Serology-enabled discovery of genetically diverse hepaciviruses in a new host

Abstract

Genetic and biological characterization of new hepaciviruses infecting animals contributes to our understanding of the ultimate origins of hepatitis C virus (HCV) infection in humans and dramatically enhances our ability to study its pathogenesis using tractable animal models. Animal homologs of HCV include a recently discovered canine hepacivirus (CHV) and GB virus B (GBV-B), both viruses with largely undetermined natural host ranges. Here we used a versatile serology-based approach to determine the natural host of the only known nonprimate hepacivirus (NPHV), CHV, which is also the closest phylogenetic relative of HCV. Recombinant protein expressed from the helicase domain of CHV NS3 was used as antigen in the luciferase immunoprecipitation system (LIPS) assay to screen several nonprimate animal species. Thirty-six samples from 103 horses were immunoreactive, and viral genomic RNA was present in 8 of the 36 seropositive animals and none of the seronegative animals. Complete genome sequences of these 8 genetically diverse NPHVs showed 14% (range, 6.4% to 17.2%) nucleotide sequence divergence, with most changes occurring at synonymous sites. RNA secondary structure prediction of the 383-base 5' untranslated region of NPHV was refined and extended through mapping of polymorphic sites to unpaired regions or (semi)covariant pairings. Similar approaches were adopted to delineate extensive RNA secondary structures in the coding region of the genome, predicted to form 27 regularly spaced, thermodynamically stable stem-loops. Together, these findings suggest a promising new nonprimate animal model and provide a database that will aid creation of functional NPHV cDNA clones and other novel tools for hepacivirus studies.

Fig 1

LIPS detection of robust anti-CHV NS3 antibody titers in horses. (A) Schematic of LIPS serological screening. Recombinant NS3 of CHV was genetically fused to the C terminus of Renilla luciferase (Ruc) and produced in Cos1 cells. The Ruc-NS3 protein extract was incubated with serum samples, Ruc-NS3 antibody complexes were then captured by protein A/G beads, and light units were measured. (B) LIPS detection of anti-NS3 CHV antibodies in different samples, including 80 dogs, 14 rabbits, 81 deer, 84 cows, 99 U.S. horses, and 4 pooled horse serum samples from New Zealand. The antibody titers from each sample are plotted in light units (LU) on a log₁₀ scale on the y axis, and equine samples positive by PCR are colored red. (C) Heat map analysis of equine and human samples for anti-CHV antibodies. Anti-CHV NS3-seronegative horse (n = 20) and -seropositive horse (n = 20) samples were analyzed for anti-CHV and anti-HCV NS3 antibodies. Antibody titers against the antigens were log₁₀ transformed, and the levels were then color coded as indicated by the scale on the right, where signal intensities range from high (red) to low (green). Each individual row represents the antibody titers in a single serum sample.

Fig 2

Phylogenetic analysis, genetic composition, and genome-wide divergence scanning of the eight NPHV genomes (A) Neighbor-joining trees of nucleotide sequences from different genome regions of NPHV and corresponding regions of HCV (genotypes 1a to 7a) and GBV-B. Trees were constructed from Jukes-Cantor-corrected pairwise distances calculated using the program MEGA version 5 (34); data sets were bootstrap resampled 500 times to indicate robustness of branching (values of ≥70% shown on branches). (B) Mean nucleotide pairwise distances (uncorrected, y axis) and ratios of synonymous to nonsynonymous Jukes-Cantor distances (dN/dS) between horse-derived hepaciviruses in different genome regions (red bars). These values were compared with equivalent calculations for GBV-C/HPgV (blue bars) and HCV (green bars). (C) Amino acid sequence divergence across the genome of horse-derived NPHV sequences (top plot) and comparison with HCV and GBV-C/HPgV (middle and bottom plots, respectively) using 300-base fragments increasing by 9 bases across each virus alignment (midpoint plotted on y axis, positions numbered using NXP-1-GBX2 as a reference sequence). Genome diagrams above each graph show gene boundaries using the same x axis scale as that on the divergence graph.

Fig 3

RNA structure analysis of NPHV 5′ UTR and complete genomes. (A) Predicted RNA structure for the 5′ UTR of NPHV based on minimum free energy predictions and comparison with homologous sequences of HCV and GBV-B (21). Bases were numbered using NXP-1-GBX2 as a reference sequence; stem-loops were numbered as in reference . Sequences homologous to targets of miRNA-122 (17) are indicated by heavy lines. (B and C) Secondary structure prediction for NPHV genome sequences using mean MFED differences (y axis) of 200- and 250-base fragments (30-base increment; midpoint plotted on x axis) for CHV and the 8 horse-derived hepacivirus sequences (B) and analysis of the 8 NPHV sequences by ALIFOLD using default parameters (C) (see reference for explanation of color coding). Due to restriction in the server, this figure was built as a composite of 6 separate overlapping 2,000-base fragments increasing by 1,500 bases (C).