Analysis of Echovirus 30 Isolates from Russia and New Independent States Revealing Frequent Recombination and Reemergence of Ancient Lineages

MATERIALS AND METHODS

We used 48 strains of E30 isolated in the course of the WHO polio surveillance program and enterovirus surveillance (Table 1). Virus isolation and identification were carried out according to a standard WHO protocol (24) by using RD and Hep-2 cell cultures. Strain serotypes were identified by neutralization test with antisera produced by RIVM (Bilthoven, The Netherlands). Virus RNA was isolated from cell culture supernatant by guanidine thiocyanate lysis and adsorption to silica (3). Reverse transcription was carried out using Moloney murine leukemia virus reverse transcriptase (Promega) and random hexanucleotide primers. PCR was performed with previously published oligonucleotides for VP1 (15) and 3D (11). All nucleotide positions are given according to the prototype E30 strain Bastianni (GenBank accession no. AF311938). After amplification, bands were excised from agarose gels, purified using QIAquick kits (Qiagen), and sequenced directly with the PCR primers. Nucleotide sequences were aligned using ClustalX software (23). Phylogenetic trees were constructed with ClustalX (neighbor-joining algorithm), with correction for multiple substitutions and excluding positions with gaps, and with 1,000 bootstrap pseudoreplicates. RNA and protein sequence distances were calculated with the PHYLIP software package (6). GenBank accession numbers for sequences are provided in Table 1.

RESULTS AND DISCUSSION

In the VP1 genome region (nucleotides [nt] 2460 to 3335), all E30 strains differed by up to 31% of nucleotide sequence (DNADIST, Kimura model) and by up to 20.5% of protein sequence (PROTDIST, Jones-Taylor-Thornton distance matrix). All strains grouped with the prototype E30 Bastianni relative to enteroviruses of other serotypes, thus confirming the serotype. We used 75 of over 150 GenBank sequences available for E30 VP1 that were reported previously (1, 15, 21), excluding very similar (less than 2% nucleotide sequence difference) strains from the same geographic region. The nucleotide sequence of the prototype E21 Farina, the closest HEV-B serotype to E30, was used to root the phylogenetic tree, but was omitted from Fig. 1. Most of the strains studied fell into a major phylogenetic group that included most E30 strains isolated worldwide since 1978 and sequenced so far. This group was reported previously in several studies (15, 21); therefore, the majority of the strains from Russia and the NIS studied here were similar to other recent E30 isolates reported elsewhere. One explanation for the observed global group of modern strains that differ from each other far less (below 13% nucleotide sequence, PHYLIP Dnadist, Kimura model) than do the E30 isolates from the 1960s and the 1970s (up to 32% of nucleotide sequence) is probably the higher fitness of this genotype, as was hypothesized previously (21). As shown by Fig. 1, most of the modern strains had the same putative ancestor relative to the strains from the 1960s and 1970s. We also suggest that an increase in economical ties and the spread of air travel have resulted in the emergence of a common epidemiological space in Europe and North America, where a more fit capsid could efficiently prevail in a matter of several years.

• Open in new tab
• Download powerpoint

FIG. 1.

Phylogenetic tree (ClustalX, neighbor joining) for complete VP1 genome region (nt 2460 to 3335). Numbers at tree nodes are a percentage of bootstrap pseudoreplicates that supported a group below; bootstrap values below 70% were omitted. The tree was rooted with E21 Farina strain (omitted). E30 sequences from GenBank have accession numbers. ▪, group 1; +, group 2; *, group 3; •, group 4; ▴, group 5.

Within the “modern” group, phylogenetic subgrouping did not always correlate with isolation location, but rather correlated with the time of isolation, which indicates a very dynamic epidemiological pattern. In general, we observed limited temporal overlap between subgroups that circulated in Russia and the NIS. Sublineages of E30 demonstrate quick and wide spreading within months over vast distances and then vanish to give space to newer lineages. It seems unlikely to us that the epidemiology of E30 is significantly driven by herd immunity pressure, as the sequence difference between most subgroups implies only minor serological differences.

Some strains in the major modern group, such as those in subgroups 1, 2, and 3 (Fig. 1), presumably arrived from Western Europe not very long before isolation in Russia. These strains originated mainly from Russia and the western NIS; however, strains of subgroup 2 could be found in Middle Asia (in Kirghizia). This observation is not unexpected and shows that the epidemiological space of the former USSR is significantly integrated into the European epidemiological space. Strains that comprise subgroup 4 seemingly diverged from the main E30 lineage over two decades ago, as they are almost an out-group within the major modern group. Viruses of subgroup 4, especially strain 23199, were similar to recent Chinese isolates and probably represent an Asian sublineage. It should be noted that we observed nonuniform rates of E30 isolation within Russia and the NIS. Most strains studied here came from Russia and Caucasia and Western countries, and only several E30 strains were among many HEV-B isolates from Middle Asia.

Rather unexpectedly, we observed a group of six strains (subgroup 5) (Fig. 1) that was strikingly different from most other modern E30 strains. These isolates were only somewhat similar to a Columbian strain isolated in 1995 and probably originated from countries where enterovirus surveillance was not performed. A close look at subgroup 5 does not allow inferring the spreading pattern of these strains. Older Georgian strains make a subset within subgroup 5; therefore, multiple introductions of different sublineages of this group likely took place (see below). Reemergence of subgroup 5 strains that fell outside the modern E30 major group demonstrates that surveillance carried out in developed countries for E30, arguably the best-studied nonpolio enterovirus, does not explicitly reflect world epidemiology of E30. We speculate that strains of this group originated from less-developed regions with less dynamic and less “global” epidemiologies, where they could have been maintained for decades. This observation not only is important for an understanding of E30 or HEV-B epidemiology but also implies that “eradicated” polio may reemerge decades later from secluded reservoirs.

Partial sequences for 3D genome fragments (nt 6468 to 6931 of E30 Bastianni) of the strains studied were aligned with available sequences of HEV-B strains of different serotypes, as E30 could be expected to recombine extensively with other HEV-B strains in the nonstructural protein (NSP) region (8, 10). Enterovirus B strains differed by up to 31.5% of nucleotide sequence and by up to 9.7% of protein sequence; thus, a majority of substitutions in this genome region were synonymous. A phylogenetic tree was created with poliovirus 1 sequence added to provide a correct root (Fig. 2). E30 strains studied here followed a grouping pattern reported previously for most modern HEV-B strains that cluster either with prototype E30 or with E1/E9, yet again supporting an observation of ubiquitous recombination in enteroviruses (10, 17, 20). In this work, however, we were mostly interested in a short-term incidence of recombination among strains isolated within only 7 years. A comparison of phylogenetic grouping in VP1 and 3D genome regions indicates many occasions of recombination over this short time frame. Only two phylogenetic groups observed in VP1, subgroups 1 and 2, were fully maintained in the 3D genome region. Within subgroup 3, strain 22696, indistinguishable from the group in VP1 and isolated close in time and location with strain 23184, was clearly recombinant in 3D. Viruses of subgroups 4 and 5, which were very similar and grouped very reliably in VP1, bore multiple marks of recombination in 3D. Strains of subgroup 4 possessed five different 3D polymerase regions (Fig. 2). These strains were hardly distinguishable in the VP1 region, which is traditionally used for typing, yet very diverse in the 3D region and, presumably, in most of the nonstructural genome region. Interestingly, strain 23199 had a 3D region very distant from those of most modern strains, again showing that rare enterovirus genome parts can persist and then reemerge in a virus that would not seem unusual from the conventional VP1 region analysis. Strains of subgroup 5 have also undergone multiple recombination events and fell into three distinct groups in the 3D genome region. Again, as was observed for subgroup 3, strains isolated at the same location in Georgia, within a 10-month interval and very similar in VP1, bore notably different nonstructural genome regions. We can thus conclude that subgroup 5 in VP1 indicates the emergence not of a new virus lineage, but of a new capsid lineage.

• Open in new tab
• Download powerpoint

FIG. 2.

Phylogenetic tree (ClustalX, neighbor-joining) for partial 3D genome region (nt 6468 to 6931). Numbers at tree nodes are a percentage of bootstrap pseudoreplicates that supported a group below; bootstrap values below 70% were omitted. The tree was rooted with poliovirus 1 (omitted). Prototype HEV-B strains do not have additional indications, studied E30 strains do not have serotypes indicated, and all modern HEV-B strains of different serotypes are given with isolation data and GenBank accession number. ▪, group 1; +, group 2; *, group 3; •, group 4; ▴, group 5.

Phylogenetic trees for structural and nonstructural genome regions indicate that different genome regions in enteroviruses circulate independently even on a scale of a few years. While it would not be correct to estimate the exact frequency of recombination in circulating E30 from our data, it is perfectly obvious that it is about once every few years, much as was recently shown for HEV-B (22). Indeed, 16 of 47 strains studied probably recombined within the last five years before being isolated and all E30 strains of the major modern group in the VP1 genome region were recombinant in 3D compared not only to the prototype strains of the 1950s but also to the HEV-B isolated in the 1980s and 1990s. Considering our results, especially those for subgroups 4 and 5, it would be incorrect to say that one can trace the circulation of individual strains; rather, one can trace the circulation of discrete capsid or nonstructural genome regions that only temporarily coexist as a distinct virus. Importantly, in the NSP region, E30 can recombine with any HEV-B strain; therefore, study of a single serotype would always produce a rather limited result that would not clearly reflect the epidemiology of NSP genome regions of a species. Flexible and highly dynamic genetics and epidemiology of circulating E30 and presumably of other enteroviruses oblige us to expect the emergence and rapid spread of strains with unusual properties. Our results also indicate that as long as enterovirus surveillance is carried out using only the VP1 genome region, either molecularly or in a neutralization test, we have a significantly reduced chance of linking the clinical manifestations and epidemiology of enterovirus infection.

ADDENDUM

After the manuscript was submitted, a study of E30 epidemiology in France also reported frequent recombination between VP1 and 3D regions (13).