Specific interaction between the hepatitis C virus NS5B RNA polymerase and the 3' end of the viral RNA

Abstract

Hepatitis C virus (HCV) NS5B protein is the viral RNA-dependent RNA polymerase capable of directing RNA synthesis. In this study, an electrophoretic mobility shift assay demonstrated the interaction between a partially purified recombinant NS5B protein and a 3' viral genomic RNA with or without the conserved 98-nucleotide tail. The NS5B-RNA complexes were specifically competed away by the unlabeled homologous RNA but not by the viral 5' noncoding region and very poorly by the 3' conserved 98-nucleotide tail. A 3' coding region with conserved stem-loop structures rather than the 3' noncoding region of the HCV genome is critical for the specific binding of NS5B. Nevertheless, no direct interaction between the 3' coding region and the HCV NS5A protein was detected. Furthermore, two independent RNA-binding domains (RBDs) of NS5B were identified, RBD1, from amino acid residues 83 to 194, and RBD2, from residues 196 to 298. Interestingly, the conserved motifs of RNA-dependent RNA polymerase for putative RNA binding (220-DxxxxD-225) and template/primer position (282-S/TGxxxTxxxNS/T-292) are present in the RBD2. Nevertheless, the RNA-binding activity of RBD2 was abolished when it was linked to the carboxy-terminal half of the NS5B. These results provide some clues to understanding the initiation of HCV replication.

FIG. 1

Expression of HCV NS5B protein in E. coli. (A) Coomassie blue staining. E. coli BL21(DE3) cells were transformed with plasmids pET15b (lanes 1 and 2) and pET15b-NS5B (lanes 3 and 4) and grown at 37°C. Plasmid pET15b-NS5B bears HCV cDNA from nt 7258 to 9030 of the coding region that encodes the full-length NS5B representing the viral polyprotein from amino acid residues 2420 to 3010. Expression and purification of the recombinant NS5B protein were performed essentially according to the procedures described by the manufacturer (Novagen). Cell lysates of soluble (lanes 2 and 4) and insoluble (lanes 1 and 3) fractions were resolved on a sodium dodecyl sulfate–8% polyacrylamide gel. Coomassie blue staining is shown. (B) Western blot analysis. The insoluble fractions of protein lysates from cells transformed with pET15b (lane 1) and pET15b-NS5B (lane 2) were immunoblotted following the procedures previously described (6) with the immunoglobulin G fraction of a rabbit antiserum that was raised against the NS5B peptide NH₂-MSYTWTGALITPCAAE-COOH. Arrowheads indicate the IPTG-induced recombinant NS5B protein.

FIG. 2

Complex formation between HCV NS5B protein and the 3′CNUX RNA. (Top) Structures of the HCV 3′-end RNAs, 3′CNUX, and 3′CNU. The heavy line that begins with nt 8736 and ends with nt 9030 represents the 3′ coding region of the HCV polyprotein. The 3′CNUX RNA consists of the 295-nt 3′ coding region and the entire viral 3′NCR encompassing a short genotype-specific region, a (U)₃₃-C(U)_n stretch, and a conserved 98-nt tail with stem-loop structure. The 3′CNU RNA contains the 295-nt 3′ coding region and a partial 3′NCR as shown. (Bottom) EMSA. The [α-³²P]UTP-labeled 3′CNUX RNA was incubated with increasing amounts (lanes 2 to 5, 0.15, 0.75, 1.5, and 3 ng, respectively) of the partially purified HCV NS5B protein or with 3 ng of the NS5B protein in the presence of various amounts of competitor RNAs as indicated (lanes 7 to 11). Reaction products were resolved in a 4% polyacrylamide gel under nondenaturing conditions. The gel was dried and subjected to autoradiography. Lanes 1 and 6 represent free 3′CNUX RNA probe to which no NS5B was added in the reaction.

FIG. 3

Preferential binding of NS5B protein to the HCV 3′CNU RNA. (Top) Structures of the HCV genomic RNA and RNA fragments used in the competition analysis. The heavy line represents the coding region of the HCV polyprotein. The 5′NCR contains a highly ordered structure 341 nt in length, and the 3′ NCR consists of the genotype-specific region, a (U)_n-C(U)_n stretch, and the conserved 98-nt tail. (Bottom) Competitive EMSA. EMSA was performed with HCV NS5B protein and the ³²P-labeled 3′CNU RNA in the presence (lanes 3 to 11) or absence (lane 2) of unlabeled competitor RNAs as indicated. Competitor RNAs used were 1- (lanes 3, 6, and 9), 5- (lanes 4, 7, and 10), and 10-fold (lanes 5, 8, and 11) molar excesses to the labeled 3′CNU RNA. Lane 1 represents free 3′CNU RNA probe to which no NS5B protein was added in the reaction. We consistently observed two bands of the 3′CNU RNA probe that should represent the same RNA fragment of different conformations. A single band was detected when the probe was analyzed on a sequencing gel containing 8 M urea (data not shown).

FIG. 4

Mapping of the binding domains of NS5B on the HCV 3′CNU RNA. (Top) Structures of the HCV RNAs used in the competitive EMSA and their characteristics in forming a complex with HCV NS5B protein. Nucleotide residues flanking the RNA termini are numbered according to the coding region of HCV polyprotein. RNA fragments that form conserved stem-loop structures (SL) are indicated by filled boxes. SL4 and SL3 contain nt 8875 to 8918 and nt 8980 to 9010, respectively, of the coding region of HCV polyprotein. SL2 contains the 3′-terminal 12 nt of the coding region plus a downstream 11 nt, and SL1 is located within the genotype-specific region (12). Binding activities are indicated by plus and minus signs. (Bottom) Competitive EMSA. The competition analysis was conducted with a gel-purified 3′CNU RNA probe and HCV NS5B protein in the presence of unlabeled 3′CN (lanes 2 to 4), 3′C (lanes 5 to 7), 3′N (lanes 8 to 10), 3′C′N (lanes 11 to 13), and a mixture of 3′C and 3′N RNA (lanes 14 to 16) at 1-, 5-, and 10-fold molar excesses to the [α-³²P]UTP-labeled 3′CNU RNA probe. Lane 1 represents the reaction in which no competitor RNA was added.

FIG. 5

Binding of HCV 3′CNU RNA to the recombinant NS5B deletion mutants. Partially purified full-length NS5A and NS5B and NS5B deletion mutants as indicated were analyzed by sodium dodecyl sulfate–15% polyacrylamide gel electrophoresis in Tricine-Tris buffer and stained with Coomassie blue (A). Equal amounts of the recombinant proteins as determined by Bio-Rad protein assay were subjected to Western blot analysis with 6× His monoclonal antibody (ClonTech) (B) and Northwestern analysis as previously described (5) with [α-³²P]UTP-labeled 3′CNU RNA in the presence of 50 mM (C) and 150 mM (D) NaCl. The control lanes represent protein lysates prepared from E. coli transformed with the pET15b plasmid. In panel C, a signal at a position beyond the molecular sizes of the recombinant NS5B proteins was detected with the control lysate. Its identity is not known. Panel E shows the structures of the NS5B recombinant proteins and summarizes their characteristics in binding 3′CNU RNA.

FIG. 6

Complex formation between NS5B deletion mutants and the 3′CNU RNA. An EMSA was performed with [α-³²P]UTP-labeled 3′CNU RNA, and equal amounts of the NS5A and NS5B recombinant proteins as indicated. The gel was dried and subjected to autoradiography.