparation of RNA templates and RNA labeling RNA templates were prepared by in vitro transcription. The plasmid T6-1, containing the full-length CSFV 3¢-UTR, and the plasmid pGEM61s-2, containing the full-length 5¢-UTR, were constructed as previously reported [18]. The DNAfragmentofthe3¢-end of the plus-strand RNA (+)3¢-UTR was amplified using primers 3¢-UTRfor and 3¢-UTRrev from T6-1; and the 3¢-end of the minus-strand RNA (–)IRES, which is complementary to the CSFV 5¢-UTR, was amplified from pGEM61s-2 using the primers 5¢-UTRfor and 5-¢UTRrev. To generate (+)3¢-UTR mutants with substitution of the 3¢ terminus cytidylate with G, A and T, PCR was performed with sense primer 3¢-UTRfor and antisense primers 3¢-UTRrev1, 3¢-UTRrev2 or 3¢-UTRrev3, respectively, using T6-1 as template. For (–)IRES mutants, PCR was perfomed using the sense primers 5¢-UTRfor1, 5¢-UTRfor2 or 5¢-UTRfor3 and anti- sense primer 5¢-UTRrev from pGEM61s-2. All the PCR amplifications were performed using pfu DNA polymerase (MBI), and the DNA fragments were recovered using a DNA purification kit. The T7 RNA polymerase promoter sequence was introduced into the primers 3¢-UTRfor and 5¢-UTRrev to initiate RNA synthesis. In vitro transcription was carried out with T7 RNA polymerase, according to the manufacturer’s instructions (Promega). After a 2 h incuba- tion at 37 °C, the DNA templates were digested twice with RNase-free DNase I. The RNA transcripts were extracted with acid phenol/chloroform (1 : 1, v/v), followed by precipitation with two volumes of ethanol and 0.4 M sodium acetate. The precipitated RNA was dissolved in diethyl pyrocarbonate-treated water and the RNA concentration determined by measuring the absorbance (A)at260nm. Occasionally, RNA transcripts were further purified by 6% PAGE containing 7 M urea. The gel fragment containing the RNA was excised and incubated overnight in 10 m M Tris/ HCl, pH 7.5, 25 m M NaCl, 1 m M EDTA. After elimination of polyacrylamide, the RNA was precipitated by ethanol. The 3¢-hydroxyl group of RNA transcripts was blocked by sodium periodate, as described previously [22]. Ten micro- grams of RNA transcript was dissolved in 100 lLof50m M sodium acetate. After the addition of 25 lL sodium perio- date (100 m M ), the mixture was incubated for 1 h at room temperature, then phenol/chloroform extracted and ethanol precipitated. Residual sodium periodate was removed by several washes using 70% ethanol. For preparation of digoxin (DIG)-labeled RNA probes or templates, the DNA fragments of (+)3¢-UTR and (–)IRES were used as templates for in vitro transcription. RNA labeling was performed according to the instruction manual supplied with the DIG RNA labeling kit of Roche Molecule Biochemicals. The mixtures were incubated for 2 h at 37 °C, and DNA templates were removed by digestion with RNase-free DNase I. In vitro RdRp The in vitro RdRp standard assay was performed in a total volume of 50 lL containing 20 m M Tris/HCl, pH 8.0, 5m M MgCl 2 ,5m M MnCl 2 ,2m M dithiothreitol, 50 m M NaCl, 0.25 m M of each NTP, 0.3 lgofRNAtemplateand  0.1 lg of purified protein. The reaction mixtures were incubated at 25 °C for 2 h and stopped by the addition of 20 m M EDTA. The RNA products were extracted with acid phenol/chloroform (1 : 1, v/v) followed by ethanol precipi- tation. Then, the precipitates were dissolved with either 20 lL of diethyl pyrocarbonate-treated water or denaturing buffer (see below). Northern blot analysis The precipitated RdRp products were dissolved in a denaturing buffer containing 95% formamide, 10 m M EDTA, 20 m M Tris/HCl, pH 8.0, at 100 °Cfor5min, and then separated by PAGE (8% gel containing 7 M urea) in 1· Tris/borate/EDTA (TBE) buffer. After electrophor- esis, the gels were transferred to a positively charged nylon membrane (Hybond) and electroblotted for 4 h at 4 °C. The membrane was dried for 2 h at 80 °C and exposed to ultraviolet irradiation (254 nm) for 2 min at 0.12 JÆsq cm )1 . Hybridization was performed overnight, at 68 °C, in 10 mL of a solution containing 50% formamide, 2% blocking reagent (Roche), 5 · NaCl/Cit and 20 lgÆmL )1 yeast tRNA, together with the appropriate DIG-labeled RNA transcripts. The excess probes were eliminated gradually by washing the membrane from low stringency (2· NaCl/Cit, 0.2% SDS) to high stringency (0.1· NaCl/Cit, 0.1% SDS) at 68 °C. Then, the bound RNA was treated with ALP- conjugated anti-DIG Ig (1 : 5000) in dilution buffer (1· blocking reagent in 0.1 M maleic acid buffer, pH 7.0) for 30 min. The reaction complexes were visualized using Nitro Blue tetrazolium/5-bromo-4-chloroindol-2-yl phos- phate, according to the manufacturers of the DIG RNA detection kit (Roche). RT–PCR and real-time quantitative RT–PCR The strand-specific oligodeoxynucleotide primers, 5¢-UTR- rev2 and 3¢-UTRfor2, complementary to the synthesized plus- and minus-strand RNAs, were used for RT-catalyzed cDNA synthesis, followed by PCR amplification of cDNAs. Six microlitres of dissolved RNA was subjected to RT with M-MLV according to the manufacturer’s instructions (Promega). PCR was performed with 5¢-UTRfor and 5¢-UTRrev2 primers for detection of synthesized plus- strand RNA. For detection of synthesized minus-strand RNA, 3¢-UTRfor2 and 3¢-UTRrev primers were used. The synthesized RNA was real-time quantified using the TaqMan assay, according to McGoldrick et al. [23] and Cheng et al. [24], with some modifications. A set of 5¢-UTR- specific primers, P1 (forward) and P2 (reverse), were used to amplify a 222-bp fragment of the 5¢-UTR (nucleotides 95–316). The fluorogenic probe was designed with the following sequence: 5¢-TACAGGACAGTCGTCAGTAG TTCGACGTGA-3¢. RNA quantification was performed as follows: the precipitated RNA was reverse transcribed to cDNA using P 2 as a primer, then 2 lL of cDNA was used 4954 G H. Yi et al. (Eur. J. Biochem. 270) Ó FEBS 2003 as the template for amplification in a 20-lL volume containing 15 pmol of each primer, 4.5 m M MgCl 2 , 1 unit of Taq polymerase (MBI) and 3 pmol of TaqMan probe (Alpha). The PCR mixtures were placed in a thermocycler (Corbett) and subjected to 45 cycles of the following reaction parameters: denaturation at 95 °Cfor30s, annealing at 60 °C for 30 s and extension at 72 °Cfor45s. RNA-binding activity by North-Western blot The protein samples were resolved by SDS/PAGE (10% gel) and transferred to nitrocellulose membrane. Mem- brane-bound proteins were renatured in buffer A (10 m M Tris/HCL, 50 m M NaCl, 5% glycerol, 5 m M MgCl 2 ,0.1% Triton-X-100, pH 7.8) containing 5 mgÆmL )1 BSA and 1m M dithiothreitol, at room temperature for 4 h. After washing with buffer A, membranes were transferred to RNA binding buffer (10 m M Tris/HCl, 100 m M NaCl, 5% glycerol, 5 m M MgCl 2 ,0.2m M dithiothreitol, 0.1% Triton- X-100, 20 lgÆmL )1 tRNA, 0.5 mgÆmL )1 BSA, pH 7.8). One microlitre of DIG-labeled RNA was added and the membranes were incubated for 2 h with gentle shaking. Membranes were washed twice with RNA binding buffer, without dithiothreitol and tRNA, and then the bound RNA was treated with ALP-conjugated anti-DIG Ig (1 : 5000). RNA–protein complexes were visualized using Nitro Blue tetrazolium/5-bromo-4-chloroindol-2-yl phosphate. Results and discussion Expression and purification of bacterial recombinant NS5BD24 and its mutant, NS5BD24GAA To isolate the function of CSFV NS5B from other viral and cellular proteins and to establish an in vitro replication system for studying the initiation of viral RNA replica- tion, NS5B protein was expressed in an E. coli system. Earlier attempts to express and purify the full-length NS5B had been hampered as a result of the poor solubility. Sequence analysis and analysis of the hydro- pathy profile of CSFV NS5B revealed that there was a conserved hydrophobic domain at the C terminus of different CSFV strains (Fig. 1A). Prompted by reports that removal of the C terminal hydrophobic domain of the RdRps of hepatitis C virus [25–27], hepatitis G virus [28], and BVDV [29] could significantly improve the solubility of the protein expressed in E. coli,theC terminal 24 amino acids of NS5B were deleted and NS5BD24 was inserted into the pET-28a vector. To facilitate protein expression and purification, additional Met–Glu residues were introduced at the N terminus for initiating translation, and a polyhistidine epitope tag (GSHHHHHH) was introduced to the C terminal of NS5BD24 for affinity purification. The fusion protein, of  75 kDa, was obtained with an imidazole elution gradi- ent of 150 to 250 m M (Fig. 1B). The protein was identified as the recombinant NS5B by Western blot analysis using CSFV-infected pig serum as primary antibody (Fig. 1C). To maximize the amount of soluble protein, expression was performed at a low temperature (18 °C) and the cell pellet was resuspended in a nonionic detergent (1% Triton-X-100) in combination with a high concentration of salt (500 m M ) and glycerol (20%). We succeeded in recovering  2 mg of soluble protein from 1 L of E. coli culture. The other mutant protein, NS5BD24GAA, was expressed and purified in parallel to the NS5BD24 protein. The sufficient amounts of soluble protein thus obtained provided the basis for further studying the characteriza- tion of the enzyme and for the development of an in vitro replication system. Fig. 1. Expression and purification of classical swine fever virus (CSFV) NS5BD24 and NS5BD24GAA fusion proteins from Escheri- chia coli. (A) Hydropathy profile (Kyte and Doolittle) of CSFV NS5B. NS5B contains a highly hydrophobic region at the C terminus. (B) Proteins were expressed and purified as described in the Materials and methods. Fractions of sample eluted from the HisTrap affinity column by a concentration gradient of imidazole were separated by SDS/PAGE (12% gel) and stained with Coomassie Brilli- ant Blue. Lane M, molecular mass markers; lanes 1–4, eluted with 150 m M imidazole buf- fer; lanes 5–8, eluted with 250 m M imidazole buffer; lanes 9–11, eluted with 350 m M imi- dazole buffer. (C) Western blot analysis of the purified protein. Lane 1, NS5BD24 protein; lane 2, pET-28a vector as a negative control; lane 3, NS5BD24GAA protein. Ó FEBS 2003 De novo RNA synthesis by CSFV RdRp (Eur. J. Biochem. 270) 4955 NS5BD24 fusion protein possessed RdRp activity To determine whether the truncated NS5B (NS5BD24) protein could direct a viral-specific sequence for initiation of RNA synthesis in vitro,the3¢-end of plus-strand (+)3¢-UTR and minus-strand (–)IRES RNA transcripts were used as templates because they were believed to contain cis-acting promoters for initiating viral RNA synthesis. As T7 RNA polymerase was able to elongate self-complementary RNA templates by template-directed RNA synthesis during in vitro transcription [30], both (+)3¢-UTR and (–)IRES RNA transcripts were purified before being used as templates. RNA templates were loaded onto the same gel as the marker and were visualized by silver staining. The RNA products synthesized by CSFV NS5BD24 were separated by denaturing PAGE (8% gel) and detected using a Northern blot assay. As shown in Fig. 2 (lane 3), the purified NS5BD24 was able to synthesize either the plus- strand or minus-strand RNA products from the respective templates. The predominant RNA products migrated similarly to the respective RNA templates (373 nucleotides for synthesized plus-strand RNA and 228 nucleotides for synthesized minus-strand RNA). When (–)IRES was used as a template, a very small amount of high molecular weight RNA products was also observed. However, no RNA products were obtained in the absence of purified protein or RNA templates, indicating that purified NS5BD24 was not contaminated with T7 RNA polymerase or RNA/DNA that could serve as a template. Furthermore, mutation of the conserved motif, GDD, to GAA almost abolished RNA synthesis (Fig. 2, lane 4), similar to the reports for hepatitis C virus RdRp [31,32]. These results showed that purified NS5BD24 fusion protein, lacking the C terminal 24 amino acids, possessed RdRp activity in vitro and that the C terminal hydrophobic domain was not necessary for RdRp activity. Evidence of de novo RNA synthesis with either plus- or minus-strand viral RNA as template For the positive strand RNA viruses, the mechanisms of initiating viral RNA synthesis are rather different. Poliovi- rus has been shown to use a uridylylated protein as primer (protein-primed) to initiate RNA synthesis [33]. Phage Qb initiated de novo RNA synthesis [34], whereas rabbit hemorrhagic disease virus (RHDV) initiated RNA synthesis by using a template-primed copy-back mechanism [35]. Besides copy-back synthesis, Dengue virus RdRp was also demonstrated to be capable of de novo initiation of RNA synthesis [36]. Previous work has shown that crude extracts of recombinant CSFV NS5B expressed in insect cells could produce two RNA products using D-RNA (an mRNA of the liver-specific transcription factor DCoH) as template. One product, which was identical in size to the input RNA template, might have resulted from a TNTase activity, while the other was determined to be a double-stranded hairpin dimer RNA synthesized by a copy-back mechanism [11]. However, our work found that the predominant RNA products were template-sized, and thus the question was whether the template-sized RNA products were generated through de novo RNA synthesis. As the template-sized RNA could be detected by complementary probes, and RNA polymerization required all four ribonucleotides as substrates (data not shown), it seemed that the template- sized RNA products probably represented RNA synthes- ized by a de novo initiation mechanism but should not result from a terminal transferase activity. If RNA products were synthesized by TNTase activity, it would have the same polarity as the input RNA template [37]. To provide further evidence of de novo RNA synthesis, the 3¢-hydroxyl group of RNA templates were blocked by treatment with sodium periodate. The migration patterns of RNA products are shown in Fig. 3A, indicating that the template-sized RNA products were truly synthesized by de novo initiation, but not by the 3¢-end elongation copy-back synthesis. It should be pointed out that a very small amount of high molecular weight RNA products were still observed with 3¢-blocked (–)IRES as template, possibly because the polymerase used the nascent RNA as template for additional rounds of RNA synthesis. Furthermore, we performed RT using strand- specific oligodeoxynucleotides as a primer that could anneal only to the 3¢-terminus of either synthesized plus- or minus- strand RNA products, followed by PCR amplification. The amplified fragments were analyzed by agarose gel electro- phoresis and stained by ethidium bromide. As shown in Fig. 3B, the expected sizes of DNA fragments (373 nucleo- tides for new synthesized plus-strand and 228 nucleotides for the minus-strand) were observed, verifying that tem- plate-sized RNA products were initiated de novo from the 3¢-terminus of the template but not by premature termin- ation or internal initiation, as suggested in reports for tomato bushy stunt virus, cucumber necrosis virus [38] and hepatitis C virus [39]. Taken together, these results strongly suggest that the purified CSFV NS5BD24 could preferen- tially initiate either plus- or minus-strand viral RNA synthesis de novo in the absence of primers and viral or Fig. 2. RNA-dependent RNA polymerase (RdRp) activity of classical swine fever virus (CSFV) NS5BD24 protein. The RNA templates were purified before use. RNA products were separated by PAGE (8% gel containing 7 M urea) and detected by Northern blot. (A) The 228- nucleotide (+)3¢-UTR RNA transcripts were used as template. (B) RdRp assay with 373-nucleotide (–)IRES RNA transcripts as tem- plate. Lane 1, absence of RNA template; lane 2, absence of NS5BD24; lane 3, presence of NS5BD24 and RNA template; lane 4, presence of NS5BD24GAA and RNA template. The position of the input RNA template is shown as Ô-Õ. 4956 G H. Yi et al. (Eur. J. Biochem. 270) Ó FEBS 2003 host factors. This result is contrary to the previous reports for CSFV RdRp [11], as well as BVDV RdRp [40], another member of the Pestivirus genus, in which the major RNA products catalyzed by RdRp were shown to be a covalently linked double-stranded molecule generated by a copy-back mechanism. At present it is unknown whether this discrep- ancy is caused by the various viral enzyme preparations or different templates used. In fact, reports have shown that the RdRp of BVDV could preferentially initiate RNA synthesis by a de novo initiation mechanism with chemically synthesized short RNA (21 nucleotides) as a template [41], although a primer extension RNA product was also observed [42]. Therefore, de novo initiation of RNA synthesis might represent the preferred mechanism used by Pestivirus RdRps in vitro. Optimal conditions for de novo RNA synthesis using (–)IRES as template To optimize conditions for de novo RNA synthesis by the NS5BD24 protein, the 3¢-end of minus-strand RNA (–)IRES, was used as template and newly synthesized RNA products were quantified by real-time quantitative RT–PCR (TaqMan assay), which allowed quantification of the starting copies of the reaction, rather than the end products, by monitoring the increment of fluorescence released [43]. We first examined the effects of time and temperature on RNA synthesis. As shown in Fig. 4A, the synthesis of RNA products occured for at least 120 min, and the preferred temperature for the RdRp assay was  25 °C (Fig. 4B). The divalent cations Mg 2+ and/or Mn 2+ are known to be required for RdRp activity. Our results showed that both Mg 2+ and Mn 2+ supported de novo RNA synthesis; however, Mn 2+ (Fig. 4D) induced RdRp activity more efficiently than Mg 2+ (Fig. 4C). Maximum activity was observed at 5–7.5 m M Mn 2+ and 7.5–10 m M Mg 2+ .The high concentrations of Mn 2+ and Mg 2+ had negative impacts on RNA synthesis. Next, we determined the optimal concentrations of template and protein for RdRp assay. As shown in Fig. 4E, the synthesis of RNA products increased with increasing amounts of enzyme to 100 ng, excess amounts of enzyme inhibited the RdRp reaction. However, a high concentration of the RNA template had no significant inhibition on RNA synthesis (Fig. 4F), which differed from the previous report for hepatitis C virus RdRp using homologous RNA as template [44]. This might be a result of the various enzymes and template used in the RdRp reaction. Analysis of initiation of de novo RNA synthesis To examine the effect of preincubation with 0.5 m M NTP on the initiation of RNA synthesis, the RdRp mixtures were first preincubated with 0.5 m M of each NTP for 30 min, thenfurtherincubatedfor90minwith0.25m M NTP as substrates. Higher activities of de novo RNA synthesis were obtained when either (+)3¢-UTR or (–)IRES template was preincubated with 0.5 m M NTP, respectively, compared to the RdRp reaction without preincubation (Fig. 5A,B). When (+)3¢-UTR was used as template, preincubation with 0.5 m M GTP and ATP resulted in higher activities, whereas higher activities were observed following preincu- bation with 0.5 m M GTP and UTP using (–)IRES as template. Previous studies have shown that a 3¢-cytidylate in the template is preferred, by several viral RdRps, for de novo initiation of RNA synthesis [41,45]. To determine whether this is also the case for CSFV RdRp, we investigated the influence of substitution of the 3¢-terminal cytidylate with guanidylate, adenylate or uridylate, on de novo RNA Fig. 3. De novo initiation of viral RNA synthesis by classical swine fever virus (CSFV) RNA-dependent RNA polymerase (RdRp). Both viral plus- and minus-strand RNA templates were treated with sodium periodate to block the 3¢-OH group and then used as template for RdRp assay. (A) Northern blot assay with (+)3¢-UTR and (–)IRES as templates. Lane 1, RdRp assay with NS5BD24GAA as a control; lane 2, RdRp assay with NS5BD24 and 3¢-blocked RNA template. (B) Synthesized RNA was subjected to RT-PCR. RT was performed using a primer complementary to the newly synthesized minus-strand (lanes 2 and 3) or plus-strand RNA (lanes 4 and 5), followed by PCR amplification. Lane 1, 100 bp DNA ladder; lanes 2 and 4, RNA template (T) used as a control; lanes 3 and 5, RT–PCR results of newly synthesized products (P); lane 6, NS5BD24 protein as a control. PCR products were electrophoresed through an agarose gel and visualized by ethidium bromide staining. The expected fragments were 228 nucleotides (newly synthesized minus-strand RNA) and 373 nucleotides (synthesized plus-strand RNA) in length. Ó FEBS 2003 De novo RNA synthesis by CSFV RdRp (Eur. J. Biochem. 270) 4957 synthesis. As shown in Fig. 5C,D, changing the 3¢ terminal C, of both plus-strand and minus-strand RNA templates, to G, A or U dramatically decreased RNA synthesis, indica- ting that a 3¢-terminal cytidylate was necessary for efficient de novo RNA synthesis. Surprisingly, RNA polymerization was slightly higher when the 3¢-terminal C was replaced with U, rather than A or G, with (+)3¢-UTR as template (Fig. 5C). For (–)IRES as template, RNA synthesis was slightly higher when the 3¢-terminal C was replaced with G, rather than with A or U (Fig. 5D). This result was in Fig. 4. Effects of reaction conditions on de novo RNA synthesis with (–)IRES as tem- plate. The synthesized RNA was quantified by real-time RT–PCR. Effects of: (A) time, (B) temperature, (C) Mg 2+ concentration, (D) Mn 2+ concentration, (E) enzyme concentra- tion and (F) template concentration. Fig. 5. Analysis of initiation of d e nov o RNA synthesis. (A) and (B) The RNA-dependent RNA polymerase (RdRp) mixtures were first preincubated with 0.5 m M of each NTP for 30 min, then further incubated for 90 min with 0.25 m M of each NTP as substrates. (A) (+)3¢-UTR as template. (B) (–)IRES as template. Lane 1, no preincubation control; lane 2, preincubation with 0.5 m M GTP; lane 3, preincubation with 0.5 m M CTP; lane 4, preincubation with 0.5 m M ATP; lane 5, preincubation with 0.5 m M UTP. (C) and (D) Effect of substitution of the 3¢-terminal cytidylate with guanidylate, adenylate or uridylate on de novo RNA synthesis. (C) (+)3¢-UTR as template. (D) (–)IRES as template. Lane 1, normal template as control; lane 2, substitution of the 3¢ C with G; lane 3, substitution of the 3¢ C with U; lane 4, substitution of the 3¢ C with A. The relative activity compared with the control is shown below each lane as a percentage. 4958 G H. Yi et al. (Eur. J. Biochem. 270) Ó FEBS 2003 agreement with those of Reigadas et al.[45]forhepatitisC virus using (–)IRES as template, but differed from those of Kao et al. [41] for BVDV with synthesized short RNA as template, who showed that changing the 3¢-terminal C to G did not direct any product synthesis. This indicated that sequences and/or structures present in other parts of the template might play an important role for efficient de novo RNA synthesis by CSFV RdRp. Specific interaction between CSFV RdRp and viral RNA templates Viral RNA synthesis requires the initiation, recognition and specific binding between RdRp and template RNA. The 3¢-end of several positive strand RNA viruses is known to specifically bind to their respective RdRps [46–50]. To determine whether purified and refolded CSFV RdRp could specifically interact with either the plus- or minus-strand viral RNA template, North-Western blot assays were performed. tRNA was included in the binding reaction mixtures to avoid any nonspecific bind- ing. Figure 6A shows that CSFV RdRp was able to interact with the 3¢-end of viral minus-strand RNA template. No detectable complexes were observed when BSA was used as a control. The specificity of the RNA– protein interaction was confirmed by template competition assays in which unlabelled homologous or heterologous RNAs were preincubated with the protein for 10 min prior to addition of the DIG-labeled RNA template. The binding to either plus- or minus-strand RNA template showed a concentration response, because the band intensity decreased as the amount of unlabelled homolog- ous RNA (cold RNA) was increased. The addition of a 50-fold molar excess of unlabelled RNA almost abolished the interaction of protein with DIG-labeled RNA tem- plate (Fig. 6B,C, lane 5). However, the unlabelled hetero- logous RNA, like yeast tRNA, showed no reduction in the intensity of the RNA–protein complexes, even at concentrations as high as 50-fold molar excess (Fig. 6B, lane 6). These results showed that the CSFV RdRp was able to specifically interact with the 3¢ end of both plus- and minus-strand viral RNA templates, and that the C terminus of NS5B was not necessary for binding activity. At present, we do not know whether other viral or cellular proteins might interact with RNA templates or RdRp to form a replication complex for initiating CSFV RNA synthesis. In fact, the binding of cellular proteins to the 3¢-end of RNA templates has been described for some viruses [51,52]. Interaction of the viral proteins NS3 and NS5 has been reported in Japanese encephalitis virus [53], Dengue virus [54] and hepatitis C virus [55,56]. Nevertheless, our recombinant NS5BD24 expressed in E. coli has several properties that resemble the functional CSFV RdRp and permit an in vitro replication system (a) it possesses RNA polymerase activity with either plus- or minus-strand RNA as template, (b) it contains RNA binding activity and (c) it can initiate de novo RNA synthesis from viral RNA templates and does not require an exogenous primer. This in vitro replication system represents a starting point in the search for cis-elements at the 3¢ end of RNA templates and possibly viral or cell proteins required for the initiation of viral RNA synthesis. Acknowledgements This work was supported by National Basic Research Developmental Projects (G1999011900) and National Natural Science Foundation of China (30170214). References 1. Wengler, G., Bradley, D.W., Collett, M.S., Heinz, F.X., Schles- inger, R.W. & Strauss, J.H. (1995) Family Flaviviridae.InClas- sification and Nomenclature of Viruses: 6th Report of the Fig. 6. Specific interaction between classical swine fever virus (CSFV) RNA-dependent RNA polymerase (RdRp) and viral RNA templates by North- Western blot assays. Proteins were transferred to nitrocellulose membranes and renatured as described in the Materials and methods. (A) Binding activity of NS5BD24 with (–)IRES RNA template. Lane 1, purified NS5BD24 protein; lane 2, BSA control. Specificity of NS5BD24 interaction with (–)IRES RNA template (B) and (+)3¢-UTR template (C). Template competition assays were performed with homologous and heterologous RNAs (cold RNA) as competitors. Increasing amounts of unlabeled RNA were preincubated with NS5BD24 for 10 min prior to the addition of DIG- labeled RNA template. The diagram was processed using Adobe PHOTOSHOP . Lane 1, no competitor; lanes 2–5, competition with fivefold, 10-fold, 20-fold and 50-fold molar excesses of unlabeled homologous RNA, respectively; lane 6, competition with 50-fold heterlogous yeast tRNA. Ó FEBS 2003 De novo RNA synthesis by CSFV RdRp (Eur. J. Biochem. 270) 4959 International Committee on Taxonomy of Viruses (Murphy, F.A., Fauquet, C.M., Bishop, D.H.L., Ghabrial, S.A., Jarvis, A.W., Martelli, G.P., Mayo, M.A. & Summers, M.D., eds), pp. 415–427. Springer-Verlag, Berlin. 2. Thiel, H.J., Plagemann, P.G.W. & Moennig, V. (1996) Pesti- viruses. In Fundamental Virology, 3rd edn. (Fields, B.N., Knipe, D.M. & Howley, P.M., eds), pp. 1059–1073. Raven Press, New York. 3. Meyers, G. & Thiel, H.J. (1996) Molecular characterization of Pestiviruses. Adv. Virus Res. 47, 53–118. 4. Rijnbrand, R., van der Straaten, T., van Rijn, P.A., Spaan, W.J., Bredenbeek, P.J. (1997) Internal entry of ribosomes is directed by the 5¢ noncoding region of classical swine fever virus and is dependent on the presence of an RNA pseudoknot upstream of the initiation codon. J. Virol. 71, 451–457. 5. Fletcher, S.P. & Jackson, R.J. (2002) Pestivirus internal ribosome entry site (IRES) structure and function: elements in the 5¢ untranslated region important for IRES function. J. Virol. 76, 5024–5033. 6. Yu, H., Grassmann, C.W. & Behrens, S.E. (1999) Sequence and structural elements at the 3¢ terminus of bovine viral diarrhea virus genomic RNA: functional role during RNA replication. J. Virol. 73, 3638–3648. 7. Rice, C.M. (1996) Flaviviridae: The viruses and their replication. In Fields Virology (Fields, B.N., Knipe, D.M. & Howley, P.M., eds), pp. 931–960. Lippincott-Raven Publishers, Philadelphia. 8. Tratschin, J.D., Moser, C., Ruggli, N. & Hofmann, M.A. (1998) Classical swine fever virus leader proteinase N pro is not required for viral replication in cell culture. J. Virol. 72, 7681–7684. 9. Moser, C., Stettler, P., Tratschin, J.D. & Hofmann, M.A. (1999) Cytopathogenic and noncytopathogenic RNA replicons of clas- sical swine fever virus. J. Virol. 73, 7787–7794. 10. Koonin, E.V. (1991) The phylogeny of RNA-dependent RNA polymerase of positive-strand RNA viruses. J. Gen. Virol. 72, 2197–2206. 11. Steffens, S., Thiel, H.J. & Behrens, S.E. (1999) The RNA-depen- dent RNA polymerases of different members of the family Flavi- viridae exhibit similar properties in vitro. J. Gen. Virol. 80, 2583– 2590. 12. Lohmann, V., Overton, H. & Bartenschlager, R. (1999) Selective stimulation of hepatitis C virus and pestivirus NS5B RNA poly- merase activity by GTP. J. Biol. Chem. 274, 10807–10815. 13. Xiao, M., Zhang, C.Y., Pan, Z.S., Wu, H.X. & Guo, J.Q. (2002) Classical swine fever virus NS5B-GFP fusion protein possesses an RNA-dependent RNA polymerase activity. Arch. Virol. 147, 1779–1787. 14. Meyers, G., Thiel, H.J. & Rumenapf, T. (1996) Classical swine fever virus: recovery of infectious viruses from cDNA constructs and generation of recombinant cytopathogenic defective interfer- ing particles. J. Virol. 70, 1588–1595. 15. Moormann, R.J.M., van Gennip, H.G.P., Miedema, G.K.W., Hulst, M.M. & van Rijn, P.A. (1996) Infectious RNA transcribed from an engineered full-length cDNA template of the genome of a pestivirus. J. Virol. 70, 763–770. 16. Ruggli, N., Tratschin, J.D., Mittelholzer, C. & Hofmann, M.A. (1996) Nucleotide sequence of classical swine fever virus strain Alfort/187 and transcription of infectious RNA from stably cloned full-length cDNA. J. Virol. 70, 3478–3487. 17. van Gennip, H.G., van Rijn, P.A., Widjojoatmodjo, M.N. & Moormann, R.J.M. (1999) Recovery of infectious classical swine fever virus (CSFV) full-length genomic cDNA clones by a swine kidney cell line express bacteriophage T7 RNA polymerase. J. Virol. Methods 78, 117–128. 18. Wu, H., Zhang, C., Zheng, C., Wang, J., Pan, Z., Li, L., Cao, S. & Yi, G. (2003) Construction of cytopathic PK-15 cell model of classical swine fever virus. Chin. Sci. Bull. 48, 887–891. 19. Widjojoatmodjo, M.N., van Gennip, H.G.P., Bouma, A., van Rijn, P.A. & Moormann, R.J.M. (2000) Classical swine fever virus E rns deletion mutants: trans-complementation and potential use as nontransmissible. J. Virol. 74, 2973–2980. 20. Behrens, S.E., Tomei, L. & De Francesco, R. (1996) Identification and properties of the RNA-dependent RNA polymerase of hepatitis C virus. EMBO J. 15, 12–22. 21. Lohmann, V., Korner, F., Herian, U. & Bartenschlager, R. (1997) Biochemical properties of hepatitis C virus NS5B RNA-dependent RNA polymerase and identification of amino acid sequence motifs essential for enzymatic activity. J. Virol. 71, 8416–8428. 22. De Francesco, R., Behrens, S.E., Tomei, L., Altamura, S. & Jir- icny, J. (1996) RNA-dependent RNA polymerase of hepatitis C virus. Methods Enzymol. 275, 58–67. 23. McGoldrick,A.,Lowings,J.P.,Ibata,G.,Sands,J.J.,Belak,S.& Paton, D.J. (1998) A novel approach to the detection of classical swine fever virus by RT-PCR with a fluorogenic probe (TaqMan). J. Virol. Methods 72, 125–135. 24. Chen,Y.,Zhang,C.,Zhou,J.,Pan,Z.,Chen,L.,Li,T.&Guo,C. (2003) Development of a fluorogenic quantitative PCR assay for rapid quantification of hog cholera lapinized virus. Virologica Sin. 18, 124–128. 25. Yamashita,T.,Kaneko,S.,Shirota,Y.,Qin,W.,Nomura,T., Kobayashi, K. & Murakami, S. (1998) RNA-dependent RNA polymerase activity of the soluble recombinant hepatitis C virus NS5B protein truncated at the C-terminal region. J. Biol. Chem. 273, 15479–15486. 26. Ferrari, E., Wright-Minogue, J., Fang, J.W.S., Baroudy, B.M., Lau, J.Y.N. & Hong, Z. (1999) Characterization of soluble hepatitis C virus RNA-dependent RNA polymerase expressed in Escherichia coli. J. Virol. 73, 1649–1654. 27. Tomei, L., Vitale, R.L., Incitti, I., Serafini, S., Altamura, S., Vitelli, A. & De Francesco, R. (2000) Biochemical characterization of a hepatitis C virus RNA-dependent RNA polymerase mutant lacking the C-terminal hydrophobic sequence. J. Gen. Virol. 81, 759–767. 28. Zhong, W., Ingravallo, P., Wright-Minogue, J., Uss, A.S., Skel- ton, A., Ferrari, E., Lau, J.Y. & Hong, Z. (2000) RNA-dependent RNA polymerase activity encoded by GB virus-B non-structural protein 5B. J. Viral Hepat. 7, 335–342. 29. Lai, V.C.H., Kao, C.C., Ferrari, E., Park, J., Uss, A.S., Wright- Minogue, J., Hong, Z. & Lau, J.Y. (1999) Mutational analysis of bovine viral diarrhea virus RNA-dependent RNA polymerase. J. Virol. 73, 10129–10136. 30. Triana-Alonso, F.J., Dabrowski, M., Wadzack, J. & Nierhaus, K.H. (1995) Self-coded 3¢-extension of runoff transcripts produces aberrant products during in vitro transcription with T7 RNA polymerase. J. Biol. Chem. 270, 6298–6307. 31. Ranjith-Kumar, C.T., Gajewski, J., Gutshall, L., Maley, D., Sarisky, R.T. & Kao, C.C. (2001) Terminal nucleotidyl transferase activity of recombinant Flaviviridae RNA-dependent RNA poly- merases: implication for viral RNA synthesis. J. Virol. 75, 8615– 8623. 32. Wang, Q.M., Hockman, M.A., Staschke, K., Johnson, R.B., Case, K.A., Lu, J., Parsons, S., Zhang, F., Rathnachalam, R., Kirk- egaard, K. & Colacino, J.M. (2002) Oligomerization and cooperative RNA synthesis activity of hepatitis C virus RNA- dependent RNA polymerase. J. Virol. 76, 3865–3872. 33. Paul, A.V., van Boom, J.H., Filippov, D. & Wimmer, E. (1998) Protein-primed RNA synthesis by purified poliovirus RNA polymerase. Nature 393, 280–284. 34. Biebricher, C.K., Eigen, M. & Luce, R. (1981) Product analysis of RNA generated de novo by Qb replicase. J. Mol. Biol. 148, 369–390. 35. Lopez Vazquez, A.L., Martin Alonso, J.M. & Parra, F. (2001) Characterisation of the RNA-dependent RNA polymerase from 4960 G H. Yi et al. (Eur. J. Biochem. 270) Ó FEBS 2003 Rabbit hemorrhagic disease virus produced in Escherichia coli. Arch. Virol. 146, 59–69. 36. Ackermann, M. & Padmanabhan, R. (2001) De novo synthesis of RNA by the Dengue virus RNA-dependent RNA polymerase exhibits temperature dependence at the initiation but not elonga- tion phase. J. Biol. Chem. 276, 39926–39937. 37. Luo, G., Hamatake, R.K., Mathis, D.M., Racela, J., Rigat, K.L., Lemm, J. & Colonno, R.J. (2000) De novo initiation of RNA synthesis by the RNA-dependent RNA polymerase (NS5B) of hepatitis C virus. J. Virol. 74, 851–863. 38. Nagy, P.D. & Pogany, J. (2000) Partial purification and char- acterization of Cucumber necrosis virus and Tomato bushy stunt virus RNA-dependent RNA polymerases: similarilties and differ- ences in template usage between tombusvirus and carmovirus RNA-dependent RNA polymerases. Virology 276, 279–288. 39. Pellerin, C., Lefebvre, S., Little, M.J., Mckercher, G., Lamarre, D. & Kukolj, G. (2002) Internal initiation sites of de novo RNA synthesis within the hepatitis C virus polypyrimidine tract. Biochem. Biophys. Res. Commun. 295, 682–688. 40. Zhong, W., Gutshall, L.L. & Del Vecchio, A.M. (1998) Identifi- cation and characterization of an RNA-dependent RNA poly- merase activity within the nonstructural protein 5B region of bovine viral diarrhea virus. J. Virol. 72, 9365–9369. 41. Kao, C.C., del Vecchio, A.M. & Zhong, W. (1999) De novo initiation of RNA synthesis by a recombinant flavivirdae RNA- dependent RNA polymerase. Virology 253,1–7. 42. Ranjith-Kumar, C.T., Kim, Y., Gutshall, L., Silverman, C., Khandekar, S., Sarisky, R.T. & Kao, C.C. (2002) Mechanism of de novo initiation by the hepatitis C virus RNA-dependent RNA polymerase:roleofdivalentmetals.J. Virol. 76, 12513–12525. 43. Heid, C.A., Stevens, J., Livak, K.J. & Williams, P.M. (1996) Real time quantitative PCR. Genome Res. 6, 986–994. 44. Sun, X.L., Johnson, R.B., Hockman, M.A. & Wang, Q.M. (2000) De novo RNA synthesis catalyzed by HCV RNA-dependent RNA polymerase. Biochem. Biophys. Res. Commun. 268, 798–803. 45. Reigadas, S., Ventura, M., Sarih-Cottin, L., Castroviejo, M., Litvak, S. & Astier-Gin, T. (2001) HCV RNA-dependent RNA polymerase replicates in vitro the 3¢ terminal region of the minus- strand viral RNA more efficiently than the 3¢ terminal region of the plus RNA. Eur. J. Biochem. 268, 5857–5867. 46. Cui, T., Sankar, S. & Porter, A.G. (1993) Binding of encep- helomyocarditis virus RNA polymerase to the 3¢ noncoding region of the viral RNA is specific and requires the 3¢ poly A tail. J. Biol. Chem. 268, 26093–26098. 47. Todd, S., Nguyen, J.H.C. & Semler, B.L. (1995) RNA–protein interactions directed by the 3¢ end of human rhinovirus genomic RNA. J. Virol. 69, 3605–3614. 48. Pata, J.D., Schultz, S.C. & Kirkegaard, K. (1995) Functional oligomerization of poliovirus RNA-dependent RNA polymerase. RNA 1, 466–477. 49. Cheng, J.C., Chang, M.F. & Chang, S.C. (1999) Specific interac- tion between the hepatitis C virus NS5B RNA polymerase and the 3¢ end of the viral RNA. J. Virol. 73, 7044–7049. 50. Agrawal, S., Gupta, D. & Panda, S.K. (2001) The 3¢ end of hepatitis E virus (HEV) genome binds specifically to the viral RNA-dependent RNA polymerase (RdRp). Virology 282, 87–101. 51. Yocupicio-Monroy,R.M.,Medina,F.,Reyes-delValle,J.&del Angel, R.M. (2003) Cellular proteins from human monocytes bind to dengue 4 virus minus-strand 3¢untranslated region RNA. J. Virol. 77, 3067–3076. 52. Luo, G. (1999) Cellular proteins bind to the poly (U) tract of the 3¢ untranslated region of hepatitis C virus RNA genome. Virology 256, 105–118. 53. Chen, C.J., Kuo, M.D., Chien, L.J., Hsu, S.L., Wang, Y.M. & Lin, J.H. (1997) RNA–protein interactions: involvement of NS3, NS5, and 3¢noncoding regions of Japanese encephalitis virus genomic RNA. J. Virol. 71, 3466–3473. 54. Kapoor, M., Zhang, L., Ramachandra, M., Kusukawa, J., Ebner, K.E. & Padmanabhan, R. (1995) Association between NS3 and NS5proteinsofdenguevirustype2intheputativeRNAreplicase is linked to differential phosphorylation of NS5. J. Biol. Chem. 270, 19100–19106. 55. Ishido, S., Fujita, T. & Hotta, H. (1998) Complex formation of NS5B with NS3 and NS4A proteins of hepatitis C virus. Biochem. Biophys. Res. Commun. 244, 35–40. 56. Piccininni, S., Varaklioti, A., Nardelli, M., Dave, B., Raney, K.D. & McCarthy, J.E. (2002) Modulation of the hepatitis C virus RNA-dependent RNA polymerase activity by the non-structural (NS)3helicaseandtheNS4Bmembraneprotein.J. Biol. Chem. 277, 45670–45679. Ó FEBS 2003 De novo RNA synthesis by CSFV RdRp (Eur. J. Biochem. 270) 4961 . ATATACGAGGTTAGTTCATTC 5¢-UTRrev TAATACGACTCACTATAGGGTGCCATGAA CAG 5¢-UTRrev2 GTGCCATGAACAGCAGAGATTTTTATAC P1 TAGCCATGCCCATAGTAGG P2 ATCAGGTCGTACTCCCATCAC 3¢-UTRfor TAATACGACTCACTATAGCGCGGGTAAC 3¢-UTRfor2. GCGCGGGTAACCCGGGATCTGAA 3¢-UTRrev GGGCCGTTAGGAAATTACCTTAGTC 3¢-UTRrev1 CGGCCGTTAGGAAATTACCTTAGTC 3¢-UTRrev2 TGGCCGTTAGGAAATTACCTTAGTC 3¢-UTRrev3 AGGCCGTTAGGAAATTACCTTAGTC Ó FEBS 2003 De novo RNA synthesis. 3¢-terminal cytidylate with guanidylate, adenylate or uridylate, on de novo RNA Fig. 3. De novo initiation of viral RNA synthesis by classical swine fever virus (CSFV) RNA- dependent RNA polymerase