The phylogenetic analysis of complete genome of EV71 circulating in mainland China in this present study showed all Chinese strains were clustered into C4 subgenotype group, except for HuB/CHN/2009 clustered into A and Xiamen/CHN/2009 clustered into B5(). Our previous studies on EV71 epidemiology showed that reported EV71 infections in mainland China had been associated with the only predominant subgenotype C4 viruses for more than 10 years, except for 2 orphan C genotype viruses were found in 1987 and 1997 respectively. The C4 subgenotype HEV71 has been the predominant endemic viruses circulating in mainland of China since 1998. However, there has been a shift of multiple subgenotypes of HEV71 circulating in other area, such as Taiwan, Australia and Malaysia. The patterns of HEV71 prevalence varied among different areas. Maybe because of the large and high density population in mainland of China, there is large newborn cohort every year becoming the susceptible population; therefore, the endemic viruses could be circulating in mainland China persistently for many years. While in other areas with small and low density population, the endemic viruses could be interrupted after circulating for a period of time with the increase of the population immunity, and then, other genotypes/subgenotypes HEV71 could be imported and becoming the endemic strains. It is also interesting that the molecular epidemiological pattern of HEV71 in mainland China appears to be similar to those of the measles virus and CVA-16 
, a single predominant endemic virus circulated in mainland China for a long time; but to be quite different from those of the rubella virus that different genotypes have co-circulated in China 
In this study, one B5 subgenotype virus was found in Xiamen (located in south of China) in 2009, which is the first B5 isolate found in China mainland. B5 subgenotype viruses were first found in Japan and Malaysia in 2003 
and later on circulating in other country or region, such as, Vienam 
and Taiwan 
.Genotype A is composed of the EV71 strain (BrCr-CA-70) identified in 1970 in the USA but was not detected afterwards until 2008 
. Yu et al 
reported the emergence of five isolates that are closely related to genotype A in central China. This was the first report of the occurrence of the modern contemporary A genotype EV71. The sequence homologies between HuB/CHN/2009 A strain and the prototype BrCr was 98.9%-98.0% for nucleotide acid and amino acid, respectively, in the full length genome. Based on the reported evolution rates calculated for HEV71 
nucleotide substitutions/site/year, the diversity between the oldest BrCr found in 1970’s and modern Chinese strains found in 2008–2009 would be at least 18%. Therefore, the occurrence of both HuB/CHN/2009 and Luan/CHN/2008 appeared weird. These A genotype viruses might be the laboratory adapted strains or the laboratory contamination. Although C4 subgenotype was the predominant strains circulating in mainland China for more than 10 years, the occurrence of orphan viruses of other genotype or subgenotypes, C, B5 and weird A, indicate the requirement of extensive surveillance of EV71 in mainland China should be strengthened.
No reported severe or fatal cases caused by C4b viruses occurred before large scale outbreak of HFMD in mainland China during 1993–2003 based on our analysis of complete genome. However, during the large scale outbreak of HFMD, increasing neurovirulence associated with C4a virus is a big concern for public health in mainland China. Our study on the analysis of complete genome, C4a viruses caused different phenotype of disease from mild to fatal. No specific lineages were associated with severe or fatal or mild cases. This indicated the viruses isolated from different phenotypic patients derived from the common ancestor and evolutes to different lineages by mutating gradually. We speculated that both viruses and host factor contributed to the phenotype of the disease.
Genomic recombinations are well known to contribute to genetic variations and evolution of enteroviruses. A range of enteroviruses of various serotypes or genotypes co-circulating in populations at some point of time was reported by different research groups 
. In the period of the enterovirus co-circulation, recombination between parts of genome of different serotype or genotype viruses may occur when different viruses infect and replicate in the same cell. This recombination process allows enteroviruses to create and maintain their genetic diversity and fitness. Several reports have shown their evidences for EV71 recombination, including intertypic and intratypic recombination between subgenotypes of EV71 and other prototype strains of HEV-A 
. But these studies reported the intratypic recombination between genotype/subgenotypes of EV71 based on the analysis of consensus sequences of genotype/subgenotypes viruses or the representative strains of genotype/subgenotype, not based on the oldest strain within the genotype or subgenotype group, which is the very important for the recombination analysis. In this study, we download all the complete genome sequences of EV71 from GenBank and searched by the published paper or the sequence submission information for the oldest strain within each subgenotype group of EV71, and then, the C4a and C4b viruses were aligned with the oldest strains of each subgenotype of EV71 and the prototype strains of HEV-A. Our comprehensive recombination analysis showed the evidence of genome recombination of genotype C4 (including C4a and C4b) sequences between EV71 genotype C structural genes and non-structural genes derived from the prototype strains of CAV16, 14 and 4, but the evidence of intratypic recombination between C4 strains and B subgenotype was not enough strong, since the B subgenotype showed the consistent high homology (80–86%) with C4a and C4b strains in 5′ end genome sequence(5′UTR-P2), while in the P3 region, higher homology was found between C4a, C4b strains and the prototype strains of CAV16, 14, 4, not B subgenotype EV71. In summary, these analyses showed the evidence of genomic recombination of C4a and C4b sequences between EV71 genotype C structural genes and non-structural genes derived from CAV16, 14, 4, not from B subgenotype EV71. This finding was inconsistent with the previous study 
Interestingly, the EV-71 isolates of subgenotype B3 shared the similar recombination pattern with C4. These isolates had high sequence similarity to EV-71 genotype B, CV-A4, CV-A14 and CV-A16/G10 at P2 genomic region (≥81%) and high sequence similarity to CV-A4, CV-A14 and CV-A16 at P3 genomic region ((≥83%), ). At the P3 genomic region, the sequence similarity of isolates of subgenotype B3 and C4 to the rest of the EV-71 genotypes was only between 75–79%. In the phylogenetic tree, the B3 strains consistently clustered with B1,2,4,5 in the genome sequences except for P3 region, where B3 segregated from other B subgenotype and clustered with CV-A4, CV-A14 and CV-A16 more closely. It is indicating that B3 shared the similar evolution pattern with C4 subgenotype viruses from the common ancestor: they likely ‘‘trap’’ sequences from other HEV-A viruses, thereby producing new individual viruses that differ from the parental strains during natural multiplication of HEV71 strains.
The clustering of isolates of subgenotype C4 with B3 and CVA4, 14, 16 at the 3' UTR genomic region was consistent with the previous clustering at the P2 and P3 genomic regions. No significant segregation (<30% bootstrap support), however, was observed for the remaining isolates and this was perhaps due to the short sequence length of the 3' UTR(~83nt). Based on these results, it appeared that sequences of genes at the 3' half of the EV-71 genome contributed to the multiple and diverse EV71 subgenotypes and these genes showed high similarity to different HEV-A viruses.
The incongruent phylogenies and simplot similarity analyses imply that recombination has played an evident role in the evolution of C4 EV71 viruses. C4a and C4b clearly contained sequences in the non-capsid region that are also present in CVA4,14,16, suggesting that these three HEV-A strains and C 4 subgenotype viruses of EV71 have a shared evolutionary history, despite their lack of similarity in the capsid region. However, the exact recombination counterpart of HEV-A could not be found because there is not sufficient data regarding the P2 and P3 sequences of the HEV-A in China or any other part of the world, but it may be assumed that genetic exchanges had occurred when the HEV71 strain co-circulated with other HEV-A during that time period in China. In this study, we have already performed the Nucleotide acid and deduced amino acid sequence identities analysis, and Phylogenetic analysis based on 5′UTR, P1, P2, P3, 3′UTR region of the genome respectively. Both identities and phylogenetic analysis indicated that both C4a and C4b viruses had much higher sequence identities with EV71 in P1 region; while in the P2 and P3 regions, both C4a and C4b viruses had much higher sequence identities with CVA-16,14,4. Although we combined the simplot analysis, the identities and phylogenetic analysis to confirm the recombination of C4 EV71 from the prototype strains of CAV16, 14 and 4, it is still a hypothesized conclusion.
Species human enterovirus A (HEV-A) which include 11 members of the coxsackievirus A (CV-A) group; CVA2–8, CVA10, CVA12, CVA14, CVA16 and human enterovirus 71 (EV-71) are associated with several human diseases 
. CVA4 cause herpangina 
and EV71, CVA14, CVA16 are highly associated with HFMD, and EV71 and CVA7 are occasionally associated with neurological diseases 
. Seiya Yamayoshi, et al reported that EV71, CVA7, CVA14 and CVA16, utilized the same cellular receptor SCARB2, a critical receptor common to all EV71 strains, for infection. Provided these SCARB2-dependent viruses sometimes co-circulate during an epidemic of HFMD 
, these viruses might have a high potential to undergo an intertypic recombination by co-infection of a SCARB2-expressing cell in vivo. In this study, we provided the evidence for the intertypic recombination of C4 subgentype between EV71 and CVA16, 14, 4. Based on the published co-infection and the cellular receptor study, we proposed that C4a and C4b viruses are intertypic recombination viruses between EV71 and other HEV-A strains derived from CVA16 or CVA14, not CVA4, since CVA4 infected cells via the different cellular receptor pathway and associated mainly with different clinical outcome, herpangina 
. The utilization of same cellular receptor between EV71 and CVA7 strains provided the possibility for the recombination between these two enteroviruses, which might appear as an emerging infectious pathogen, however, and may have unexpectedly high virulence, since both of them associated with neurological diseases. Careful and continuous surveillance of these viruses for the potential recombination is important for public health.
In this study, we provided the evidence confirming that these recombination C4 viruses have been occurred in China since 1998 and persistently circulated in China more than 14 years, and evolved into 2 major evolution lineages, C4a and C4b viruses during the decade. More and more severe neurological diseases and fatal cases have been caused by the intertypic recombinant C4a viruses throughout mainland China since 2007. A total of 5,034,764 HFMD cases including 61,582 severe and 1,894 fatal cases were reported to NNDRS during 2008–2011 in mainland of China 
, with increase of the severe and fatal cases year by year. The recombination C4a EV71 viruses have been associated with more than 80% of the severe cases and 92% of the fatalities since 2008 
. The reason for the epidemic of large-scale outbreaks of HFMD with increasing morbidity and mortality has been one of the most important issues in biomedical research in recent years. In the present study, except for the surveillance gap in mainland China during 1999–2000 and 2004–2006, we obtained the full scope of the EV71 epidemic based on the complete genome analysis. Our present and previous data indicated that C4b viruses had been circulating in mainland China for 5 years since 1998 (C4b was prevalent from 1998 to 2003 but has now disappeared from mainland China), which rarely caused severe disease and death in HFMD patients, and replaced by C4a viruses causing the nationwide outbreak with higher morbidity and mortality and caused many severe and fatal HFMD patients since 2007. The acquisition of a segment of genes from CV-A16 or CVA14 by EV71 C4b viruses could have rendered the virus more fit to adapt to a new environment especially the host immunity. And then, C4b viruses continued to adapt to its hosts via nucleotide mutations, while such mutations often lead to the changes in the pathogenicity, patterns of prevalence, and clinical manifestations of HEV71 infection, complicating clinical diagnosis, et al. Hence it could be the reason why C4b was rapidly replaced by C4a viruses which remained higher prevalence with increasing neurovirulence and transmissibility in mainland China since 2007. Additionally, it is possible that the C4a EV71 strains currently circulating in mainland China associated with higher morbidity and mortality obtained an unidentified neurovirulence determinant(s) in the viral genome and became more neurovirulent than those that circulated previously. Because it has been reported that an increase in neurovirulence levels can be caused by point mutations or by genetic recombination between avirulent poliovirus vaccine strains and nonpolio enteroviruses 
. Other research groups have identified the genetic differences between EV71 isolates from patients with severe and mild clinical disease and try to test the virulence phenotype of these recombinant viruses in the mouse model of infection via a reverse genetic approach. And our research team is currently using reverse genetics methods to identify the key nucleotide and amino acid differences between evolutionary branches C4a and C4b in order to discover the important determinants of neurovirulence and the possible reasons for the repeated outbreaks of HFMD in China in recent years. Our comprehensive study on the complete genome of C4 subgenotype of EV71 is significant for the prevention and control, vaccine development of EV71 in the world. However, genetic studies alone are not significant enough for vaccine development. It is critical to study antigenic variations for selecting vaccine strains