Environmental Surveillance of Norovirus Genogroups I and II for Sensitive Detection of Epidemic Variants

Abstract

Sewage samples have been investigated to study the norovirus concentrations in sewage or the genotypes of noroviruses circulating in human populations. However, the statistical relationship between the concentration of the virus and the number of infected individuals and the clinical importance of genotypes or strains detected in sewage are unclear. In this study, we carried out both environmental and clinical surveillance of noroviruses for 3 years, 2013 to 2016. We performed cross-correlation analysis of the concentrations of norovirus GI or GII in sewage samples collected weekly and the reported number of gastroenteritis cases. Norovirus genotypes in sewage were also analyzed by pyrosequencing and compared with those identified in stool samples. The cross-correlation analysis found the peak coefficient (R = 0.51) at a lag of zero, indicating that the variation in the GII concentration, expressed as the log₁₀ number of copies per milliliter, was coincident with that in the gastroenteritis cases. A total of 15 norovirus genotypes and up to 8 genotypes per sample were detected in sewage, which included all of the 13 genotypes identified in the stool samples except 2. GII.4 was most frequently detected in both sample types, followed by GII.17. Phylogenetic analysis revealed that a strain belonging to the GII.17 Kawasaki 2014 lineage had been introduced into the study area in the 2012-2013 season. An increase in GI.3 cases was observed in the 2015-2016 season, and sewage monitoring identified the presence of GI.3 in the previous season (2014-2015). Our results demonstrated that monitoring of noroviruses in sewage is useful for sensitive detection of epidemic variants in human populations.IMPORTANCE We obtained statistical evidence of the relationship between the variation in the norovirus GII concentration in sewage and that of gastroenteritis cases during the 3-year study period. Sewage sample analysis by a pyrosequencing approach enabled us to understand the temporal variation in the norovirus genotypes circulating in human populations. We found that a strain closely related to the GII.17 Kawasaki 2014 lineage had been introduced into the study area at least 1 year before its appearance and identification in clinical cases. A similar pattern was observed for GI.3; cases were reported in the 2015-2016 season, and closely related strains were found in sewage in the previous season. Our observation indicates that monitoring of noroviruses in sewage is useful for the rapid detection of an epidemic and is also sensitive enough to study the molecular epidemiology of noroviruses. Applying this approach to other enteric pathogens in sewage will enhance our understanding of their ecology.

Keywords: Norovirus; massive parallel sequencing; molecular epidemiology; rapid detection of epidemic; sewage; virological surveillance.

FIG 1

Concentrations of norovirus GI and GII in sewage and numbers of gastroenteritis cases reported in the study area. The norovirus GI and GII detection limits are 1.5 and 1.6 log₁₀ copies/ml, respectively.

FIG 2

Distribution of cross-correlation coefficients between the number of gastroenteritis cases reported in the study area (A) or in Japan (B) and the concentration of norovirus GI (A-1 and B-1) or GII (A-2 and B-2) in sewage. The lag of the time that the norovirus concentration changed behind the time that the number of gastroenteritis cases changed was defined as “+.” The two horizontal lines represent the 95% confidence interval for the correlation.

FIG 3

Genotypes identified in clinical samples in each norovirus season. The period of each norovirus season is September to August, except in the 2012-2013 (November to August) and 2015-2016 (September to March) seasons.

FIG 4

Phylogenetic trees of sequences obtained from sewage and stool samples (A, GII.17; B, GI.3). The name of each sequence obtained consists of the sampling date (year-month-day), followed by S (sewage) or P (patient). The name of each reference sequence consists of the year, followed by the strain name, country, and accession number. The values at the nodes represent bootstrap rates of >70%.

FIG 4

Phylogenetic trees of sequences obtained from sewage and stool samples (A, GII.17; B, GI.3). The name of each sequence obtained consists of the sampling date (year-month-day), followed by S (sewage) or P (patient). The name of each reference sequence consists of the year, followed by the strain name, country, and accession number. The values at the nodes represent bootstrap rates of >70%.