Prevalence and genotype distribution of group A rotavirus circulating in Shanxi Province, China during 2015–2019

Shanxi Province is located in the middle-area of China (latitude 34°34′–40°44′N and longitude 110°14′–114°33′E), consisting of 11 cities (80 counties) and a population of about 37 million.

Fecal specimens were collected from children < 10 years old who were hospitalized for diarrhea in Children’s Hospital of Shanxi (CHS) between Jan 1, 2015, and Dec 31, 2019. Oral informed consent was obtained from their patients or guardians. All enrolled children have obvious symptoms of gastroenteritis (defined as ≥3 loose stools over a 24 h period), and only those who have not received any medications and medical tests before hospitalization were eligible for enrolment. Patients exceed the age and diagnosed with chronic diarrhea by clinicians were excluded from this study. Specimens were collected in a sterile sampling cup, keeping low temperature (4 °C), and were sent to the virus microbiology laboratory of Taiyuan Center for Disease Control and Prevention (Taiyuan CDC) for rotavirus nucleic acid testing within 24 h. Upon receipt, each sample was allocated a unique laboratory code and entered into a gastroenteritis information database (Excel) in Taiyuan CDC.

Fecal suspensions (10%, w/v) were prepared of 0.01 M phosphate-buffered saline (PBS) (pH = 7.2), vortexed, and centrifuged (8000 g at 4 °C for 8 min). Total RNA was extracted from 60 μL of the fecal suspension using a MagMax-96 Viral RNA Isolation Kit (Thermo Fisher Scientific, Foster City, CA), according to the manufacturer’s instructions. RNA-positive control and negative control (PBS) was included in the extraction procedure in each batch, and the quality of extracted RNA was checked through a NanoDrop 1000 Spectrophotometer (NanoDrop Technologies, Houston, TX, USA). Rotavirus nucleic acid testing and genotyping were kept physically separated. All steps of sample preparation and RNA extraction were done in a biosafety cabinet.

The real-time quantitative reverse transcription PCR (qRT-PCR) was applied by using a Rotavirus (Group A, B, and C) Multiple Real-time PCR Kit (TaqMan probe) (S-SBIO, Taizhou, China). PCR cycling parameters were set up according to the instruction: 50 °C for 30 min, 95 °C for 5 min, followed by 45 cycles of 94 °C for 10 s, 55 °C for 40 s in a CFX96 Real-time Thermal Cycler (Bio-Rad, Hercules, CA). A positive result was defined as a threshold cycle (Ct) value < 35, and positive internal control was defined as a Ct value < 30. The qRT-PCR negative samples were not under the scope of our study.

For VP7 amplification, the reverse transcription PCR was carried out using a PrimeScript One Step RT-PCR Kit. (TaKaRa, Dalian, China). A 1062-bp fragment was amplified with the consensus forward primer Beg9 (GGCTTTAAAAGAGAGAATTTCCGTCTGG) and the reverse primer End9 (GGTCACATCATACAATTCTAATCTAAG). The amplification conditions included: 50 °C for 30 min, 94 °C for 2 min followed by 32 cycles of 94 °C for 30 s, 55 °C for 30 s, 72 °C for 90 s, and a final extension of 72 °C for 5 min. For the VP4 amplification, the semi-nested reverse transcription PCR was carried out using a PrimeScript One Step RT-PCR Kit. The first amplification was carried out using primers VP4F2 (TTTATAGACAGCTTCTCACTAATTC) and VP4R3 (TATGTGCAGTTACTTGTTCACC) under the condition of 50 °C for 30 min, 94 °C for 2 min, followed by 10 cycles of 94 °C for 30 s, 55 °C for 30 s, 72 °C for 90 s. Then 1 μl of the first PCR products was used for amplification with the nested primers VP4F4 (TTTATAGACAGCTTCTCACTAATTC) and VP4R3 (TATGTGCAGTTACTTGTTCACC). The amplification conditions included 94 °C for 5 min, followed by 30 cycles of 94 °C for 30 s, 55 °C for 30 s, 72 °C for 1 min, and a final extension of 72 °C for 5 min. A portion (5 ul) of the reaction mixture with loading buffer, followed electrophoresis in 1.5% agarose gel (Takara, Dalian, China), was visualized by Goldview (Transgen, Beijing, China). The amplicons were purified and sequenced by Sangon Biotech (Shanghai, China). All the PCR runs included the positive control and non-target control (reagent blank) to avoid false-positive results. All the primers were synthesized by Sangon Biotech (Shanghai, China).

The resulting sequences were prepared and aligned by BioEdit (version 7.2.5) with the Clustal-W program. Based on the sequence data, the genotype assignment was accomplished using BLAST (http://​blast.​ncbi.​nlm.​nih.​gov/​Blast.​cgi) and RotaC v2.0 (http://​rotac.​regatools.​be). Phylogenetic trees were constructed using the Neighbor-joining method with MEGA (version 5.0) and bootstrap analysis was performed with 1000 replications. The reference sequences used in plotting the phylogeny trees from the GenBank database are shown in Table 1.

The nucleotide sequences used in this study were submitted in GenBank (https://​www.​ncbi.​nlm.​nih.​gov/​genbank/​) under accession numbers MT710743 - MT710824 (VP7) and MT710825 - MT710927 (VP4), which are shown in Table S2.

Data analyses were performed by IBM SPSS Statistics software (Version 19.0). Univariate logistic regression was performed to ascertain the equality of means of RVA positive or negative for variables: sexes, symptoms, and sample types (Table 2). Statistical significance was defined as p < 0.05.

During the period 2015–2019, a total of 961 hospitalized children younger than 10 years of age, 609 boys and 352 girls, confirmed providing fecal specimens for rotavirus testing. RVA was detected in 183 stool samples via qRT-PCR with an overall positivity rate of 19.0% (183/961). Rotavirus B and C were not detected. The maximum number of samples was tested in May 2017, and the largest number of RVA-positive cases occurred in February 2019 (Fig. 1). No statistically significant difference was found between sex and RVA-infection (p > 0.05) (Table 2). Types of fecal specimen collected in this study included watery stool, loose stool, pus and blood stool, and mucus stool, no statistically significant difference existed between four types positive for RVA (Table 2). Apart from the symptom of diarrhea, some patients were accompanied by other related symptoms such as vomiting, fever, abdominal pain, and dehydration. There was a statistically significant difference between RVA-infection and fever (> 38.5 °C) (p < 0.05). Whereas, no significant difference (p > 0.05) was found between RVA-infection and the other three symptoms (vomiting, abdominal pain, and dehydration) (Table 2). Twenty RVA-positive patients coinfected with other diarrhea associated viruses: 16 with norovirus, 2 with sapovirus, and 1 with astrovirus. One was coinfected with more than two viruses (norovirus and astrovirus). Six were coinfected with other pathogenic bacteria: 2 with Salmonella, 3 with EAEC, and 1 with ETEC (Table S1).

Overall, most cases (292) were found in the age group of 12–23 months which was also the group with the highest average RVA-positive rate (23.7%) (Fig. 2a). On the contrary, the lowest rate (8.5%) was observed in the group > 60 months with a statistically significant difference (p < 0.05, Fig. 2a). More than one-fifth of gastroenteritis cases were found in July–August, but the RVA-positive rate remained low and with a statistically significant difference (p < 0.05, Fig. 2b). 44.9% of the RVA-positive cases were found in December, although gastroenteritis cases were rare in the cold season (November–January) (Figs. 1 and 2b). All the patients came from the cities or regions of Shanxi Province, the highest prevalence (29.5%) was observed in Yangquan, followed by 22.4% in Xinzhou, 21.8% in Lvliang, 21.1% in Northern Shanxi, and 16.9% in Taiyuan (Fig. 2c). The patients from Southern Shanxi had the lowest average positive rate (11.6%, Fig. 2c).

One hundred fifty-nine samples were genotyped for VP7, and 183 samples were successfully genotyped for VP4. G9 was the prevalent G-genotype, representing 76.0% (139/183) of all positive samples, followed by G3 (7.1%, 13/183), G2 (3.3%, 6/183), and G1 (0.5%, 1/183) (Fig. 3a). P[8] (95.1%, 174/183) was the prevalent P-genotype, followed by P[4] (4.9%, 9/183) (Fig. 3b). One hundred fifty-nine strains were confirmed both G- and P-genotype. G9P[8] was the most commonly detected genotype (76.0%, 139/183), followed by G3P[8] (7.1%, 13/183), G2P[4] (3.3%, 6/183), G1P[8] (0.5%, 1/183), and G9P[4] (0.5%, 1/183) (Fig. 3c).

Due to the long distance between the patients and the designated hospital, cases from two northern cities (Datong and Shuozhou) and four southern cities (Linfen, Yuncheng, Changzhi, and Jincheng) were less compared with the central cities of Shanxi Province, therefore, cases from these cities were divided into two independent parts (Northern Shanxi and Southern Shanxi) for next analyzing. G9P[8] was the dominant genotype and occupied more than 80% of RVA positive cases in most cities or regions of Shanxi Province (Fig. 4). The only G1P[8] and G9P[4] were detected in Taiyuan (2016) and Southern Shanxi (2018), respectively (Fig. 4). G2P[4] was detected in Taiyuan, Xinzhou, Lvliang, and Yanquan; G3P[8] was detected in Taiyuan, Xinzhou, Lvliang, and Southern Shanxi. Strains that cannot be typed (GntP[8] and GntP[4]) were found in most areas in Shanxi province, except Yangquan (Fig. 4).

The G1-G2-G3 VP7 tree was constructed based on 18 Shanxi strains sequenced in this study and 13 representative members (Fig. 5, Table 1). The G1 lineage contained the only Shanxi strain (SX/2016/073) and two representative strains (WZ202 and Kerala-RV01) detected in China and Indonesia with 95.9 and 99.8% nucleotide similarity, respectively. SX/2016/073 shared 90.8% nucleotide similarity and 92.8% amino acid similarity with the RotaTeq G1 strain RotaTeq-WI79–9 (Fig. 5, G1). Five Shanxi G2 strains (SX/2015/136, SX/2017/267, SX/2019/270, SX/2014/140, and SX/2016/011) were closely related to a Japanese strain (Tokyo 17–10) detected in 2017, with 99.8–99.9% nucleotide similarities. The other Shanxi G2 strain (SX/2018/393) was related to a South Korea strain (Seoul-710) with a nucleotide similarity of 94.9%. All the Shanxi G2 strains shared 92.7–92.9% nucleotide similarity and 95.1–95.5% amino acid similarity with the RotaTeq G2 strain RotaTeq-SC2–9 (Fig. 5, G2). Ten Shanxi G3 strains (SX/2017/001, SX/2017/021, SX/2017/080, SX/2017/172, SX/2017/020, SX/2017/473, SX/2017/082, SX/2017/047, SX/2018/089, and SX/2018/146) were closely related to the representative strains (E2432 and Tokyo 17–08) with high nucleotide similarities (99.3–99.6%). The other Shanxi G3 strain (SX/2017/271) was related to 1CR7 detected in South Korea (2017), with 99.3% nucleotide similarity. All the Shanxi G3 strains shared 81.8–94.1% nucleotide similarity and 92.5–97.7% amino acid similarity with the RotaTeq vaccine G3 strain RotaTeq-SC2–9 (Fig. 5, G3).

Sixty-three Shanxi G9 strains and 10 representative members were selected for G9 phylogenetic analyses (Fig. 6, Table 1). Except for SX/2015/069, 62 Shanxi G9 strains fell into 2 minor lineages (I and II). Fifty-five Shanxi G9 strains (belong to lineage 1) were closely related to each other with 99.2–99.9% nucleotide similarity, clustering with other strains isolated in Japan (Tokyo 18–43), USA (VU12–13-101), Thailand (T152), and China (SC6, Hu/JS2013, BJ-Q794) during 2011–2018 (Fig. 6, lineage I). Seven Shanxi G9 strains (SX/2015/061, SX/2015/065, SX/2015/081, SX/2015/059, SX/2015/132, SX/2015/156, and SX/2016/002) belong to lineage II were closely related to each other with 99.4–99.9% nucleotide similarity, clustering with a Chinese strain (km15119) isolated in Yunnan Province in 2016 (Fig. 6, lineage II). The SX/2015/069 was separate from other Shanxi G9 strains, but still has 99.2% nucleotide similarity with the reference strain HU/JS2013 (Fig. 6).

The P[4]-VP4 tree was constructed based on 9 Shanxi strains and 2 representative members (Fig. 7, Table 1). Eight Shanxi strains (SX/2015/136, SX/2019/271, SX/2019/284, SX/2018/400, SX/2019/270, SX/2016/041, SX/2017/267, and SX/2015/140) were closely related to WZ189 detected in Zhejiang Province (China) in 2017, with high nucleotide similarities (99.7–99.9%). The other Shanxi strain SX/2018/393 was closely related to the Hu/13–146 detected in Shanghai (China), with 98.0% nucleotide similarity (Fig. 7).

The P[8]-VP4 tree was constructed based on 93 Shanxi strains and 6 representative members (Fig. 8, Table 1). The Shanxi G9 strains fell into 2 minor lineages (I and II, Fig. 8). Eighty-seven Shanxi P[8] strains (belong to lineage 1) were closely related to each other with 96.9–99.9% nucleotide similarity, clustering with a Japanese strain (Tokyo 18–50) and 3 China’s strains (Z1602, km15119, SC1) isolated in previous studies during 2011–2018. The Shanxi P[8] strains belong to lineage I shared 91.9–93.0% nucleotide similarity and 93.1–94.0% amino acid similarity with the RotaTeq vaccine P1a[8] strain RotaTeq-WI79–4 (Fig. 8, lineage I). The other 6 Shanxi P[8] strains (SX/2016/124, SX/2016/172, SX/2016/112, SX/2015/144, SX/2015/117, and SX/2015/069) belong to lineage 2 were closely related to a Chinese strain Hu/JS-2012 detected in Jiangsu Province in 2012, with 98.9–99.3% nucleotide similarities (Fig. 8, lineage II). The Shanxi P[8] strains belong to lineage II shared 87.6–87.7% nucleotide similarity and 88.4–88.8% amino acid similarity with the RotaTeq vaccine P1a[8] strain RotaTeq-WI79–4 (Fig. 8).