INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Human enteroviruses (family Picornaviridae) consist of more than 60 serotypes that include polioviruses, coxsackievirus types A and B, echoviruses, and the undesignated enterovirus types 68 to 71 (19, 20, 21, 23, 30). The enteroviral genome consists of a positive single-stranded RNA molecule that is about 7,500 nucleotides long and that contains an approximately 750- nucleotide 5' noncoding region (5'NCR), a single open reading frame, and a short 3' noncoding region (3'NCR). A small, basic protein (VPg) is covalently attached to the 5' end of the genome, while the 3' end is modified by polyadenylation. Enteroviral 5'NCRs are highly conserved, as they appear to play vital roles in viral translation, virulence, and possibly, encapsidation (7, 10, 22, 28, 29, 32).

Polioviruses, the most pathogenic of all enteroviruses, include three distinct serotypes, designated type 1, type 2, and type 3, that were originally defined by their patterns of reactivity with neutralizing antibodies (3). Polioviruses are the main causative agents of poliomyelitis but have also been associated with seasonal undifferentiated febrile illness, particularly during summer outbreaks (13, 20), and enteroviral meningitis (2). Poliomyelitis, a life-threatening acute paralytic disease, is being effectively controlled by the inactivated poliovirus vaccine (26) and the oral poliovirus vaccine (OPV) (24, 25). OPV contains all three live, attenuated poliovirus serotypes from sequential passage in monkey tissues. Vaccination with live OPV strains (Sabin types 1, 2, and 3) generally mounts a long-lasting immune response that protects the organism from future viral infections with wild-type poliovirus strains (24, 25). However, rare reversion of live OPV vaccine strains may occasionally cause vaccine-associated paralytic poliomyelitis (4, 5, 6, 7, 11, 16). Detailed typing of all polioviruses isolated from patients with poliomyelitis is therefore essential to public health polio surveillance programs aiming to eradicate wild-type polioviruses.

The World Health Organization-recommended poliovirus serotyping procedure allows only intertypic differentiation but not intratypic differentiation of clinical poliovirus isolates (17, 33). Recent advances in molecular virology by highly efficient PCR amplification methods have provided new alternatives to poliovirus detection and typing (15, 31). Thus, PCR genotyping of polioviruses includes serotype-specific PCR primers (9), genotype Sabin-specific PCR primers (34), and restriction fragment length polymorphism (RFLP) analysis (1, 4), which may potentially allow their inter- and intratypic differentiation. In practice, however, the published RFLP method failed to type or even detect a small but nevertheless significant fraction of clinical poliovirus isolates (31; A. Georgopoulou, P. Markoulatos, N. Spyrou, and N. C. Vamvakopoulous, unpublished data). These findings prompted our present attempt to develop an improved reverse transcriptase (RT)-PCR-RFLP method for direct genotyping of vaccine and wild-type poliovirus strains.

We report here on our contribution toward this goal. More specifically, a 440-bp RT-PCR fragment was amplified from the highly conserved 5'NCRs of enteroviruses. The product was then enzymatically digested with a selected set of four restriction endonucleases, generating inter- and intratypic poliovirus-specific patterns (type-specific RFLPs), readily discernible from those of the 55 reference nonpoliovirus enterovirus strains so far tested as RFLP controls for poliovirus-specific RFLPs. To account for potential genetic drift within the recognition sequence of the proposed minimum set of four restriction endonucleases commonly used (namely, HaeIII, DdeI, NcoI and AvaI), among clinical poliovirus isolates, two additional restriction endonucleases, HpaII and StyI, were used as replacements. Our method correctly typed all 21 (100%) clinical poliovirus isolates tested so far.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Virus stocks, clinical samples, and cells. Table 1 summarizes the types and origins of the poliovirus reference strains used in this study, along with the RFLP fragment sizes in base pairs that were generated by single digestion with the six restriction endonucleases employed. Clinical specimens were obtained from the collection of clinical samples maintained by the Department of Virology of the Hellenic Pasteur Institute (Table 2). They had been stored at 80°C for periods ranging from a few months to several years prior to reinoculation and were typed by seroneutralization with rabbit polyclonal antibodies (National Institute for Public Health and the Environment, Bilthoven, The Netherlands), but their identities were not revealed prior to genotyping. Reference virus strains and infected clinical samples were propagated in Vero (African green monkey kidney) cells, L20 cells (a recombinant murine cell line that contains the human receptor for polioviruses) (9), and HEp-2 cells (a cell line derived from a human epidermoid carcinoma of the larynx), which were grown in round-bottom plastic tubes. The tubes were incubated at 37°C for a period of 1 to 4 days. When a complete cytopathic effect was observed, the tubes were frozen at 20°C, thawed, and centrifuged at 3,500 rpm in a Beckman GPR centrifuge at 4°C for 20 min. The supernatants were discarded, and the pellets were used for RNA extraction. To account for potential false-positive results, negative controls were routinely included in all experimental procedures.

Extraction of viral RNA. Viral RNA was extracted primarily from infected cell pellets by using a commercially available kit (Snap-O-Sol; Biotecx, Houston, Tex.). Total cellular RNA was resuspended in 20 µl of RNase-free sterile distilled water (Sigma Chemical Co., St. Louis, Mo.) and stored at 80°C until RT-PCR amplification analysis. Virus-infected cell culture supernatants were used as an alternative source of viral RNA templates.

Oligonucleotide primers. The primers used in this study were selected from the highly conserved 5'NCR of the known enteroviral sequences by using available primer design computer programs (8). The sequence of downstream primer UC₅₃ was 5'-TTGTCACCATAACCAGCCA-3', while the sequence of upstream primer UG₅₂ was 5'-CAAGCACTTCTGTTTCCCCGG-3'. Primer UG₅₂ was identical to previously described primer 1 (35). Primer UC₅₃ differs in one internal position from primer 3, indicated by the boldface letter, and is one base shorter at the 5' end (35). This primer pair amplified all 58 reference enterovirus strains examined so far except unassigned enterovirus type 68 to 71 strains and coxsackievirus type A1, A19, and A22 strains, which have not yet been tested. The originally described primers have been used for enterovirus detection (35). Their use for genotyping of polioviruses was attempted here for the first time. The primers were synthesized by Genosys (Europe, Cambridge, United Kingdom) and were adjusted to a concentration of 7 nmol/µl in sterile distilled water and stored at 20°C.

Reverse transcription and PCR. cDNA synthesis and PCR reagents were obtained from Amersham Life Sciences (Cleveland, Ohio). Reverse transcription was performed in 20-µl reaction mixtures containing 1 µl (25 U) of RNase inhibitor, 1 µl of a solution with downstream primer UC₅₃, 5 µl of a solution with extracted RNA, 4 µl of 5× RT buffer, 2 µl of a 100 mM mixture of the four deoxynucleoside triphosphates, 10 U of avian myelobastosis virus RT, and 6 µl of RNase-free sterile distilled water (Sigma). The reaction mixture was incubated at 37°C for 60 min, and the avian myelobastosis virus RT enzyme was heat inactivated by incubation at 95°C for 5 min. Enzymatic amplifications were performed in 100-µl reaction mixtures containing 92 µl of the PCR reaction mixture (10 µl of 10× PCR buffer with 1.5 mM MgCl₂, 8 µl of a 10 mM mixture of the four deoxynucleoside triphosphates, 2 U Taq DNA polymerase, 73 µl of RNase-free sterile distilled water), 4 µl of cDNA, and 2 µl of each primer by using 40 cycles of denaturation (94°C, 30 s), annealing (45°C, 30 s), and primer extension (72°C, 1 min) in a Perkin-Elmer GeneAmp PCR system 9600 thermal cycler. The PCR amplicons were analyzed by electrophoresis in 2.5% Tris-borate-EDTA-agarose minigels containing 1 µg of ethidium bromide per ml, visualized, and recorded in a FOTO/PHORESIS I system (FOTODYNE, Hartland, Wis.), as described previously (12).

RFLP analysis of PCR products. Aliquots of 20 µl of the RT-PCR amplicons were digested singly with 20 U of the various restriction enzymes used, including HaeIII, HpaII, NcoI, StyI, BstOI, and AvaI (Promega Corporation, Madison, Wis.) and DdeI (New England Biolabs, Beverly, Mass.), in a final volume of 30 µl at 37°C for 2 h according to the manufacturers' recommendations. The results were analyzed in 3% agarose gels (Metaphor FMC Bioproducts, Rockland, Maine) containing 1 µg of ethidium bromide per ml, visualized and recorded as described above. Restriction analysis and multiple alignments with poliovirus sequences submitted to GenBank were performed with the Gene Runner, version 3.00, program (Hastings Software, Inc.).

Computer analysis. The RFLP patterns were analyzed and the fragment sizes were calculated with the Gel Pro Analyzer program (Media Cybernetics, Silver Spring, Md.). The fragment sizes reported were calculated relative to the known fragment sizes of the HaeIII digest of phage X174 DNA (GIBCO BRL, Life Technologies, Gaithersburg, Md.), which was included as a marker in all electrophoresis runs. The size fluctuations between repetitive runs were almost undetectable.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

RT-PCR amplification of poliovirus 5'NCR target. RT-PCR amplification of all reference poliovirus strains (wild-type and vaccine strains) and all 21 clinical samples used in this study (Tables 1 and 2, respectively) with the UC₅₃-UG₅₂ primer pair (see Materials and Methods) generated the expected enterovirus-specific 440-bp product, shown in the uncut lanes 1 in Fig. 1A and B and Fig. 2A and B. As a testament to the conserved nature of the selected 5'NCR RT-PCR poliovirus amplification target, all 55 reference nonpoliovirus enterovirus strains used as RFLP controls for poliovirus-specific RFLP patterns (see below) also generated the expected 440-bp product (data not shown).

View larger version (47K): [in this window] [in a new window]   FIG. 1.   RT-PCR-RFLP analysis of two representative reference poliovirus strains (Table 1) with restriction endonucleases HaeIII, DdeI, HpaII, NcoI, StyI, and AvaI. (A) Sabin 1 type strain; (B) Mahoney strain. Lanes M, HaeIII digest of X174 DNA used to calculate the apparent sizes (in base pairs) of the various fragments indicated in Table 1; lanes 1, uncut RT-PCR product; lanes 2 to 7, restriction fragments produced by digestion with HaeIII, DdeI, HpaII, NcoI, StyI and AvaI, respectively.

View larger version (54K): [in this window] [in a new window]   FIG. 2.   RT-PCR-RFLP analysis of two representative clinical poliovirus samples (Table 2) with restriction endonucleases HaeIII, DdeI, HpaII, NcoI, StyI, and AvaI. (A) Virus from clinical sample with patient code number 6902, genotyped as Sabin type 1; (B) virus from clinical sample with patient code number 1085, genotyped as the Mahoney strain. Lanes M, HaeIII digest of X174 DNA used to calculate the apparent sizes (in base pairs) of the various fragments indicated in Table 2; lanes 1, uncut RT-PCR product; lanes 2 to 6, restriction fragments produced by digestion with HaeIII, DdeI, HpaII, NcoI and StyI, respectively. In panels A and B, lanes 7 contain the restriction fragments produced by digestion with AvaI.

Genotyping of reference poliovirus strains. The 440-bp RT-PCR products of Sabin poliovirus types 1 to 3 and the Mahoney, MEF, and Saukett wild-type poliovirus strains were digested initially with the restriction endonucleases HaeIII, DdeI, HpaII, NcoI, and StyI. These enzymes were selected from the published sequences of poliovirus genomes including those of the Sabin type 1, 2, and 3 vaccine strains and the Mahoney wild-type strain. The restriction fragments were analyzed by agarose gel electrophoresis (Fig. 1A and B), and their computer-calculated sizes are shown in Table 1. Their different RFLP patterns allowed complete intertypic differentiation between poliovirus type 1, 2, and 3 strains and intratypic differentiation between poliovirus type 2 (Sabin type 2 and MEF) and type 3 (Sabin type 3 and Saukett) strains. Differentiation of the Sabin poliovirus type 1 strain from the Mahoney wild-type strain was achieved by the additional use of the AvaI restriction endonuclease, as shown in both Fig. 1A and B and Table 1.

Genotyping of clinical samples. To assess the diagnostic value of the method in clinical practice, 21 clinical enterovirus isolates that had been typed by seroneutralization without revealing their identity prior to genotyping (Table 2) were subjected to RFLP analysis by analogy to reference strains (Fig. 2A and B). Table 2 summarizes their RFLP-generated, computer-calculated sizes. Direct comparison of the RFLP patterns of the clinical samples (Table 2) and those of the reference poliovirus strains (Table 1) led to proper genotypic assignment of all polioviruses from clinical samples.

Proposed minimum diagnostic enzymatic combination: suggested use of alternative enzymes. As already mentioned, complete inter- and intratypic differentiation of polioviruses can be achieved with only four of the six informative restriction endonucleases. With the exception of AvaI, which is indispensable for intratypic differentiation of poliovirus type 1 strains and which cannot be replaced, the five remaining endonucleases may be interchanged, as outlined in Fig. 3, to allow proper genotyping in cases of noninformative RFLP patterns caused by recombination and/or sporadic point mutation events in viruses in clinical samples. Such diagnostic demands have been encountered and are being investigated.

View larger version (12K): [in this window] [in a new window]   FIG. 3.   Diagnostic enzymatic combinations. The proposed enzyme replacements are shown in boldface and italics.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The accurate typing of virus isolates is critical for the management of polio. The RFLP developments in the PCR methodology have contributed greatly to these diagnostic demands (1, 4). Their implementation, however, has revealed certain weaknesses that call for definite improvements (31; Georgopoulou et al., unpublished data). Central to our quest for an improved RFLP analysis-based protocol for inter- and intratypic differentiation of polioviruses was the selection of the most highly conserved segment within the 5'NCR of the poliovirus enterovirus genome as a target for RT-PCR amplification (7, 10, 22, 28, 29, 32), along with the selection of the proper primer pair, the UC₅₃-UG₅₂ primer pair (35). Our working hypothesis was that even large genomic alterations in other segments of the viral genome would reflect only minor changes in the conserved 5'NCR or that the selected target would not tolerate extensive genetic drift due to strict functional limitations for virus viability (7, 10, 22, 27, 28, 29, 32).

RT-PCR amplification of the selected target-primer combination produced a 440-bp fragment from all poliovirus strains, 55 nonpoliovirus reference strains, and all clinical poliovirus isolates tested so far. The remaining nonpoliovirus reference strains, namely, coxsackie virus types A1, A19, and A22 and unassigned enterovirus types 68 to 71, have not yet been tested by this RT-PCR amplification protocol. Diagnostic RFLP analysis of poliovirus-specific RT-PCR fragments fully differentiated all reference poliovirus strains and all clinical poliovirus isolates analyzed so far. The criterion used for the selection of informative restriction enzymes for poliovirus-specific RFLP analysis was generation of distinct poliovirus-specific reference RFLP patterns when the patterns are among themselves and with those of the 55 nonpoliovirus reference enterovirus controls. Of the seven enzymes tested by use of these criteria, only HaeIII, HpaII, DdeI, StyI and NcoI produced highly informative RFLP patterns with all poliovirus and non-poliovirus enteroviruses examined. AvaI was used solely with polioviruses, while BstOI was noninformative for the reasons described below.

The sequence in the 5'NCR of Sabin type 1 (GenBank accession number V01150) (18) was aligned with the sequence kindly provided by Françis Delpeyroux (F. Delpeyroux, personal communication). This alignment revealed a difference at nucleotide position 355 (T in the GenBank sequence but C in the sequence provided by Françis Delpeyroux). Since the recognition sequence of BstOI is CC/(A/T)GG, where the boldface letter corresponds to position 355, and given that a BstOI restriction site appeared in our experiments, it was concluded that the correct nucleotide at this position was cytosine and not thymidine. Another difference between the two sequences was detected at nucleotide position 26 (G in the GenBank sequence of Sabin type 1 but A in the sequence provided by Françis Delpeyroux), but this difference was not investigated any further.

With the exception of Sabin poliovirus type 1, all other viral genotypes or serotypes for which published sequence information was available, including the Sabin type 2 and Sabin type 3 vaccine strains and the Mahoney wild-type strain, gave the expected RFLP patterns.

A minimum number of four restriction enzymes, HaeIII, DdeI, NcoI and AvaI, was sufficient for complete inter- and intratypic differentiation of polioviruses. HpaII and StyI were also examined, however, to compensate for potential genetic drift that altered the recognition sites of HaeIII, DdeI, or NcoI in atypical clinical samples. The only irreplaceable member of the minimal diagnostic set of four restriction endonucleases was AvaI. The eight conditional enzymatic combinations summarized in Fig. 3 will thus extend the provisional genotyping powers of the method to a broader spectrum of variant vaccine or wild-type poliovirus strains. Such variant strains have been identified and will be the subject of a future report.

Multiple poliovirus infections will require subcloning adaptations of RT-PCR products prior to RFLP analysis for genotypic assignment by the present method. The method will theoretically fail to type or even identify potentially viable poliovirus variants with extensive genomic alterations and/or major recombinations within the 5'NCR RT-PCR amplification target. In practice, however, no such virus has yet been found. With the hope that polioviruses severely deformed in the functionally sensitive 5'NCR to the extent that they will escape detection by this method will be extremely rare or even nonexistent, we foresee improved applicability of the method for inter- and intratypic poliovirus enterovirus diagnostic genotyping with the aim of eradicating wild-type polioviruses.