New complete genome sequences of human rhinoviruses shed light on their phylogeny and genomic features

Abstract

Background: Human rhinoviruses (HRV), the most frequent cause of respiratory infections, include 99 different serotypes segregating into two species, A and B. Rhinoviruses share extensive genomic sequence similarity with enteroviruses and both are part of the picornavirus family. Nevertheless they differ significantly at the phenotypic level. The lack of HRV full-length genome sequences and the absence of analysis comparing picornaviruses at the whole genome level limit our knowledge of the genomic features supporting these differences.

Results: Here we report complete genome sequences of 12 HRV-A and HRV-B serotypes, more than doubling the current number of available HRV sequences. The whole-genome maximum-likelihood phylogenetic analysis suggests that HRV-B and human enteroviruses (HEV) diverged from the last common ancestor after their separation from HRV-A. On the other hand, compared to HEV, HRV-B are more related to HRV-A in the capsid and 3B-C regions. We also identified the presence of a 2C cis-acting replication element (cre) in HRV-B that is not present in HRV-A, and that had been previously characterized only in HEV. In contrast to HEV viruses, HRV-A and HRV-B share also markedly lower GC content along the whole genome length.

Conclusion: Our findings provide basis to speculate about both the biological similarities and the differences (e.g. tissue tropism, temperature adaptation or acid lability) of these three groups of viruses.

Figure 1

Whole-polyprotein phylogenetic tree. Whole-polyprotein, maximum likelihood phylogenetic tree shows a closer relation between HRV-B and HEV than between HRV-A and HRV-B. The figure indicates the percentage of bootstraps (out of 1000) that supports the corresponding clade. The sequence of simian picornavirus 1 (SV-2) was used as an outgroup. The branch lengths are measured in substitutions per site.

Figure 2

Protein and amino-acid similarity comparison between HRV-A, HRV-B and HEV. A) Schematic representation of HEV and HRV genome organization showing boundaries of encoded proteins. B) Protein similarity comparison. For each protein, the following is shown: – top row: a simplified tree representation of the relationships between HRV-A (A), HRV-B (B) and HEV (E), according to the corresponding ML tree (see additional file 1). Three cases are possible: HRV-B closer to HRV-A; HRV-B closer to HEV; and undecided (none of the above clearly more likely than the other). – bottom row: an all-versus-all sequence identity matrix (darker colorus indicate higher identity percentage). The similarity values are given in additional file 2. C) Close-up of the identity matrix for VP1. D) Bootscanning. The whole polyprotein alignment was divided into windows of 200 aa starting every 50 aa, and a 100-bootstrap ML tree was computed on each window. The black curve indicates the degree of support (as a percentage of bootstrap replicates) for either the "HRV-B closer to HRV-A" (upper half) or "HRV-B closer to HEV" (lower half) topology at each position along the whole genome (see Methods for details). The background colour reflects the log likelihood of the tree at each position which is a measure of overall confidence in the tree. Darker colour indicates higher confidence.

Figure 3

Alignments and conserved secondary structures for cis-acting 2C replication elements conserved within HRV-B and HEV. A) Multiple sequence alignment across all considered genomes that shows consensus secondary RNA structure (in dot bracket format, see first row); sequences are colour-coded according to RNA structure conservation; the sequence conservation profile for each group is shown in grey bars beneath the alignments. B) Secondary structure of the conserved cre 2C, colour-coded according to the different types of base pairs in the corresponding alignment columns. The more different the types of base pairs existing for two pairing alignment columns, the more evolutionary conservation of the structure (cp. compensatory and consistent mutations).

Figure 4

Local GC composition of HRV-A, HRV-B, and HEV. Average GC percentage computed over a sliding window of 600 nt and a step of 10 nt along whole-genome multiple alignments of HRV-A, HRV-B, and HEV, respectively (thick lines). The shaded areas represent one standard deviation above and below the average.