Prevalence and Genotypes of Human Noroviruses in Tropical Urban Surface Waters and Clinical Samples in Singapore

Abstract

The prevalence and genotypes of norovirus genogroup I (GI) and GII in tropical urban catchment waters and an estuarine bay were studied. A comparative analysis was performed with environmental isolates of noroviruses and concurrently identified clinical isolates in Singapore during gastroenteritis outbreaks between August 2006 to January 2007. Noroviruses in environmental water samples were concentrated by using ultrafiltration techniques and then analyzed by reverse transcription-seminested PCR assay targeting the partial capsid region of noroviruses and DNA sequencing. Among the 60 water samples collected, noroviruses were detected in 43 (71.7%) of these samples. Of these 43 norovirus-positive samples, the coexistence of both GI and GII strains was identified in 23 (53.5%) water samples. The phylogenetic analysis revealed multiple genotypes of noroviruses GI and GII in environmental water samples. GI and GII strains were clustered into seven and nine (including two unclassified) genotypes, respectively. The major norovirus genotypes in environmental water samples were GI/2 and GI/4 and GII/4. Genotyping of the 21 norovirus-positive clinical samples showed that all of the strains belonged to the GII/4 cluster. The environmental and clinical norovirus GII/4 isolates showed high levels of nucleotide sequence identity to each other and to the novel GII/4 variant associated with global epidemics of gastroenteritis during 2006. This study suggests the emergence and circulation of multiple novel norovirus GI and GII genotypes in water environments. Further comprehensive surveillance of water environments for noroviruses and routine clinical reporting is warranted.

Noroviruses have been identified as etiologic agents of acute gastroenteritis across all age groups worldwide. Phylogenetic characterization of sequence information of the genes encoding the viral RNA-dependent RNA polymerase (RdRp) and the capsid protein divides the norovirus genus into five genogroups (47). Each genogroup can be further divided into different genotypes, of which genogroup I (GI) and GII genotypes are mostly human pathogens. Noroviruses are highly infectious and transmitted primarily via the fecal-oral route. Food-borne outbreaks of acute gastroenteritis associated with noroviruses have been frequently reported worldwide, including Singapore (17, 24, 29, 45, 46). There have been a number of documented waterborne outbreaks of noroviruses originating from contaminated drinking water (10, 19, 26, 30, 33) and recreational waters (7, 13). However, the occurrence of noroviruses in tropical urban water catchments and recreational bays has not been widely studied. Noroviruses are included in the latest U.S. Environmental Protection Agency's contaminant candidate list (41), a list of emerging contaminants that may pose a public health risk in water environments.

Human noroviruses cannot be cultivated in routine cell culture or in animal models (6). However, developments in molecular techniques in the last two decades have facilitated their detection in clinical and environmental samples (3). Reverse transcription-PCR (RT-PCR) is currently the most widely used assay for detection of noroviruses in environmental water (18). Moreover, this method coupled with nucleotide sequencing techniques is able to gather valuable information on the norovirus genotypes occurring in the environment, thus providing epidemiological information of norovirus infections in the community. The RT-PCR primers that target the viral RdRp gene in open reading frame 1 (ORF1) or capsid gene in ORF2 have been designed to detect and genotype various norovirus strains (20, 43, 44). The concentration of noroviruses from large volumes of water is necessary due to the low ambient concentrations of viruses in environmental waters (9, 21, 23). Noroviruses have been detected in wastewater (40, 42) and drinking water and surface waters (8, 23, 40), but reports of genotyping of noroviruses from environmental waters are limited (21, 42).

Noroviruses in the urban environment may be transported by stormwater runoff, combined and sanitary sewer overflows, and discharge of wastewater treatment plant effluent (2). The significance of stormwater runoff in urban catchments and waterways has received increased attention in recent years. Public health and recreational impacts such as the incidence of waterborne diseases related to recreational water contact and contamination of drinking water supplies are issues of concern.

Little is known about the occurrence of noroviruses in the tropical aquatic environment, since most studies have been conducted in temperate and cold regions. To date, there have been no reports of epidemiological data and molecular characterization of noroviruses in Singapore, and therefore, seasonality is unknown. The only reported study of norovirus activity in the country were a series of 14 reported outbreaks of norovirus gastroenteritis occurring from December 2003 to January 2004, involving 350 people, and these were linked to oyster consumption (29). Singapore General Hospital has been performing norovirus PCR as a routine diagnostic test since 2004 and has observed that increased norovirus activity and outbreaks tend to occur in the second half of the year from July to January (unpublished data). This period coincides with the monsoon seasons in Singapore, which are from July to September and November to January.

We investigated here the occurrence of norovirus GI and GII in surface waters from urban catchments and the receiving estuarine bay. DNA sequencing and phylogenetic analysis of a partial capsid gene were used to characterize the norovirus isolates. Norovirus strains obtained from environmental waters were compared to clinical strains collected from human stools and rectal swabs in the same geographic region during gastroenteritis outbreaks.

MATERIALS AND METHODS

Environmental water samples.

Between June 2006 and June 2007, water samples were collected from major rivers (upstream and downstream) and canals receiving stormwater runoff located in the central part of Singapore city (Fig. 1). All sampling sites were located in highly urban settings. The estuarine bay which is fed by the discharges from these waterways was also sampled from March to June 2007. Ten-liter (rivers and canals) and 20-liter (estuarine bay) water samples were collected with autoclaved and sample-rinsed carboys and transported to the laboratory on ice for immediate processing.

FIG. 1.

Map of the water catchment and estuarine bay located in the central of Singapore city showing sampling sites.

Concentration of noroviruses from water samples by TFF.

Noroviruses from water samples were concentrated by using tangential flow filtration (TFF) system with a 30-kDa membrane cassette (Pall Corp.). Water samples were prefiltered by using an AcroPak 200 capsule filter with a 0.2-μm-pore-size membrane (Pall Corp.). The TFF system was assembled as recommended by the manufacturer and connected to a peristaltic pump (model 77410-05; Cole Parmer, Chicago, IL). The system was sanitized with a 0.1 N sodium hydroxide solution and thoroughly flushed with sterile ultrapure water. The water sample was processed until the sample volume remained approximately 200 to 300 ml in the carboy. To enhance the recovery of viruses, elution was performed after the concentration step by the recirculation of sterile glycine (100 ml, 0.05 M, pH 7.0) through the system for 5 min. After completion of each water sample, the system was sanitized and flushed with sterile ultrapure water as recommended by the manufacturer.

A secondary concentration of water samples was done by an Amicon Ultra centrifugal filter device (Millipore). Briefly, 60 to 120 ml of the TFF-concentrated sample was loaded to the filter device and centrifuged at 4,000 × g for approximately 10 to 30 min. The final volume of each concentrated water sample was approximately 500 μl. The catchment and estuarine bay water samples were 3000- to 6000-fold and 6000- to 12 000-fold concentrated, respectively, before noroviruses isolation.

RNA extraction, PCR amplification, cloning, and sequencing for noroviruses in water samples.

Viral RNA was extracted from 140 μl of the concentrated sample by using a QIAamp Viral RNA minikit (Qiagen, Germany) according to the manufacturer's instructions. Viral RNA was eluted in 80 μl of DNase-, RNase-free water and stored at −80°C until the seminested RT-PCR was performed.

Amplification of the capsid gene of noroviruses was performed as previously described (40) with minor modification. cDNA was synthesized at 42°C for 60 min in a 10-μl RT reaction as follow: 5 μl of template RNA, 1.5 mM MgCl₂, 200 μM concentrations of each deoxynucleoside triphosphate, 0.5 μl of ImProm-II reverse transcriptase, 10 U of RNase inhibitor (Promega Corp., Madison, WI), and 1 μM concentrations of primer G1-SKR (5′-CCAACCCARCCATTRTACA-3′) for GI and primer G2-SKR (5′-CCRCCNGCATRHCCRTTRTACAT-3′) for GII. Then, 10 μl of the cDNA sample was mixed with 1× PCR buffer, 1.5 mM MgCl₂, 200 μM concentrations of each deoxynucleoside triphosphate, 0.2 μM concentrations of primers COG1F (5′-CGYTGGATGCGNTTYCATGA-3′) and G1-SKR for GI and primers COG2F (5′-CARGARBCNATGTTYAGRTGGATGAG-3′) and G2-SKR for GII, and 1.25 U of AmpliTaq Gold DNA polymerase (Applied Biosystems). After initial incubation at 95°C for 10 min, PCR was performed for 40 cycles (94°C, 1 min; 50°C, 1 min; and 72°C, 2 min), with a final extension step at 72°C for 15 min. At the end of the amplification, 2 μl of the first-round PCR product was added to a seminested PCR mix containing the same reagents as the first round of PCR but with the primers G1SKF (5′-CTGCCCGAATTYGTAAATGA-3′) and G1SKR for GI and the primers G2SKF (5′-CNTGGGAGGGCGATCGCAA-3′) and G2SKR for GII. The amplification products were analyzed on 1.7% agarose gels and visualized under UV illumination after being stained with ethidium bromide.

Standard precautions were applied when conducting the seminested RT-PCR, such as UV sterilization of PCR equipment and the working environment, using aerosol-resistant tips, separate locations for sample preparation, and amplification. Negative controls using PCR-grade water were included in each reaction set. Diluted sewage samples (positive with noroviruses by RT-PCR and DNA sequencing) were included as positive controls.

The seminested PCR products were sequenced directly from both directions with the ABI BigDye terminator cycle sequencing kit and a Prism 3100 genetic analyzer. When it was impossible to obtain the sequence because of the presence of multiple virus strains, PCR products were cloned into the pGEM-T Vector System II (Promega Corp., Madison, WI) and the construct was transformed into JM109 competent cells. At least three clones were randomly selected and sequenced.

Clinical samples.

Stool samples and rectal swabs from several gastroenteritis outbreaks in Singapore between August 2006 and January 2007 were collected and analyzed at the Molecular Laboratory, Department of Pathology, Singapore General Hospital. In addition, stool samples from sporadic gastroenteritis cases during this period were also included in the study.

RNA extraction, PCR amplification, and sequencing for noroviruses in clinical samples.

RNA from stool samples and rectal swabs was extracted by using a QIAamp stool minikit (Qiagen, Germany) and a QIAamp viral RNA minikit (Qiagen), respectively, according to the manufacturer's instructions. Extracted RNA was first tested by real-time TaqMan RT-PCR for noroviruses and to determine genogroup (GI or GII). This was performed on the Roche Lightcycler 1.2 (Roche Diagnostics GmbH, Mannheim, Germany) using the protocol described by Kageyama et al. (16) with modifications. These primers target the ORF1-ORF2 junction. For GI PCR, 4 μl of extracted RNA was added to 6 μl of PCR amplification mix, comprising 3.75 μl of Roche Lightcycler RNA Master HybProbe (Roche Diagnostics), 0.75 μl of manganese acetate, 4 pmol each of the primers COG1F and COG1R (5′-CTTAGACGCCATCATCATTYAC-3′), 2 pmol each of the probes RING1(a)-TP (6-FAM-AGATYGCGATCYCCTGTCCA-TAMRA) and RING1(b)-TP (6-FAM-AGATCGCGGTCTCCTGTCCA-TAMRA), and 0.3 μl of nuclease-free water. For GII PCR, 4 μl of extracted RNA was added to 6 μl of PCR amplification mix, comprising 3.75 μl of Roche Lightcycler RNA Master HybProbe (Roche Diagnostics), 0.75 μl of manganese acetate, 3 pmol each of the primers COG2F and COG2R (5′-TCGACGCCATCTTCATTCACA-3′), 1.75 pmol of the probe RING2-TP (6-FAM-TGGGAGGGCGATCGCAATCT-TAMRA), and 0.575 μl of nuclease-free water. Positive and negative controls were used for each run. In addition, for each sample, an inhibition control was set up by adding 4 μl of the RNA extract, 1 μl of norovirus GI- or GII-positive control and 6 μl of the amplification mix. The thermal cycling protocols for both GI and GII PCR were as follows: 20 min at 61°C for RT and then 2 min at 95°C for initial denaturation, followed by 60 cycles of three steps consisting of 5 s at 95°C, 12 s at 56°C, and 20 s at 72°C, with a cooling step at 40°C for 30 s.

Extracted RNA that was positive for norovirus was subsequently amplified as previously described (17), targeting the capsid gene. This yielded amplicons of ∼468 bp in size, which were then sequenced using the ABI Avant 3100 (Applied Biosystems).

Phylogenetic analysis.

Nucleotide sequences were edited and aligned by using the BioEdit sequence alignment editor (http://www.mbio.ncsu.edu/BioEdit/bioedit.html). The sequences were compared to norovirus prototypes and to other reference sequences available in the GenBank database. Phylogenetic distances were determined by the Kimura two-parameter method using the DNADIST program of the PHYLIP package, version 3.67 (http://evolution.genetics.washington.edu/phylip.html). A phylogenetic tree was generated by the NEIGHBOR program. The robustness of the grouping was determined by bootstrap resampling of the multiple sequence alignments (100 sets) with the programs SEQBOOT, DNADIST, NEIGHBOR, and CONSENSE. The output graphics of the trees were created with the TREEVIEW package (32). Bootstrap values of 70 or higher were considered statistically significant for the grouping.

Nucleotide sequence accession numbers.

The original nucleotide sequences of the environmental and clinical norovirus isolates described here have been deposited in the GenBank database under the following accession numbers: FJ788285 to FJ788350 for norovirus GII and FJ788351 to FJ788386 for norovirus GI.

RESULTS

Occurrence and distribution of noroviruses in urban surface water.

A total of 60 water samples were analyzed for noroviruses by using the seminested RT-PCR assay. Noroviruses were detected in 43 (71.7%) of these samples (Table 1). Of these 43 norovirus-positive samples, 4 (9.3%) samples contained only GI strains and 16 (37.2%) samples contained only GII strains. The coexistence of both GI and GII strains was identified in 23 (53.5%) water samples. All norovirus-positive environmental samples were further analyzed by sequencing to determine the genotypes.

TABLE 1.

Quantity of water sample collected at each sampling site and characteristics of norovirus-positive samples

Water source	Total no. of samples collected	Norovirus-positive samples
		Sample code for sequencinga	Genogroup(s)b	Genotype(s)
River 1
Upstream (R1U)	8	R1U-150806	GII*	GII/4, 12
		R1U-031006	GI + GII	GI/1, GII/4
		R1U-011106	GII	GII/4
		R1U-201106	GII	GII/4
		R1U-111206	GI* + GII	GI/4, GII/4
Downstream (R1D)	9	R1D-220606	GII*	GII/3
		R1D-150806	GII	GII/4
		R1D-011106	GI	GI/4
		R1D-201106	GI + GII	GI/2, GII/4
		R1D-111206	GI + GII	GI/4, GII/4
		R1D-271206	GII	GII/14
River 2
Upstream (R2U)	11	R2U-220606	GI*	GI/4, 8
		R2U-240706	GII	GII/3
		R2U-150806	GI + GII*	GI/4, GII/4
		R2U-031006	GI + GII	GI/2, GII/4
		R2U-011106	GII	GII/16
		R2U-201106	GI* + GII*	GI/4, GII/4 (2 variants), 6
		R2U-111206	GI* + GII*	GI/2 (2 variants), GII/6 (2 variants), 14
		R2U-271206	GII	GII/4
Downstream (R2D)	9	R2D-220606	GI	GI/8
		R2D-240706	GI + GII*	GI/10, GII/3, UG2
		R2D-150806	GI + GII	GI/4, GII/4
		R2D-031006	GII	GII/4
		R2D-011106	GI* + GII	GI/2, GII/16
		R2D-201106	GII	GII/4
		R2D-111206	GI* + GII	GI/2, GII/4
River 3
Upstream (R3U)	2	R3U-251006	GI* + GII	GI/2 (2 variants), 4, 8 (2 variants), GII/16
		R3U-200607	GII	GII/12
Downstream (R3D)	2	R3D-251006	GI + GII*	GI/2, GII/16
Canal 1 (C1)	11	C1-220606	GI + GII*	GI/4, GII/UG1
		C1-040706	GII	GII/3
		C1-180706	GII*	GII/3
		C1-240706	GI* + GII	GI/4, 10, GII/3
		C1-150806	GI + GII	GI/1, GII/4
		C1-011106	GI* + GII	GI/2, 14, GII/4
		C1-201106	GI* + GII	GI/2, GII/4
		C1-111206	GI + GII	GI/2, GII/4
		C1-271206	GI + GII	GI/2, GII/4
Canal 2 (C2)	1	C2-230507	GI + GII	GI/10, GII/16
Canal 3 (C3)	1	C3-040407	GII	GII/13
Estuarine bay (E)	6	E-140307	GI* + GII*	GI/3, GII/4
		E-110407	GI*	GI/2 (2 variants)
		E-090507	GII*	GII/4

^aThe sample code for sequencing is defined as the sampling location followed by the day, month, and year of isolation.

^b*, norovirus genotypes were determined by cloning and sequencing.

In order to test the possible PCR inhibition in concentrated environmental samples, RNA extracted from norovirus-negative samples was further diluted 10- and 100-fold, and the assay was repeated. Of the 17 norovirus-negative environmental samples, only one sample (R1D-271206) was positive for norovirus GII after 10-fold dilution of extracted RNA.

Noroviruses in clinical samples during gastroenteritis outbreaks.

A total of 120 stool and rectal swabs were received during the study period. There were eight gastroenteritis outbreaks attributed to noroviruses. The outbreaks were from five nursing homes, two schools, and one hospital (a mental health institution). Thirty-five samples were positive for GII while none was positive for GI. Thirty samples positive for GII were selected for sequencing and of these, twenty-one were sequenced successfully (Table 2).

TABLE 2.

Characteristics of sequenced norovirus-positive clinical samples

Outbreak/sporadic	Sample code for sequencing	Collection date (day/mo/yr)	Origin	Age (yr) of patienta	Genogroup/genotype
Outbreak in 100-bed nursing home	CL4097-3108	31/08/06	Resident	92	GII/4
	CL4098-3108	31/08/06	Resident	84	GII/4
	CL4099-3108	31/08/06	Staff	U	GII/4
Outbreak in 282-bed nursing home	CL4006-1309	13/09/06	Staff	32	GII/4
	CL4007-1409	14/09/06	Resident	85	GII/4
	CL4008-1409	14/09/06	Staff	32	GII/4
	CL4009-1409	14/09/06	Resident	87	GII/4
	CL4010-1409	14/09/06	Resident	79	GII/4
Outbreak in secondary school	CL4015-2909	29/09/06	Food handler	45	GII/4
	CL4016-2909	29/09/06	Food handler	42	GII/4
	CL4018-2909	29/09/06	Food handler	59	GII/4
	CL4021-0310	3/10/06	Food handler	66	GII/4
Outbreak in mental health hospital	CL4025-0310	3/10/06	Resident	23	GII/4
Outbreak in primary school	CL4029-0310	3/10/06	Student	U	GII/4
	CL4031-0610	6/10/06	Student	U	GII/4
	CL4033-0610	6/10/06	Food handler	67	GII/4
	CL4035-0610	6/10/06	Food handler	68	GII/4
Sporadic	CL4056-1110	11/10/06	Child	6	GII/4
Outbreak in nursing home	CL4095-2010	20/10/06	Resident	76	GII/4
	CL4096-2010	20/10/06	Resident	75	GII/4
	CL4097-2010	20/10/06	Resident	70	GII/4

^aU, unknown age.

Phylogenetic analysis of the norovirus capsid gene in clinical and environmental water samples.

A neighbor-joining tree constructed from an alignment of the 291-base nucleotide sequence for norovirus GI (Fig. 2) and the 301-base nucleotide sequence for norovirus GII (Fig. 3) of the capsid gene revealed multiple genotypes in environmental water samples. Seven different GI genotypes were detected in water samples: GI/1, GI/2, GI/3, GI/4, GI/8, GI/10, and GI/14. The major GI genotypes in the present study were GI/2 (15 isolates) and GI/4 (10 isolates). Norovirus GII strains were clustered into seven genotypes (GII/3, GII/4, GII/6, GII/12, GII/13, GII/14, and GII/16) and two unclassified genotypes (GII/UG1 and GII/UG2). Norovirus GII/4 was the most prevalent GII genotype in environmental waters (24 isolates). On the basis of BLAST searches and pairwise comparisons, the most closely related norovirus to the unclassified GII/UG1 and GII/UG2 was the YURI strain, Japan (89.3% nucleotide identity), and the Leverkusen strain, Germany (92.4% nucleotide identity), respectively. Genotyping of the 21 norovirus-positive clinical samples showed that all of the strains belonged to the GII/4 cluster. Of the 21 nucleotide sequences isolated from the clinical samples, 20 were identical except for CL4025-0310. The environmental and clinical norovirus GII/4 isolates showed high levels of nucleotide sequence identity (95 to 100%) to each other. All genotype clusters, including two unclassified GII genotypes were statistically supported by bootstrap values of >70%.

FIG. 2.

Phylogenetic tree based on partial nucleotide sequences of the capsid gene of environmental norovirus GI isolates. Nucleotide sequences isolated from environmental water samples are indicated in boldface and designated according to the sampling location, date, and the clone number for different genotypes. Numbers at each branch indicate bootstrap values for the clusters supported by that branch. Bootstrap values of ≥70 are shown.

FIG. 3.

Phylogenetic tree based on partial nucleotide sequences of the capsid gene of clinical and environmental norovirus GII isolates. Nucleotide sequences isolated from environmental water samples and bootstrap values were as described in the legend to Fig. 2. Nucleotide sequences of clinical samples are indicated as CL, following by the sample code. The number of identical nucleotide sequences obtained from clinical samples is listed in parentheses.

In the present study, distinct nucleotide sequences could not be identified by direct sequencing of several PCR products recovered from environmental water samples. Cloning of these PCR products and sequencing of three to five individual clones resulted in identification of different norovirus genotypes in a single water sample (Table 1).

Noroviruses isolated from clinical and environmental water samples in the present study were compared to strains from different geographic locations. Most of the GI and GII isolates in the present study were closely related to the reference strains isolated from Japan, Korea, China, Australia, and New Zealand (Fig. 2 and 3). Two GI isolates from water samples (R1U-031006 and C1-150806) showed high nucleotide sequence identity (>98%) to the outbreak strains from Sweden. Phylogenetic analysis showed that the GII/4 cluster could be further divided into three small clusters (Fig. 3), which were closely related to the Hunter virus and two novel GII/4 variants, 2006a and 2006b, from Australia and New Zealand as identified by Tu et al. (39). These GII/4 reference strains are associated with the global increase in norovirus-associated gastroenteritis epidemics in 2006 (39). The GII/4 strains from clinical specimens examined here clustered only with the 2006b virus.

DISCUSSION

Despite the potential public health hazard posed by noroviruses, data regarding the occurrence of these pathogens in environmental water sources in tropical countries, such as Singapore, is lacking. In the present study, the prevalence of noroviruses in surface waters from highly urbanized catchments and the receiving estuarine bay was studied. The clinical significance of the environmental norovirus strains was determined by comparing these isolates with the clinical strains obtained during gastroenteritis outbreaks from the same period.

The results, based on direct molecular detection and characterization of noroviruses, confirmed the presence of these pathogens in urban water environments. The present study thus demonstrates that tangential flow filtration combined with PCR assay is an effective method to simultaneously concentrate and detect both norovirus GI and GII in environmental water samples. A concentration step is critical for detecting pathogenic viruses in environmental waters since their ambient concentrations are usually less than the limit of detection. Recently, ultrafiltration methods such as TFF have shown promise in recovering bacteriophages and polioviruses from environmentally seeded water samples (11, 27, 31); however, the concentration of different groups of human enteric viruses, including noroviruses, in field samples has not been fully investigated.

The high prevalence of noroviruses in urban rivers and canals in the present study was expected, since there are many old buildings within the catchment, and the sewers were laid more than 40 years ago. The current ongoing sewer rehabilitation program is critical to improving the water quality of these urban catchments by preventing wastewater leakage. The high rate of detection of noroviruses in the urban catchment waters could be due, in part, to the stormwater runoff during the wet weather season from October to December 2006. It is interesting that norovirus outbreaks in Singapore tend to occur during this time of the year.

The genetic diversity of noroviruses in environmental waters has been reported by others from temperate regions (21-23). The present study provides additional evidence for the great genetic diversity of both norovirus GI and GII in tropical water environments. Genotyping of 21 clinical samples collected during the same period as the environmental surveys clustered all norovirus strains to GII/4. This is consistent with other clinical molecular epidemiological studies that showed GII, specifically GII/4, as the major causative agent of acute gastroenteritis, especially when associated with outbreaks (4, 5, 12, 15). In contrast, genotyping of environmental water samples showed a wide spectrum of genotypes of GI and GII, in addition to the GII/4 strains. The occurrence of multiple genotypes of noroviruses other than GII/4 in environmental waters is important since these strains may be silently circulated in the community. These genotypes have been reported as causative agents of gastroenteritis outbreaks (17), although their isolation from surface waters has rarely been reported. The present results suggest that phylogenetic analysis of both environmental and clinical isolates is a useful approach in understanding the circulation of human noroviruses within the community.

Correlations were observed between the clinical norovirus GII/4 isolates and GII/4 strains isolated from environmental samples during the same period. A total of 14 environmental isolates (from August to December 2006) in the 2006b cluster were closely related to clinical isolates. Particularly, six GII/4 environmental isolates (R2U-150806, C1-150806, R1D-150806, C1-011106, R1U-111206, and R1D-111206) were identical to 20 clinical isolates.

The prevalence of multiple genotypes of GI (seven different genotypes) in water environments corroborates the observations of a study conducted in the urban rivers in Korea (22). In the Korean study, GI strains were clustered into eight genotypes, with GI/1 and GI/13 as the major genotypes. The majority of GI genotypes in the present study were GI/2 and GI/4. Norovirus GI strains have been reported to be more often implicated in waterborne outbreaks than GII (10, 26, 33), highlighting the significance of the presence of these strains in recreational waters.

Interestingly, our noroviruses GII/4 isolates from clinical and environmental samples (2006a and 2006b clusters) showed the highest levels of capsid alignment with strains isolated from Australia, New Zealand, Europe, and China during the global gastroenteritis epidemic period in 2006 (39), implying the possible epidemiological relationships of these isolates. In addition, most of the environmental isolates of other norovirus genotypes such as GI/2, GI/4, GI/3, GI/8, GI/14, GII/3, GII/13, and GII/UG1 in the present study appeared more closely related to the clinical strains from Asian countries, especially Japan, Korea, and China. This may indicate that noroviruses in Asia have evolved in a region-specific manner. Our results suggest that noroviruses could be introduced from individuals traveling between geographic regions.

Human noroviruses were detected in downstream waters of urban rivers and the receiving estuarine bay, indicating urban runoff as a source of viral contamination. The present findings from the field study suggest the environmental stability of both norovirus GI and GII in water, which allows them to be transported from the pollution source to the receiving water. A norovirus-associated outbreak among a party of canoeists in contact with contaminated recreational waters has been reported in the United Kingdom (7). Moreover, exposure of individuals to multiple genotypes of noroviruses may lead to mixed norovirus infections, which provide the opportunity for recombination to occur between different strains. It has been shown that recombination is not a rare phenomenon among noroviruses and thus contributes to the genetic diversity of these viruses (1). The coexistence of both GI and GII strains in most of the water samples may facilitate the emergence of novel intergenogroup recombinants (28, 34). The high prevalence of noroviruses in water environments suggests the possible role played by water as a vector for viral transmission during recreational activities.

A recent study showed that norovirus can be disinfected by adequate free chlorine with provision of proper pretreatment processes before chlorination (36). Norovirus contamination of drinking water can also be controlled by advanced water treatment processes such as UV and gamma radiation, ozone disinfection, and membrane technology (14, 25, 37, 38). Thus, the presence of noroviruses may represent little public health impact if receiving waters are intended for drinking, especially in areas with advanced water treatment. However, further field studies are needed to evaluate the effectiveness of both conventional and advanced water treatment processes for removing or inactivating noroviruses in drinking water.

In conclusion, we provide here valuable preliminary data on the occurrence and molecular characterization of noroviruses in urban catchments and receiving estuarine bay in Singapore. The prevalence of multiple genotypes of noroviruses in environmental waters may indicate the emergence of these genotypes in clinical infections in the same geographical region. Rapid detection and routine clinical reporting of noroviruses need to be implemented. Further comprehensive surveillance of water environments, including sewage, in Singapore is currently being conducted to identify and quantify noroviruses and other enteric viruses in waters. The present study highlights the importance of international collaborations and exchange of information on outbreak data, considering that novel norovirus variants are emerging and circulating globally. It has been suggested that global waterborne disease surveillance and reporting networks are needed (35).