INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Hepatitis C virus (HCV) is the primary causative agent of parenterally transmitted non-A, non-B hepatitis and affects a significant part of the world's population. HCV infection frequently leads to chronic hepatitis, cirrhosis of the liver, and possibly hepatocellular carcinoma (11, 45). HCV, a member of the family Flaviviridae (44), has been a difficult virus to study due to the lack of an appropriate tissue culture system and an adequate, simple, and low-cost animal model. The RNA genome of HCV has recently been cloned, and the single-stranded, plus-polarity RNA genome of the virus is approximately 9,500 nucleotides long and is flanked by untranslated regions (UTR) at both the 5' and 3' ends (22, 28, 48). The 5' UTR of HCV RNA (341 nucleotides) is highly structured and contains an internal ribosome entry site which extends to nucleotide 370, partially overlapping the structural protein (core) coding sequences (8, 20, 21, 51, 53). The secondary structure of the 5' UTR appears to be highly conserved among various HCV strains, and structural similarity to members of the pestivirus family has been reported using biochemical approaches (7, 22). The 5' UTR of the viral RNA is followed by a single large open reading frame that encodes a polyprotein of approximately 3,000 amino acids, which is subsequently proteolytically processed by cellular signal peptidases and two HCV-encoded proteases to produce mature structural and functional proteins (16-18). The 3' UTR of approximately 200 nucleotides contains three distinct regionsa short region with sequence heterogeneity preceding a poly(U) and/or poly(UC) region of variable length followed by a highly conserved sequence (X region) of approximately 100 nucleotides (6, 24, 30, 48, 55). This X region of conserved sequence forms three stable stem-loop structures, SL-I, SL-II, and SL-III. The major nonstructural proteins include two proteases, NS2 and NS3, followed by NS4A, NS4B, NS5A, and NS5B (reviewed in reference 12). NS4A (a 54-amino-acid polypeptide), however, seems to act as a cofactor for NS3 activity, and the central domain has been implicated as essential for this role of NS4A (3, 4, 15, 33, 34, 49). Almost nothing is known about HCV RNA replication; however, various laboratories have recently demonstrated synthesis of full-length complementary RNA (and dimeric RNA), which can be achieved in vitro by the NS5B protein (38, 40, 60).

The HCV NS3 protein has been the subject of intense study due to its associated protease and helicase activities. The C-terminal 450 amino acids of the NS3 protein constitute a polynucleotide-stimulated NTPase activity (25, 42, 46), a 3'-5' unwinding activity (19, 25, 47), and a single-stranded-RNA binding activity (19, 47). The N-terminal one-third of NS3 contains a serine protease activity and is responsible for the downstream cleavages in the nonstructural region (31, 50). The three-dimensional structure of NS3 has been solved by X-ray crystallography, and structure-based mutagenesis of NS3 has identified important amino acid residues required for helicase and ATPase activities (9, 29, 37, 59). Recent studies have shown that coexpression of the NS4A protein directs the NS3 protein, which is diffusely distributed in the cytoplasm and nucleus in the absence of NS4A, to the endoplasmic reticulum (54).

Many single-stranded positive-strand RNA viruses encode their own helicases (and NTPases). They are thought to play a role in viral RNA replication. Our previous work has shown that the poliovirus-encoded NTPase (and helicase) 2C specifically interacts with the 3'-terminal sequences of viral negative-strand RNA (2). In addition to its specific RNA binding ability, it is also able to interact with cellular cytoplasmic membranes (5, 10, 13, 14). Since poliovirus RNA synthesis takes place in cytoplasmic membranes, it is believed that the poliovirus 2C protein anchors the negative strand to the cytoplasmic membrane, thus allowing initiation of positive-strand RNA synthesis to occur. Moreover, initiation of positive-strand synthesis is likely to require an unwinding activity to melt the double-stranded structure at the 3' end of the negative strand formed by initial copying of the input viral plus-strand RNA. The rationale stated above has prompted us to examine the interaction of intact NS3 and the pNS3 (with the helicase domain only) with the 3'-terminal sequence of the negative-strand RNA of HCV. We show that in fact both NS3 and pNS3 interact specifically with the 3'-terminal sequences of HCV negative-strand RNA. This interaction is impaired by deletion of the 3'-terminal sequences of the negative strand. The results presented here suggest that a stem-loop structure within the 3'-terminal sequence is important for interaction with NS3. Unlike the poliovirus-encoded helicase (2C), which interacts with the 3' UTR of the negative strand but not with the 3' sequences of the positive strand, NS3 (and pNS3) interacts with both the positive- and negative-strand 3' UTR sequences. Initial mutagenesis suggests that an intact higher-order structure at the 3'(+) UTR is necessary for the interaction of NS3 with the 3' terminus of the positive-strand RNA. Thus, our results show that NS3 specifically binds to the 3' ends of both the positive- and negative-strand RNAs of HCV. The implication of NS3 binding to the 3'-terminal sequences of viral positive- and negative-strand RNA in HCV RNA replication is discussed below.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Expression and purification of HCV NS3 and pNS3. The recombinant clones used for expression of the viral wild-type NS3 and the truncated version, pNS3 (with an intact helicase domain but a deleted protease domain) were constructed by PCR amplification using the HCV 1969 cDNA clone (a kind gift from Genevieve Inchauspe, INSERM, Lyon, France) as a template. The coding sequences were ligated into the pET-21b expression vector (Novagen), following standard molecular biology protocols, at the BamHI and HindIII sites using the primer pairs RB-1 and RB-2 (for NS3) and RB-3 and RB-2 (for pNS3) incorporating appropriate enzyme cleavage sites (Table 1).

UTR cloning and probe preparation. The viral 5' UTR sequence encompassing the first 127 nucleotides was PCR amplified from the HCV 1969 cDNA clone using RB-4 and RB-5 as the forward and reverse primers. These oligonucleotides incorporated the appropriate restriction sites, and the resulting PCR product was gel purified, sequentially digested with HindIII and EcoRI, and ligated into the corresponding sites of the transcription vector pGEM-3 (Promega) using standard protocols. The nucleotide sequences of the recombinant clones used throughout were confirmed by dideoxy sequencing (Amersham). The cloning strategy was chosen so that the positive- or the complementary negative-strand transcripts [5'(+) UTR₁₂₇ or 3'() UTR₁₂₇, respectively] could be generated following linearization, using either SP6 or T7 RNA polymerase in the presence of [³²P]UTP (3,000 Ci/mmol; Amersham) in an in vitro transcription reaction. The labeled transcripts were purified by denaturing gel electrophoresis before they were used in binding reactions. The 5' UTR mutants (I to XXI) were obtained using the same vector and were generated with the appropriate forward primers, as indicated in Table 1, and the same reverse primer used for the wild-type 5' UTR₁₂₇ (RB-5). For the rescue experiment using the mutants V-[A] to V-[F] (see Fig. 7), the fragment for cloning was derived with mutant V DNA as a template. For mutant VII, the forward primer used was the same as for the wild type (RB-4), and the reverse primer was RB-27. The similar-size RNA from the coding region used for specificity analysis (see Fig. 4B) was obtained from a clone constructed using the primer pairs RB-28 and RB-29.

RNA binding, UV-cross-linking analysis, and probe stability analysis. The binding reaction (25 µl) contained binding buffer III, 20 mM dithiothreitol, 2 mM ATP, 15 µg of yeast tRNA, RNasin (30 U), and approximately 100 ng of purified protein. The binding reaction was preincubated with purified NS3 or pNS3 in a 30°C water bath for 5 min, following which the ³²P-labeled RNA probe (200,000 cpm/25-µl sample) was added. The competitor RNAs were preincubated with the protein for 10 min prior to the addition of the probe. The incubation was continued for an additional 15 min to facilitate the binding of the probe RNA to the protein. At termination, samples were cross-linked by UV irradiation as detailed in references 1 and 2, and samples were analyzed by SDS-14% PAGE. The signal intensity from each lane was quantitated with a laser densitometer scanner (Molecular Dynamics), and the data were analyzed with the Image-QuaNT software program. Each reaction was performed two or three times for reproducibility.

In vitro transcription and translation. The standard protocol for in vitro transcription was followed. Briefly, the expression clone, pET-NS3 or pET-pNS3, was linearized using HindIII, purified, and used as a template for the synthesis of capped mRNA. The mRNAs obtained were used in translation reactions with rabbit reticulocyte lysate (Promega) according to the manufacturer's instructions. In vitro-synthesized proteins were labeled with 40 µCi of [³⁵S]methionine (specific activity, >100 Ci/mmol; Amersham) and resolved along with UV-cross-linked protein samples on SDS-PAGE gels. In selected experiments, the coupled transcription-translation system (Promega) was also used.

Immunoprecipitation. The in vitro-translated protein or the protein-nucleotidyl complexes were immunoprecipitated with the same concentrations of monoclonal antibody to NS3 (AUSTRAL Biologicals, San Ramon, Calif.) or a nonspecific antibody targeted toward CREB. Three separate UV-cross-linked samples were pooled for immunoprecipitations. The reaction mixtures were incubated at 4°C for 3 h with 1× radioimmunoprecipitation assay buffer (20 mM Tris-HCl [pH 7.5], 0.5% deoxycholate, 1% NP-40, and 150 mM NaCl), 0.5 mM phenylmethyl sulfonyl fluoride, and 5 mg of bovine serum albumin. Immune complexes were adsorbed on protein A-Sepharose beads (5 mg/reaction; Pharmacia) for 1 h at 4°C, following which the beads were washed five times with radioimmunoprecipitation assay buffer to reduce background binding and washed once with phosphate-buffered saline. The proteins were finally eluted with Laemmli sample buffer and resolved on an SDS-14% PAGE gel, followed by fixing and autoradiography.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

HCV protease/helicase NS3 interacts specifically with the 3' terminus of HCV negative-strand RNA. In order to examine the interaction of NS3 with the 3'-terminal sequence(s) of HCV negative-strand RNA, both NS3 and the truncated protein (pNS3) were expressed in E. coli BL21(DE3) cells. Both proteins contained a T7 tag at the N terminus and six histidine residues in tandem at the C terminus. The additional amino acids contributed by the epitopes increase the mass of the expressed proteins by approximately 2 kDa. The bacterially expressed proteins were purified by Co²⁺ immobilized affinity chromatography, and the purified polypeptides were analyzed by SDS-PAGE followed by Coomassie blue staining. As shown in Fig. 1B, both NS3 and pNS3, with estimated molecular masses of 75.6 and 55.6 kDa, respectively, were purified to near homogeneity (approximately 98%). To reconfirm that the proteins seen in the Coomassie blue-stained gel were indeed NS3 and pNS3, Western blot analysis was performed with available antibodies against the N-terminal T7 tag. The signals obtained in the immunoblot analysis corresponded correctly with the expected migrations of the polypeptides.

View larger version (67K): [in this window] [in a new window]   FIG. 1.   HCV-encoded NS3 protein expression. (A) Schematic representation of HCV genome organization marking the relative position of NS3 protein. The numbers correspond to nucleotide positions in the HCV 1969 cDNA clone. (B) On the left is a Coomassie blue-stained gel of purified NS3 and pNS3. The positions of the NS3 (lane 1) and pNS3 (lane 2) proteins are indicated, along with the protein marker lane (M). On the right is a Western blot analysis of the purified proteins resolved on a 14% gel; lanes 1 and 2 correspond to the same protein samples run on the Coomassie blue-stained gel. The numbers on the right correspond to the migrations of the molecular mass marker proteins and are marked in kDa.

View larger version (25K): [in this window] [in a new window]   FIG. 2.   HCV-encoded NS3-RNA interaction. (A) Schematic representation of the viral positive-strand UTR used in the present study. (B) Predicted secondary structure of the first 127 nucleotides of the sequence of the positive strand, 5'(+) UTR127 (adapted from reference 20). (C and D) Analysis of purified pNS3 and NS3 binding to the 5' positive- or 3' negative-strand UTR127 RNA probe. The binding reaction mixture contained either 5'(+) UTR127 (lanes 1 and 2) or 3'() UTR127 (lanes 3 and 4) probe. Lanes 1 and 3 are controls with no protein added while in lanes 2 and 4 approximately 100 ng of pNS3 (C) or NS3 (D) was added. The migration of the [35S]methionine-labeled in vitro-translated wild-type or truncated NS3 is shown in lanes R, and the numbers correspond to the migrations of rainbow molecular mass markers (Amersham) in kDa. The relative position of the UV-cross-linked RNA-protein complex is indicated.

View larger version (42K): [in this window] [in a new window]   FIG. 3.   Characterization of the interaction between pNS3 and 3' UTR127 RNA. (A) Effect of binding buffer composition on nucleoprotein complex formation. Reaction mixtures containing pNS3 were incubated with RNA probes under various binding buffers as detailed in Materials and Methods. The protein was analyzed for each binding buffer condition at three concentrations (approximately 50, 100, and 150 ng/25-µl reaction volume), and UV-cross-linked complexes were resolved by SDS-14% PAGE. Lane R represents in vitro-translated [35S]methionine-labeled pNS3 protein, and the numbers on the left indicate the positions of the molecular mass marker proteins in kDa. (B) Effect of nucleoprotein inhibitor ATA on complex formation. Reactions with (lane 2) or without (lane 1) pNS3 and reactions in the presence of increasing concentrations (2.5, 5.0, and 10 µM) of ATA and pNS3 (lanes 3 to 5) are shown. (C) Effect of heat denaturation of pNS3 on RNA interaction. The pNS3 protein was preincubated for 10 min at 30, 40, 65, and 90°C (lanes 1 to 4) prior to its addition to the binding reaction and then was processed as described above. (D) Immunoprecipitation analysis of the UV-cross-linked RNA-protein complex. Purified pNS3 UV cross-linked to 3'() UTR127 was immunoprecipitated as detailed in Materials and Methods. [35S]methionine-labeled in vitro-translated (IVT) pNS3 protein was analyzed directly (lane 1) or following immunoprecipitation with anti-NS3 (lane 3) or a nonspecific antibody (lane 2). Lanes 4 to 7 contain UV-cross-linked RNA-protein complex either loaded directly (lanes 4 and 5) or following immunoprecipitation with the control (lane 6) or anti-NS3 (lane 7) antibody. The reaction mixtures in lanes 5 to 7 contained pNS3, but lane 4 had no added pNS3. The portion of the gel containing lanes 6 and 7 was overexposed to visualize the pNS3 band.

Specificity of RNA-protein interaction. To determine if the binding of pNS3 to the 3'() UTR₁₂₇ RNA was specific, competition binding assays were performed in which unlabeled homologous and heterologous RNAs were added to the binding reaction during formation of the ³²P-labeled UTR-pNS3 complex. While the addition of 20- and 50-fold molar excesses of unlabeled 3'() UTR₁₂₇ resulted in approximately 60 to 70% reduction in binding, the intensity of the labeled complex was reduced to almost 10% by the addition of a 100-fold molar excess of unlabeled homologous RNA (Fig. 4A, lanes 2 to 4). No significant reduction in the intensity of the RNA-protein complex was noted in the presence of 20-, 50-, and 100-fold molar excesses of cold, heterologous, similar-size RNA from hepatitis A virus (Fig. 4A, lanes 5 to 7). Additionally, when an unrelated similar-size labeled RNA fragment from the HCV negative strand was used in the binding reaction, only approximately 4% as much RNA-pNS3 complex formation was observed as with the wild type 3'() UTR₁₂₇ probe (Fig. 4B, compare lane 2 with lane 4). Finally, two mutant 3'() UTR₁₂₇ RNAs that were defective in binding NS3 and NS3 (mutants V and VI; see below) were unable to compete with 3'() UTR₁₂₇ probe in the binding assay over a range of concentrations (Fig. 4C). Taken together, these results suggest that the interaction of pNS3 with the 3'() UTR₁₂₇ sequence is specific.

View larger version (28K): [in this window] [in a new window]   FIG. 4.   Specificity of pNS3 binding to the negative-strand RNA probe. (A) Cold homologous RNA corresponding to the 3'-terminal 127 bases of the negative strand and a heterologous, similar-size RNA sequence from hepatitis A virus were used in the competition assay as described in Materials and Methods. Reactions were performed without (lane 1) or with 50 (lane 2), 100 (lane 3), or 250 (lane 4) ng of cold homologous RNA or 50 (lane 5), 100 (lane 6), or 250 (lane 7) ng of heterologous RNA. (B) Specific binding to negative-strand RNA. A similar-size RNA obtained from the coding sequence of the HCV negative strand, labeled and processed as described for the UTR127 probe, was added to the binding reaction containing the pNS3 protein. Lane R represents the [35S]methionine-labeled in vitro-translated pNS3 protein. Protein binding to 3'() UTR127 probe without (lane 1) and with (lane 2) added pNS3 is shown. Lanes 3 and 4 shows binding to RNA derived from the coding sequence in the absence (lane 3) and presence (lane 4) of pNS3. (C) Competition using inactive mutant V and VI UTR127RNAs. The cold mutant RNAs were added to binding reactions as for panel A. Reactions with no cold competing RNA (control) (lane 1), with homologous RNA (lane 2), and with increasing concentrations (50, 100, and 250 ng/25-µl reaction volume) of mutant V or VI RNA (lanes 3 to 8) are shown. Homologous RNA was used at the same concentration as for lane 3 in panel A.

3'-Terminal sequences of the negative-strand UTR₁₂₇ RNA are important for protein binding. To determine sequences within the 3' UTR RNA (127 nucleotides) of HCV required for NS3 binding, we initially deleted 10 and 25 nucleotides from the 3' terminus. These mutants were termed 10 (mutant I) and 25 (mutant II). Compared to the wild-type 3'() UTR₁₂₇ Fig. 5B, lane 2, both mutants I and II were defective in RNA binding (lanes 4 and 6). While deletion of the first 10 nucleotides from the 3' terminus of the negative-strand (10; mutant I) (the minus signs preceding the nucleotide numbers denote the position in the negative-strand RNA starting from the 3' end) resulted in almost 80% decrease in RNA binding (lane 4), removal of the first 25 nucleotides (25; mutant II) reduced binding by approximately 85% (lane 6). In both mutants I and II, a stem-loop structure within the first 20 nucleotides from the 3' end was deleted (Fig. 5A). Therefore, the initial results suggested that the predicted stem-loop spanning nucleotides 5 through 20 may be important for pNS3-RNA interaction. To determine if the stem-loop was somehow involved in NS3 binding, the stem was destabilized by replacing the five guanosine residues with cytosines (Fig. 5A, mutant IV). As expected, like mutants I and II, mutant IV was also defective in pNS3 binding (18% binding remaining [Fig. 5B, lane 8]). It therefore appeared that the stem within the stem-loop structure adjacent to the 3' terminus of the negative-strand UTR₁₂₇ RNA was important for NS3 binding. Deletion of the last 27 nucleotides from the UTR RNA (nucleotides 101 through 127; mutant VII) did not significantly alter protein binding (Fig. 5B, compare lanes 16 and 18 [mutant VII]). Similar results were obtained when pNS3 was replaced by NS3 in the binding reaction with the above-mentioned mutants (data not shown).

View larger version (35K): [in this window] [in a new window]   FIG. 5.   Analysis of HCV-encoded pNS3 interaction with negative-strand 3' UTR127 mutant RNA probes. (A) Schematic illustration of the mutant RNAs used in the binding assay. Mutants with deleted (mutants I, II, and VII) or altered (mutants IV, V, and VI) bases compared to the wild-type RNA are indicated by shading and  or *, respectively. The intervening sequences are indicated by dots. (B) UV-cross-linking analysis of pNS3 protein binding to mutant RNAs. Lanes 1, 2, 9, 10, 15, and 16 represent binding to wild-type (wt) RNA and the remaining lanes show binding to the mutant RNAs as indicated above each set of lanes. The odd-numbered lanes served as control reactions with no added protein, while the even-numbered lanes represent binding reactions containing approximately 100 ng of purified pNS3. Lane R is the [35S]methionine-labeled in vitro-translated pNS3, and the numbers on the left and right indicate the positions of the rainbow molecular mass marker proteins in kDa. The lower gels show the stability of the RNA probes in the absence and presence of pNS3 protein, the details of which are discussed in Materials and Methods. The relative binding of pNS3 protein to mutant RNAs compared to the control wild-type RNA is shown as percent control.

View larger version (63K): [in this window] [in a new window]   FIG. 6.   Analysis of NS3 and pNS3 binding to the 3' negative-strand UTR127 mutant RNAs. (A) Schematic illustration of the mutant RNAs used in binding analysis. The altered bases in the mutant RNAs compared to the wild-type RNA are shaded and marked by asterisks. The intervening sequences are indicated by dots. (B) Specific RNA binding activity of NS3 and pNS3 proteins using the mutant RNA probes. Reactions showing binding to either wild-type (wt) or mutant (mutant VIII, IX, IX-A, and IX-B) RNA probes are indicated above the gel. The odd-numbered lanes contain control reactions with no added proteins, while the even-numbered lanes represent binding reactions containing either the pNS3 or NS3 protein. Lanes R represent [35S]methionine-labeled in vitro-translated (IVT) pNS3 or NS3, respectively, and the numbers on the left indicate the positions of the molecular mass marker proteins in kDa. Only the relevant IVT proteins are shown along with the appropriate binding reactions. The bottom gel shows the result of probe stability analysis with NS3 protein, the details of which are discussed in Materials and Methods. Quantitation of RNA-protein complexes is indicated as percent control.

G-C pairs at positions 7, 8, and 9 in the 3' UTR₁₂₇ RNA-proximal stem are important for protein binding. To determine if all five G-C pairs present in the stem were important for NS3 binding, we introduced mutations into the binding-inactive backbone of mutant V to restore NS3 binding. It is important to note that in mutant V RNA, the positions of the five Gs and Cs in the stem were flipped so that the Cs were 3' proximal in the mutant instead of the Gs being 3' proximal, as in the wild-type 3'() UTR₁₂₇. Replacing three C-G pairs at positions 6, 7, and 8 in the mutant V backbone with the wild-type G-C sequence (mutant V-[A]) restored only 26% binding compared to the wild-type (Fig. 7B, lanes 5 and 6) while having three G-C pairs at positions 8, 9, and 10 (mutant V-[C]) did not bring back any significant level of pNS3 binding (lanes 7 and 8). Similar results were obtained when pNS3 was replaced with NS3 in the binding reaction (Fig. 7B, middle). Restoring three G-C base pairs at positions 7, 8, and 9 (mutant V-[B]), however, brought back 60 to 85% of the binding activity compared to the wild type (lanes 9 and 10). Mutant V-[D], in which only the 7 position had a G-C reversion, retained only 2% of the binding activity, and introducing two G-C base pairs at positions 7 and 8 brought the binding activity to only 8% of the control (lanes 19 and 20; mutant V-[E]). Interestingly, replacing two G-C pairs at positions 7 and 9, however, restored 50 to 70% of the binding activity compared to that of the wild-type UTR₁₂₇ RNA (lanes 17 and 18; mutant V-[F]). These results suggested that the G-C base pairs at positions 7, 8, and 9 in the negative strand may be important for protein binding.

View larger version (51K): [in this window] [in a new window]   FIG. 7.   Rescue analysis of NS3 and pNS3 binding to mutant V RNA. (A) Schematic illustration of the various mutant probes used for RNA-protein binding analysis. All of the probes contain mutant V RNA as a backbone. The reversion to the wild-type sequence of specific C-GG-C base pairs in the 3'-proximal stem is marked by shading and an asterisk. The intervening sequences are indicated by dots. (B) Restoration of specific nucleoprotein complex between mutant RNA probes V-[A] to-[E] and NS3 (full length and truncated). Reactions containing either wild-type (wt) or mutant RNA are indicated above the gels. The odd-numbered lanes served as controls without the added proteins, while the even-numbered lanes represent binding reactions containing approximately 100 ng of either purified pNS3 or NS3 protein. Lane R represents the [35S]methionine-labeled in vitro-translated (IVT) pNS3 (top) or NS3 (middle) protein, and the numbers on the left and right indicate the positions of migration of the molecular mass marker proteins in kDa. Only the relevant IVT proteins are shown along with the appropriate binding reactions. The bottom gel shows the results of probe stability analysis with the NS3 protein, the details of which are discussed in Materials and Methods. Quantitation of the RNA-protein complex is indicated as percent control for both NS3 and pNS3 binding.

View larger version (77K): [in this window] [in a new window]   FIG. 8.   Analysis of NS3 and pNS3 protein binding to 3' negative-strand UTR127 RNA mutants with a single G-CC-G change. (A) Schematic illustration of the mutant RNA probes used in binding analysis. The base pair sequences in the mutant RNAs that are altered compared to the wild-type sequence are marked by shading and asterisks. The intervening sequences are indicated by dots. (B) Specific RNA binding activity of NS3 and pNS3 proteins to mutant RNA probes. Reaction lanes showing binding to either wild-type (wt) or mutant (X, XI, and XII) RNA probes are indicated above the gels. The odd-numbered lanes contain control reactions with no added proteins, while the even-numbered lanes represent binding reactions containing approximately 100 ng of either purified pNS3 or NS3 protein. Lane R represents the [35S]methionine-labeled in vitro-translated (IVT) pNS3 or NS3 protein, and the numbers on the left indicate the positions of migration of the molecular mass marker proteins in kDa. Only the relevant IVT proteins are shown along with the appropriate binding reactions. The bottom gel shows the result of probe stability analysis with NS3 protein, the details of which are discussed in Materials and Methods. Quantitation of the RNA-protein complex is indicated as percent control for both NS3 and pNS3 binding.

View larger version (56K): [in this window] [in a new window]   FIG. 9.   Analysis of NS3 binding to additional 3'UTR127 RNA mutants with specific G-CA-U changes. (A) Schematic illustration of the mutant RNA probes used in the binding analysis. The base pair sequence in the mutant RNAs that have been altered compared to the wild-type sequence are marked by shading and asterisks. The intervening sequences are indicated by dots. (B) Specific RNA binding activity of NS3 protein to mutant RNA probes. Reaction lanes with binding to either wild-type (wt) or mutant (XIX, XX, and XXI) RNA probes are indicated above the gels. The odd-numbered lanes contain control reactions with no added protein while the even-numbered lanes represent binding reactions containing 100 ng of purified NS3 protein. Lane R represents the [35S]methionine-labeled in vitro-translated NS3 protein, and the numbers on the right indicate the migrations of the molecular mass marker proteins in kDa. The lower gel represents the result of probe stability analysis with and without NS3 protein in the binding reaction, as in the experiments discussed above. The relative binding of NS3 protein to mutant RNAs compared to the wild-type RNA is shown as percent control.

NS3 also interacts with the 3'-terminal sequences of the positive-strand RNA. Since initiation of negative-strand synthesis is likely to occur at or near the 3' terminus of the positive-strand RNA, we examined whether NS3 is capable of interacting with the 3'-terminal sequences of the positive-strand RNA. The 3' UTR sequence of genomic RNA was cloned into the transcription vector, and RNA was transcribed from the clone, similar to the procedure with the negative-strand UTR₁₂₇ probe. The ³²P-labeled 3'(+) UTR RNA was incubated with purified recombinant NS3, and the resulting RNA-protein complexes were analyzed by UV cross-linking. As shown in Fig. 10B, NS3 readily formed complexes with the 3'(+) UTR of HCV (lanes 3 and 4). The intensity of the protein-nucleotidyl complex was 70% of that with 3'() UTR₁₂₇ RNA (Fig. 10B, compare lane 4 with lane 2). The 5'-terminal sequences of the negative strand [5'() UTR] that are complementary to the 3'(+) UTR sequence showed very little binding to NS3 [approximately 3% compared to the 3'(+) UTR binding (Fig. 10B, lanes 5 and 6)]. These results suggested that NS3 is capable of interacting with the 3'-terminal sequences of both positive- and negative-strand RNA.

View larger version (36K): [in this window] [in a new window]   FIG. 10.   Analysis of NS3 binding to the 3' UTR of the positive strand. (A) Schematic representation of the viral 3' UTR RNA and the derived mutants used in analysis. The mutants with deleted RNA sequences (mutants A to D) are shown. (B) UV-cross-linking analysis of NS3 binding to 3' UTR RNA of the positive strand (lanes 3 and 4) and the corresponding complementary region of the negative-strand 5' UTR probe (lanes 5 and 6). The specific interaction of NS3 with the negative-strand 3' UTR127 probe was included as a control (CONT) (lanes 1 and 2) (see Materials and Methods). (C) Binding analysis using mutant RNA probes (A, B, C, and D). NS3 binding to the mutant RNA probes (lanes 5 to 12) is compared to its binding to the wild-type 3'(+) UTR probe (lanes 1 and 2). Reactions with mutant RNA probes are indicated above the gel. The reactions in lanes 1 and 2 are similar to those in panel B. The odd-numbered lanes in each set contain control reactions with no added protein, while the even-numbered lanes represent binding reactions containing approximately 100 ng of purified NS3 protein. The positions of UV-cross-linked complexes are indicated. (D) Specificity of NS3 interaction with 3' UTR probe RNA. Cold homologous RNA corresponding to the 3' UTR and a heterologous viral RNA with poly(U) sequence were added to reactions as detailed in Materials and Methods. Reaction mixtures contained no competitor RNA (control reaction [lane 1]); 50 (lane 2), 250 (lane 3), or 500 (lane 4) ng of cold heterologous RNA; or similar amounts of homologous RNA (lanes 5 to 7). Lane R is the [35S]methionine-labeled in vitro-translated NS3 protein. The positions of the proteins are marked.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

We have shown in the present study that the HCV-encoded protease/helicase protein NS3 interacts with the 3'-terminal sequences of the viral positive-and negative-strand RNA. Specifically, the helicase portion of NS3 is responsible for RNA-protein interaction, as indicated by the similar binding profiles of the full-length NS3 protein and the truncated protein lacking the protease domain (pNS3). The interactions of NS3 with the 3' UTR₁₂₇ of the negative-strand RNA and the 3' UTR of the positive-strand RNA appear to be specific, since the formation of the nucleoprotein complexes is successfully competed by the unlabeled homologous RNA but not by similar-size RNAs derived from hepatitis A virus or poliovirus. The specificity of NS3-RNA interaction is also validated by the fact that only the 3'() UTR₁₂₇ and 3'(+) UTR sequences, but not the corresponding complementary sequences from the 5'(+) and 5'() termini, interact with NS3. The interaction of NS3 with the 3'-terminal sequences of the negative-strand UTR RNA appears to depend on a stem-loop structure near the 3' terminus of the RNA. Additional mutagenesis confined to the stem-loop structure spanning nucleotides 5 through 20 suggests that an intact double-stranded structure of the stem and the orientation of three G-C pairs (at positions 7, 8, and 9) within this stem are important elements for specific interaction of NS3 with the negative-strand RNA.

In the past, various investigators have reported on the RNA binding activity of the helicase portion of NS3. The majority of the RNA binding studies reported to date have used partially single-stranded or single-stranded RNA, including homopolymeric RNAs. Kanai et al. demonstrated binding of NS3 to poly(U) Sepharose resin with an apparent dissociation constant of 2 × 10⁷ M (26). Other investigators have used filter binding (35), gel retardation (19), and fluorescence quenching (43) assays to assess the RNA binding activity of NS3. Except for some preference for poly(U), none of these studies detected any specific interaction of NS3 with HCV RNA. The preference for poly(U) can be explained by the fact that poly(U) stretches are present at the 3' UTR terminus in HCV positive-strand RNA. Our binding results, along with those of the competition analysis, clearly demonstrate that the requirements of NS3 interaction with the 3'(+) UTR are much more complex than the mere presence of the poly(U) stretch. In fact, the 5' half of the 3'(+) UTR, which contains the poly(U) stretch, is unable to interact with NS3 (or pNS3) in the absence of the 3' half of the UTR under our assay conditions (Fig. 10C). In contrast to the relatively stringent requirement for 3'() UTR₁₂₇ binding, the interaction of NS3 with the 3'(+) UTR appears to require an intact structure of the entire 3'(+) UTR RNA under the assay conditions used in this study. To our knowledge, the results presented in this paper represent the first demonstration of a specific interaction of HCV NS3 protein with the 3'-terminal sequences of the HCV RNA. The importance of the 5' and 3' UTR sequences in viral infectivity have recently been confirmed in the chimpanzee model (32, 56-58). It should be pointed out that the 3'-terminal sequences of the negative- and positive-strand UTR contain some extra nucleotides that are contributed by the multiple cloning site. These sequences most likely do not play any role in NS3 (or pNS3) recognition, as several of the internal mutations within the viral UTR sequences result in total blockage of RNA-protein interaction even though the mutant RNAs still contain the extra 3'-terminal nucleotides.

All pestiviruses and flaviviruses contain conserved helicase motifs in their NS3 proteins, suggesting an important role of the helicase in the life cycles of these viruses. Other positive-strand RNA viruses, such as poliovirus, rhinovirus, and coxsackie virus, also encode a distinct but homologous protein (called 2C) with RNA binding and NTPase activities and putative helicase activity (39, 41). We have shown previously that the poliovirus-encoded 2C protein specifically interacts with the 3' UTR sequences of the viral negative-strand RNA (2). Both the precursor, 2BC, and the mature protein, 2C, are also capable of interacting with the cellular membranes (5, 10, 13, 14). It has been postulated that 2C may anchor viral negative-strand RNA to the cytoplasmic membrane so that initiation of positive-strand RNA synthesis can occur at the 3' end of negative-strand RNA (2). Given the functional similarity between poliovirus 2C and the HCV helicase (NS3 and pNS3), it is tempting to speculate that NS3 might have a similar role in HCV RNA synthesis. Recent results have suggested that interaction of NS3 with NS4A directs NS3 to the endoplasmic reticulum (54). Moreover, since NS3 appears to interact with the 3' UTR of the positive- as well as the negative-strand RNA, a likely scenario might be that the NS3-NS4A complex bound to either the positive- or negative-strand 3' UTR may anchor RNA-protein complexes to the cytoplasmic membrane. The polymerase NS5B (and possibly NS5A) can then join this complex by protein-protein and/or protein-RNA interaction and initiate positive- or negative-strand RNA synthesis. In fact, direct interaction of HCV NS5B with NS3 and NS4A has been reported (23). Future experiments will be directed to see if NS3-UTR interaction is influenced by NS4A either alone or when present as the NS3-NS4A fusion polypeptide. Also, in the closely related dengue virus, another member of the family Flaviviridae, both NS3 and NS5 have been shown to interact in vivo in CV-1 and HeLa cells (27). It is also possible that one or more host cell proteins could play a role in the membrane association of the viral RNA replication complex through participation of a membrane-associated protein, hVAP-33 (52). This SNARE-like protein containing a membrane-spanning domain has been shown to interact with both NS5A and NS5B. As reliable systems to study HCV RNA replication become available, it will be possible to address mechanistic questions to assess the roles of various viral proteins in RNA replication.

The results presented here do not exclude the possibility that NS3 also interacts with other regions of viral positive- or negative-strand RNA. However, since initiation of RNA synthesis is likely to start at the 3' terminus of positive- and negative-strand RNA, it is likely that the interaction presented here is important for viral RNA replication. It is also important to note that specific binding of a protein to an RNA sequence and/or structure does not necessarily provide functional relevance to the RNA-protein interaction. Additional studies must be conducted to assess the importance of the mutations in both the 3'() and 3'(+) UTR that alter NS3 binding. Such studies may now be possible with the advent of the HCV replicon system that has recently been reported (36). Future work will be directed towards evaluation of the 3'(+) and 3'() UTR mutations in viral RNA synthesis.

In summary, we have demonstrated specific interaction of the HCV NS3 (helicase domain) with the 3'() and 3'(+) UTR sequences. The initial mutagenesis studies reported here have confirmed the requirement for the specific structure and sequence in NS3 binding. We suggest that the binding of NS3 to the 3'-terminal sequences of positive- and negative-strand RNA plays an important role in HCV RNA replication.