Totally thirty-four E18 strains were isolated. Eighteen E18 strains were obtained from patients with VE and VM in Hebei Province during 2015. Sixteen E18 strains were obtained from patients with HFMD in Shandong, Hebei, Shaanxi, Heilongjiang, Jiangsu, and Yunnan provinces in China between 2015 and 2016. The VP1 gene sequences of all E18 strains isolated from patients with HFMD were determined. Two E18 strains isolated from patients with HFMD and four E18 strains isolated from patients with VE were chosen for complete genome sequencing.
In total, 16 complete VP1 gene sequences and six genome sequences were acquired and submitted to the GenBank database (Table 1). In total, 100 complete VP1 gene sequences from E18 were used for analysis, including 84 retrieved from public databases dated before January 1, 2018. As indicated in our previous study, the phylogenetic tree revealed that E18 strains could be divided into genotypes A, B, and C with the support of the high confidence values (posterior value > 99%). Genotype C can be divided into subgenotypes C1 and C2. All E18 VP1 sequences obtained from patients in China after 2015 belonged to the subgenotype C2 branch (Chen et al. 2017). Except for HeB15-54462, all E18 VP1 sequences obtained from China after 2015 clustered together. There were no evident differences in phylogenetic analysis of the VP1 sequences between HFMD and VE/VM strains (Fig. 1).
Table 1. Sequences of the VP1 gene and complete genome of E18 strains in this study.
Figure 1. Phylogenetic tree (PT) based on the complete VP1 gene sequences of echovirus 18 (E18). The PT was constructed with BEAST software (100 million generations) using a lognormal relaxed clock, a constant size tree prior, and the GTR + G substitution model. The posterior probabilities from the MrBayes analyses supporting the trees are indicated at the nodes; only values over 95% are shown. Squares indicate China E18 strains after 2015. The filled squares indicate sequences of E18 isolated from HFMD cases, whereas the hollow squares indicate sequences of E18 isolated from VE/VM cases. Filled circles indicate the E18 prototype strain Metcalf. The scale bar indicates years. The strain name, year of sampling, and GenBank accession numbers are shown.
As shown in Supplementary Table S3, there were 72 amino acid substitutions between the prototype strain (Metcalf) and the clinical strains of subgenotype C in this study. Three amino acid substitutions (M104L, Y215F, and I216V) were only found in Chinese strains. The substitution of M104L occurred only in the Hebei strains, including strains from VE/VM and HFMD cases. For amino acids 6, 10, 262, and 275, almost all Chinese strains were consistent with the prototype strain, whereas most of the strains from other countries changed (R6K, N10D, V262T/A, and D275E). However, no significant differences between HFMD strains and VE/VM strains were found. Compared with the prototype strain Metcalf, as the E18 strains from VE/VM, one amino acid substitution at residue 84 (R→N/S) in the BC loop was also found in all the strains from HFMD. Further investigations are required to determine whether these amino acid changes significantly affected the prevalence of E18.
In total, 17 complete genome sequences of E18 were used for genomic analysis, including six from this study and 11 retrieved from the GenBank database. Phylogenetic analysis showed that these isolates were grouped into subgenotypes C1 (Germany) and C2 (one from Korea and nine from China) based on the complete VP1 region, except the prototype strain Metcalf. Eight of nine Chinese complete genome sequences were clustered together closely, whereas the HeB15-54462 strain had a closer phylogenetic relationship with the kor05-ECV18-054cn strain from Korea (Fig. 1).
No deletions or insertions were observed in the coding regions. Nucleotide substitutions in the clinical E18 strains were scattered all over the genome. For complete genomes of the genotype C strains, pairwise nucleotide sequence identities were 81.6%–99.9% among the 16 isolates and 98.8%–99.9% and 83.5%–99.7% within groups C1 and C2, respectively. A comprehensive comparison of nucleotide sequences based on each gene was also performed (Table 2). The sequences of structural protein-encoding regions (VP1–4) were highly conserved, whereas the P1 gene nucleotide sequence identities within the genotype C were more than 86.5%. Comparatively, nonstructural genes (2A, 2B, 2C, 3A, 3B, 3C, and 3D) exhibited much more diversity, particularly in the 2B region, with the lowest identity of 74.7% (Table 2).
Table 2. Pairwise nucleotide sequence identities based on corresponding regions of 17 complete genomic sequences of E18.
Phylogenetic relationship analyses based on protein-coding regions were performed with MEGA 5.0 software. The grouping of the strains in phylogenetic dendrograms based on P1, P2, and P3 regions was different (Fig. 2), suggesting that potential intratypic recombination had occurred between C1 and C2 strains.
Figure 2. Phylogenetic trees (PTs) showing the relationships between the E18 isolates and EV-B prototype strains. These PTs were constructed by the different genomic regions of 16 complete E18 genomic sequences and EV-B prototype sequences. The neighborjoining trees were reconstructed based on the P1, P2, and P3 regions. Triangles indicate the subgenotype C1. Filled circles indicate the subgenotype C2. Squares indicate the E18 prototype Metcalf strain. The percentages of bootstrap replicates (percentage of 1000 pseudoreplicate datasets) supporting the trees are indicated at the nodes; only values over 80% are shown.
In addition, potential multiple intertypic recombination events between the genome sequences of these E18 strains and other serotypes of species EV-B were also observed. Phylogenetic analyses based on nucleotide sequences of P1 showed that all 16 sequences exhibited the closest phylogenic relationship with the E18 prototype strain Metcalf (Fig. 2). However, based on the 5' UTR, 3' UTR, P2, and P3 regions, subgenotype C1 and C2 sequences clustered with different EV-B serotypes, but not near the E18 prototype strain (Fig. 2, Supplementary Figure S1). The different topologies of the phylogenetic trees between P1 and P2–3 regions suggested that multiple intertypic recombinations of these circulating E18 strains with other EV-B serotypes may have occurred in noncoding or nonstructural protein-coding regions.
To further depict the recombination events, similarity plots and bootscan analyses were performed using Simplot software v3.5.1 with the EV-B prototype strain as reference sequences. All the strains in this study were analyzed (Fig. 3). The results revealed that multiple recombination events occurred between the genomic sequences of representative strains and other EV-B serotypes. As indicated in the results, these sequences showed highest similarity and the closest phylogenetic relationship with the E18 prototype strain Metcalf in the P1 region. However, in some segments of the 5' UTR, P2, and P3 regions, the representative strains showed lower similarity with E18 Metcalf (Fig. 3A, 3C, 3E, and 3G). The bootscanning results further confirmed the phylogenetic relationships (Fig. 3B, 3D, 3F, and 3H). As illustrated in our previous study (Chen et al. 2017), two apparent crossing sites in the 5' UTR and 2A gene regions revealed intertype recombination events that were consistent with the deduction described above. According to the bootscanning results, it was hard to confirm the donor sequences.
Figure 3. Similarity plot and results of bootscanning analyses of all C2 genotypes of E18 and other EV-B prototype strains based on the fulllength genomes. Six strains in this study were chosen as representative strains. A sliding window of 500 nucleotides moving in 20-nucleotide steps was used in this analysis. A, B HeB15-54462; C, D HeB15-54498; E, F E18-221, E18-291, and E18-398; G, H E18-393. A, C, E, G Results of similarity plot analyses; B, D, F, H results of bootscanning analyses.
VP1 Sequence Analysis
Complete Genome Analysis
Phylogenetic Analysis of P1, P2, and P3 Coding Regions
Table S1. List of nucleotide sequences of primer for amplification of the whole genome sequences of E18
Table S2. List of the 84 complete VP1 encoding sequences and 11 complete genome sequences of E18 downloaded from GenBank.
Table S3. Amino acid changes in VP1 gene of E18 strains belonged to subgenotypes C1 and C2 comparing with the prototype strain Metcalf.
Figure S1. Phylogenetic trees showing the relationships between the E18 isolates and EV-B prototypes. These were constructed by the different genomic regions of 17 complete E18 genomic sequences and other EV-B prototype sequences. The neighbour-joining trees were reconstructed based on the 5′ UTR and 3′ UTR regions, respectively. Triangles indicated the subgenotype C1. Filled circle indicated the subgenotype C2. Squares indicated the E18 prototype Metcalf strain. The percentage of bootstrap (percentage of 1000 pseudoreplicate datasets) replicates supporting the trees are indicated at the nodes.