Norovirus GII.P16/GII.2 Strain in Shenzhen, China: A Retrospective Study

Fecal specimens were collected from the 203 outbreaks and epidemiological data were collected through the AGE outbreak surveillance system in Shenzhen. The genotypes were determined by sequencing analysis. To gain a better understanding of evolutionary characterization of NoV in Shenzhen, the molecular evolution was analyzed by time-scale evolutionary phylogeny and amino acid mutations.

the genome since the domain encodes the receptor binding domain, which responsible for histoblood group antigen (HBGA) binding, and important epitopes targeted by antibodies that inhibit binding [14,15].
In this report, the retrospective study aimed to illustrate the genotypic diversity of NoV strains in outbreaks and genetic characteristics of GII.P16/GII.2 strain in the Shenzhen, China, from September 2015 to August 2018.

Sample and RNA extraction
All the fecal specimens were derived from the AGE outbreaks which were inspected to Shenzhen Center for Disease Control and Prevention (Shenzhen CDC) by districts Center for Disease Control and Prevention (districts CDC). The outbreaks were identi ed as more than 5 cases within 3 days. The viral RNA of fecal specimens were extracted by Viral Nucleic Acid Extraction Kit II (Geneaid, Taiwan, China) and stored at -80℃. And the NoV outbreaks were con rmed by two positive samples in per outbreak using real-time reverse transcription polymerase chain reaction (qRT-PCR) [24] (S1 Table).
Reverse transcription and PCR Sequence ampli cation was used one-step RT-PCR Kit (QIAGEN, Germany). And before October 2016, the primer set G1SKF/G1SKR and COG2F/G2SKR were used for VP1 typing to detect GI(330bp) and GII(387bp), respectively [25](S1 Table). After October 2016, the primer set MON432/ G1SKR and MON431/G2SKR were used to augment both of partial RdRp region and VP1 gene of GI (543 bp) or GII (557 bp), respectively [26](S1 Table). And genotypes were con rmed by the Blast and an automated online NoV genotyping tool offered by the Netherlands National Institute for Public Health and the Environment (http://www.rivm.nl/mpf/norovirus/typingtool) [27]. Reverse transcription of viral RNA via SuperScript III kit (invitrogen, USA). Then cDNA were used Polymerase Chain Reaction (PCR) to retrieve nearly full-length sequence. The sequence of primers were shown in S1Table. The F of GII.P16/GII.2 strain represented forward primer and R represented reverse primer.

Phylogenetic analysis of RdRp region and VP1 gene
To evaluate the evolution of the NoV GII.P16/GII.2 strain in Shenzhen, the full-length RdRp region or VP1 gene from this study and GenBank were collected and phylogenetic trees of the genes were constructed by the MCMC method. The best substitution model was selected by MEGA6.0 using the BIC method. The analysis of phylogenetic data would be con rmed until the effective sample sizes greater than 200 by the Tracer. The nal result would be visualized by the FigTree v1.4.0 program.

Recombination variant
To evaluate the impact the intergenic recombination of nonstructural region and capsid region, the amino acid (aa) mutations of nonstructural region and capsid region among different genotypes were analyzed by MEGA6.0.

Statistical analysis
The difference between GII.2 NoV detection rates in dominant setting distribution was compared using Fisher's Exact Test in SPSS Statistics v.22.0 (IBM Corp., Armonk, NY, USA), and a P-value less than 0.05 was considered statistically signi cant.

Phylogenetic Analysis of RdRp region and VP1 gene
In this study, all the GII.2 NoV were GII.P16/GII.2 strain. To examine the evolution of the strains, 52 fulllength RdRp region sequences of GII.P16/GII.2 strain from Shenzhen and 95 reference sequences from GenBank were collected to analyze. The best substitution model was TN93 (Tamura-Nei)+G (Gamma).
After MCMC chains were run for 1.0×10 8 steps for the RdRp region sequences and the rst 10% state was buin-in, effective sample sizes greater than 200 were accepted. The MCC tree shown that the evolutionary rate of the RdRp region of Gll.P16/GII.2 strain was estimated as 2.1×10 -3 substitutions/site/year (95% HPD interval,1.7×10 -3 -2.5×10 -3 substitutions/site/year). The common ancestors of GII.P16/GII. recombination strain), 2 aa sites (A644P, A1521V) achieved substitution in GII.P16/GII.2 recombination strains and 1 aa site (S/T753T) retured to anterior aa site. RdRp region could be divided into three highly conserved ingredients according to function and structure, including the ngers, thumb, and palm subdomains ,which could be organized into Motifs A to G [11]. The result shown 1310 aa site (Motifs G) was substituted (Table 4).

HBGA-Binding Pro le, Epitopes Predicted and Epitope A to E sites of GII.P16/GII.2 recombination strain
To explore HBGA-Binding pro le, epitopes predicted and epitope A to E sites of GII.P16/GII.2 recombination strain, 72 full-length VP1 gene sequences from this study and 65 reference sequences, including GII.Pc/GII. 2 (1976-1978 (5.3%) parsim-infomative sites were revealed but there were no mutations in the amino acids of the HBGA-Binding pro le, epitopes predicted and epitope A to E of GII.P16/GII.2 strain (S2 Table) Discussion In this study, NoV-associated AGE outbreaks in Shenzhen, China, from September 2015 to August 2018 were analyzed. There were 203 NoV outbreaks were reported to Shenzhen Center for Disease Control and Prevention. The NoV infection was initially described as "winter vomiting disease" due to its seasonal preference [31], monthly distribution also indicated that the peak of the outbreak in Shenzhen was during the November to March. Previous study have found a link between NoV's increased number and climate or weather [32,33]. Especially, rain, humidity, and temperature changes are important factors that in uence the seasonal increase in NoV outbreaks, and it is suspected that meteorological factors may have signi cant in uence on the activity and transmission of NoV. The peak in this study was in December, when Shenzhen began to turn cold, and March, when temperature began to turn warm, which is suspected that climate changes have an impact on NoV transmission. The NoV outbreak usually occurs in hospitals, nursing homes, schools, child care centers, hotels and other semi-enclosed places [34][35][36]. A study in United States reported 3,960 NoV outbreaks between 2009 and 2013 and found that long-term care homes were the most frequent sites of NoV outbreaks [37]. Another study from Qin [38] shown that middle schools were the most important setting of NoV outbreaks in China, followed by primary schools between 2006 and 2016. In this study, we classi ed the outbreak settings into 12 categories, and the results shown that most outbreaks were occurred in child care centers, followed by primary school. This suggests that schools remain the focus of NoV outbreaks in Shenzhen, but that the current high incidence is among children of lower school age. Combining on the results of monthly distribution of NoV outbreaks in Shenzhen, we suspect that the decrease in the number of NoV outbreaks in January and February is related to school holidays. When the scale of the outbreak was analyzed, the average number of people involved per outbreak in Shenzhen was 9, smaller than the 18 persons reported in the United States [37]. Shenzhen is one of the cities where the economy are most develpoed This may bene t from local high public health system and high effective handling of public health emergencies in Shenzhen (http://www.szemo.gov.cn). In genotype detection, both GI and GII genogroups were detected, furthermore, 15 capsid types and 15 polymerase types were found. Among the genotypes, the most was GII.2, followed by GII.3. GII.4 Sydney2012 was only account for 3.4%. In 2016, NoV outbreaks were associated with GII.P16/GII.2 strains were reported in multiple regions in China [39,40]. In this study, we identi ed the GII.2 strain were NoV GII.P16/GII.2, and was recombinant strain as other regions in China [39,40]. This study found the rst outbreak identi ed as GII.P16/GII.2 recombinant strain was September 30, 2016 in Shenzhen and then the GII.P16/GII.2 strain caused steep rise in acute gastroenteritis in Shenshen in the following months. In current study of NoV evolution, the recombination was thought to be important and common for virus evolution. Most recombination often occurs within ORF1/ORF2 overlapping regions or near the RdRp region, resulting in different capsid and RdRp genotypes. In the study, we calculated the evolutionary rates of RdRp region and VP1 gene, which were 2.1 × 10 − 3 substitutions/site/year and 2.7 × 10 − 3 substitutions/site/year, respectively, indicating that the polymerase and capsid regions of NoV GII.P16/GII.2 strains had evolved independently, which was consistent with the results of previous studies [41].The evolution rate of NoV GII.2 was much lower than that of GII.4 NoV The result of sequances alignment shown that important sites of VP1, including HBGA-Binding pro le, epitopes predicted and epitope A to E sites, were not mutated. This suggested that the reason of prevalence of NoV GII.P16/GII.2 strains in population was different from that of the previous pandemic NoV GII.4, which mainly due to changes in capsid region leading to changes in blocking antibody epitopes to cause population among people [43,44]. Parra [45] analyzed the GII.2 capsid sequences over a 40-year period and found only small differences, which our results agree with, indicating that GII.2 strain is more genetically stable than GII.4 strain. As the same time, lack of variation in antigen regions of strains may also explain their short duration. These results suggested that the presence of a structure other than the VP1 contributes signi cantly to the prevalence of GII.P16/GII.2 stain [19]. This could learn something from the epidemic reasons of GII.P17/GII.17 that caused the outbreak of acute gastroenteritis in many countries in the winter of 2014-2015. Tohm [46] summarized the epidemic reasons of GIIP17/GII.17 in the population and believed to have a relationship with unstructured region. And amino acid substitutions were found in the nonstructural regions, including P48, NTPase, P22 and RdRp in this study. These nonstructural polymerases played important roles in norovirus replication, which cougld destroy host cells and promote virus synthesis by interfering with intracellular protein transport, vesicle misorientation and golgi disintegration [8][9][10][11]. The results of this experiment suggested that the unstructured region may provide more materials for virus replication, accelerate the apoptosis of host cells' golgi bodies and enhance the tness by changing the interaction mode. Another study also reported that GII.P16/GII.2 strain had a higher viral load than that of GII.Pe/GII.4 and GII.P17/GII.17 in patients [47]. But not all changes in unstructured region would cause epidemic. The study of Tohma calculated the amino acid substitution sites in the RdRp region of GII.P2/GII.2 and found that the replacement rate of GII.P2 was higher than that of GII.P16 [19]. however, no great GII.P2/GII.2 outbreaks were found. It indicated that GII.P16 played a crucial role in GII.P16/GII.2 epidemic and the further research of mechanism was limited.
This study shown the GII.P16/GII.2 outbreaks had reduced in Shenzhen, while the continuous surveillance to monitor genotypes is still necessary to identify new variants in time.The limitations of this study were as followed: rst, genotyping was only successful for 150 (73.9%) in our study within positive NoV cases. Second, our study lack of clinical information and epidemiological data are incomplete within outbreaks. In future studies, epidemiological surveillance should be better perfected and molecular analysis for different NoV genotypes would be developed.

Conclusions
In conclusion, this study reported the epidemiological patterns and genetic characteristics of NoV in Shenzhen from September 2015 through August 2018 and the main cause was GII.P16/GII.2 strain. This study also provided the evidence that the NoV GII.P16/GII.2 strain was static in Shenzhen.