The comparative genomics of human respiratory syncytial virus subgroups A and B: genetic variability and molecular evolutionary dynamics

Abstract

Genomic variation and related evolutionary dynamics of human respiratory syncytial virus (RSV), a common causative agent of severe lower respiratory tract infections, may affect its transmission behavior. RSV evolutionary patterns are likely to be influenced by a precarious interplay between selection favoring variants with higher replicative fitness and variants that evade host immune responses. Studying RSV genetic variation can reveal both the genes and the individual codons within these genes that are most crucial for RSV survival. In this study, we conducted genetic diversity and evolutionary rate analyses on 36 RSV subgroup B (RSV-B) whole-genome sequences. The attachment protein, G, was the most variable protein; accordingly, the G gene had a higher substitution rate than other RSV-B genes. Overall, less genetic variability was found among the available RSV-B genome sequences than among RSV-A genome sequences in a comparable sample. The mean substitution rates of the two subgroups were, however, similar (for subgroup A, 6.47 × 10(-4) substitutions/site/year [95% credible interval {CI 95%}, 5.56 × 10(-4) to 7.38 × 10(-4)]; for subgroup B, 7.76 × 10(-4) substitutions/site/year [CI 95%, 6.89 × 10(-4) to 8.58 × 10(-4)]), with the time to their most recent common ancestors (TMRCAs) being much lower for RSV-B (19 years) than for RSV-A (46.8 years). The more recent RSV-B TMRCA is apparently the result of a genetic bottleneck that, over longer time scales, is still compatible with neutral population dynamics. Whereas the immunogenic G protein seems to require high substitution rates to ensure immune evasion, strong purifying selection in conserved proteins such as the fusion protein and nucleocapsid protein is likely essential to preserve RSV viability.

Fig 1

RSV protein sequence variability. The number of substitutions per site for RSV-A (black bars) and RSV-B (red bars) and the protein sequence variability (percent numbers in red) in each RSV-B protein were calculated per strain relative to the consensus. The numbers within parentheses indicate percentages of protein sequence variability between the consensus sequences of type A and type B RSV strains.

Fig 2

Substitution hot spots in specific RSV G protein domains and consequences for glycosylation. (A) Schematic representation of the RSV G protein and its specific domains: the transmembrane domain (TM), heparin binding domain (HBD), N-terminal cytosolic domain (Cytosol), C-terminal ectodomain (Ecto), immunogenic domain (ID; aa 159 to 198), CX3C chemokine motif (CX3C; aa 182 to 186), and the region homologous to the fourth subdomain of TNFr (TNFr; aa 171–186). Dashed lines represent mucin-like regions. (B and C) The percentage of strains with predicted sites for N-glycosylation (B) and O-glycosylation (C) within the G protein at a certain amino acid position for both RSV-A (black; 37 strains) and RSV-B (red; 36 strains).

Fig 3

Amino acid sequence divergence in the G protein of RSV-B strains. Insertions (gray, underlined), deletions (−), and variable C-terminal ends are indicated.

Fig 4

Plot of the root-to-tip divergence as a function of sampling time for the RSV-A and RSV-B genomes. RSV-A is indicated by black triangles, and RSV-B is indicated by red squares; the corresponding regression lines are plotted in the same colors.

Fig 5

RSV-B whole-genome-based phylogeny. The distribution of Dutch-Belgian strains (blue) and Milwaukee strains (green) is indicated. The node bars depict the credibility intervals for nodes showing a posterior probability support > 95% (blue) or < 95% (yellow).

Fig 6

Bayesian skyline plot. The estimated change in effective population size over time for both the full-genome (blue) and G gene (purple) data sets is indicated. The thick lines represent the mean estimate, whereas the transparent areas represent the 95% highest-posterior-density intervals.

Fig 7

Comparison of RSV-A and RSV-B evolutionary rate partitions. The mean rate of evolutionary changes indicated by the numbers of substitutions per site for each year was estimated per gene and for the combined noncoding sequence parts for the RSV-A (black triangles) and RSV-B (red squares) data sets.

Fig 8

RSV nucleotide sequence variability in the whole genome. The number of substitutions per site for the RSV-A (black bars) and RSV-B (red bars) genomes and the nucleotide sequence variability (percent numbers in red) in each RSV-B gene were calculated per strain relative to the consensus. The numbers within parentheses indicate percentages of nucleotide sequence variability between the consensus of type A and type B RSV strains.