In the present study,the complete genomes of four common (4/EV71/Wenzhou/CHN/2014,15/ EV71/Wenzhou/CHN/2014,116/EV71/Wenzhou/ CHN/2014,and 120/EV71/Wenzhou/CHN/2014) and two virulent (11/EV71/Wenzhou/CHN/2014 and 109/EV71/Wenzhou/CHN/2014) enterovirus 71 (EV71) isolates were sequenced and described. They are 7405 bp in length and belong to EV71 sub-genotype C4 (C4a cluster). Nucleotide sequence alignment revealed six nucleotide variations (G^P151→T^P151,G^P199→A^P199,G^P261→T^P261,A^P328→C^P328,G^P422→A^P422,and G^P437→T^P437) in the two virulent isolates within the 5’UTR of the IRES element. RNA secondary structure predictions of IRES and FCE indicated that the common isolates shared similar structures,which were different from those of the virulent isolates. Moreover,the G^P114→C^P114 and G^P151→T^P151 mutations in the virulent isolates contributed to the formation of the unique RNA secondary structures in SL II. Furthermore,nucleotide/amino acid sequence alignments of 82 EV71 isolates indicated that six sites (T^P488 and C^P577 in the 5’UTR; Asn^P57 in 2A; Ile^P56 in 3C; C^P10 and A^P47 in the 3’UTR) are potentially associated with the neurovirulence of EV71. Finally,the 3D structures of 2A were analogous,whereas the structures of VP1 and 3C were variable.

Enterovirus 71 (EV71) is a non-enveloped virus containing a positive-sense single-stranded RNA genome (approximately 7.4 kb) and belongs to the Enterovirus genus of the family Picornaviridae. Structurally,six stem loop (SL) structures,SL I to SL VI,are contained in the EV71 5’UTR. Functionally,the 5’UTR is divided into three regions: a 5’ terminal cloverleaf (1 to 89-101 bp),an internal ribosome entry site (IRES) element (123-126 bp to 602-605 bp,ranging from SL II to SL VI),and a hypervariable region (the last 120 bp)^[1]. This type of positive-sense single-stranded RNA virus utilizes a cap-independent translation initiation mechanism to regulate the expression of its viral genome^[2]. Jerrey et al. found that the functional cis-acting element (FCE,spanning SL IV to SL VI within the 5’UTR) enabled Poliovirus mRNA to translate in a cap-independent fashion and that the major determinant of this FCE is located in a domain (nts 320-631)^[3]. Unlike other hand-foot-mouth disease (HFMD) pathogens [such as Coxsackie virus A (CVA)],EV71 infection is more likely to cause severe neurological complications,including brain stem encephalitis,aseptic encephalitis,meningitis,poliomyelitis-like paralysis,neurogenic pulmonary edema,and even death^[4]. Some studies have reported that the nucleotide mutations within the 5’UTR (such as the stem loop II,SL II,nts 105-179) and some amino acid variations are associated with EV71 neurovirulence^[5-6]. Generally,common EV71 isolates cause mild and self-limited herpes outbreaks on the skin and mucosae,whereas virulent EV71 isolates (‘virulent’ particularly refers to neurovirulent) lead to severe neurological complications and even death^[7].

In the present study,patients diagnosed with either mild HFMD (with herpes only on the skin and mucosa) or severe HFMD (with both herpes and meningitis) were recruited from the Second Affiliated Hospital of Wenzhou Medical University. Throat swab samples or cerebrospinal fluid samples were filtered through a 0.22-μm sterile filter,and the filtered suspension was added to either RD cells or Vero cells,which were maintained in RPMI-1640 medium supplemented with 10% FBS,100 U/mL penicillin,and 100 μg/mL streptomycin. After an infection period of 1.5 h in a 37 °C/5% CO₂ incubator,the cell supernatant was discarded,and fresh RPMI-1640 medium without FBS was added to the cells. After a 4-d incubation,the supernatant was blind inoculated into freshly prepared cells several times until the cells exhibited a cytopathic effect (CPE),as scored by ‘+++’ (the majority of cells had shrunk and become dislodged from the bottom of the culture bottles). Based on EV71-specific primers (Table S1 in the website of BES,www.besjournal.com),RT-PCR assays were conducted to amplify six overlapping gene segments (638 bp,1575 bp,1433 bp,1624 bp,2286 bp,and 786 bp),which span the complete genome of EV71,and the PCR products were sequenced. The complete genome sequences and deduced amino acid sequences of six EV71 isolates were deposited in the GenBank database. The GenBank Accession Numbers of four common isolates (4/EV71/ Wenzhou/CHN/2014,15/EV71/Wenzhou/CHN/2014,116/EV71/Wenzhou/CHN/2014,and 120/EV71/ Wenzhou/CHN/2014) were KT008669,KT345960,KT008672,and KT345959,respectively. Two virulent isolates (11/EV71/Wenzhou/CHN/2014 and 109/EV71/Wenzhou/CHN/2014) were given the following accession numbers: KT008670 and KT008671,respectively. In this study,six nucleotide variations (G^P151→T^P151,G^P199→A^P199,G^P261→T^P261,A^P328→C^P328,G^P422→A^P422,and G^P437→T^P437) within the IRES region were observed simultaneously in the two virulent EV71 isolates when compared with the four common isolates (Figure 1A). RNA secondary structure predictions for IRES,which includes domain I to domain V,were carried out. The results showed that the four common isolates shared similar RNA structures,whereas the two virulent isolates had a large circular stem loop structure between domain II and domain III. This structural difference between the common and virulent isolates may have an impact on RNA translation and neurovirulence (Figure 2). In order to learn more about the subtle RNA structure variations,we also performed RNA secondary structure predictions for two specific functional regions: SL II and FCE. The results indicated that the variation of G^P114→C^P114 in virulent isolate 11/EV71/Wenzhou/CHN/2014 contributed to the formation of an additional stem loop structure and that the mutation of G^P151→T^P151 in both virulent isolates producedan enhanced stem loop structure (with two more nucleotides) (Figure 3A). Moreover,the variat ions of G^P114→C^P114 and G^P151→T^P151 resulted in higher minimum free energy (-12.4 kcal/mol and -17 kcal/mol,respectively) and contributed to the formation of relatively stable secondary structures (Figure 3A). The predicted secondary structures of the FCE of the four common isolates were similar,whereas the structures of two virulent isolates were variable (Figure 3B). Wen et al.demonstrated that the 5’UTR RNA secondary structures of virulent EV71 isolates differed from those of common isolates^[8]. However,they did not perform a detailed analysis of additional subtle variations in key regions within the 5’UTR. In this study,instead of repeating previous studies on predictions of the RNA secondary structure of the whole 5’UTR,we selected the IRES for RNA secondary structures analysis and specifically studied two special regions (SL II and FCE) that were previously reported to be associated with neurovirulence and viral mRNA translational efficiency,respectively. In agreement with previous studies on the structure of the whole 5’UTR,the secondary structures of SL II and FCE were similar among the four common isolates but differed from those of the two virulent isolates. Furthermore,sequence alignment of the 3’UTR (84 nts) suggested that the nucleotides in the 3’UTR were relatively conserved and that only four nucleotide positions were variable. The variation of T^P10→C^P10 or A^P10 was found in two virulent isolates when compared with four common isolates (Figure 1B).

Next,76 EV71 isolates along with the previous six EV71 isolates were divided into ‘Virulent’ and ‘Common’ groups (Table S2 in the website of BES,www.besjournal.com). The nucleotides/amino acids that exhibited large differences between the ‘Virulent’ and ‘Common’ groups were picked and subjected to Chi-square tests. For example,in the 488th position of the 5’UTR,there were two ‘T’s in the ‘Common’ group but seven ‘T’s in the ‘Virulent’ group. The results of the Chi-square tests showed that this difference was statistically significant,and it was concluded that this difference was associated with the neurovirulence of EV71 (χ²=4.204,P= 0.04,α=0.05) (Table 1). As a result,two alternative nucleotides (T^P488 and C^P577) in the 5’UTR,two alternative nucleotides (C^P10 and A^P47) in the 3’UTR,and three alternative amino acids (Gln^P22 in VP1: Asn^P57 in 2A; Ile^P56 in 3C) were considered to be relevant to neurovirulence. To expound the impact of these neurovirulence-related amino acids on the structures of the corresponding proteins,3D structure prediction software for proteins (https://swissmodel.expasy.org/interactive/xTemLr/models/) was used to predict the structures of VP1,2A,and 3C. We wanted to determine whether,like the predicted RNA structures of the 5’UTR,the 3D structures of these isolates differed between the common and virulent isolates. Unfortunately,the 3D structures were not noticeably different between the common and virulent isolates. 2A had almost the same structure for all six isolates,whereas both VP1 and 3C were variable among the six isolates. The on-line software (ProMod3 Version 1.0.0.) used for the analysis is the most commonly used software for the 3D structure analysis of proteins,and all sequence identities between our amino acid sequences and the models built on-line were above 90%,with several sequence identities reaching 100% (seq identities ＞50% were conclusive; for more information,please refer to the SWISS-MODEL Homology Modeling Reports in the supplementary materials) (Figure S2 in the website of BES,www.besjournal.com). As a result,these 3D results are probably attributed to the following factors: (1) the limited number of samples on 3D structure predictions; (2) a single amino acid change in a pivotal site may contribute more to EV71 neurovirulence than the structure of the corresponding protein; and (3) a crucial amino acid may determine EV71 vitality when binding to host cells. In summary,further studies are needed to determine how these variations affect the neurovirulence of EV71.

Researchers have explored the association of certain nucleotides/amino acids with the neurovirulence of EV71. For example,Liu et al. reported four neurovirulence-related amino acid substitutions,His^P22→Gln^P22,Glu^P98→Lys^P98,Glu^P145→Gly^P145,and Ala^P289→Thr^P289,in VP1^[9]. Val^P263→Ile^P263 in 3D of virulent isolates was demonstrated to be a unique amino acid variation that distinguished the virulent isolates from common isolates^[10]. Among the seven nucleotides/amino acids identified in the present study,Gln^P22 in VP1 has already been described^[9]. However,neither the four common isolates nor the two virulent isolates had this neurovirulence-associated amino acid. Of the six newly identified neurovirulence-associated nucleotides/amino acids,five variations were unique to the two virulent isolates. It should be emphasized that we searched and analyzed the whole genome of almost all virulent EV71 isolates (C4 subgenotype) that had been described in either published papers or the GenBank database. Moreover,all analyzed EV71 isolates were carefully chosen. All common isolates were from patients with herpes only on the skin and mucosae and without any neurological symptoms. Virulent isolates were isolated from patients with not only herpes but also at least one of the following severe neurological complications: brain stem encephalitis,aseptic encephalitis,meningitis,poliomyelitis-like paralysis,and neurogenic pulmonary edema. Most importantly,the number of EV71 isolates (either common or virulent isolates) utilized in the present study was higher than in any previous EV71 genome sequence analysis of C4. Therefore,all of these factor make our findings more convincing.

Taken together,we demonstrated the genetic characteristics of six native EV71 isolates and identified six novel nucleotide/amino acid sites (T^P488 and C^P577 in the 5’UTR; Asn^P57 in 2A; Ile^P56 in 3C; C^P10 and A^P47 in the 3’UTR) that were concluded to be associated with the neurovirulence of EV71. Furthermore,the RNA structure predictions of IRES,SL II,and FCE suggested that there may be a relationship between EV71 neurovirulence and the RNA secondary structures of IRES,SL II,and FCE in the 5’UTR. The results of the present study add to the knowledge on EV71 origin,evolution,and neurovirulence and will aid in developing safe and effective vaccines. In the future,further studies should be conducted to determine whether and how these amino acid/nucleotide variations,as well as changes to the corresponding RNA and protein structures,affect neurovirulence using cell culture systems,animal models,and molecular biology techniques.

Acknowledgement

We would like to thank the participants for their cooperation in this study.