Rapid emergence and predominance of a broadly recognizing and fast-evolving norovirus GII.17 variant in late 2014

Abstract

Norovirus genogroup II genotype 4 (GII.4) has been the predominant cause of viral gastroenteritis since 1996. Here we show that during the winter of 2014-2015, an emergent variant of a previously rare norovirus GII.17 genotype, Kawasaki 2014, predominated in Hong Kong and outcompeted contemporary GII.4 Sydney 2012 in hospitalized cases. GII.17 cases were significantly older than GII.4 cases. Root-to-tip and Bayesian BEAST analyses estimate GII.17 viral protein 1 (VP1) evolves one order of magnitude faster than GII.4 VP1. Residue substitutions and insertion occur in four of five inferred antigenic epitopes, suggesting immune evasion. Sequential GII.4-GII.17 infections are noted, implicating a lack of cross-protection. Virus bound to saliva of secretor histo-blood groups A, B and O, indicating broad susceptibility. This fast-evolving, broadly recognizing and probably immune-escaped emergent GII.17 variant causes severe gastroenteritis and hospitalization across all age groups, including populations who were previously less vulnerable to GII.4 variants; therefore, the global spread of GII.17 Kawasaki 2014 needs to be monitored.

Figure 1. High activity of norovirus GII.17 during the 2014–2015 winter in Hong Kong.

(a) Temporal distribution of hospitalized norovirus GII.4 and GII.17 cases during the study period from March 2014 to May 2015. Black arrow denotes first detection (27 September 2014) of the novel GII.17 Kawasaki 2014. (b) Age distribution of norovirus GII.4 (n=163) and GII.17 (n=128) cases. Red horizontal lines and purple rectangular shades indicate medians and interquartile ranges, respectively. ****P<0.0001 (Mann–Whitney U-test).

Figure 2. Emergence of norovirus GII.17 Kawasaki 2014 during the 2014–2015 winter in Hong Kong.

Maximum likelihood phylogenetic inference of complete and near complete VP1 nucleotide sequences was performed. Gaps in alignment were neglected in the analysis and the final data set contained 1,586 nucleotide positions. Shown are trees with the highest log likelihood. Statistical evaluation was performed by 1,000 bootstrap replications and percentages of clustering (≥80%) are shown at nodes. Magenta and blue squares denote GII.17 Kawasaki 2014 and earlier GII.17 sequences obtained in this study, respectively. Green triangle denotes the first case of GII.17 Kawasaki 2014 in Hong Kong. Purple circles indicate cases selected for saliva-binding assay. Scale bars indicate the number of substitutions per site. Sequence nomenclature: year–month of collection followed by unique sequence identifier (for GII.17 sequences acquired in this study); GenBank accession number followed by unique sequence identifier, country of origin and year of collection (for GII.17 sequences downloaded from GenBank). Country/region abbreviations: AU, Australia; CH, Switzerland; CN, China; GF, French Guiana; JP, Japan; NP, Nepal; TW, Taiwan; US, the United States of America.

Figure 3. Genetic signatures in inferred antigenic epitopes of norovirus GII.17 Kawasaki 2014.

(a) Time-ordered SimPlot analysis of all complete and near complete norovirus GII.17 VP1 nucleotide sequences obtained in this study and GenBank. (b) Three-dimensional VP1 structure of the novel GII.17 Kawasaki 2014 predicted by homologous modelling based on the crystal structure of a GII.10 strain (PDB number 3V7A). Surface-exposed mutations and insertions, compared with other earlier local and regional GII.17 VP1 sequences collected in 2013–2014, are coloured in purple and magenta, respectively. (c) List of amino-acid changes in the P2 domain of the novel GII.17 Kawasaki 2014 compared with other earlier local and regional GII.17 VP1 sequences collected in 2013–2014. Surface-exposed mutations and insertions are highlighted as in b. Asterisk denotes residue under episodic diversifying (positive) selection. Number sign denotes residue insertion in the histo-blood group antigens binding interface. Parentheses refer to ‘in the vicinity of' which is defined as 1 to 2 residues adjacent to antigenic epitopes inferred from those of GII.4 using multiple sequence alignment. Residues are numbered with reference to the novel GII.17 Kawasaki 2014. JP, Japan; nt, nucleotide.

Figure 4. Rapid evolution of norovirus GII.17 VP1.

(a) Pairwise GII.17 VP1 nucleotide and amino-acid distance comparison. (b) Root-to-tip regression analysis to estimate GII.17 VP1 nucleotide substitution rates without a molecular clock assumption. Magenta circles indicate the novel GII.17 Kawasaki 2014. Green and blue dotted lines denote best fit lines for GII.17 from 1978 to 2002 and 2002 to 2015, respectively. Grey dotted lines denote best fit lines for both GII.4 and GII.17 from 1970s to 2015. Substitution rates are expressed as the number of nucleotide substitutions per site per year. Complete and near complete GII.17 VP1 sequences were used. GII.4 VP1 sequences used: CHDC5191/USA/1974 (GenBank accession number: JX023286; variant: ancestral); CHDC2094/USA/1974 (FJ537135; ancestral); calicivirus/USA/1995 (AJ004864; US-95/96); Farmington Hills/USA/2004 (JQ478408; Famington Hills 2002); Hunter 284E/AU/2004 (DQ078794; Hunter 2004); Yerseke38/NL/2006 (EF126963; Yerseke 2006); Sakai2/JP/2006 (AB447448; Asia 2003); Kumanoto4/JP/2006 (AB447462; Den Haag 2006); Apeldoorn317/NL/2007 (AB445395; Apeldoorn 2007); New Orleans/USA/2010 (JN595867; New Orleans 2009); NSW0514/AU/2012 (JX459908; Sydney 2012); CUHK-NS-264/HKG/2014 (JX459908; Sydney 2012); CUHK-NS-508/HKG/2015 (KT780371; Sydney 2012). (c) Upper panel. Number of accumulated VP1 nucleotide differences of local GII.17 Kawasaki 2014 compared with the first strain (magenta arrow) detected on 27 September 2014 in Hong Kong against time (in weeks) since first emergence. Lower panel. Number of accumulated VP1 nucleotide differences of local GII.4 Sydney 2012 collected between March 2014 and May 2015 compared with the earliest local strain NS-264 (purple arrow) collected on 23 March 2014 during the study period. Magenta and purple lines denote the LOWESS fit-lines for GII.17 Kawasaki 2014 and GII4 Sydney 2012, respectively. The earliest local GII.17 Kawasaki 2014 (NS-405) and GII.4 Sydney 2012 (NS-264) strains were excluded in LOWESS data fitting and Spearman's correlation calculation.

Figure 5. Broad saliva-binding profile of norovirus GII.17 Kawasaki 2014.

Unfiltered stool suspensions were used. Virus binding was made using in-house saliva-coated wells, followed by norovirus detection with a commercial diagnostic kit. Testing was performed using saliva of two individuals for each secretor blood group A, B and O and weak secretor. Dashed horizontal line indicates lower detection limit of the assay. HBGA, histo-blood group antigen; Neg, negative controls; NS, not significant; OD, optical density; Se, secretor; Se^w, weak secretor. *P<0.05 (Mann–Whitney U-test).