Rapid evolution of pandemic noroviruses of the GII.4 lineage

Abstract

Over the last fifteen years there have been five pandemics of norovirus (NoV) associated gastroenteritis, and the period of stasis between each pandemic has been progressively shortening. NoV is classified into five genogroups, which can be further classified into 25 or more different human NoV genotypes; however, only one, genogroup II genotype 4 (GII.4), is associated with pandemics. Hence, GII.4 viruses have both a higher frequency in the host population and greater epidemiological fitness. The aim of this study was to investigate if the accuracy and rate of replication are contributing to the increased epidemiological fitness of the GII.4 strains. The replication and mutation rates were determined using in vitro RNA dependent RNA polymerase (RdRp) assays, and rates of evolution were determined by bioinformatics. GII.4 strains were compared to the second most reported genotype, recombinant GII.b/GII.3, the rarely detected GII.3 and GII.7 and as a control, hepatitis C virus (HCV). The predominant GII.4 strains had a higher mutation rate and rate of evolution compared to the less frequently detected GII.b, GII.3 and GII.7 strains. Furthermore, the GII.4 lineage had on average a 1.7-fold higher rate of evolution within the capsid sequence and a greater number of non-synonymous changes compared to other NoVs, supporting the theory that it is undergoing antigenic drift at a faster rate. Interestingly, the non-synonymous mutations for all three NoV genotypes were localised to common structural residues in the capsid, indicating that these sites are likely to be under immune selection. This study supports the hypothesis that the ability of the virus to generate genetic diversity is vital for viral fitness.

Figure 1. Comparison of the amino acid sequence of the six NoV RdRps used in this study to other representative GII strains.

The top six sequences are the RdRps characterised in this study (GII.4 US95_96, rGII.4, GII.4 2006a, GII.4 2006b, GII.7 and rGII.b). The remaining 19 strains are reference strains and are named according to their GII genotype followed by their year of isolation (GII.4 strains only) and then their strain name. The alignment illustrated that a K291T substitution first appeared in the GII.4 lineage after 2001 and was unique for the GII.4 pandemic strains.

Figure 2. The effect of mutations at residue 291 on NoV GII.4 RdRp kinetic activity.

A K291T mutation in wildtype (W+) NoV/US95_96-RdRp lead to a 20.2% increase in RdRp activity. Open bars show US95_96^291K-RdRp: 0.168 s⁻¹±0.018 [n = 6], US95_96^291T-RdRp: 0.202 s⁻¹±0.019 [n = 6)], p-value = 0.008. Whereas, a mutation T291K lead to a 22.2% reduction in activity for NoV/2006a-RdRp compared to the wildtype K291T. Hashed bars show 2006a^291T-RdRp: 0.351 s⁻¹±0.037 [n = 3], 2006a^291K-RdRp: 0.273 s⁻¹±0.020 [n = 4], p-value = 0.029.

Figure 3. Phylogenetic analysis of the amino acid sequence of the P2 domain from GII.4 (A), GII.b/GII.3 and GII.3 (B) and GII.7 (C) strains circulating between 1987 and 2008.

The phylogenetic tree was generated using the Neighbour-Joining method by comparison of the P2 amino acid sequence (152 aa for GII.4 and 7, and 160 aa for GII.3) obtained for 84 NoV strains. GenBank accession numbers are included in the figure and follow the strain name. The percentage bootstrap values in which the major groupings were observed among 1000 replicates are indicated. The branch lengths are proportional to the evolutionary distance between sequences and the distance scale, in nucleotide substitutions per position, is shown.

Figure 4. Rate of evolution for the GII.4, GII.7, GII.b/GII.3 and GII.3 strains.

The rate of evolution for each genotype was determined by calculating the number of nucleotide substitutions in ORF2 compared to the oldest strain in each lineage. The number of changes was then plotted against the year of that strains detection. The rate of evolution was equivalent to the gradient of the line (GII.4 = 6.30±0.39, r² = 0.84; GII.b = 4.03±0.80, r² = 0.68; GII.3 = 3.09±0.95, r² = 0.49; GII.7 = 3.82±0.25, r² = 0.99) divided by the length of the capsid gene (1623 bp for GII.4 and GII.7, and 1647 bp for GII.3 and GII.b/GII.3).

Figure 5. Hypervariable residues in GII.3, GII.4 and GII.7 are localised to common regions on the surface of the capsid P2 domain.

The structure of the GII.4 P1 and P2 domain was solved previously (PDB ID 2OBS [52]) while the structure of the P domain was predicted for GII.3 and GII.7 in this study. The location of the hypervariable residues are indicated numerically and are coloured pink for all three genotypes. Residues occupying similar regions are depicted by the same coloured circle. The previously published hypervariable residues in the P2 domain of GII.4 were localised to six main regions on the surface of the P2 domain ,. GII.3 had hypervariable residues in four of these regions and GII.7 had hypervariable residues in two of these regions.