In all 293 HA sequences, 93.9% (275) of EA SIVs were EA H1N1 SIVs and 6.1% (18) were EA H1N2 SIVs. From 2001 to 2018, 275 EA H1N1 SIVs sequences were collected from nineteen provinces/municipalities of China (Fig. 1A). Hong Kong contributed the most EA H1N1 SIVs sequences, followed by Guangdong and Shandong provinces. The first EA H1N1 sequence was reported in Hong Kong in early 2001. Subsequently, more EA H1N1 SIVs sequences were reported and peaked during 2009–2011 (Fig. 1B). Human cases infected with EA H1N1 SIVs were reported from Jiangsu, Hebei, Hunan, Yunnan, and Fujian provinces (Table 1). These results indicated that the EA H1N1 SIVs were currently prevalent in pigs and sporadically cause human infections in China.
Figure 1. The description analysis of EA H1N1 SIVs sequences in China during 2001–2018. A Geographic distributions of EA H1N1 SIVs sequences in China. EA H1N1 SIVs affected regions are highlighted in blue. Blue from light to dark indicated the virus EA H1N1 SIVs number increased from 1 to 90. The red circles indicated human cases of infection with EA H1N1 SIVs. B Temporal distribution of EA H1N1 SIVs sequences in swine in China during 2001–2018.
Year Province Strain name Genotype References 2011 Jiangsu A/Jiangsu/1/2011 1 Yang et al. (2012) 2012 Hebei A/Hebei-Yuhua/SWL1250/2012 1 Wang et al. (2013) 2015 Hunan A/Hunan/42443/2015 3 Zhu et al. (2016) 2015 Yunnan A/Yunnan-Longyang/SWL1982/2015 3 Zhu et al. (2019) 2015 Yunnan A/Yunnan-Wuhua/SWL1869/2015 3 Zhu et al. (2019) 2016 Fujian A/Fujian-Cangshan/SWL624/2016 5 Xie et al. (2018)
Table 1. Human cases infected with EA H1N1 SIVs in China.
To determine the evolution of EA H1N1 SIVs in China during 2001–2018, phylogenetic analyses of all eight gene segments were performed. Phylogenetic tree of hemagglutinin (HA) showed that the EA H1N1 SIVs isolated from pigs in China formed a monophyletic group (Fig. 2A, red branches). We used the swine H1 clade classification tool to classify the clade of HA (Anderson et al. 2016) and found that all the HA sequences of EA H1N1 SIVs were classified as 1C.2.3. In contrast, the other genes of EA H1N1 SIVs, including the neuraminidase (NA), basic polymerase 2 (PB2), basic polymerase 1 (PB1), polymerase (PA), nucleoprotein (NP), matrix (M), and nonstructural protein (NS), were demonstrated distinct diversity (Fig. 2B–2H). The NA genes were derived from EA H1N1 and A(H1N1)pdm09. Origins of the PB2, PB1, PA, NP and M genes included EA H1N1, A(H1N1)pdm09 and TR H1N2. The NS genes were derived from EA H1N1, A(H1N1)pdm09, and TR H1N2 respectively. These results indicated that dynamic reassortments occurred between the EA H1N1 and co-circulated SIVs in pigs.
Figure 2. Phylogenies of HA (A), NA (B), PB2 (C), PB1 (D), PA (E), NP (F), M (G), and NS (H) of EA H1N1 SIVs from 2001 to 2018. The phylogenetic analysis was performed by MEGA 7.0 with maximum likelihood (ML) method. The bootstrap was 1000. The red indicated the sequences of EA H1N1 SIVs. The black indicated the reference sequences.
Reassortant genotypes of the EA H1N1 SIVs could then be defined based on the clade distributions of their internal gene segments. The EA H1N1 SIVs isolated from both pigs and humans were classified into 11 distinct genotypes, from genotype 1 to genotype 11. Genotype 1 viruses accounted for 55.3% of the EA H1N1 SIVs. And the eight gene segments of this genotype viruses were exclusively originated from EA H1N1 (Fig. 3). EA H1N1 and TR H1N2 viruses produced double reassortment viruses and generated genotypes 2, 6, and 10 (Fig. 3). The introductions of the A(H1N1)pdm09 into pigs continuously provided their internal genes to EA H1N1 SIVs and generated novel genotypes 4, 7, and 8 (Fig. 3). Reassortments of EA H1N1, A(H1N1)pdm09 and TR H1N2 generated triple reassortment viruses and formed genotypes 3, 5, 9, and 11 (Fig. 3). These results indicated that the EA H1N1 SIVs circulated in China exhibited a high genotypic diversity.
Figure 3. Genotypes of EA H1N1 SIVs identified in China. The name of representative viruses are showed (left). All the eight gene segments of EA H1N1 SIVs are at the top of the graph. Origin of each gene segment is indicated by a colored block for representing the different swine influenza virus lineages. The 11 distinct genetic constellations are labeled as genotype 1 to 11. EA H1N1, Eurasian avian-like H1N1; A(H1N1)pdm09, 2009 pandemic H1N1; TR H1N2, triple reassortment H1N2.
Genotype 1 was widely prevalent from 2001 to 2013 (Fig. 4A). From 2009 to 2013, the genotypes 2 and 4 were co-circulated with genotype 1 (Fig. 4A). Since 2013, genotypes 3 and 5 have gradually replaced genotypes 1, 2, and 4 in swine populations (Fig. 4A). Guangdong Province had the largest number of genotypes, with a total of 7 genotypes (Fig. 4B). In addition, Hong Kong and Guangxi ranked second and third, respectively (Fig. 4B). This result indicated that genotype 3 and 5 have become predominant in pig population.
Figure 4. Development and prevalence of EA H1N1 genotypes during 2001–2018. A 11 distinct genotypes are listed on the left. The colored circles represent the corresponding genotypes and the isolated time. B The distribution of EA H1N1 genotypes in each province. The size of the circle indicated the number of EA H1N1 SIVs. The different colors represent the different genotypes which refer to A.
We next analyzed the molecular characteristics which were associated with mammalian adaptations, receptor binding ability, virulence or transmission and antiviral resistance of all EA H1N1 SIVs (Table 2).
Gene Phenotypic characteristic(s) Mutation Percentage HAa Altered the receptor specificity E190D E (0%) D (85.9%) G225E G (12.7%) E (84.1%) NAb Resistance to NA inhibitors H274Y H (100%) Y (0%) N294S N (100%) S (0%) PB2 Altered the virulence in mice D9N D (98.1%) N (1.1%) L89V L (0.4%) V (98.9%) E158G E (100%) G (0%) Mammalian host adaption D256G D (100%) G (0%) Enhance viral polymerase activity T271A T (68%) A (30.8%) Enhance the 627 k and 701 N function K526R K (94.4%) R (5.6%) Restored the polymerase activity M535L M (100%) L (0%) Enhance the viral polymerase activity, Q591K Q (66.2%) K (0.4%) increase the virulence in mice E627K E (98.5%) K (0.4%) D701N D (30%) N (70%) Altered the virulence in mice A676T A (2.2%) T (96%) PB1 Enhance the viral polymerase activity L473V L (0.4%) V (99.6%) Altered the virulence in mice R198K R (0.8%) K (99.2%) PA Enhance the viral polymerase activity L336M L (72.9%) M (26.8%) Enhance the viral polymerase activity, increase the virulence in mice K356R K (74.4%) R (25.7%) Species-associated signature positions S409N S (17.5%) N (81.0%) NP Enhance the viral polymerase activity A150R A (0%) R (98.9%) Enhance the viral polymerase activity, N319K N (85.6%) K (14.4%) altered the virulence in mice Q357K Q (66.7%) K (31.5%) M1 Increase the transmission in guinea pigs P41A P (0%) A (98.6%) Altered the virulence in mice T215A T (0%) A (100%) M2 Resistance to adamantine derivatives S31N S (0.7%) N (98.9%) NS1 Altered the virulence in mice D92E D (97.1%) E (1.5%) Altered the antiviral response in host N205S N (10.2%) S (54.2%) G210R G (12.7%) R (56.7%) aThe H3 numbering system was used.
bThe N2 numbering system was used.
Table 2. Prevalence of key molecular markers in EA H1N1 SIVs in China.
The influenza virus HA gene was a major factor that determines the receptor binding property and the host range. It was well known that E190D and G225E could increase the receptor-binding affinity to human type a-2, 6- linked sialic acid receptors (Tumpey et al. 2007). And majority EA H1N1 SIVs contained HA-190D (85.9%) and 225E (84.1%) mutations. Studies have shown that adaptive mutations of influenza A viruses were identified mostly in viral polymerase complexes (Chen et al. 2007). Mutations PB2-E627K and D701N not only enhanced the viral polymerase activity but also increased the virulence of H7N9 and H5N1 avian influenza viruses in mammals (Chen et al. 2007; Steel et al. 2009; Zhu et al. 2015). The substitution PB2-D701N could also enhance viral replication and pathogenicity of EA H1N1 viruses in mice (Liu et al. 2018). 70.0% of the EA H1N1 SIVs posed N at position 701 in the PB2 protein. Whereas, 98.5% of all EA H1N1 SIVs at 627 site were E. Another PB2-Q591K mutation was shown to enhance polymerase activity, replication and virulence in mice in H5N1 influenza virus (Yamada et al. 2010). The percentage of 591K residue in EA H1N1 SIVs was 0.4%. In addition to key adaptive signatures in PB2 protein, several other mutations in PB1, PA and NP were implicated in enhanced viral polymerase activity including PB1-L473V (Xu et al. 2011), PA-K356R (Xu et al. 2016), NP-Q357K (Zhu et al. 2019) and so on. Theses mutations in EA H1N1 SIVs were PB1-L473V (99.6%), PA-K356R (25.7%), and NP-Q357K (31.5%), respectively. Furthermore, it was also reported that M1- P41A mutation in EA H1N1 SIVs increased the transmission in guinea pigs (Campbell et al. 2014). We found that 98.6% was A at position 41 in M1 protein in EA H1N1 SIVs. In our study, we analyzed the percentage of these mutations in all eight gene segments of EA H1N1 SIVs. We found that some molecular markers were already widespread in EA H1N1 SIVs, such as PB1-L473V (99.6%), and M1-P41A (98.6%). There were still some molecular markers including PB2-E627K (0.4%) and PB2- K526R (5.6%) that account for a very small proportion in EA H1N1 SIVs. Therefore, continuous surveillance needs to be implemented.
Antiviral drugs M2 and neur-aminidase inhibitors played important role in influenza treatment (Babu et al. 2001; Hay et al. 1985). It was known that M2-S31 N mutation was resistant to amantadine and rimantadine by changing in the transmembrane channel domain (Shiraishi et al. 2003). We found that 98.9% of EA H1N1 SIVs posed 31N in the M2 protein. It suggested that the EA H1N1 SIVs were resistant to amantadine and rimantadine widely. H274Y and N294S mutations caused resistance to neuraminidase inhibitors (Ives et al. 2002). However, these substitutions were not observed in all EA H1N1 SIVs, which suggested the neur-aminidase inhibitors could be used for human infections with EA H1N1 viruses.
The EA H1N1 SIV outside European countries was detected in Hong Kong in 2001. Currently, EA H1N1, CS H1N1, TR H1N2 and A(H1N1)pdm09 influenza viruses co-circulated in pigs in China (Chen et al. 2013). The cocirculation of these lineage viruses in pigs resulted in an increased number of novel reassortment viruses. A total of 11 different genotypes were identified. Genotype 1 viruses have all of their eight gene of avian-origin, which were widely prevalent from 2001 to 2013. Higher genetic diversity was detected between 2009 and 2014, with all genotypes were identified. This might have been caused by strengthened surveillance of SIVs since A(H1N1)pdm09 virus pandemic. During this period, genotype 1, 2, and 4 were prevalent in China, nevertheless, genotype 7–11 were detected only one time. Since 2013, genotype 3 and 5 have become increasingly prevalent and have a selective advantage over original EA H1N1 viruses. Although Hong Kong contributed the largest number of EA H1N1 viruses and had the second largest number of genotypes, most of the swine slaughtered in Hong Kong come from provinces in Chinese mainland (Vijaykrishna et al. 2011). This indicates most of swine influenza viruses detected in Hong Kong were imported from Chinese mainland.
The prevalence of EA SIVs in pigs could cause the human infections. In China, human infections with EA H1N1 SIVs were reported occasionally. The first human infection with EA H1N1 SIVs was identified in Jiangsu Province in 2011 (Yang et al. 2012). Thereafter, another 3-year-old boy was identified as EA H1N1 SIV case in Hebei Province in 2012 (Wang et al. 2013), which belonged to genotype 1. In 2015, one and two human cases with EA H1N1 SIVs infections were reported in Hunan and Yunnan provinces, respectively (Zhu et al. 2016, 2019), which were classified as genotype 3. Fujian Province reported its first EA H1N1 human case in 2016 (Xie et al. 2018), which was listed as genotype 5. These were consistent with the prevalence of EA H1N1 SIVs circulating in pig populations during the same period.
The available full genome sequences may have some bias. Hong Kong contributed the largest number of EA H1N1 sequences in the public database. It might be due to systematic surveillance since 1998 in Hong Kong (Vijaykrishna et al. 2011). Since 2007, swine influenza virus surveillance began to been implemented in Guangdong, Guangxi, Shandong and other provinces in Northern China (He et al. 2018; Liu et al. 2009; Sun et al. 2016; Yang et al. 2016; Zhu et al. 2011). Therefore, although all EA H1N1 sequences from the public database have exclusively been included for analysis, they might not cover all of the evolutionary image of the viruses.
In our study, some important molecular signatures were analyzed. The specific amino acid mutations of HA could switch the receptor preference from avian-type receptors to human-type receptors. It was noted that majority of all EA H1N1 SIVs posed 190D (85.9%) and 225E (84.1%), which were preferentially bind to the human-type receptors and caused human infections. Some adaptive mutations including PB2-T271A, PB2-Q591K, PB2-E627K, and PB2-D701 N were able to enhance viral polymerase activity and further facilitated pathogenicity in mice (Bussey et al. 2010; Yamada et al. 2010; Zhu et al. 2015). The EA H1N1 viruses beard 271A (30.8%), 591K (0.4%), and 701N (70.0%) in the PB2 protein, respectively. Through the analysis of antiviral resistance molecular characteristics, we found that EA H1N1 SIVs were resistant to amantadine and rimantadine. However, they were sensitive to neur-aminidase inhibitors. Hence, we should pay more attention to these molecular signatures and identify their effects on pathogenicity, transmission and antiviral resistance of EA H1N1 SIVs.
Taken together, our finding suggested that dynamic reassortments among EA H1N1 SIVs and other swine influenza viruses were continuously occurring in pigs. Occasionally, it has caused human infections, therefore, we should strengthen the monitoring of EA H1N1 SIVs.