The NS5A protein of bovine viral diarrhea virus contains an essential zinc-binding site similar to that of the hepatitis C virus NS5A protein

Abstract

The recent demonstration that the NS5A protein of hepatitis C virus (HCV) contains an unconventional zinc-binding site with the format Cx(17)CxCx(20)C and the presence of a similar sequence element in the NS5A proteins of members of the Pestivirus genus has led to the hypothesis that the NS5A protein of the pestivirus bovine viral diarrhea virus (BVDV) is a zinc-binding protein. A method for the expression and partial purification of BVDV NS5A was developed, and the partially purified protein was analyzed for zinc content by atomic absorption spectroscopy. BVDV NS5A was found to coordinate a single zinc atom per protein molecule. Mutation of any of the four cysteines of the predicted zinc-binding motif eliminated zinc coordination. Furthermore, analysis of mutations at these cysteine residues in the context of a BVDV replicon system indicated that these residues were absolutely essential for RNA replication. The recently determined crystal structure of the N-terminal zinc-binding domain of the HCV NS5A protein, combined with secondary structure predictions of the region surrounding the mapped BVDV zinc-binding region, indicates that the BVDV zinc-binding motif fits the general template Cx(22)CxCx(24)C and likely comprises a three-stranded antiparallel beta-sheet fold. These data highlight the similarities between the Hepacivirus and Pestivirus NS5A proteins and suggest that both proteins perform a not-yet-defined function in RNA replication that requires coordination of a single zinc atom.

FIG. 1.

(A) Diagrammatic comparison of HCV and BVDV NS5A proteins. The HCV NS5A protein is divided into three domains based on the presence of two low-complexity sequence blocks (designated LCS I and LCS II, gray shading) predicted to be interdomain connecting loops. The region believed to constitute domain I (amino acids 1 to 213) contains the amino-terminal membrane-anchoring helix (designated anchor, black bar), as well as a zinc ion coordination motif (CPC tripeptide and four cysteines involved in zinc binding are shown). Domain II (amino acids 250 to 342) and domain III (amino acids 356 to 447) constitute the carboxyl-terminal half of HCV NS5A. No domain-mapping information is available for BVDV NS5A. The amino-terminal region of BVDV NS5A contains a putative membrane-anchoring helix (designated Anchor, black bar) and four conserved cysteine residues (C34, C57, C59, and C84) centered around a CPC tripeptide sequence (C57, P58, and C59). The position and spacing of these cysteine residues suggest that this region of BVDV NS5A constitutes a zinc-binding site similar to that characterized for HCV NS5A. The regions trailing this potential metal coordination site are poorly characterized for BVDV. (B) Sequence alignments of Hepacivirus and Pestivirus NS5A zinc-binding site coding regions. Residues comprising the zinc-binding site previously identified in the HCV NS5A protein are conserved across a diverse range of pestiviruses (gray shading). The numbering of these residues (top of alignment) corresponds to the amino acid sequence of the NS5A protein of the NADL isolate of BVDV. Note the difference in spacing of the outermost conserved cysteine residues between HCV and the pestiviruses and the presence of the central CPC tripeptide sequence. Virus names and respective NCBI protein identifications or GenBank accession numbers: HCV-Con 1, HCV genotype 1b sequence (AJ238799); HCV-H77, HCV infectious clone 1a H77 (AF009606); GBV-B, human GB virus B (9628102); BVDV-NADL, BVDV strain NADL (P19711); BVDV-Oregon, BVDV strain Oregon (AAC40704); BVDV-Osloss, BVDV strain Osloss (AAA02769); BVDV-SD1, BVDV strain SD1 (Q01499); BVDV-CP7, BVDV strain CP7 (Q96662); BVDV-2, BVDV strain 2 (044731); Giraffe, novel pestivirus isolated from a giraffe (620053); Reindeer, novel pestivirus isolated from a reindeer (620051); CSFV-Alfort, classical swine fever virus strain Alfort (AAB50409); CSFV-Brescia, classical swine fever virus strain Brescia (AAA43843). CSFV-Riems, classical swine fever virus strain Riems (AAA86908); CSFV KC/CS, classical swine fever virus vaccine strain KC/CS (AAC98302); BDV-1, border disease virus 1 (620062); BDV-C143 (AAB60887), border disease virus strain C143; BDV-X818, border disease virus strain X818 (AAC16444). (C) Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of a representative sample of the partially purified BVDV NS5A (lane BVDV 5A) protein used in the experiments described in this report. The gel was stained with Coomassie brilliant blue R-250. The values on the left of the gel correspond to masses of the molecular size marker (lane MW) in kilodaltons.

FIG. 2.

Confirmation of lethal phenotypes of BVDV mutants by immunofluorescence microscopy. BVDV replicons constructed in the noncytopathic Jiv backbone have the ability to spontaneously revert to a cytopathic phenotype. To be certain that the zinc-binding site mutants were indeed lethal in BVDV replication (assayed by colony formation under drug selection) and not merely a reversion to a cytopathic replicon, analysis of BVDV replication, or the lack thereof, was monitored by immunofluorescence microscopy of cells transfected with these replicons. Detection of the BVDV NS3 protein (red staining) with mouse anti-NS3 monoclonal antibody 184 (19) and an anti-mouse Alexa 594 conjugate is shown. Nuclei were counterstained for clarity with Hoechst 333421 (blue staining). In all cases, the zinc-binding site mutations in replicons appeared negative for NS3 expression (C34, C57, C59, and C84). Background NS3, representing translation of the BVDV polyprotein without RNA replication, was similar in the zinc-binding site mutants and BVDV noncytopathic (JIV-) and cytopathic (NADL) replicons cured of active RNA replication by treatment with the polymerase (POL) inhibitor VP32947 (4) (right side).

FIG. 3.

Comparison of HCV and BVDV NS5A zinc-binding site structures. (A) Secondary structure content of the region surrounding the zinc-binding sites of HCV and BVDV strain NADL NS5A. Sequences and secondary structures of HCV are based on the crystal structure of the Con1 isolate (Protein Data Bank accession number 1ZH1) (51). BVDV secondary structure assignments were determined by computer prediction with the PSIPRED algorithm (35). β strands are indicated by black arrows; α helices are indicated by black boxes. Strands and helices are numbered according to the nomenclature designated in the HCV NS5A crystal structure. Conserved cysteine residues involved in zinc coordination are shaded gray. (B) Crystal structure of the HCV NS5A zinc-binding site region (51). Amino acids 36 to 85 of the NS5A protein are shown, with residues following amino acid 85 removed for clarity. The three β strands comprising an antiparallel β sheet are indicated (B1, B2, and B3). The small α helix, designated H2, is also shown. The zinc atom (yellow sphere) and conserved cysteine side chains that interact with the metal ion (highlighted in red and numbered) are shown. (C) Model of the zinc-binding region of BVDV NS5A. Theoretical topology of the zinc-binding region of BVDV created by combining the secondary structure predictions in panel A with the HCV NS5A structural information in panel B. The three predicted β strands are shown (numbered B1, B2, and B3) and arranged to generate an antiparallel β sheet similar to that seen in HCV NS5A. The locations of the zinc atom (yellow) and coordinating cysteine residues (C34, C57, C59, and C84) are shown. A putative helix (light gray cylinder, labeled H2) is shown to represent a possible helix in this region of BVDV NS5A.