Functional properties of a monoclonal antibody inhibiting the hepatitis C virus RNA-dependent RNA polymerase.

The hepatitis C virus (HCV) RNA-dependent RNA polymerase (RdRp), represented by nonstructural protein 5B (NS5B), has recently emerged as a promising target for antiviral intervention. Here, we describe the isolation, functional characterization, and molecular cloning of a monoclonal antibody (mAb) inhibiting the HCV RdRp. This mAb, designated 5B-12B7, binds with high affinity to a conformational epitope in the palm subdomain of the HCV RdRp and recognizes native NS5B expressed in the context of the entire HCV polyprotein or subgenomic replicons. Complete inhibition of RdRp activity in vitro was observed at equimolar concentrations of NS5B and mAb 5B-12B7, whereas RdRp activities of classical swine fever virus NS5B and poliovirus 3D polymerase were not affected. mAb 5B-12B7 selectively inhibited NTP binding to HCV NS5B, whereas binding of template RNA was unaffected, thus explaining the mechanism of action at the molecular level. The mAb 5B-12B7 heavy and light chain variable domains were cloned by reverse transcription-PCR, and a single chain Fv fragment was assembled for expression in Escherichia coli and in eukaryotic cells. The mAb 5B-12B7 single chain Fv fragment bound to NS5B both in vitro and in transfected human cell lines and therefore may be potentially useful for intracellular immunization against HCV. More important, detailed knowledge of the mAb 5B-12B7 contact sites on the enzyme may facilitate the development of small molecule RdRp inhibitors as novel antiviral agents.

The hepatitis C virus (HCV) RNA-dependent RNA polymerase (RdRp), represented by nonstructural protein 5B (NS5B), has recently emerged as a promising target for antiviral intervention. Here, we describe the isolation, functional characterization, and molecular cloning of a monoclonal antibody (mAb) inhibiting the HCV RdRp. This mAb, designated 5B-12B7, binds with high affinity to a conformational epitope in the palm subdomain of the HCV RdRp and recognizes native NS5B expressed in the context of the entire HCV polyprotein or subgenomic replicons. Complete inhibition of RdRp activity in vitro was observed at equimolar concentrations of NS5B and mAb 5B-12B7, whereas RdRp activities of classical swine fever virus NS5B and poliovirus 3D polymerase were not affected. mAb 5B-12B7 selectively inhibited NTP binding to HCV NS5B, whereas binding of template RNA was unaffected, thus explaining the mechanism of action at the molecular level. The mAb 5B-12B7 heavy and light chain variable domains were cloned by reverse transcription-PCR, and a single chain Fv fragment was assembled for expression in Escherichia coli and in eukaryotic cells. The mAb 5B-12B7 single chain Fv fragment bound to NS5B both in vitro and in transfected human cell lines and therefore may be potentially useful for intracellular immunization against HCV. More important, detailed knowledge of the mAb 5B-12B7 contact sites on the enzyme may facilitate the development of small molecule RdRp inhibitors as novel antiviral agents.
The hepatitis C virus (HCV) 1 is a major cause of chronic hepatitis, liver cirrhosis, and hepatocellular carcinoma worldwide (1). A protective vaccine does not exist to date, and therapeutic options are limited (2,3). HCV contains a single-strand RNA genome of positive polarity and ϳ9600 nucleotides in length that encodes a polyprotein precursor of ϳ3000 amino acids (aa) (see Refs. 4 and 5 for recent reviews). The polyprotein precursor is co-and post-translationally processed by cellular and viral proteases to yield the mature structural and nonstructural proteins. HCV RNA replication proceeds via synthesis of a complementary (Ϫ)-strand RNA using the genome as a template and the subsequent synthesis of genomic RNA from this (Ϫ)-strand template. The key enzyme responsible for both of these steps is the virally encoded RNA-dependent RNA polymerase (RdRp), represented by nonstructural protein 5B (NS5B).
The HCV RdRp has been shown to be essential for viral replication in vitro (6) and in vivo (7,8), and it has been extensively characterized both at the biochemical (9 -12) and structural (13)(14)(15)(16) levels. HCV NS5B contains motifs shared by all RdRps and possesses the classical fingers, palm, and thumb subdomains. As a unique feature of the HCV RdRp, extensive interactions between the fingers and thumb subdomains result in a completely encircled active site. The HCV RdRp has emerged as a promising target for antiviral drug development. In this context, it has recently been validated as an antiviral target in the related pestiviruses (17).
In this study, we describe the isolation, functional characterization, and molecular cloning of a mAb that specifically and efficiently inhibits the HCV RdRp. The mechanism of enzyme inhibition was elucidated at the molecular level. Hence, this mAb may serve as a unique molecular probe for future mechanistic studies toward the elucidation of the HCV RdRp reaction pathway and may provide a new framework for the development of small molecule RdRp inhibitors as novel antiviral agents.

EXPERIMENTAL PROCEDURES
Production and Purification of RdRps-Recombinant HCV and classical swine fever virus (CSFV; kindly provided by Jon-Duri Tratschin, Institute of Virology and Immunoprophylaxis, Mittelhä usern, Switzerland) NS5B were produced in a recombinant baculovirus system and purified by affinity chromatography as described (10,18). Purified poliovirus 3D polymerase was a gift of Eckard Wimmer and Aniko Paul (State University of New York, Stony Brook, NY).
Establishment of mAbs-Eight-week-old female BALB/c mice were immunized with recombinant HCV NS5B in its functionally active, native conformation. Spleen cells from immunized mice were fused with the X63-Ag8.653 myeloma cell line (American Type Culture Collection, Manassas, VA). Hybridomas were selected, and supernatants were screened by ELISA essentially as described (19). Hybridomas immunoreactive with recombinant and cellularly expressed NS5B protein were cloned twice by limiting dilution. mAb isotypes were determined with reagents from Amersham Biosciences, Inc. Ascites fluid was produced from hybridomas 5B-3B1.5.3 and 5B-12B7.54.1 (Eurogentec, Seraing, Belgium). mAbs were purified from ascites fluid by protein G affinity chromatography. For in vitro RdRp inhibition assays, mAbs were dia-lyzed for 16 h at 4°C against buffer containing 100 mM NaCl and 10 mM Tris-HCl (pH 7.2). Biotinylation was performed using the FluoReporter Biotin-XX labeling kit (Molecular Probes, Inc., Eugene, OR). Reactivity of biotinylated mAbs was revealed with horseradish peroxidase-conjugated streptavidin (Molecular Probes, Inc.).
Indirect Immunofluorescence Microscopy and Western Blotting-Indirect immunofluorescence microscopy and Western blotting were performed as previously described (22).
In Vitro Transcription-Translation-The TNT T7 coupled reticulocyte lysate system (Promega, Madison, WI) was used essentially following the manufacturer's recommendations. Reactions were routinely performed for 90 min at 30°C in the presence of 0.8 mCi/ml [ 35 S]methionine (Amersham Biosciences, Inc., Buckinghamshire, United Kingdom).
Immunoprecipitation-Cells were lysed in buffer containing 150 mM NaCl, 50 mM Tris-HCl (pH 8.0), 1% Nonidet P-40, and protease inhibitors (Complete ® protease inhibitor mixture, Roche Molecular Biochemicals, Mannheim, Germany). Lysates were precleared for 4 h at 4°C with protein G-Sepharose (Amersham Biosciences, Inc.), followed by immunoprecipitation with 2 l of 5B-12B7 ascites fluid or 5 g of anti-FLAG mAb M2 (Sigma)/500 l for 16 h at 4°C. Immunocomplexes were collected by the addition of protein G-Sepharose for 2 h at 4°C. Finally, beads were washed with the lysis buffer, resuspended in 2ϫ sample loading buffer, boiled for 5 min, and analyzed by SDS-PAGE. In some experiments, cells were homogenized in a Dounce homogenizer in hypotonic buffer containing 10 mM Tris-HCl (pH 7.5) and 2 mM MgCl 2 , and immunoprecipitation was performed from post-nuclear supernatants as described above.
Epitope Mapping-The mAb 5B-3B1 epitope was mapped by random DNase I fragment expression library screening using the NovaTope system (Novagen, Madison, WI).
In Vitro RdRp Assays-Polymerase assays contained 500 ng of HCVspecific in vitro transcript corresponding to a functional replicon RNA (6); 200 ng of HCV NS5B, 400 ng of CSFV NS5B, or 20 ng of poliovirus 3D polymerase; various concentrations of mAbs as specified under "Results"; 1 Ci of [␣-32 P]CTP adjusted to a 10 M final concentration; a 500 M concentration of each of the remaining NTPs; 10 units of RNasin (Promega); and assay buffer (5 mM dithiothreitol, 20 mM Tris-HCl (pH 7.5), 12.5 mM MgCl 2 , 10 mM KCl, and 1 mM EDTA) in a total volume of 25 l. After a 5-min incubation of the enzyme with the mAb at 4°C, the RdRp reaction was initiated by the addition of NTPs and RNA, followed by a 1-h incubation at 27°C. The reaction was terminated by the addition of 1 ml of 10% trichloroacetic acid and 0.5% tetrasodium pyrophosphate and 100 g of salmon sperm DNA. After a 30-min incubation at 4°C, samples were filtered through glass microfiber GF/C filters (Whatman, Kent, United Kingdom). Filters were washed five times with 1% trichloroacetic acid and 0.1% tetrasodium pyrophosphate, and bound radioactivity was measured after the addition of Rotiszint 2200 (Roth, Karlsruhe, Germany) in a liquid scintillation counter (Beckman Instruments). Background values obtained with an inactive RdRp mutant were subtracted.
Determination of Equilibrium Dissociation Constants (K d )-HCV NS5B protein lacking the carboxyl-terminal 21 aa (NS5B⌬C21) was expressed in Escherichia coli and purified to homogeneity as described previously (12). This recombinant protein has nine tryptophan residues and thus has strong protein fluorescence at 330 nm when it is excited at 283 nm. This property was utilized to determine the equilibrium dissociation constant (K d ) of NS5B for a nucleotide (GTP, CTP, ATP, or UTP) or an 8-mer RNA template (GR-1, 5Ј-AGAGAGCC-3Ј) by following the quenching of intrinsic protein fluorescence. Measurements were performed at 23°C on a PTI fluorescence spectrophotometer (Photon Technology International, Lawrenceville, NJ). The excitation wavelength was set at 283 nm, and the emission wavelength was set at 330 nm. The binding buffer included 20 mM Tris-HCl (pH 7.5), 10 where I o is the fluorescence of free protein and I total is the fluorescence loss due to the full occupation of all the binding sites by a ligand or

5Ј-GCAGGATCCACCATGGCCAAGAACGAGGTTTTCTGCG-3Ј
NS5B299fwd EcoRI substrate. Nonlinear least-squares fit of the data was performed using Kaleidagraph (Synergy Software, Reading, PA). In the experiments, a control sample containing fluorescence intensity similar to that of tryptophan was titrated in the same way as the protein sample. The fluorescence loss due to the addition of free ligand was determined. The percentage of fluorescence loss was used to compensate for the measured protein fluorescence intensity in determining the real I value. To study the effect of mAb 5B-12B7 on the NS5B/substrate (RNA/NTP) interaction, the mAb (54 nM; the NS5B-binding site concentration is 108 nM) was mixed with NS5B (53 nM) before the addition of a substrate. The K d of NS5B for the substrate in the presence of mAb 5B-12B7 was calculated as described above.
Cloning of mAb Variable Domains and Assembly of Single Chain Variable Domain Fragment (scFv) Constructs-Total cellular RNA was extracted from early passage 5B-12B7.54.1 hybridoma cells using RNAzol (Biotecx Laboratories, Houston, TX). First-strand cDNA synthesis with an oligo(dT) primer was performed using the first-strand cDNA synthesis kit (Amersham Biosciences, Inc.). The heavy chain variable domain (VH) was amplified by PCR using the degenerate primers VH-BACK and VH-FOR (Table I) with three initial cycles of 2 min of denaturation at 95°C, 2 min of annealing at 42°C, and 1 of min extension at 74°C, followed by 28 cycles of 2 min of denaturation at 95°C, 2 min of annealing at 56°C, and 1 min of extension at 74°C. The light () chain variable domain (VK) was amplified under the same conditions using the primer pair VK-BACK and VK-FOR (Table I). The VH amplification product was digested with HindIII and BstEII and ligated into the HindIII-BstEII sites of pWW152 (23,24) to yield plasmid pWW12B7VH. The VK amplification product was digested with PvuII and XbaI and ligated into the PvuII-XbaI sites of pWW152 to yield plasmid pWW12B7VK. The pWW152 vector contains HindIII and BstEII sites for the subcloning of murine VH cDNA fragments, followed by a synthetic sequence encoding the 15-aa linker (GGGGS) 3 , and PvuII and XbaI sites for the subcloning of murine VK cDNA fragments. Eight clones each of the VH and VK domains were sequenced. Subsequently, the HindIII-BstEII fragment of pWW12B7VH-6 was ligated into the HindIII-BstEII sites of pWW12B7VK-2 to yield the scFv construct pWW12B7.
For bacterial expression, the 5B-12B7 scFv sequence was isolated from pWW12B7 as a HindIII-XbaI fragment and fused in frame to the E. coli phoA alkaline phosphatase gene in plasmid pSW602, which is derived from the expression vector pFLAG-1 (IBI Biochemicals, New Haven, CT). The resulting pSW602-12B7 construct encodes a fusion protein consisting of the E. coli OmpA signal peptide, a FLAG tag, a hexahistidine tag, the 5B-12B7 scFv sequence, and E. coli PhoA under the control of an isopropyl-␤-D-thiogalactopyranoside-inducible tac promoter (23). Plasmid pSW602-12B7 was transformed into the phoAnegative E. coli strain CC118 (25). Bacteria were grown in LB medium to A 600 ϭ 0.7 at 37°C before induction with 0.5 mM isopropyl-␤-Dthiogalactopyranoside for 90 min. Subsequently, bacteria were harvested by centrifugation, and periplasmic extracts were prepared following the manufacturer's recommendations (IBI Biochemicals).
For expression of scFv in mammalian cells, the cytomegalovirus promoter-driven expression construct pCMV12B7FLAG with a carboxylterminal FLAG tag was generated by PCR amplification of 5B-12B7 scFv from pWW12B7 using primers 12B7VH6fwd and 12B7VK2FLAGrev (Table I), followed by digestion of the amplification product with EcoRI and XbaI and ligation into the EcoRI-XbaI sites of pcDNA3.1. As a nonrelevant control, the c-Myc-specific 9E10 scFv construct was amplified from pWW152-9E10 2 using primers 9E10fwd and 9E10FLAGrev (Table  I), followed by digestion of the amplification product with EcoRI and XbaI and ligation into the EcoRI-XbaI sites of pcDNA3.1 to yield plasmid pCMV9E10FLAG. In addition, the HCV nonstructural protein 4A-specific E6 scFv construct (a gift of Cinzia Traboni, Istituto di Ricerche di Biologia Molecolare P. Angeletti, Pomezia, Italy) was ligated into the XbaI site of pcDNA3.1/Zeo (Invitrogen) in the correct or reverse orientation, yielding constructs pCMVZeoSCFVE6FLAG and pCMVZeoSCFVE6FLAGrev, respectively.

RESULTS
Characteristics of mAbs-Recombinant HCV NS5B protein was used in its functionally active, native conformation as an antigen to raise murine mAbs. Screening of ϳ500 hybridomas resulting from two separate fusions allowed the isolation and cloning of six NS5B-specific mAbs, designated 5B-2A5, 5B-2A7, 5B-3B1, 5B-3F3, 5B-4C1, and 5B-12B7. Characteristics of these mAbs are summarized in Table II. mAbs 5B-3B1 and 5B-12B7 were further investigated. mAb 5B-3B1 reacted strongly in ELISA and immunoblotting (Fig. 1B), suggesting that it recognizes a linear epitope. By contrast, mAb 5B-12B7 functioned well in immunofluorescence (Fig. 1A) and immunoprecipitation (Fig. 1B) assays, but not in immunoblot assays, indicating a conformation-sensitive nature of the recognized epitope.
A representative immunofluorescence analysis of full-length NS5B and of a carboxyl-terminally truncated protein representing NS5B aa 1-392 (NS5B392) is shown in Fig. 1A. As described previously (20), the full-length protein was found in a staining pattern characteristic for the endoplasmic reticulum. By contrast, the NS5B392 construct, which lacks the NS5B membrane insertion sequence, showed diffuse cytoplasmic and nuclear staining.
Immunoprecipitation and Western blot experiments using mAbs 5B-12B7 and 5B-3B1 are shown in Fig. 1B. To perform such experiments with two mAbs of murine origin, mAb 5B-3B1 was biotinylated, and its reactivity was revealed with horseradish peroxidase-conjugated streptavidin. mAb 5B-12B7 efficiently immunoprecipitated NS5B from tetracycline-regulated cell lines inducibly expressing NS5B in the context of the entire HCV polyprotein. Moreover, NS5B expressed individually either as a full-length molecule or as a carboxyl-terminal truncation was immunoprecipitated efficiently. Most important, this mAb recognized and efficiently bound to NS5B in the functional context of the HCV replication complex present in 9-13 cells harboring selectable subgenomic replicons, even when such cells were lysed without the use of detergents. This suggests the interesting possibility of using mAb 5B-12B7 to isolate and study NS5B under native conditions in the context of a functionally active replication complex.
Competitive inhibition experiments were performed to explore whether the mAbs recognized distinct or closely related NS5B epitopes. As shown in Fig. 1C, binding of biotinylated mAbs 5B-3B1 and 5B-12B7 to recombinant NS5B was not affected by excess non-biotinylated mAbs 5B-12B7 and 5B-3B1, respectively. These results clearly demonstrate that these mAbs recognize distinct epitopes on HCV RdRp. In addition, none of the additional four NS5B-specific mAbs listed in Table  II were found to compete with binding of mAb 5B-3B1 or 5B-12B7 to NS5B (data not shown).
Epitope Mapping-Strong reactivity in immunoblot assays suggested that mAb 5B-3B1 recognizes a linear epitope. A random DNase I fragment expression library screening strategy was therefore performed to map this epitope. Three overlapping cDNA clones reactive with mAb 5B-3B1, designated 6-1-1, 8-1-1, and 15-3-2, resulted from this approach. Alignment of the amino acid sequences encoded by these cDNA clones allowed us to map the minimal epitope to NS5B aa 372-382 ( Fig. 2A). This epitope is located at the palm-thumb subdomain boundary, as can be visualized on the three-dimensional structure of NS5B (Fig. 2B). 2 W. Wels, unpublished data.

mAb Inhibiting HCV RdRp
The observation that mAb 5B-12B7 did not react in immunoblot assays suggested that it recognizes a conformation-sensitive epitope on NS5B. Such epitopes are difficult to map, particularly in the case of proteins with a complex three-dimensional structure such as HCV RdRp. Using a comprehensive set of amino-and carboxyl-terminal NS5B deletion constructs (Fig.  3A), we examined the reactivity of this mAb by immunoprecipitation analyses. As shown in Fig. 3B, all constructs yielded stable proteins by in vitro transcription-translation. All truncated proteins were efficiently immunoprecipitated by a mouse polyclonal antiserum raised against recombinant NS5B (data not shown). mAb 5B-12B7 efficiently immunoprecipitated the full-length NS5B protein, carboxyl-terminal deletions to aa 392 (NS5B⌬C21 and NS5B392, but not NS5B299), and amino-terminal deletions to aa 139 (NS5B11-591, NS5B46 -591, and NS5B139 -591, but not NS5B299 -591) (Fig. 3C). These results allowed us to map the mAb 5B-12B7 epitope to HCV aa 139 -392. Accordingly, this mAb efficiently immunoprecipitated a NS5B fragment corresponding to aa 139 -392, which constitute the entire palm subdomain and a small portion of the fingers subdomain. mAb 5B-12B7 efficiently immunoprecipitated re-combinant NS5B proteins with amino acid substitutions at residues important for enzymatic activity, viz. D220G, D225G, G283R, T286V, T287K, N291K, G317A, D318H, D319E, and R345K (10), indicating that these amino acid residues are not critical for mAb binding (data not shown).
Inhibition of HCV RdRp Activity by mAb 5B-12B7-Given the highly specific interaction of mAb 5B-12B7 with NS5B under native conditions, we reasoned that this mAb might inhibit the RdRp activity. To explore this possibility, constant amounts of highly purified HCV NS5B were used for an in vitro RdRp assay in the presence of various concentrations of mAb 5B-12B7 that was purified from ascites fluid by protein G affinity chromatography. Incorporation of radioactivity into newly synthesized RNA was measured after trichloroacetic acid precipitation onto glass fiber filters by liquid scintillation counting, and background values as determined by analogous assays with a purified inactive HCV RdRp (ϳ1200 cpm) were subtracted. Control reactions without mAbs routinely yielded 80,000 -90,000 cpm. As shown in Fig. 4, RNA synthesis was efficiently blocked by very low concentrations of mAb 5B-12B7. Assuming that the majority of the antibodies isolated from ascites fluid correspond to mAb 5B-12B7, we calculated a virtually complete inhibition of HCV RdRp at a 1:1 molar ratio of mAb to enzyme. This effect was specific because no inhibition was found with mAb 5B-3B1 (directed against the linear NS5B epitope) and mAb 1B6 (directed against the HCV nonstructural protein 3 serine protease domain) (26).

FIG. 1. Characteristics of mAbs 5B-12B7 and 5B-3B1.
A, indirect immunofluorescence microscopy. U-2 OS cells were transiently transfected with pCMVNS5Bcon (NS5B) or pCMVNS5Bcon392 (NS5B392), encoding a carboxyl-terminally truncated protein comprising NS5B aa 1-392, followed by indirect immunofluorescence microscopy using mAb 5B-12B7 as described under "Experimental Procedures." B, immunoprecipitation/Western blot analysis. In lanes 1 and 2, UHCVcon-57.3 cells, which inducibly express the entire polyprotein derived from a functional HCV genotype 1a consensus cDNA, were cultured for 24 h in the presence (57.3 ϩ tet) or absence (57.3 Ϫ tet) of tetracycline. In lanes 3-5, U-2 OS cells were not transfected (U-2 OS) or were transiently transfected with pCMVNS5Bcon (NS5B) or pCMVNS5Bcon392 (NS5B392). Cells were lysed in buffer containing 1% Nonidet P-40 and subjected to immunoprecipitation with mAb 5B-12B7 as described under "Experimental Procedures." In lanes 6 and 7, HuH-7 cells or an HuH-7 cell line harboring a selectable subgenomic HCV replicon (9)(10)(11)(12)(13) was lysed in hypotonic buffer without the use of detergents as described under "Experimental Procedures," followed by immunoprecipitation with mAb 5B-12B7. Immunoprecipitates were separated by 12% SDS-PAGE and analyzed by Western blotting using biotinylated mAb 5B-3B1, followed by detection with horseradish peroxidase-conjugated streptavidin. Molecular mass standards are indicated on the left. C, competitive inhibition experiments. The reactivity of 0.1 g of biotinylated mAb 5B-3B1 (5B-3B1*) or 5B-12B7 (5B-12B7*) per 96-well with NS5B bound to the solid phase was examined in the presence of 0, 1, 5, or 10 g of the indicated mAb (corresponding to a 0-, 10-, 50-, or 100-fold molar excess). ELISA reactivity revealed by horseradish peroxidaseconjugated streptavidin is expressed as percent A 490 of reactions performed in the absence of competing mAb. Values represent the means of duplicate determinations. To analyze whether mAb 5B-12B7 can also affect the RdRp activities of other (ϩ)-strand RNA viruses, we chose two other enzymes: NS5B of CSFV, which is closely related to HCV and has very similar biochemical properties (18), and the more distantly related poliovirus 3D polymerase. Both enzymes were not inhibited even at very high concentrations of mAb 5B-12B7 (Fig. 4), corroborating the specificity of this mAb for HCV RdRp.
Molecular Mechanism of RdRp Inhibition by mAb 5B-12B7-The recombinant HCV NS5B⌬C21 protein has nine tryptophan residues and thus has strong protein fluorescence at 330 nm when it is excited at 283 nm. This property was utilized to determine the equilibrium dissociation constant (K d ) of NS5B for nucleotide or 8-mer RNA ligands. We performed fluorescence quenching experiments by titration of the intrinsic protein fluorescence with successive addition of a ligand. In the experiments performed in the absence of mAb 5B-12B7, NS5B fluorescence was quenched by all five ligands tested, including four nucleotides (GTP, CTP, ATP, and UTP) and the RNA template GR-1 (Fig. 5A), but it was not quenched by a 5Јunphosphorylated dinucleotide, GG (data not shown). The K d values calculated from the quenching curves shown in Fig. 5 are summarized in Table III. Fluorescence quenching experiments were then performed to investigate the effect of mAb 5B-12B7 on the NS5B/RNA/NTP interaction. In the experiments performed in the presence of mAb 5B-12B7 (Fig. 5B), the molar ratio of NS5B to mAbbinding site concentration was set at 1:2 to assure that all the NS5B bound to mAb 5B-12B7. In this context, the apparent inhibition constant (K i ) of mAb 5B-12B7 for NS5B polymerase activity was found to be Ͻ2 nM (data not shown). Thus, NS5B and mAb 5B-12B7 should be in a fully associated form in our fluorescence quenching experiments. The addition of mAb 5B-12B7 to the NS5B solution increased the total protein fluorescence by 1.7-2-fold. Control experiments showed that the mAb 5B-12B7 fluorescence was not quenched by the addition of GTP or CTP at the concentrations used (data not shown). Therefore, the fluorescence quenching experiments could be used to probe the effect of the mAb on the interaction between NS5B and nucleotide or RNA ligands and therefore to deduce the molecular mechanism of action and to functionally map mAb 5B-12B7.
Among the five ligands tested for NS5B bound to mAb 5B-12B7, protein fluorescence was quenched by all except CTP at the concentrations used (Fig. 5B). The K d values of NS5B for ligands in the presence of mAb 5B-12B7 were calculated from FIG. 3. mAb 5B-12B7 epitope. A, schematic representation of NS5B sequences present in the expression constructs pCMVNS5Bcon (1), pCMVNS5Bcon⌬C21 (2), pCMVNS5Bcon392 (3), pCMVNS5Bcon299 (4), pCMVNS5Bcon203 (5), pCMVNS5Bcon139 (6), pCMVNS5Bcon299 -392 (7), pCMVNS5Bcon11-591 (8), pCMVNS5Bcon46 -591 (9), pCMVNS5Bcon139 -591 (10), pCMVNS5Bcon299 -591 (11), and pCMVNS5Bcon139 -392 (12). The locations of mAb 5B-3B1 and 5B-12B7 epitopes are illustrated at the top. Constructs that reacted with mAb 5B-12B7 in immunoprecipitation assays are highlighted in gray. B, in vitro transcription-translation. The expression constructs shown in A were in vitro transcribed and translated using a coupled rabbit reticulocyte lysate system as described under "Experimental Procedures" and separated by 15% SDS-PAGE. [ 35 S]Methionine-labeled translation products were detected by autoradiography. Molecular mass standards are indicated on the right. C, immunoprecipitation of the reactions shown in B using mAb 5B-12B7 as described under "Experimental Procedures." Immunocomplexes were separated by 15% SDS-PAGE. [ 35 S]Methionine-labeled translation products were detected by autoradiography. Molecular mass standards are indicated on the right. Samples loaded in B represent about one-third of the material used for the immunoprecipitations loaded in C.

FIG. 4. Inhibition of HCV RdRp activity by mAb 5B-12B7.
Constant amounts of HCV or CSFV NS5B or poliovirus 3D polymerase (3D pol ) were used for in vitro RdRp assays with a heteropolymeric template RNA and different concentrations of mAbs as specified. After 1 h at 27°C, incorporation of radioactivity into newly synthesized RNA was determined by trichloroacetic acid precipitation and liquid scintillation counting. The RdRp activity of a given enzyme in the absence of antibody was set at 100%. Background values were determined in RdRp assays with an enzymatically inactive HCV NS5B protein (D318N) (10). The molar ratio of mAb to enzyme is plotted versus the percentage of RdRp activity. Values represent the means of duplicate determinations. the respective quenching curves (Table III). When the RNA oligonucleotide GR-1 was used to titrate the fluorescence of the NS5B⅐mAb 5B-12B7 complex, the presence of stoichiometric mAb 5B-12B7 appeared to interfere slightly with the NS5B/ RNA template interaction by increasing the K d by 4-fold. However, the RNA oligonucleotide still bound to NS5B at submicromolar concentrations, suggesting that the mAb did not directly affect the binding of RNA to NS5B. Interestingly, the four nucleotides had very distinct binding profiles when NS5B was bound to mAb 5B-12B7. Judged from the K d values in the presence or absence of mAb 5B-12B7 (Table III), the presence of the mAb did not interfere with the NS5B/GTP interaction. This suggests that there may be a discrete binding site for GTP other than the NTP-binding site for elongation, possibly the primer-binding site for de novo initiation (27). However, the K d was increased by 10-and 21-fold for UTP and ATP, respectively, implying that binding of mAb 5B-12B7 to NS5B interferes with the binding of these two nucleotides to NS5B. Interestingly, in the presence of mAb 5B-12B7, the addition of CTP failed to quench the NS5B fluorescence, indicating that the binding site observed in the absence of mAb 5B-12B7 was totally blocked by the mAb.
5B-12B7 scFv Binds to NS5B Both in Vitro and in Transfected Cells-The mAb heavy and light chain variable domains were cloned, and a scFv was assembled to assess the potential of mAb 5B-12B7 to bind to NS5B intracellularly. To this end, total cellular RNA was prepared from early passage 5B-12B7.54.1 hybridoma cells. Subsequently, reverse transcription was performed using an oligo(dT) primer, and PCR was performed using degenerate primers designed to hybridize to the partially conserved 5Ј-and 3Ј-regions of the heavy and light chain variable domains. PCR products corresponding to the VH and VK domains were inserted into the pWW152 vector, which contains the coding sequence for a flexible (GGGGS) 3 linker. This construct was then subcloned into the appropriate vectors for expression in E. coli and in eukaryotic cells. The nucleotide sequence determined by sequencing eight clones each of the VH and VK domains and the deduced amino acid sequence of the 5B-12B7 scFv are illustrated in Fig. 6.
For bacterial expression, the 5B-12B7 scFv sequence was fused to the E. coli phoA alkaline phosphatase gene. 5B-12B7 scFv-PhoA fusion protein was expressed in E. coli strain CC118 as described under "Experimental Procedures" and enriched by preparation of periplasmic extracts. As a control, a similar fusion protein containing the c-Myc-specific 9E10 scFv construct was produced. Binding of recombinant scFv-PhoA proteins to immobilized NS5B was determined by ELISA. As shown in Fig. 7A, specific binding of bacterially expressed 5B-12B7 scFv-PhoA fusion protein to NS5B was observed, whereas 9E10 scFv-PhoA showed only background binding.
For expression in eukaryotic cells, scFv was subcloned under the control of a cytomegalovirus promoter, and a FLAG tag was fused to the carboxyl terminus by PCR. The potential of this construct to bind to NS5B intracellularly was assessed by cotransfection with full-length and carboxyl-terminally truncated NS5B proteins. In contrast to the full-length protein, the NS5B392 construct was not membrane-associated in the cell and was expressed at very high levels (Fig. 1A). As shown in Fig. 1B, this construct was efficiently immunoprecipitated by parental mAb 5B-12B7. As non-relevant controls, the scFv constructs 9E10 (against c-Myc) and E6 (against HCV non-FIG. 5. Molecular mechanism of RdRp inhibition by mAb 5B-12B7. Fluorescence quenching curves of NS5B were determined by the addition of nucleotide or 8-mer RNA oligonucleotide ligands in the absence (A) or presence (B) of mAb 5B-12B7 as described under "Experimental Procedures." The fluorescence intensity (I) at a given ligand concentration was used to calculate the equilibrium dissociation constant (K d ) as described under "Experimental Procedures" (Table III). structural protein 4A) were used. In addition, a construct in which the E6 sequence was cloned in reverse orientation served as a control for the background of the immunoprecipitation/ Western blot procedure. As shown in Fig. 7B, the 5B-12B7 scFv efficiently immunoprecipitated the NS5B392 construct and, to a lesser extent, the full-length protein, whereas the other scFv constructs showed no or only background binding. Taken together, these results clearly demonstrate that mAb 5B-12B7 retains its capacity to specifically bind to its target antigen as a scFv fragment.

DISCUSSION
In this study, we have described a mAb, designated 5B-12B7, that specifically and efficiently inhibits the HCV RdRp. This mAb binds with high affinity to a conformational epitope represented by NS5B aa 139 -392, which constitute the entire palm subdomain and a small portion of the fingers subdomain. The catalytically active center of the enzyme resides in the palm subdomain. mAb 5B-12B7 completely inhibited HCV RdRp activity in vitro at a 1:1 molar ratio. Inhibition was highly specific since RdRps of related viruses, viz. CSFV NS5B and poliovirus 3D polymerase, were unaffected by this mAb. In addition, the activity of the structurally closely related bacteriophage ⌽6 RdRp (28) was not affected by mAb 5B-12B7, 3 which further demonstrates the specificity of inhibition. Interestingly, mAb 5B-12B7 was found to block NTP binding to NS5B selectively, whereas binding of an RNA substrate was not affected. These observations suggest that the mAb 5B-12B7 epitope overlaps with the NTP (but not the RNA)-binding site.
By contrast, mAb 5B-3B1, which recognizes a distinct linear epitope located at NS5B aa 372-382, did not inhibit RdRp activity. As shown in Fig. 2B, the mAb 5B-3B1 epitope is located far away from the tunnel for nucleotide entry and the RNA-binding groove. Interestingly, mAb 5B-3B1, which reacts very strongly in ELISA and immunoblot assays, failed to stain full-length NS5B in immunofluorescence analyses. This property was independent of the coexpression of other HCV structural and nonstructural proteins, suggesting that in cells the mAb 5B-3B1 epitope is masked by homo-or heterotypic protein/protein interactions. Consistent with this notion, mAb 5B-3B1 reacts well in immunofluorescence analyses of carboxylterminally truncated NS5B (NS5B392, consisting of aa 1-392), which must somehow expose the linear epitope (data not shown). Interestingly, three different mAbs directed against the palm subdomain of HCV RdRp have recently been reported to lack polymerase inhibitory activity (29). This underscores the unique properties of mAb 5B-12B7.
De novo RNA synthesis is often associated with RNA polymerase-catalyzed reactions (30,31). It involves two discrete nucleotide-binding sites: one for the initiating nucleotide and the other for the elongating nucleotide. It has been shown that HCV NS5B can initiate RNA synthesis via a de novo mechanism and that GTP is likely the initiating nucleotide (32,33). In a de novo initiated synthesis, the first initiating nucleotide must serve as a primer and bind to a site different from that of an elongating nucleotide. It was recently reported that a conserved arginine residue in the thumb subdomain plays an important role in forming the primer-binding site for the initiating nucleotide (27). In this report, we demonstrate that GTP is the only nucleotide whose binding to NS5B is not affected by mAb 5B-12B7. This supports the notion that GTP is the initiating nucleotide and that its binding to the primer site is not blocked by this mAb. Clearly, further studies are needed to address the differential binding activities of various nucleotides. In this context, mAb 5B-12B7 may serve as an excellent molecular probe for future mechanistic studies toward the elucidation of the HCV RdRp reaction pathway.
In a different set of experiments, we compared the inhibitory effect of mAb 5B-12B7 on wild-type NS5B and an enzymatically active derivative of NS5B in which the unique ␤-hairpin in the thumb subdomain was deleted (15,27). It was found that both were inhibited equally by the mAb, suggesting that the binding site of mAb 5B-12B7 is not in the proximity of the ␤-hairpin, which is on the opposite site of the tunnel for NTP entry (data not shown).
Cloning of the mAb 5B-12B7 VH and VK domains allowed the assembly of a scFv that retained NS5B binding activity in vitro and in transfected cells. This construct may therefore be useful for intracellular immunization strategies against HCV. A similar approach has been successfully explored, for example, for the inhibition of human immunodeficiency virus (34) and hepatitis B virus replication (35). More important, identification of the exact mAb 5B-12B7 contact sites on NS5B may facilitate the development of small molecule RdRp inhibitors as novel antiviral agents. Finally, the well characterized mAbs described here represent highly valuable tools to further investigate the HCV RdRp and its role during formation of the viral replication complex.