Gastroenteritis Outbreaks Associated with the Emergence of the New GII.4 Sydney Norovirus Variant during the Epidemic of 2012/13 in Shenzhen City, China

Noroviruses (NoVs) are the leading cause of gastroenteritis outbreaks in humans worldwide. Since late 2012, a new GII.4 variant Sydney 2012 has caused a significant increase in NoV epidemics in several countries. From November of 2012 to January of 2013, three gastroenteritis outbreaks occurred in two social welfare homes (Outbreaks A and B) and a factory (Outbreak C) in Shenzhen city of China. Feces and swabs were collected for laboratory tests for causative agents. While no bacterial pathogen was identified, all three outbreaks were caused by NoVs with detection rates of 26.2% (16/61) at Outbreak A, 35.2% (38/108) at Outbreak B), and 59.3% (16/27) at Outbreaks C. For Outbreak B, 25 of the 29 symptomatic individuals (86.2%) and 13 of the 79 asymptomatic individuals (16.5%) were found NoV-positive. For Outbreak C, an asymptomatic food handler was NoV-positive. All thirteen NoV sequences from the three outbreaks were classified into genogroup II and genotype 4 (GII.4), which we identified to be the GII.4 Sydney 2012 variant. The genome of two isolates from Outbreaks A and B were recombinant with the opening reading frame (ORF) 1 of GII.4 Osaka 2007 and ORF2 and 3 of the GII.4 New Orleans. Our study indicated that the GII.4 Sydney 2012 variant emerged and caused the outbreaks in China.


Introduction
Noroviruses (NoVs), members of the Norovirus genus in the Caliciviridae family, are recognized as the most common viral cause of gastroenteritis outbreaks throughout the world [1][2][3][4], wherein they are responsible for > 90% of non-bacterial outbreaks of gastroenteritis worldwide [5]. In the United States, NoVs are estimated to cause 21-million illnesses each year, which results in 71,000 hospitalizations and 800 deaths [6,7]. In developing countries, NoVs are estimated to claim over 200, 000 deaths annually in children < 5 years of age [5]. Norovirus outbreaks occur frequently in semi-closed institutions, such as hospitals, nursing homes for the elderly, schools, prisons, restaurants, hotels, and cruise ships [8][9][10][11][12][13]. NoVs transmit via the consumption of contaminated food or water, through person-to-person contact, or by exposure to aerosols from vomitus [14][15][16]. NoVs are highly contagious owing to their ability to infect at low doses, their high stability in the environment, and by the limited short-term immunity hosts amount [17]. Accordingly, outbreaks of NoVs are extremely difficult to control.
NoVs are non-enveloped RNA viruses that contain a single-stranded, positive-sense, and polyadenylated RNA genome of~7.5 kb in length [18]. The NoV genome consists of three open reading frames (ORFs 1-3). ORF1 encodes a polyprotein that is processed post-translation into six non-structural proteins, including an RNA-dependent RNA polymerase (RdRp) [19]. ORF2 and ORF3 encode the major (capsid protein, VP1) and minor (VP2) structural proteins, respectively [20]. The genomes of NoVs are highly diverse. Each belongs to one of six genogroups (GI-GVI) and each genogroup can be further stratified into one of more than thirty-six genotypes [21][22][23]. While GI, GII, and GIV NoVs can infect humans, the GII.4 genotype is attributed to causing both outbreaks and many sporadic cases [24].
The recently emerged GII.4 variant, Sydney 2012, was firstly identified in Australia early in 2012, but also lead the increase of acute gastroenteritis outbreaks in the United States, France, Japan, United Kingdom, the Netherlands, New Zealand, and Hong Kong [30][31][32][33][34][35]43]. Furthermore, the Sydney 2012 variant was also responsible for three gastroenteritis outbreaks in two social welfare homes and a factory in Shenzhen, China, in 2012/13. Herein, we undertake epidemiological and laboratory investigations into these three gastroenteritis outbreaks caused by Sydney 2012 variant.

Ethics statement
The present study was approved by the Institutional Review Board of Shenzhen CDC. Written consent was obtained from both the symptomatic and asymptomatic individuals who participated in this study before collecting their stool samples and medical data.

Epidemiological investigation
Cases were defined with ! 3 loose stools and /or vomiting in a 24-hour period during the outbreaks. A standardized questionnaire was developed for data collection of demographic data (sex and age), illness onset (symptoms, duration of symptoms) and potential risk factors (water and food consumption, patient contacts). We then collected stool and vomitus from the study set and took environmental swabs of vegetables and the chopping boards, bench surfaces, and cooking utensils used in the preparation of raw and cooked foods for laboratory diagnosis.

Conventional reverse transcription (RT)-PCR
The capsid region used for NoV genotyping was amplified by conventional RT-PCR according to the protocol and primers (CoG2F, G2-SKR, G1-SKF, G1-SKR) reported by Yan et al. [45]. Thirteen stool samples determined to be NoV-positive by real-time PCR were submitted for RT-PCR.

Amplification of full-length NoV genomes
The complete genomes of two of the earliest NoVs-JB031230049 and JB031230054-were amplified. Initially, cDNA was reverse transcribed from viral RNA (9.5 μL) using a modified oligo(dT) 20 primer (V 3 NT 20 in Table 1) and a SuperScript III first-strand synthesis system (Invitrogen, Carlsbad, CA). The total, 50-μL reaction mixture contained 5 μL 5× Ex Taq Buffer, 1 μL of each primer (20 pmol), 1 μL dNTP mix (10 mM), 0.5 μL Ex Taq enzyme (Takara, Japan), 5 μL cDNA, and 36.5 μL RNase-free water. CoG2F [45]and V 3 NT 20 [46] primers were used to amplify the fragment from the ORF2 to the 3' Poly A. The cycling conditions were configured as follows: initial denaturation at 94˚C for 3 min, followed by 40 amplification cycles (94˚C for 15 s, 62˚C for 3 min, and 72˚C 30s) and a final extension step at 72˚C for 15 min. The ORF1 was divided into five overlapping fragments, which were amplified using the P290 and G2-SKR primers and four pairs of newly designed primer sets ( Table 1). The conditions were as follows: initial denaturation at 94˚C for 3 min, then 35 cycles (94˚C for 45 s, 55˚C for 45 s, and 72˚C 1 min), and a final extension at 72˚C 10 min. The fragment covering the ORF1/ ORF2 overlap and amplified by the P290 [47] and G2-SKR [45] primers was also used to examine potential recombination.

DNA Sequencing and phylogenetic analysis
All PCR products were excised from the gel and purified using a QIAquick gel extraction kit (Qiagen, Hilden, Germany). All purified products were sequenced with the primers listed in Table 1 using the Big-Dye terminator cycle sequencing kit and the ABI Prism 310 Genetic Analyzer (Applied Biosystems Inc., Foster City, CA). The resulting NoV sequences were analyzed using CLUSTAL X (Version 1.83) followed by phylogenetic analysis using MEGA version 4.1. The statistical significance of the inferred phylogenies was estimated using a bootstrap analysis of 1,000 pseudoreplicate data sets. SimPlot software (Version 1.3) was used to align and compare sequences to identify potential recombination(s) among known genotypes of NoVs. The nucleotide sequences generated in this study were deposited in the Gen-Bank (KJ995534-KJ995550 for short sequences within the capsid region and KJ955492-KJ955493 for complete genome sequences).

Epidemiological investigation
Among the three outbreaks of gastroenteritis in Shenzhen, China, during 2012/13 winter season, two arose in social welfare homes in the Luohu (Outbreak A) and Nanshan (Outbreak B) districts, while the epicenter of Outbreak C was in a factory in the Baoan district. Twenty-eight individuals were sick with acute gastroenteritis in the Outbreak A, including 22 residents and 6 healthcare workers. The index case occurred on November 28, 2012, with the outbreak peaking on November 30 (Fig 1). The most frequent symptoms were diarrhea (71.43%), followed by vomiting (60.71%), nausea (35.71%), abdominal pain (7.14%) and fever (3.57%; Table 2). No hospitalizations resulted from Outbreak A. Of the 28 infected persons, 19 were living on the fourth floor, 6 on the fifth, 2 on the third, and 1 on the sixth. The index case was an 82-year-old female living on the fourth floor. In Outbreak B, 39 individuals were sick with gastroenteritis, including 33 residents, 5 healthcare workers, and 1 staff. The index case occurred on December 14, 2012. Outbreak B peaked between 23 and 25 of December (Fig 1). The most frequent symptoms were diarrhea (94.87%), followed by vomiting (82.05%), nausea (5.13%), abdominal pain (43.59%) and fever (2.56%; Table 2). We analyzed the attack rates among males and females stratified into age groups. Among the 33 sick residents, the attack rates were 18.33% (11/60) for males and 16.30% (22/135) for females, though this difference was not statistical significant (X 2 = 0.12, P = 0.73). The ages of the 33 patients spanned from 48 to 92 years of age. The infection rate was highest in those > 80 years of age (54.55%, 18/33), followed by those aged 70-79 (27.27%, 9/33). The index case was an 83-year-old female living on the fifth floor of the healthcare center. She had 6-8 episodes of diarrhea on December 14 during her leave starting on December 8. She returned to the center on December 15 and had no symptoms of diarrhea on December 16. Thereafter, she visited entertainment rooms on the other floors, which culminated with at the highest attack rate on the fifth floor (31.25%; Table 3). Seventy-six individuals were sick with acute gastroenteritis in Outbreak C. The index case occurred on January 16, 2013. Outbreak C peaked between January 16 and 17 (Fig 1). The common symptoms reported were pain (83.6%), nausea (5.13%), diarrhea (77.0%), vomiting (59.0%), and fever (19.7%; Table 2). Among the 414 staff, the attack rates were 15.63% (10/64) in males and 18.86% (66/350) in females-a different that was not statistically significant. The factory was a six-story building. The highest attack rate was 71.43% (5/7) on the sixth floor, followed by 25

Phylogenetic analysis
Thirteen sequences from the three outbreaks were obtained and analyzed, including two from Outbreak A, six from Outbreak B, and five from Outbreak C. Phylogenetic analysis using a collection of reference sequences representing a variety of NoV genotypes showed that all thirteen   (Fig 3). To eliminate the possibility that the polymerase and capsid encoding genes were from two separate NoVs that co-infected the same patients, a fragment of~1100 nucleotides in length, spanning both ORF1 and ORF2, was amplified from a stool sample. Simplot analyses of the junction regions between ORF1 and ORF2 confirmed the presence of recombinant strains and indicated a potential cross-over site where the recombination between ORF1 and ORF2 is likely to have occurred (Fig 4).

Discussion
The NoV-associated gastroenteritis outbreaks are common in semi-closed settings, including hospitals, cruise ships, military camps, schools, and elderly care homes [8][9][10][11][12][13]. Over a quarter (27%) of NoV outbreaks, for example, are epicentered about healthcare settings [50], while long-term care facilities (elderly care homes, etc.) account for 50% of these NoV outbreaks occurring in a healthcare setting [50]. In our study, two of the three NoV outbreaks occurred in two social welfare centers that were funded by the local government as long-term care facilities for the elderly, while the remaining outbreak occurred in a factory that was a semi-close setting where many employees dwelled in communal dormitories. A NoV outbreak is declared in accordance with Kaplan's criteria, wherein (1) > 50% of affected persons must be afflicted with vomiting, (2) the mean or median incubation period is 24-48 hr, (3) that the mean or median illness duration is 12-60 hr, and (4) that no bacterial pathogens are isolated from stool cultures [51]. The epidemiological investigation in this study revealed that the vomiting rates of all three outbreaks were greater than 50% (60.7% for Outbreak A, 82.1% for Outbreak B, and 59.0% for Outbreak C). Asymptomatic infections are estimated to occur in about one-third of all studies investigating outbreaks involving volunteers. We detected 16% (13 out of 79) of asymptomatic persons who were NoV-positive in Outbreak C, including a food handler who could have been the source of this outbreak. Ozawa et al. reported that the prevalence of NoV detection in food handlers was 19% in different foodcatering settings in Japan and that 73% of symptomatic and 7% of asymptomatic food handlers were positive for the NoV [52].
The evolution of GII.4 NoV variants is potentially driven by several factors [53]. In a similar manner to the influenza A virus, a new GII.4 variant tends to emerge every 2-3 year through antigenic drift. Furthermore, recombination is likely to be an important factor for the emergence of the GII.4 variants in a resemblance to the reassortment of genetic material by the influenza virus [54]. Intra-genotype recombination has frequently been reported in some GII.4 variants [41,55]. In this study, two genomes of the GII.4 Sydney 2012 variant from two outbreaks were analyzed for recombination. We found that ORF1 was derived from an Osaka 2007 virus, while the ORFs 2&3 were from the viruses responsible for Apeldoorn 2008 and New Orleans 2009 (Fig 3).
In conclusion, the emerging GII.4 Sydney 2012 variant caused three gastroenteritis outbreaks in the winter season of 2012/13 in Shenzhen in China. These outbreaks occurred during the same time frame when the GII.4 Sydney 2012 variant caused global epidemics. Two outbreaks occurred in social welfare centers for the elderly and were most likely due to an asymptomatic food handler. Our study highlights first the susceptibility and vulnerability of the elderly in the closed environment of a nursing home setting and secondly the need to improve sanitation practices by food handlers.