VIROLOGY

184,645-654

(199 1)

The Natural Genomic Variability of Poliovirus Analyzed by a Restriction Fragment Length Polymorphism Assay JEAN BALANANT, Unit4

SOPHIE GUILLOT, ADINA CANDREA, de Virologie

Mgdicale,

lnstitut

Received

Pasteur, March

FRANCIS DELPEYROUX,

25 rue du Docteur 12, 199 1; accepted

Roux, June

20,

75724

Paris

cedex

AND

RADU CRAINIC’

15, France

199 1

The genomic variability of poliovirus was examined by analyzing the restriction fragment length polymorphism of a reverse-transcribed genomic fragment amplified by the polymerase chain reaction. The fragment was a 480-nucleotide sequence of the poliovirus genome coding for the N-terminal half of the capsid protein VPl, including antigenic site 1. The identification of a pair of generic primers flanking this fragment allowed its amplification in practically all the poliovirus strains tested so far (more than 150). By using the restriction enzymes Haelll, Ddel, and Hpall, strain-specific restriction profiles could be generated for the amplified genomic fragment of each of the six reference poliovirus strains tested: one representative wild poliovirus of each of the three serotypes (Pi/Mahoney. PS/Lansing, and P3/Finland/ 23127/84) and the three Sabin vaccine strains. When 21 poliovirus field isolates previously identified as Sabin vaccinerelated were tested, they showed restriction profiles identical to those of the originating homotypic Sabin virus, demonstrating the conservation of these profiles during virus replication in humans. These profiles could thus be used as markers for Sabin-derived genotypes. To compare the distribution of poliovirus genotypes in nature before and after the introduction of poliovirus vaccines, the restriction profiles of the amplified genomic fragment of a total of 72 strains of various geographic and temporal origins were determined. Strains isolated before the introduction of polio vaccines displayed a wide diversity of genotypes. In contrast, wild (Sabin unrelated) strains isolated after vaccine introduction, during a single epidemic in a particular geographic area, showed identical or very similar restriction profiles, indicating the circulation of predominant regional genotypes. Our results indicate that the assay we developed for the analysis of the restriction fragment length polymorphism of the poliovirus genome may be used to identify and characterize poliovirus genotypes circulating in nature. 0 1991 Academic PESS. I~C.

INTRODUCTION

tered. Nor is there a satisfactory explanation for the outbreaks of poliomyelitis that have occurred over the last decade in populations in which the disease had been controlled for many years, such as Finland in 1984 (Magrath eta/., 1986) and Israel in 1988 (Slater et a/., 1990). The characterization of PV isolates is important both for analyzing PV natural variation and circulation in the human population and for mapping out a strategy for poliomyelitis eradication. Since the introduction of the OPV, many attempts have been made to assess whether field isolates originate from the Sabin OPV strains. Phenotypic markers, such as the ability to grow at supraoptimal temperature (Lwoff, 1959) or at low pH, were used for many years to identify PV strains. PV antigenicity was first tested using strainspecific (McBride, 1959; Nakano and Gelfand, 1962) or cross-adsorbed neutralizing antisera (Van Wezel and Hazendonk, 1979). The introduction of monoclonal antibodies (MAbs) (Osterhaus et al., 1981; Crainic et a/., 1982; Ferguson et a/., 1982; Humphrey et al., 1982) significantly improved the antigenic characterization of PV. The first type of genomic analysis used to compare PV strains was Tl oligonucleotide mapping of viral

The effective control of poliomyelitis was achieved by the introduction, about 30 years ago, of two polio vaccines, inactivated (IPV) (reviewed by Salk and Salk, 1987) and oral (OPV) (reviewed by Sabin, 1985). The OPV contains live, attenuated Sabin strains (Sabin and Boulger, 1973; Sabin 1985) one of each of the three serotypes, which differ phenotypically and genotypitally from the wild polioviruses (PV) in circulation prior to its use. The massive administration of the OPV in large areas of the world considerably altered the distribution of PV variants in nature. Contrary to what might have been expected, once diffused in the human population, the vaccine strains not only failed to completely replace wild viruses, but they themselves underwent variation during intra- and interhuman passages. Despite the overall success of vaccination in diminishing the morbidity of poliomyelitis, several problems still trouble our understanding of the epidemiology of the disease. For example, it is not understood why a few paralytic cases persistently arise in some regions in which the OPV has been systematically adminis’ To whom

reprint

requests

should

be addressed. 645

0042-6822/91

$3.00

CopyrIght 0 1991 by Academic Press, Inc. All rights of reproduction I” any form reserved

646

BALANANT

RNA (Minor, 1980; Nottay eT al., 1981; Kew and Nottay, 1984). Nucleic acid hybridization has also been used to detect and identify enteroviruses for diagnostic purposes, applied either directly to isolated viral genomes (Petitjean et al., 1990; Kew, personnal communication) or to enzymatically amplified subgenomic fragments (Chapman et al., 1990). An elegant method to determine PV strain genealogy has been developed (Rico-Hesse eta/., 1987) which involves comparing the nucleotide sequences of a subgenomic viral RNA fragment and constructing genealogic dendrograms. DNA fingerprinting of hypervariable minisatellite genes, developed recently for genealogic studies of human populations, is based on the similarities of restriction fragment length polymorphism (RFLP) patterns of genes of related individuals (Jeffreys et a/., 1985). By analogy with DNA fingerprinting we reasoned that it would be possible to identify particular poliovirus genotypes by examining restriction patterns of a highly variable genome segment of poliovirus. In the case of RNA viruses, this approach became feasible only after a method was found to efficiently amplify DNA. By combination of the polymerase chain reaction (PCR) (Saiki eta/., 1985; Mullis and Faloona, 1987) with reverse transcription it is possible to produce a large number of DNA copies from RNA and to compare restriction patterns of different strains of single-stranded RNA viral genomes. In this paper we present evidence that the polymorphism of a poliovirus genome region, consisting of a 480-nucleotide sequence (residues 2402 to 2881) which codes for a part of the VP1 capsid polypeptide including the antigenic site 1, can be used to characterize poliovirus strains. Through this approach we have developed an assay to obtain information on PV natural genomic variability and the genealogy of circulating strains. MATERIAL

AND METHODS

Viruses Six reference laboratory poliovirus strains, representative of each of the three serotypes (PV-1, PV-2, and PV-3), were used: Pi/Mahoney (1 M), P2/Lansing (2L), P3/Finland/23127/84 (3F) and the three attenuated Sabin vaccine strains (Sl , S2, and S3). The sequences of the entire genome of each of these strains have been determined (Kitamura et a/., 1981; Nomoto et a/., 1982; Racaniello and Baltimore, 1981; Stanway et a/., 1984; Toyoda et a/., 1984; Hugues et a/., 1986; La Monica et al., 1986). Poliovirus isolates used in this study were kindly supplied by D. Magrath (World Health Organization, Geneva), P. Minor (National Institute for Biological Standards and Control, London), 0.

ET AL.

Kew (Center for Disease Control, Atlanta), A. AubertCombiescu (Cantacuzino Institute, Bucarest), B. Le Guenno (Institut Pasteur, Dakar), and R. Handcher (Chaim Sheba Medical Center, Tel-Aviv University, TelHashomer). Viruses were grown and titrated in HEp-2c cells. Viral RNA was obtained either from cell-free supernatants of infected cells after a complete cytopathic effect or from the cytoplasm of cells infected at high multiplicity (> 10 tissue culture infectious doses 50TCID,,-per cell) after one cycle of growth. Viral RNA extraction The supernatant of infected cells (1.5 ml), containing about 1O8 TCID,, of virus, was clarified by centrifugation at 13,000 rpm for 10 min. Following treatment with 200 pg/ml Proteinase K (Apligene, France) for 15 min at 37”, sodium dodecyl sulfate was added to a final concentration of 0.39/o and the mixture was further incubated for 15 min at 37” and then for 30 min at 50”. Viral RNA was purified by one phenol extraction, one phenol-chloroform extraction, and one chloroform extraction and precipitated from 0.3 M Na acetate by ethanol overnight at -20”. Purified RNA was dissolved in 30 ~1sterile distilled water containing 5 U/PI RNA guard RNase inhibitor (Pharmacia), aliquoted in 6-111samples, and kept at -30” until used. The following alternative technique was employed when larger quantities of cDNA were required. Cytoplasmic extracts were prepared from 4 X 10” HEp-2c cells 8 hr post-infection with the virus at a multiplicity of infection of 20 TCID,,. Cells were washed with cold phosphate-buffered saline (PBS) and treated on ice for 10 min with 200 ~1 of lysis buffer (10 mM Tris-hydrochloride, pH 7.4, 140 m/l/l NaCI, 1 mM EDTA, pH 8.0, and 0.5% NP-40). Lysed cells were scraped into the lysis buffer and centrifuged for 2 min at 10,000 rpm, and the supernatant was recovered. Sodium dodecyl sulfate was added to a final concentration of l%, and the RNA was purified by two phenol-chloroform extractions and one chloroform extraction, dissolved, aliquoted, and stored as above. Primers The primers were chosen so as to fit the criterion of universality for poliovirus strains. For this, the nucleotide sequences of the entire genomes of the six reference laboratory poliovirus strains were aligned and compared (data from Genbank, released 63.0 databank, treated on a MV 8000 computer at Pasteur Institute, Paris, and with a locally developed sequence treatment by S. Crainic). Oligonucleotide primers were purchased from the Organic Chemistry Unit (Pasteur Institute, Paris). Two constant sequences, each of

POLIOVIRUS

,-

51

F-

VARIABILITY

647

VP1

UGI 1M

GENOMIC

I I

s3,-

$1

--

I

5’ L

2406

2500

- - _ _

2700

2600

--__

Nuclcotide

--__ --__

--__

NUMBER

,,-----

“Ps+

--

___---1

r 0

1 1

1 Pl I 1

2800 __-_--

-

t, 3

I 4

Pi

_.---

1

* 2900

---

t-, ,

, 0

P3

, ,

k bases

“i

7’

FIG. 1. The genome region of poliovirus used for RFLP assay. The 480-base RNA segment between the downstream (UCl) and the upstream (UGl) primers codes for the N-terminal half of the capsid protein VPl, including antigenic (Ag) site 1. The positions of the Haelll, Ddel, and Hoall restriction sites are shown on the genomes of Pi/Mahoney (1 M), Pl/Sabin (Sl), P2/Lansing (2L), P2/Sabin (S2), P3/Finland/23127/84 (3F), and PB/Sabin (S3), six representative poliovirus strains with known nucleotide sequence.

about 20 nucleotides, were found at positions 24022421 and 2761-2881 of the poliovirus (Pi/Mahoney) genome, delimiting a segment which varied in size between 472 and 480 nucleotides as a function of the strain (Fig. 1). This region codes for the N-terminal half of the VP1 capsid polypeptide, including the antigenic site 1 coding segment, i.e., amino acid positions 93 to 104 of VP1 (Van der Wet-f era/., 1983; Wichowski et a/., 1983) corresponding to base positions 2756 to 2791 of Pi/Mahoney. The “downstream” primer (UCl) has the sequence 5’-GAATTCCATGTCAAATCTAGA, and the “upstream” primer (UGl) has the sequence 5’-TTTGTGTCAGCGTGTAATGA. Primers were adjusted to a concentration of 50 pmol/pl in sterile distilled water and stored at -30”.

Reverse transcription

and enzymatic

amplification

The purified viral RNA (6 ~1) was diluted in an equal volume of water, and 50 pmol of downstream UC1 primer was added. The mixture was incubated for 3 min at 80” and annealed for 5 min at 42”. After the addition of 2 ~1 of a dNTP mixture containing 10 mM

each of dATP, dCTP, dGTP, and dTTP (Genofit, Switzerland), 4 ~1 of 5X reverse transcription buffer [250 mlLl Tris (pH 8.3 at 42”) 30 mM MgCI,, 50 mM dithiothreitol, and 200 rnM KCI], and 1 ~1 (9 U) avian reverse transcriptase (Promega Biotec), the mixture was further incubated at 42” for 30 min. Reverse transcription was terminated by 5 min boiling and quick cooling on ice. When RNA was purified from cell-free supernatants, the amplification was performed on the whole volume of the cDNA-containing reverse transcriptase mixture. The 20 ~1 of cDNA was adjusted to 100 ~1 final volume containing 10 mll/l Tris-hydrocloride (pH 8.4 at 22”) 50 mM KCI, 1.5 ml\/l MgCI,, 0.1 mg/ml gelatin, 0.1% Triton X-l 00, and 50 pmol of each UC1 and UGl. The mixture was subjected to one thermal cycle (Hybaid thermal reactor, UK) of 5 min denaturation at 95”, followed by 5 min annealing at 45”. After a short centrifugation, 2.5 U of Taql polymerase (Genofit) was added, the reaction mixture was covered with 100 ~1 of light mineral oil (Sigma), and elongation was carried out for 2 min at 70”. Reaction mixtures were then subjected to 30 automated cycles of 10 set denaturation (94’) 1

648

BALANANT

min annealing (45”) and 1 min elongation (70”). When the viral RNA was obtained from infected cell cytoplasmic extracts, only l/l 0 (2 ~1) of the cDNA solution was used in the amplification reaction. All the other steps were done as described above for RNA prepared from virus-containing cell supernatants. The amplification product was analyzed by electrophoresis on 2% agarose minigels. A gel electrophoresis band of about 480 bases was considered to be the product of a specific amplification of the targeted poliovirus genome region.

Restriction

mapping

Generally, 10 ~1of the DNA solution obtained by PCR was digested by either 10 U of Haelll (Promega), 5 U of Ddel (BRL), or 14 U of /-/pall (Promega), adjusted to a final volume of 20 ~1with the corresponding buffer. The digestions were incubated for 2 hr at 37”. DNA fragments were resolved by migration in a 3% agarose gel containing 1 pg/ml ethidium bromide, and gels were examined under uv light. In some cases, when the restriction patterns of two viruses were apparently similar, their similitude was checked by parallel close migration or by comigration of the two samples.

RESULTS The polymorphism of a 480-nucleotide of the poliovirus genome

ET AL.

rately gave more information concerning strain relatedness. For this reason, we used only the three individual digestions in further analyzing poliovirus isolates.

RFLP analysis, a marker for Sabin vaccine strain relatedness Since most of the restriction patterns of the Sabin viruses differed from those of the homotypic representative wild viruses (Fig. 2) we thought that they could be used as markers of Sabin relatedness. Therefore, we tried to determine whether these Sabin-specific restriction profiles are conserved during replication of the Sabin viruses in humans. Seven PV isolates of each of the three serotypes, previously determined to be of Sabin origin, were examined. The results presented in Fig. 3 indicate that all tested strains had RFLP patterns identical to those of the homotypic Sabin virus, regardless of which enzyme was used. Moreover, with rare exceptions, strains previously characterized as unrelated to Sabin strains, isolated before (Fig. 4) or after (Fig. 5) the introduction of OPV, had RFLP patterns distinct from those of the homotypic Sabin strain. This experiment indicated that the RFLP assay we designed, incorporating restriction profiles as strain-related markers, could be used to determine the Sabin relatedness of poliovirus isolates.

segment

To determine whether the region consisting of residues 2402 to 2881 of the poliovirus genome had the properties necessary to be effective as a polymorphic gene fragment for poliovirus strain identification, we determined the RFLP patterns of the fragment from representative poliovirus strains of each of the three serotypes. We included in this test wild strains generally used in the preparation of IPV, as well as the attenuated, Sabin vaccine viruses. The three restriction enzymes used in the test, Haelll, Ddel, and //pall, were selected by a computer-assisted analysis of the six poliovirus strains with known nucleotide sequence, so as to generate strain-specific restriction patterns. From the results of this test (Fig. 2) it can be seen that practically every strain had a specific restriction pattern. All the Sabin vaccine strains could be differentiated from the homotypic wild reference strains, except P3/Sabin, which was indistinguishable from its P3/Leon/37 parent (see below). The patterns generated by Haelll and Ddel were more strain-specific than those by Hpall. However, as will be shown further, /-/pall was able to reveal distinctions between strains that were not detected by the other two enzymes. Double digestion with Ddel and /-/pall facilitated the visualization of RFLP differences, but digestion with each enzyme sepa-

Genomic diversity among polioviruses before the introduction of vaccines

isolated

Polioviruses have been divided into three serotypes on the basis of cross-immunity (Bodian et a/., 1949) but numerous variants exist within each serotype, distinguishable by phenotypic and genotypic analyses. To see whether this diversity is reflected by the polymorphism of the genome segment that we examined, we determined the RFLP patterns of a collection of poliovirus strains isolated from paralytic cases prior to the introduction of the OPV. Many of these strains had already served to demonstrate poliovirus variability (W.H.O., 1981; Ferguson eta/., 1986). The results presented in Fig. 4 show that wild strains isolated before the introduction of the OPV had a wide variety of RFLP profiles. This was especially evident after digestion with Haelll: 16 different restriction profiles were found in the 22 PV-2 and PV-3 strains examined. However, the similarity of the restriction patterns of some of the strains isolated during the same decade in geographically distant regions revealed relationships among them. Examples shown in Fig. 4 are P2/Japan/56 and P2/lndia/56 (Haelll), P2/Egypt/52 and P2/Venezuela/ 59 (Ddel), and P3/Bombay/57, P3/lndia/58 and P31 Canada/52 (Ddel and //pall). Strains isolated a few years after the OPV had been

POLIOVIRUS

GENOMIC

VARIABILITY

Hae ill

Hpa II

NWUC 1Y Sf 2Y 21 s2 3sh 311 53 3f NW

YIUC”

649

51 2Y 2L s2 33k 3Lt 53 3F YW

FIG. 2. The RFLP patterns of the reference poliovirus strains belonging to the three different serotypes PV-1, PV-2, and PV-3. Genomic RNA of Pi/Mahoney (1 M). Pl/Sabin (Sl), P2/MEF-1 (2M), P2/Lansing (2L), P2lSabin (S2), PB/Saukett (3Sk), P3/Leon/37 (3Le), PBISabin (S3). and P3/Finland/23127/84 (3F) were extracted, reverse-transcribed, and amplified as described under Materials and Methods. Each of the resulting DNA fragments (approximately 480 bp) was separately digested with Haelll, Ddel, or Hpall, and also with both Ddel and Hpall. The digestion products were run on agarose gels in parallel with molecular weights markers (pBR322/Mspl) and with an uncut (UC) DNAfragment derived from Pi/Mahoney

introduced were found to be related to wild viruses in circulation before the use of vaccine. This was the case for the two P3/UK/62 strains isolated from paralytic cases, which were shown to be related to P3/ Bombay/57 and P3/lndia/58 by the /-/pall patterns and to P3/Canada/52 by both the Ddel and the Hoall patterns (Fig. 4). The diverse RFLP patterns occasionally observed among strains isolated at short intervals in the same geographic region (e.g., P2/Los Angeles/55 and /56, P2/lndia/56 and 157, and the two P3/UW62 strains in Fig. 4) indicate cocirculation of different viral variants in delimited geographic areas. Genealogic relationships could be established between strains used in the preparation of vaccines. For instance, two of the strains used to prepare IPV (Pl/ Mahoney and Pi/France/l 342) had identical patterns. It was surprising that the restriction patterns of Pl/ Brunenders, a strain used by some producers of IPV, appeared to be related to Pl/Sabin (OPV) but distinct from Pi/Mahoney which is the parent of Sabin 1 vaccine virus. The P2/MEF-l/42 strain, used worldwide in

the preparation of IPV, appears to be closely related to P2/Lansing/USA/37 by the Haelll and Hpall profiles. However, the two strains could be distinguished by their Ddel profiles. P2/Nicaragua/57 appeared to be a direct derivative of P2/Lansing/37. The P3/Sabin virus was found to have restriction profiles identical to those of its P3/Leon/37 parent irrespective of the restriction enzyme used. This was predictable since none of the 10 nucleotide differences between the genomes of these two strains (Stanway et a/., 1983) is located in the restriction sites examined. A relationship could be revealed between PB/Sabin (OPV) and P3/Saukett (IPV) by their Haelll and /-/pall profiles. Recent epidemics prototype strains

are characterized

by dominant

Since the RFLP assay could be used to determine poliovirus strain relatedness, we tried to see how the mass administration of OPV has influenced the distribution of poliovirus genotypes in human populations.

BALANANT

650

PV-1

PV-2

I

the

Legend

ET AL.

PV-3

I

I

to Fig. 2

For this purpose, we determined the RFLP profiles of a number of wild poliovirus strains isolated from recent poliomyelitis cases that occurred under different epidemiological conditions. The results of some representative tests (24 strains) are presented in Fig. 5. A striking uniformity was observed among homotypic strains isolated from the same geographic area during the same period of time. This was true whether isolates came from sporadic paralytic cases (Pi/Sweden/77, Pl/ France/82) or from outbreaks in endemic regions (Pl/ SenegaV86, Pi/Romania/80) or areas in which poliomyelitis had apparently been eliminated (P3/Finland/ 84). lntraepidemic variants could sometimes be detected by the RFLP pattern obtained with a particular enzyme (e.g., Pi/Senegal/79/86, Pl/lsrael/6449/88, and P3/Finland/84). In some cases intraepidemic variant strains were isolated either late in the epidemic or in areas which were geographically distant from the main epidemic regions. Examples are strain 79, which was isolated approximately 2 months after strains 1

and 12 in the Senegal/86 epidemic (B. Le Guenno, personal communication), and strains 6449 and 6466, which were isolated in different regions of Israel during the 1988 epidemic (Slater et al., 1990). Relationships were observed among predominant genotypes of different origin inside a given serotype. For instance, the two Pi/France/82 isolates appeared to be related to Pi/Sweden/l 0703/77 by the RFLP patterns obtained with Ddel and /-/pall. Similarly, a relationship was observed among the main group of P3/ Finland/84 strains (23127 in Fig. 5 being representative), the two P3/France/82 isolates, and the P3/ Romania/455.5/80 strain, indicated by their similar Ddel restriction profiles. Some of the strains isolated from paralytic cases in recent epidemics had RFLP patterns similar to those of the homotypic Sabin vaccine virus. This was the case with the two P2/Romania/598/80 strains isolated from stool and the CNS of the same child (Fig. 3), the strain P3/Romania/549.1/80 (Fig. 5) and one type 1 poliovirus

POLIOVIRUS

PV-1

1

GENOMIC

PV-2

VARIABILITY

651

PV-3

FIG. 4. The RFLP patterns of wild PV-1, PV-2, and PV-3 strains isolated before the introduction of vaccines. The RFLP profiles of P3/UK/62 strains, isolated a few years after the OPV was introduced, and those of Sabin OPV strains are presented for comparison. The genomic vrral RNA was treated as described in the Legend to Fig. 2

isolated in Romania in 1980 (not shown). For the PV-2 strains this result was in agreement with their previous characterization, by antigenic and sequence analyses, as Sabin-derived viruses (Couderc et a/., 1989).

DISCUSSION The diversity of naturally circulating strains of poliovirus, resulting from its high degree of genomic variability, has been amply documented both by phenotypic (antigenicity, reproductive capacity at supraoptimal temperatures, etc.) and bygenotypic analyses (oligonucleotide fingerprinting, genome sequencing) (Minor et al., 1980; Kew et a/., 1981; Crainic et a/., 1983; RicoHesse et al., 1987). During poliovirus replication in humans, point mutations become fixed under the steady selective pressure of neutralizing antibodies and other local conditions in the gut, such as temperature and host cell. To study the genetic variability of circulating poliovirus, we devised an assay, using RFLP analysis, which can be used to assess the relationships among

poliovirus isolates. Our analysis was limited to a single genomic segment of 480 nucleotides. In choosing this segment, we reasoned that reliable information on phenotypic variation could best be obtained by analyzing a variable region involved in poliovirus antigenicity, thus subjected to the natural selective pressure of virusneutralizing antibodies. Therefore, we selected a genome segment encoding a capsid protein (VPl) involved in the constitution of an antigenic determinant, antigenic site 1 (Blonde1 et a/., 1983; Wichowski et a/., 1983; Van der Wet-f et al., 1983; Chow et al., 1985; Hogle et a/., 1985). Using a computer-assisted search of the sequenced genomes of six reference PV strains, we were able to identify two nucleotide sequences within the capsid protein coding region fitting the criteria of generic primers for poliovirus. The specificity of the amplified genomic fragments was suggested by their constant size, by their uniqueness, by the reproducibility of their RFLP patterns, and by the correspondence of the experimental RFLP patterns of the reference viruses with

652

BALANANT

PVSweden /77

France /a2

Senegal It36

1

ET AL.

PV-2 Rumania 180

Israel 188

Rumania /80

PV-3 Rumania 180

France /a2

in different to Fig. 2.

geographic

FInlandI

FIG. 5. The RFLP patterns of PV-1, PV-2, and PV-3 strains isolated from introduction of vaccines. The genomic viral RNA was treated as described

paralytic cases in the Legend

regions

of the world

after the

those predicted from their nucleotide sequences. The accuracy of the amplification reaction was assessed by comparing nucleotide sequences of the PCR-amplified DNA of four of the reference poliovirus strains with the corresponding published sequences. In each case, the nucleotide sequence of the amplified DNA was identical to that of the original virus. Taking advantage of the specificity of reaction, we could lower the stringency in both reverse transcription and PCR, so as to increase the efficiency of amplification. In this way, it was possible to amplify the cDNA of all the PV strains that we have tested so far, i.e., more than 150 strains. The RFLP patterns of the strains studied were generally in good agreement with neutralization epitope maps (Crainic et a/., 1983, 1984; Couderc et al., 1989; results not shown) in establishing poliovirus strain relatedness. However, strain relatedness was more systematically recognized with the RFLP test than with antigenic analysis with neutralizing monoclonal antibodies. This is not surprising because, while mutations detectable by such an antigenic analysis are selected

during the multiplication of poliovirus in the human gut, the probability that a mutation detectable by the RFLP assay will be selected is very low. The average fixation rate of nonselectable mutations has been estimated to be about one to two base substitutions over the entire genome per week of in viva PV multiplication (Nottay et al., 1981; Kew and Nottay, 1984). These considerations, along with the results presented in Fig. 3, suggest that the RFLP test can be employed as a reliable assay in tracing the Sabin origin of PV isolates. A wide variety of RFLP patterns was found in strains isolated before the introduction of the vaccine. Despite this heterogeneity, close examination of the RFLP patterns obtained with one or two restriction enzymes revealed relationships among isolates of the same serotype. Furthermore, poliovirus strains isolated after vaccine introduction during some recent epidemics in particular geographic areas had identical or very similar RFLP patterns, indicating the circulation of particular regional PV genotypes. It was sometimes possible to demonstrate differences between cocirculating vi-

POLIOVIRUS

GENOMIC

ruses (e.g., P2/Romania/598/80 and P2/Romania/78/ 80 in Figs. 3 and 5, respectively) or to detect variants that had arisen during the short duration of an outbreak (e.g., P3/Finland/84, Pi/Senegal/86, or PlAsraeV88 in Fig. 5). The results of the RFLP analysis of the P3/Finland/84 strains coincided with those of a detailed antigenie characterization for which monoclonal antibodies directed against antigenic site 1 were used (Hovi, 1989). Relationships could be detected between some new epidemic strains and previously circulating viruses (e.g., between P3/Finland/84, P3/Roumania/80, and P3/France/82 in Fig. 5 and P3/Japan/57 in Fig. 4). Together with the antigenic analysis with monoclonal antibodies and the genome sequencing data, the RFLP test can provide valuable information on the circulation and prevalence of different poliovirus variants in nature. Such information is of basic interest for the epidemiological surveillance of poliomyelitis, especially in countries where the circulation of wild polioviruses has not yet been eliminated. Beyond the results we have obtained so far, the assay described here may also be used to uncover previously unrecognized epidemiological links between poliomyelitis outbreaks in different geographic areas around the world. ACKNOWLEDGMENTS We gratefully thank Helena Kopecka for permanent support and advice during this work. We thank Florian Horaud, Karine Crainic, and Ami Vonsover for helpful discussions and critical reading of the manuscript. We thank Karen Pepper for the revision of the English version. This work was supported by grants from the Direction Scientifique des Applications de la Recherche of the lnstitut Pasteur, Paris and from the lnstitut National de la Sante et de la Recherche MBdicale (CRE No. 88.1006).

REFERENCES BLONDEL, B., AKACEM, O., CRAINIC, R.. COUILLIN, P., and HORODNICEANU. F. (1983). Detection by monoclonal antibodies of an antigenie determinant critical for poliovirus neutralization present on VP1 and on heat-inactivated virions. Virology 126, 707-710. BODIAN, D., MORGAN, I. M., and HOWE, H. A. (1949). Differentiation of types of poliomyelitis viruses. Ill. The grouping of fourteen strains into three basic immunological tyeps. Am. J. Hyg. 49, 234-245. CHAPMAN, N. M., TRACY, S., GAUNTT, C. J., and FORTMULLER, U. (1990). Molecular detection and identification of enteroviruses using enzymatic amplification and nucleic acid hybridization. J. C/in. Microbial. 28, 843-850. CHOW, M.. YABROV, R., BITTLE, J., HOGLE, J., and BALTIMORE, D. (1985) Synthetic peptides from four separate regions of the poliovirus type 1 capsid protein VP1 induce neutralizing antibodies. Proc. Natl. Acad. Sci. USA 82, 9 1 O-9 14. COUDERC, T., GUINGUENE, B., HORAUD, F., AUBERT-COMBIESCU, A., and CRAINIC, R. (1989). Molecular pathogenesis of type 2 poliovirus in mice. Eur. J. Epidemiol. 5, 270-274. CR~INIC, R.. COUILLIN, P., CABAU, N., Boui, A., and HORODNICEANU. F. (1982). Determination of type 1 poliovirus subtype classes with

VARIABILIlY

653

neutralizing monoclonal antibodies. Dev. Biol. Stand. 50, 229234. CRAINIC, R., COUILLIN, P., BLONDEL, B., CABAU, N., Boui, A., and HORODNICEANU, F. (1983). Natural variation of poliovirus neutralization epitopes. lnfec. lmmun. 41, 12 17-l 225. CRAINIC, R., BLONDEL, B., AUBERT-COMBIESCU, A., COUILLIN, P., CANDR~A. A., Bouf, A., and HORODNICEANU, F. (1984). Neutralization epitope patterns on poliovirus strains isolated from paralytic cases. Dev. Biol. Stand. 57, 165-l 75. FERGUSON, M., QI, YA-HUA, MINOR, P. D., MAGRATH, D. I., SPITZ, M., and SCHILD, G. C. (1982). Monoclonal antibodies specific for the Sabin vaccine strain of poliovirus 3. Lancet ii, 122-l 24. FERGUSON, M., MAGRATH, D. I., MINOR, P. D.. and SCHILD, G. C. (1986). WHO collaborative study on the use of monoclonal antibodies for the intratypic differentiation of poliovirus strains. Bull. WHO 64,239-246. HOGLE, M. J., CHOW, M., and FILMAN, D. J. (1985). Three-dimensional structure of poliovirus at 2.9 A resolution. Science 229, 13581365. HOVI, T. (1989). The outbreak of poliomyelitis in Finland in 19841985: Significance of antigenic variation of type 3 polioviruses and site specificity of antibody responses in antipolio immunization. Adv. Virus Res. 37, 243-275. HUGHES, P. J., EVANS, D. M. A., MINOR, P. D., SCHILD, G. C., ALMOND, J. W., and STANWAY, G. (1986). The nucleotide sequence of a type 3 poliovirus isolated during a recent outbreak of poliomyelitis in Finland. 1. Gen. Viral. 67, 2093-2 102. HUMPHREY, D. D., KEW, 0. M., and FEORINO, P. M. (1982). Monoclonal antibodies of four different specimens for neutralization of type 1 polioviruses. Infect. Immunol. 36, 84 l-843. JEFFREYS. A. J., WILSON, V., and THEIN, S. L. (1985). Individual-specific ‘fingerprints’ of human DNA. Nature (London) 316, 76-79. KEW, M. 0.. NOTTAY, B. K., HATCH, M. H., NAKANO, J. H., and OBIJESKI, J. F. (1981). Multiple genetic changes can occur in the oral poliovaccines upon replication in humans. J. Gen. Viral. 66, 337-347. KEW, M. O., and NOTTAY, B. K. (1984). Molecular epidemiology of poliovlruses. Rev. Infect. Dis. 6 (Suppl. 2), S499-S504. KITAMUFIA, N., SEMLER, B. L., ROTHBERG, M. G., LARSEN, G. R., ADLER, C. J., DORNER, A. H., EMINI, A. E., HANECAK. R., LEE, J. J., VAND DER WERF, S., ANDERSON, C. W., and WIMMER, E. (1981). Primary structure, gene organization and polypeptide expression of poliovirus RNA. Nature (London) 291, 547-553. LA MONICA, N., MERIAM, C., and RACANIELLO, V. R. (1986). Mapping of sequences required for mouse neurovirulence of poliovirus type 2 Lansing. J. Viral. 57, 515-525. LWOFF, A. (1959). Factors influencing the evolution of viral diseases at the cellular level in the organism. Bacterial. Rev. 23, 109-l 24. MAGRATH. D. I., EVANS, D. M. A., FERGUSON, M., SCHILD, G. C., HORAUD, F., CR~INIC, R., STENVICK, M., and Hovr, T. (1986). Antigenic and molecular properties of type 3 poliovirus responsible for an outbreak of poliomyelitis in a vaccine population. J. Gen. Viral. 67, 899-905. MCBRIDE, W. D. (1959) Antigenic analysis of polioviruses by kinetic studies of serum neutralization. Virology 7, 45-58. MINOR, P. D. (1980). Comparative biochemical studies of type 3 poliovirus. J. Viral. 34, 73-84. MULLIS, K. B., and FALOONA, F. A. (1987). Specific synthesis of DNA in vitro vra polymerase-catalysed chain reaction. ln “Methods in Enzymology” (R. Wu. Ed.), Vol. 155, pp. 335.-350. Academtc Press, San Diego. NAKANO, J. H.. and GELFAND, H. M. (1962). The use of a modified Wecker technique for the serodifferentiation of type 1 polioviruses related and unrelated to Sabin’s vaccine strain. I. Standardization and evaluation of the test. Am. J. Hyg. 75. 363-376.

654

BAIANANT

NOMOTO, A., OMATA, T., TOYODA, H., KUGE, S., HORIE, H., KATAOKA, Y., GENBA, Y., NAKANO, Y., and IMURA, N. (1982). Complete nucleotide sequence of attenuated poliovirus Sabin 1 strain genome. Proc. Natl. Acad. Sci. USA 79, 5793-5797. NOTTAY, B. K., KEW, 0. M., HATCH, M. H., HEYWARD, J. T., and OBIJESKI, I. F. (1981). Molecular variation of type 1 vaccine-related and wild polioviruses during replication in humans. Virology 108,405423. OSTERHAUS, A. D. M. E., VAN WEZEL, A. L., UZTDEHAAG, F. G. C. M., HAZENDONK, T. G., VAN ASTEN, J. A. A. M., and VAN STEENIS, B. (1981). Monoclonal antibodies to poliovirus: Production of specific monoclonal antibodies to the Sabin vaccine strains. Intervirology 16,218-244. PETITJEAN, F., QUIBRIAC, M., FREYMUTH, F., FUCHS, F., LXONCHE, N., AYMARD, M., and KOPECKA, H. (1990). Specific detection of polioviruses in clinical samples by molecular hybridization using poliovirus subgenomic riboprobes. J. C/in. Microbial. 28, 307-311. RACANIELLO, V. R., and BALTIMORE, D. (1981). Molecular cloning of poliovirus cDNA and determination of the complete nucleotide sequence of the viral genome. Proc. Nat/ Acad. Sci. USA 78,48874891. RICO-HESSE, R., PALLANCH, M. A., NOTTAY, B. K., and KEW, 0. M. (1987). Geographic distribution of wild poliovirus type 1 genotypes. Virology 160, 31 l-322. SABIN, A. B., and BDULGER, L. R. (1973). Hystory of Sabin attenuated poliovirus oral live vaccine strains. J. Biol. Stand. 1, 1 15-l 18. SABIN, A. B. (1985). Oral poliovirus vaccine: History of its development and use and current challenge to eliminate poliomyelitis from the world. J. Infect. Dis. 151, 420-436. SAIKI, R. K., SCHARF, F., FALOONA, F., MULLIS, K. B., HORN, G. T., ERLICH, H. A., and ARNHEM, N. (1985). Enzymatic amplification of fl-globin genomic sequences and restriction site analysis for diagnosis of sickle cell anemia. Science 230, 1350-l 354. SALK, J., and SALK, D. (1987). Control of influenza and poliomyelitis with killed virus vaccines. Science 195, 834-847.

ET AL. SIATER, P. E., ORENBTEIRN, W. A., MORAG, A., AVNI, A., HANDSCHER, R., GREEN, M. S.. COSTIN, C., YARROW, A., RISHPON, S., HAVKIN, O., BEN-ZYI, T., KEW, 0. M., REY. M., EPSTEIN, I., SWARTS, T. A., and MELNICK, J. L. (1990). Poliomyelitis outbreak in Israel in 1988: A report with two commentaries. Lancet 335, 1192-l 198. STANWAY, G., CANN, A. J., HAUPTMAN. R., HUGHES, P., CLARKE, L. D., MOUNTFORD, R. C., MINOR, P. D., SCHILD, G. C., and ALMOND, J. W. (1983). The nucleotide sequence of poliovirus type 3 Leon 12 al b: Comparison with poliovirus type 1. Nucleic Acids Res. 11, 56295643. STANWAY, G., HUGHES, P., MOUNTFORD, R. C., REEVE, P., MINOR, P. D., SCHILD, G. C., and ALMOND, J. W. (1984). Comparison of the complete nucleotide sequences of the genomes of the neurovirulent poliovirus P3/Leon/37 and its attenuated Sabin vaccine derivative PYLeon 12a,b. Proc. Natl. Acad. Sci. USA 81, 1539-1543. TOYODA, H., KOHARA, M., KATAOKA, Y., SUGANUMA, J., OMATA, T., IMURA, N., and NOMOTO, A. (1984). Complete nucleotide sequence of all three poliovirus serotype genomes. Implications for genetic relationship, gene function and antigenic determinants. J. Mol. Biol. 174, 561-585. VAN DER WERF, S., WYCHOWSKI, C., BRUNEAU. P., BLONDEL, B., CFIAINIC, R., HORODNICEANU, F., and GIRARD, M. (1983). Localization of a poliovirus type 1 neutralization epitope in viral capsid polypeptide VP1 Proc. Nat/. Acad. Sci. USA 80, 5080-5084. VAN WEZEL, A. L., and HAZENDONK, A. G. (1979). lntratypic serodifferentiation of poliovmyelitis strains by strain-specific antisera. Intervirology 11, 2-8. WICHOWSKI, C., VAN DER WERF, S., SIFFERT, O., CFZAINIC, R., BRUNEAU, P., and GIRARD, M. (1983). A poliovirus type 1 neutralization epitope is located within amino acid residues 93 to 104 of viral capsid polypeptide VP1 . Eur. Mol. Biol. Organ. J. 2, 201 g-2024. W.H.O. (1981). Markers of poliovirus strains isolated from cases temporally associated with the use of live poliovirus vaccine. Report on a WHO collaborative study. J. Biol. Stand. 9, 163-l 84.