Abstract
Of 58 enterovirus strains isolated from Japanese travellers returning from Asian countries, eight were non-serotypable with existing antisera. By sequencing a part of the VP1 region, six of these strains were typed as echovirus 9, enterovirus (EV)-73, EV-79 or EV-97. The nucleotide identity of the VP1 region of isolate T92-1499 to all enterovirus prototypes was <70 %. The VP1 sequence of isolate TN94-0349 was closely related to coxsackievirus (CV)-A9 (73.3 % nucleotide identity), but the virus could not be neutralized with a serum raised against the prototype CV-A9 strain. On the basis of complete molecular comparisons, T92-1499 and TN94-0349 were identified as EV-98 and EV-107, respectively, by the ICTV Picornavirus Study Group. Serum neutralization tests of Japanese individuals revealed a seroprevalence rate of 11 % for EV-73, and even lower seroprevalence rates, 1.0–3.8 %, were found for the other new enteroviruses, suggesting that prior circulation of these viruses in Japan was unlikely.
Viruses of the genus Enterovirus in the family Picornaviridae are antigenically divided into poliovirus (PV) 1–3, coxsackievirus A (CV-A) 1–A24, coxsackievirus B (CV-B) 1–B6, echovirus (E-) 1–34 and enterovirus (EV-) 68–71 (Melnick, 1996). Virus isolation is still the most common method for identification of enterovirus infection, and isolates have traditionally been classified by antigenic types, based on a serum neutralization assay. However, identification using a neutralization test requires an expensive antiserum stock panel covering every known type of enterovirus. Isolates of ‘new serotypes’ that are not of a known human enterovirus serotype would obviously also present difficulties in typing.
Recent developments in molecular detection technology and available reference nucleotide sequence data make it possible for a virus infection to be confirmed by molecular detection and identification (Oberste et al., 1999). Based on the comparison of available nucleotide sequences, human enteroviruses are classified into four species, i.e. Human enterovirus A (HEV-A), HEV-B, HEV-C (including poliovirus) and HEV-D (King et al., 2000). Furthermore, recent development of molecular typing systems for enteroviruses has revealed a correlation between the enterovirus VP1 sequence and serotype, providing a scientific basis for new enterovirus types, EV-73 to EV-97, and EV-99 to EV-101, to be able to be primarily identified by VP1 nucleotide sequences (Pallansch & Roos, 2007).
During an investigation of possible ‘imported’ enteroviruses by serially isolating them from stool samples collected at the quarantine section of a local international airport, we obtained a total of 58 enterovirus isolates. Eight of these 58 isolates were not ‘typable’ with existing antisera against known enterovirus serotypes. In this study we applied the molecular technology cited above to type these ‘imported’ enterovirus isolates. We recognized two of the isolates as new enterovirus types. Furthermore, seroprevalence of these new enteroviruses among the local Japanese population was estimated by testing for a neutralizing antibody.
The 58 viruses were isolated from faecal samples of travellers who reported gastrointestinal symptoms to the quarantine centre of Nagoya International Airport between January 1989 and December 1998. Eight of these isolates could not be identified by a neutralization test using the intersecting serum panel obtained from the National Institute of Infectious Diseases (NIID), Tokyo. Isolates TN97-0482 and TN94-0349 were further tested for neutralization using high-titre monovalent antisera from NIID directed against the type obtained by sequencing. Antiserum against isolate TN94-0349 was also prepared by immunizing guinea pigs with purified virus particles prepared by caesium chloride density-gradient centrifugation.
For amplification of a part of the VP1 gene, primer 187 as forward and primer 011 as reverse (Oberste et al., 2000) were used. In order to determine complete sequences of the new enteroviruses, four sets of primers were used, three of which have been described previously (Caro et al., 2001; Ishiko et al., 2002; Oberste et al., 2000). For amplification of the 3CD region, a new set of primers was designed, based on the sequence of HEV-B reported in the public databases, as follows: EVB-3CP (sense: 5′-CAYGTTGGTGARAATGGNCA-3′) and EVB-3DN (anti-sense: 5′-ACRTGRTCTTGRGTGTTYTTKGGA-3′). Specific, non-degenerate primers were designed from preliminary sequences to close gaps between the originally obtained PCR products. The DNA sequence was determined by a model 4000 automated DNA sequencer (Li-Cor). The nucleotide and deduced amino acid sequences of these new enteroviruses were compared with one another and with those of all HEV-B serotypes by using the Genetyx program (Genetix).
The sequences described here have been deposited in GenBank/DDBJ under accession nos AB426608–AB426615. The complete sequences of T92-1499 (EV-98) and TN94-0349 (EV-107) have been provided to Dr Nick Knowles, chair of the Picornavirus Study Group of the International Committee on Taxonomy of Viruses, in support of a proposal.
Serum samples were obtained from children (aged 7 months to 14 years) who visited the outpatient ward of a local hospital and were diagnosed as having a common cold, febrile rash, aseptic meningitis or other common diseases. Samples were also collected from healthy adults (aged 18 years or over) at the time of blood donation in 2003. Informed consent for utilizing their serum samples for viral antibody tests was obtained either from the individuals themselves or, in the case of minors, from their parents. Serum antibody titres against 100 50 % tissue culture infective doses (TCID50) as determined on RD cells were obtained by serial dilution as described previously (Yamashita et al., 1993).
Of the 58 isolates, 50 were identified as enteroviruses by the neutralization test (Table 1⇓). The other eight isolates were successfully amplified with EV RT-PCR primers to determine the partial VP1 gene sequence (Oberste et al., 2000). The partial VP1 nucleotide sequence of isolate TN97-0487 was 78.3 % identical to the E-9 prototype strain (Hill) sequence and showed 88.5 % amino acid identity. The sequence analyses agreed well with the results obtained from a neutralization test using high-titre (100 units) monovalent antiserum. TN97-0487 was more similar to Barty, an E-9 prime strain, with which it shared 83.5 % nucleotide identity (94.8 % amino acid identity), and contained a C-terminal extension to the capsid protein VP1 with an RGD (arginine-glycine-asparatic acid) motif.
Enterovirus isolates of Japanese travellers from Asian countries
The partial VP1 nucleotide sequences of two isolates (T97-1831 and TAS92-1482) were 75.6 and 82.6 % identical to EV-73 prototype CA55-1988 (89.9 and 95.7 % amino acid identity). The partial VP1 nucleotide sequences of a further two isolates (TS94-0534 and NH95-0601) were 86.7 and 92.6 % identical to EV-79 strain USA/CA79-10384 isolated in the USA (amino acid identity 94.6 and 99.1 %), respectively. One isolate (DT94-0227) from a traveller returning from India and Thailand was also 85.3 % (97.5 % amino acid) identical to EV-97 strain BAN-99-10355, isolated in Bangladesh, in the nucleoside sequence of the complete VP1 region.
Two isolates (T92-1499 and TN94-0349) could not be identified as known types of enteroviruses by molecular identification. The nucleotide identity of the complete VP1 sequence of T92-1499 (EV-98) from a traveller returning from Thailand was lower than 70 % to all other enterovirus prototypes. In addition, all of the phylogenetic trees based on the sequences of VP2, VP3 and VP1 showed that T92-1499 (EV-98) was independent from all other types within the species HEV-B (see Supplementary Fig. S1, available with the online version of this paper). The complete VP1 nucleotide sequence of isolate TN94-0349 (EV-107) from a traveller returning from Thailand and Nepal was 73.3 % (79.9 % amino acid) identical to CV-A9 Griggs strain and 67.9–61.6 % (72.6–61.2 % amino acid) identical to the other HEV-B members. TN94-0349 was not neutralized with a high-titre antiserum raised against CV-A9 (obtained from NIID, Tokyo). Furthermore, antiserum raised against TN94-0349 did not neutralize the Griggs strain of CV-A9 (Table 2⇓). As shown in Fig. 1⇓, the RGD motif , which was found within the C-terminal extension of the VP1 of CV-A9 strains, including the prototype (Chang et al., 1992; Santti et al., 2000), was absent in the deduced VP1 of TN94-0349.
Alignment of the VP1 peptides of TN94-0349 (EV-107) and CV-A9 Griggs strain. The RGD motif found in CV-A9 is underlined.
Cross-neutralization test between the prototype coxsackievirus A9 (Griggs strain) and isolate TN94-0349
Although TN94-0349 formed a cluster with the CV-A9 prototype strain in the VP1, VP2 and VP3 regions, it was a marginal decision (Supplementary Fig. S1, Supplementary Table S1). In the VP2 and VP3 regions isolate TN94-0349 diverged by 25.4 and 27.4 %, respectively, from CV-A9 Griggs strain and by 31.8–26.3 % and 35.1–29.9 %, respectively, from other HEV-B members. Isolate TN94-0349 did not form a cluster with CV-A9 in phylogenetic trees of the other regions. These two isolates were recognized in the list of new enteroviruses (as EV-98 and EV-107) kept on the relevant Picornavirus Study Group () site.
The complete sequences of NH95-0601 (EV-79) and DT94-0227 (EV-97) were also determined. DT94-0227 (EV-97) formed a cluster with the known serotype of EV-97 in each of the regions except 2A. On the other hand, NH95-0601 (EV-79) was independent from the known serotype of EV-79 in regions that represent non-structural proteins. The four isolates completely sequenced in this study [NH95-0601 (EV-79), DT94-0227 (EV-97), T92-1499 (EV-98) and T94-0349 (EV-107)] formed a cluster with CV-B4, E-30, EV-74, EV-75, EV-77, EV-85, EV-86, EV-87, EV-88 and EV-97 in a phylogenetic tree based on the 3D region (Supplementary Fig. S1).
Supplementary Table S2 provides the distribution of serum neutralizing antibodies in the different age groups of local individuals for these newly identified enteroviruses. Of 210 serum samples tested, 24 (11.4 %) were positive for neutralizing antibody against T97-1831 (EV-73) at a titre of 1 : 8 or higher. However, lower prevalence rates, 1.4–3.8 %, were found for all the other viruses (EV-79, EV-97, EV-98 and EV-107) tested.
EV isolates that are resistant to neutralization tests with existing antisera are naturally considered as new enterovirus serotypes. A proposed new type of human enterovirus is based on the genetic relationships between the virus strains of the same species. In pairwise comparisons of complete VP1 sequences, enteroviruses of the same serotype were clearly distinguished from those of heterologous serotypes, and the limits of intraserotypic divergence appeared to be 25 % in nucleotide sequence difference or 12 % in amino acid sequence difference (Oberste et al., 2000). The nucleotide sequence of the TN94-0349 (EV-107) VP1 region differed by 26.7 % from the CV-A9 prototype strain Griggs. Every new enterovirus (EV-73 to EV-101) recently reported based on sequence data appeared to be at least 30 % different in nucleotide sequence from known viruses. A best-match nucleotide sequence identity of between 70 and 75 % may provide a tentative identification, pending confirmation by other means, such as neutralization with monospecific antisera (Oberste et al., 2000). Based on the cross-neutralization test, TN94-0349 (EV-107) was apparently not CV-A9 (Table 2⇑). We therefore conclude that TN-94-0349 (EV-107) is a unique serotype different from CV-A9 with a nucleotide sequence identity of 73.3 % in the VP1 region.
T92-1499 has been recognized as an isolate of a new human enterovirus serotype, ‘enterovirus 98’, subject to approval by the appropriate taxonomic authority. Our isolate T92-1499 (EV-98) is included in species HEV-B but is distinguished from the established type of the species not only in VP1, but also in every region from 5′ NTR to 3′ NTR. In 2001, a new enterovirus, enterovirus 73, was proposed based on molecular identification. Since then, 32 additional new types, EV-74 to EV-97, and EV-99 to EV-106, have also been identified by VP1 nucleotide sequences. In fact, five out of six isolates that were not a new serotype in this study indeed belonged to one of these types, i.e. EV-73, EV-79 and EV-97. By determining the complete nucleotide sequences of these isolates, we hope to provide further evidence that complete capsid sequence determination may be useful in identification of new serotypes in the future (Oberste et al., 2001). The isolates characterized in the late 1990s, including E-30, EV-74 and EV-75, are known to be closely related through phylogenetic analysis of the 3D genome region (Lukashev et al., 2005). NH95-0601 (EV-79), DT94-0227 (EV-97), T92-1499 (EV-98) and TN94-0349 (EV-107) formed a cluster with CV-B4, E-30, EV-74, EV-75, EV-77, EV-85, EV-86, EV-87, EV-88, EV-97 and EV-100 in the 3D region (Supplementary Fig. S1). Our four isolates mentioned above isolated between 1992 and 1995 could be included with the E-30/EV-74/EV-75-like strains described by Lukashev et al. (2005).
In this study, low seroprevalence rates of 1.4–3.8 % were found for the new enteroviruses (EV-79, EV-97, EV-98 and EV-107) (Supplementary Table S2). Seroepidemiological studies are necessary to assess these new types of enteroviruses for their impact on public health, as well as the surveillance of infectious agents. Based on the seroprevalence described above, these newly identified enterovirus types appear not to have circulated previously in our areas.
Acknowledgments
We sincerely thank the staff of the quarantine centre at Nagoya International Airport, Japan, for providing us with stool specimens for virus isolation. This work was supported in part by a Grant-in-Aid for Research on Emerging and re-emerging Infectious Diseases from the Ministry of Health and Labour and Welfare.