Eight Years of Experience with Molecular Identification of Human Enteroviruses

Abstract

We have successfully typed 1,121 human enterovirus (HEV) isolates during the last 8 years by adapting partial VP1 sequencing to routine identification of HEV isolated from diverse clinical and environmental specimens. The isolates include 48 of the 59 traditional nonpoliovirus HEV serotypes and members of 8 newly discovered types, which would have remained untypeable by neutralization using the conventional cross-sectional pools of antisera.

Human enteroviruses (HEV) (family Picornaviridae) are small single-stranded RNA viruses that are grouped into four species, HEV-A, HEV-B, HEV-C, and HEV-D, by molecular characteristics. These species comprise altogether 62 antigenically distinct serotypes and at least 30 additional recently described genetically distinguishable types (18) (http://www.picornaviridae.com/enterovirus/enterovirus.htm). The identification of HEV types is essential in studies associating the HEV types with certain diseases, and it traditionally includes isolation of the virus in susceptible cell lines followed by neutralization either with type-specific antisera or with intersecting pools of antisera. The neutralization assay itself is laborious and time-consuming and may be further complicated by the aggregation of virus particles or by the existence of virus mixtures. Besides, the identification by neutralization is limited to only those serotypes against which potent monotypic antisera are available, while some of the currently circulating HEV strains with significant antigenic drifting and, especially, the new HEV types remain unidentified.

During the last decade, a variety of molecular technique-based methods have been developed for the identification of HEV types (e.g., see references 1, 3, 4, 6, 9, and 12). The sequencing of the (partial) VP1 capsid protein coding region, which is based on the existence of the most prominent serotype-specific immunogenic epitopes in VP1 leading to an absolute correlation between the VP1 sequence and the serotype (11), has been chosen for HEV identification in many laboratories (7, 10, 13, 14). Alternatively, molecular typing may include sequencing of coding regions for other capsid proteins (6, 9), whereas typing by sequencing the nontranslating genomic segments or regions coding for the nonstructural proteins is not widely used because of the discordant phylogeny due to the frequent intraspecies recombination in these regions (15).

The partial VP1 sequencing method described by Oberste et al. (12) was introduced to our laboratory in 1999 and used alternatively with the neutralization assay during the years 2000 to 2003 and since then exclusively. At the beginning, a part of the isolates were typed by both methods with 100% congruity (Table 1). Sixty percent of the viruses typed during the 8-year study period were originally isolated in Finland, and these included HEV from the environmental surveillance of polioviruses and clinical HEV isolates sent to the national enterovirus reference laboratory from other Finnish laboratories. The rest of the molecularly typed isolates were received as untypeable nonpolio enteroviruses (NPEV) from a number of National Polio Laboratories of the WHO Polio Laboratory Network supporting the Global Poliovirus Eradication Initiative. Altogether, 45% of the HEV isolates originated from environmental samples and 55% from a variety of clinical specimens, e.g., stool, skin or mouth vesicles, and nasopharyngeal swaps. The cell lines used in isolation included human rhabdomyosarcoma (RD), human colorectal adenocarcinoma (CaCo-2), human lung carcinoma (A549), human larynx epidermoid carcinoma (Hep-2C), human cervical carcinoma (HeLa), human fetal fibroblast (HFF), the green monkey kidney cell lines GMK, Vero, and LLC, and the recombinant mouse L cells expressing the human poliovirus receptor (L20B).

TABLE 1.

Numbers of HEV isolates typed by partial VP1 sequencing or concurrently by neutralization during the years 2000 to 2007

Yr	No. of isolates:		Congruity of typing results (%)
	Typed by sequencing	Concurrently typed by neutralizing assay
2000	54	31	100
2001	57	11	100
2002	74	17	100
2003	72	19	100
2004	110	0
2005	106	0
2006	234	0
2007	414	0

RNA was extracted from 100 μl of the infected cell cultures by commercial procedures (RNeasy Total RNA kit [Qiagen GmbH, Hilden, Germany] or E.Z.N.A. Total RNA kit [Omega Bio-Tek Inc., Doraville, GA]) according to the manufacturer's instructions. One microliter of RNA was used in a single-tube reverse transcription (RT)-PCR assay. During the first 2 years of the study, two distinct regions of VP1 were amplified with two sets of primers as described in reference 12. Later on, the RT-PCR was first performed with a single set of primers only, forward and reverse primers 292 and 222, respectively (13). The alternative primers 040, 011, and 012 were used with 8.1% of the isolates, because a very small amount or no amplicon was obtained with the two primary primers or because of the poor quality of the first sequences. The RT-PCR products were analyzed using electrophoresis in a 2% agarose gel, and the amplicons were purified with commercial reagents (Qiagen Gmbh, Hilden, Germany) before use in sequencing (ABI Prism BigDye terminator cycle sequencing; Applied Biosystems). The quality of the electropherograms was examined using the Vector NTI Advance software program (Invitrogen) and the sequences compared to the sequence database in GenBank by BLAST. For an acceptable result, the length of the compared sequence had to be more than 100 nucleotides. The sequences that had more than 75% nucleotide identity with the query sequence determined the type of the isolate.

Altogether, 1,121 HEV isolates were molecularly typed during the reported 8-year period. The isolates comprised representatives of each of the four HEV species: 70.4% were members of HEV-B, 16.6% members of HEV-A, 12.1% members of HEV-C, and 0.9% members of HEV-D, respectively. The frequency distribution of the 59 “classical” NPEV serotypes is shown in Fig. 1. The five most common serotypes identified were echoviruses 11 and 6 and coxsackievirus (CV) A24, CV-B5, and CV-B4. Due to the incoherency of the isolate collection, this frequency distribution may not directly reflect the distribution of NPEV in Finland or other countries in the study region. However, four of the five most common serotypes (echoviruses 6 and 11, CV-B4, and CV-B5) were also among the most abundant serotypes isolated from sewage or patients in Finland during the years 1971 to 1992 (5), and they were also among the 10 most frequently reported serotypes in the United States in 1970 to 2005 (8). The most commonly reported NPEV in the United States, echovirus 9 (8), was infrequently found in Finland during this study period and during the previous study years 1971 to 1992 (5). The only HEV-C serotype that was common in the studied NPEV collection was CV-A24. The CV-A24 strains were isolated mainly from stool specimens from children who had immigrated to Finland and could most likely be regarded as importations rather than evidence of endemic circulation of CV-A24 in Finland. Eleven of the fifty-nine NPEV serotypes remained undetected. The lack of these types in the HEV isolate collection may be due to their poor propagation in cell lines used or their infrequent circulation in the catchment population. The isolates of the genetically characterized “newer” types EV-73 to EV-102 are shown in Table 2. Some isolates had less than 70% partial VP1 nucleotide identity to any sequence in the GenBank and were regarded as untypeable and supposed to represent new enterovirus types. The complete VP1 sequences of three new HEV candidates were sent to the Picornavirus Study Group, assigned a candidate prototype number (EV-94, EV-96, and EV-97), and subsequently characterized further (16, 17).

FIG. 1.

The numbers of different “traditional” nonpoliovirus serotypes of HEV typed by partial VP1 sequencing during the years 2000 to 2007. CV, coxsackievirus; E, echovirus; EV, enterovirus.

TABLE 2.

Human enterovirus isolates typed as EV73 to EV102 by partial VP1 sequencing

Type	Isolate code	Country	Yr of isolation	Specimen type	GenBank accession no. (reference)
EV74	FIN05-6	Finland	2005	Stool	EU481512 (this study)
EV75	FIN04-13	Finland	2004	Stool	EU481514 (this study)
EV75	FIN04-1243	Finland	2004	Stool	EU481515 (this study)
EV75	EGY05-E124	Egypt	2005	Sewage	EU481513 (this study)
EV76	KAZ00-14550	Kazakhstan	2000	Stool	EF364396 (15)
EV76	EGY04-E442	Egypt	2004	Sewage	EU481516 (this study)
EV76	FIN05-E17421	Finland	2005	Sewage	EU481519 (this study)
EV76	EGY06-E488	Egypt	2006	Sewage	EU481517 (this study)
EV76	EGY07-E14	Egypt	2007	Sewage	EU481518 (this study)
EV89	EGY04-E481	Egypt	2004	Sewage	EU481520 (this study)
EV90	LVA02-10337	Latvia	2002	Stool	EF392677 (15)
EV90	LVA04-E1347	Latvia	2004	Sewage	EU481522 (this study)
EV90	SVK04-E1341	Slovakia	2004	Sewage	EU481523 (this study)
EV90	EGY04-E425	Egypt	2004	Sewage	EU481521 (this study)
EV94	EGY04-E210	Egypt	2004	Sewage	DQ916376 (14)
EV94	EGY04-E430	Egypt	2004	Sewage	DQ916377 (14)
EV94	EGY04-E435	Egypt	2004	Sewage	DQ916378 (14)
EV94	EGY04-E438	Egypt	2004	Sewage	DQ916379 (14)
EV94	EGY05-E23	Egypt	2005	Sewage	EU481524 (this study)
EV96	SVK03-24	Slovakia	2003	Stool	EF364404 (15)
EV96	FIN04-7	Finland	2004	Stool	EF364398 (15)
EV96	FIN05-2	Finland	2005	Stool	EU481525 (this study)
EV96	FIN05-5	Finland	2005	Stool	EF364399 (15)
EV96	FIN05-10	Finland	2005	Stool	EF364400 (15)
EV96	FIN05-12	Finland	2005	Stool	EF364401 (15)
EV96	FIN05-14	Finland	2005	Stool	EF364402 (15)
EV96	FIN06-7	Finland	2006	Stool	EF364403 (15)
EV96	FIN06-9	Finland	2006	Stool	EU481511 (this study)
EV97	FIN03-2875	Finland	2003	Stool	EF364397 (15)

As judged by the partial VP1 sequences, 37 isolates subjected to typing as NPEV were actually human rhinoviruses (HRV). The approximately 350-nucleotide-long sequences were not regarded as sufficient for the exact typing of HRV, but they were adequate enough to assign all of these isolates to species HRV-A. Besides being successful in amplifying HEV and HRV, the primers 222 and 292 also detected bovine enteroviruses (BEV). Altogether, 81 viruses isolated in L20B cells from environmental specimens collected in Finland, Croatia, Slovakia, and Egypt were classified as BEV. The partial VP1 nucleotide identities between the typed isolates and the BEV sequences in GenBank varied from 73 to 88%, confirming the species identification, while no serotype assignment was attempted.

In conclusion, molecular typing by partial VP1 sequencing was reliable and significantly shortened the hands-on time in routine HEV typing. Altogether, 1,121 HEV isolated in a variety of cell lines from diverse clinical and environmental specimens were successfully typed. The typed isolates included 48 of the 59 traditional NPEV serotypes and also 8 new types, which would have remained untypeable by neutralization. We believe that in our isolate collection no HEV remained untyped. One primer pair, 222 and 292, amplified almost all HEV isolates, and all the remaining ones were amplified by using primers 040, 011, and 012. Selected viruses that produced a cytopathic effect in enterovirus-accessible cells but were negative with both sets of VP1 primers were subsequently shown to be negative with primers from the conserved regions in the 5′ nontranslating region (2), as well. Some of the latter isolates were further typed as adenoviruses or parechoviruses.

Nucleotide sequence accession numbers.

The accession numbers for the sequences deposited in GenBank are EU481511 to EU481525.