Rotavirus RNA polymerases resolve into two phylogenetically distinct classes that differ in their mechanism of template recognition

Abstract

Rotaviruses (RVs) are segmented double-stranded RNA viruses that cause gastroenteritis in mammals and birds. Within the RV genus, eight species (RVA-RVH) have been proposed. Here, we report the first RVF and RVG sequences for the viral RNA polymerase (VP1)-encoding segments and compare them to those of other RV species. Phylogenetic analyses indicate that the VP1 RNA segments and proteins resolve into two major clades, with RVA, RVC, RVD and RVF in clade A, and RVB, RVG and RVH in clade B. Plus-strand RNA of clade A viruses, and not clade B viruses, contain a 3'-proximal UGUG cassette that serves as the VP1 recognition signal. VP1 structures for a representative of each RV species were predicted using homology modeling. Structural elements involved in interactions with the UGUG cassette were conserved among VP1 of clade A, suggesting a conserved mechanism of viral RNA recognition for these viruses.

Figure 1

Maximum likelihood phylogenetic trees comparing the nucleotide sequences of VP1 (+)RNA (left) or the deduced amino acid sequences of RdRP VP1 (right) for several RV isolates. The species of isolates are indicated. Scale bar represents 0.05 substitutions per site.

Figure 2

Comparison of predicted VP1 structures. (A) Schematic bar representation of VP1 colored by domain (N-terminal, yellow; C-terminal, pink) and subdomain (fingers, blue; palm, brick red; thumb, green). The length of each domain for RVA SA11 VP1 (upper) and each predicted domain for RVG 03V0567 VP1 (lower) is indicated. Locations of predicted insertions in RVG VP1 that are greater than three amino acids in length, based on homology modeling, are indicated with cyan bars and numbered. Open triangles indicate deletions of three or more amino acids. (B) Ribbon drawings comparing structures of the N-terminal (left), polymerase (center), and C-terminal (right) domains of RVA SA11 VP1 to those of VP1 homology models. The models included in the upper panel are of Bristol (RVC), 05V0049 (RVD), and 03V0568 (RVF) VP1, and in the lower panel is the model of 03V0567 (RVG) VP1. RVA SA11 VP1 is colored as in panel (A). Homology models are colored gray. The locations of predicted insertions in RVG VP1 greater than three amino acids in length are colored cyan and numbered as in panel (A).

Figure 3

Viral (+)RNA recognition by clade A and clade B VP1. (Upper) Close up of the template entry tunnel of RVA SA11 VP1 (colored ribbon) with bound RNA oligonucleotide (orange sticks) in two slightly different views. PDB 2R7W was used to make the images. RNA bases are labeled and numbered in the 3′ to 5′ direction. RVA SA11 VP1 subdomains are colored as in Figure 2. Side chains of VP1 residues that interact with the RNA are shown and labeled. The predicted structure of RVG 03V0567 VP1 (gray ribbon) is overlaid on the RVA structure. Parts of the structure have been hidden for clarity. (Lower) Structure-based sequence alignment of specific regions of VP1, made using PDB 2R7Q and the RVG 03V0567 homology model. RVA SA11 residues that interact with RNA, and the aligning residues in the RVG VP1 model, are colored red. The initial amino acid number in the series and the RNA nucleotide with which the RVA SA11 VP1 residues interact are indicated above the alignment. The identities of the aligning amino acids and aligning nucleotides for RVG 03V0567 are indicated below the alignment.

Figure 4

Consensus alignments for the polymerase domain of clade A and clade B VP1. (Left) Ribbon drawing of the polymerase domain of RVA SA11 VP1 (PDB 2R7W). Conserved RdRP motifs are colored light blue (motif F), pink (motif A), lavender (motif B), green (motif C), teal (motif D), and yellow (motif E), and the 'priming loop' is colored gold. A bound UGUGACC RNA oligonucleotide (yellow sticks) is shown for orientation. (Right) Alignments of the RVA SA11 VP1 polymerase domain sequence and consensus sequences for clade A (SA11, Bristol, 05V0049, and 03V0568) and clade B (Bang373, DB176, 03V0567, B219, and J19) polymerase domains. Alignments are based primarily on predicted VP1 structures and PDB 2R7Q, with regions of discrepancy aligned by hand. Tildas indicate insertions. Asterisks indicate amino acids that differ within a clade. Single-letter codes for amino acids that are conserved within a clade are in black text, and amino acids that are conserved between both clades are in red text. Fingers (blue), palm (brick red), and thumb (green) subdomains are indicated by colored bars. Amino acid sequences of RdRP motifs are highlighted with colors similar to those in the left panel and labeled.