Development of an improved RT-LAMP assay for detection of currently circulating rubella viruses.

Rubella virus is the causative agent of rubella. The symptoms are usually mild, and characterized by a maculopapular rash and fever. However, rubella infection in pregnant women sometimes can result in the birth of infants with congenital rubella syndrome (CRS). Global efforts have been made to reduce and eliminate CRS. Although a reverse transcription-loop-mediated isothermal amplification (RT-LAMP) assay for detection of rubella virus has been reported, the primers contained several mismatched nucleotides with the genomes of currently circulating rubella virus strains. In the present study, a new RT-LAMP assay was established. The detection limit of this assay was 100-1000PFU/reaction of viruses for all rubella genotypes, except for genotype 2C, which is not commonly found in the current era. Therefore, the new RT-LAMP assay can successfully detect all current rubella virus genotypes, and does not require sophisticated devices like TaqMan real-time PCR systems. This assay should be a useful assay for laboratory diagnosis of rubella and CRS.

IL-28B (IFN-λ3) and IFN-α synergistically inhibit HCV replication.

Genetic variation in the IL-28B (interleukin-28B; interferon lambda 3) region has been associated with sustained virological response (SVR) rates in patients with chronic hepatitis C treated with peginterferon-α and ribavirin. However, the mechanisms by which polymorphisms in the IL-28B gene region affect host antiviral responses are not well understood. Using the HCV 1b and 2a replicon system, we compared the effects of IFN-λs and IFN-α on HCV RNA replication. The anti-HCV effect of IFN-λ3 and IFN-α in combination was also assessed. Changes in gene expression induced by IFN-λ3 and IFN-α were compared using cDNA microarray analysis. IFN-λs at concentrations of 1 ng/mL or more exhibited concentration- and time-dependent HCV inhibition. In combination, IFN-λ3 and IFN-α had a synergistic anti-HCV effect; however, no synergistic enhancement was observed for interferon-stimulated response element (ISRE) activity or upregulation of interferon-stimulated genes (ISGs). With respect to the time course of ISG upregulation, the peak of IFN-λ3-induced gene expression occurred later and lasted longer than that induced by IFN-α. In addition, although the genes upregulated by IFN-α and IFN-λ3 were similar to microarray analysis, interferon-stimulated gene expression appeared early and was prolonged by combined administration of these two IFNs. In conclusion, IFN-α and IFN-λ3 in combination showed synergistic anti-HCV activity in vitro. Differences in time-dependent upregulation of these genes might contribute to the synergistic antiviral activity.

A continuing high incidence of subacute sclerosing panencephalitis (SSPE) in the Eastern Highlands of Papua New Guinea.

The aims of this descriptive study were to confirm the high incidence of subacute sclerosing panencephalitis (SSPE) previously reported from Papua New Guinea (PNG) and to relate SSPE to previous measles vaccination and measles illness. From February 1997 to April 1999 we diagnosed a total of 55 patients with SSPE at Goroka Base General Hospital in Eastern Highlands Province (EHP) of PNG. The diagnosis was based on high cerebrospinal fluid and serum measles virus antibody titres with progressive neurological disorder and myoclonic jerks. Of these 55 patients 42 were from EHP, including 32 whose onset was in the 2-year period 1997-1998. The annual incidence of SSPE in EHP in these 2 years was 98 per million population under 20 years of age, the highest ever reported. This incidence was more than ten times higher than the highest incidence in the prevaccine era reported from elsewhere. The mean age of onset of SSPE was 7.7 years (range 2.8-14.8 years) and the interval between measles and the onset of SSPE, where known, had a mean of 5.9 years and a range of 2.5-11.1 years. Among the SSPE patients 19 had a documented history of measles vaccination. Eight of these 19 also had documentation of previous measles illness; of these, seven were vaccinated after the development of measles and one was vaccinated 20 days before measles illness. Two non-SSPE children received vaccination twice which was documented and subsequently developed measles which was also substantiated by documentation. Two patients with SSPE yielded amplified nucleotide sequences of measles virus that were different from any of the vaccine strains. We found no evidence to implicate measles vaccination in the development of SSPE.

Effectiveness and safety of mutant Escherichia coli heat-labile enterotoxin (LT H44A) as an adjuvant for nasal influenza vaccine.

The effectiveness and safety of mutant Escherichia coli heat-labile enterotoxin, LT H44A (His to Arg substitution at position 44 from the N-terminus of the A1 fragment of the A subunit) as an adjuvant for nasal influenza vaccine were examined. (1) When 0.2 microg of LT H44A, together with 0.2 microg of influenza A/PR/8/34 virus (PR8, H1N1) vaccine, was administered intranasally into BALB/c mice (twice, 4 weeks apart), anti-PR8 hemagglutinin (HA) IgA and IgG antibody (Ab) responses were induced at levels that were sufficient to provide either complete protection against infection with a small volume of PR8 virus suspension or partial protection against infection with a lethal dose of the suspension. The dose of the mutant LT and vaccine used here (0.2 microg/ 20 g doses mouse) corresponded to the estimated dose per person, i.e. 0.1 mg/10 kg body weight. (2) Using these vaccination conditions, no additional total IgE Ab responses were induced. (3) The mutant was confirmed to be less toxic than the native LT when the toxicity was analyzed either using Y1 adrenal cells in vitro (1/483 EC(50)) or by an ileal loop test. (4) One hundred micrograms of the mutant, administered intranasally or intraperitoneally into guinea-pigs (Heartley strain, 0.3-0.4 kg), caused no body-weight changes 7 days after administration, although 100 microg of the native LT administered intraperitoneally caused death in all guinea-pigs due to diarrhea within 2 days. The intranasal administration of 100 microg of the mutant resulted in almost no pathological changes in the nasal mucosa 3 days after administration. These results suggest that LT H44A, which can be produced in high yields in an E. coli culture (about 5 mg/l), could be used as one of the effective and safe adjuvants for nasal influenza vaccine in humans.

Mutants of cholera toxin as an effective and safe adjuvant for nasal influenza vaccine.

The effectiveness and safety of mutants of cholera toxin (CT) as an adjuvant for nasal influenza vaccine was examined. Four CT mutants, called CT7 K (Arg to Lys), CT61F (Arg to Phe), CT112 K (Glu to Lys), and CT118E (Glu to Gln), were produced by the replacement of one amino acid at the A1-subunit using site-directed mutagenesis. All these mutants were confirmed to be less toxic than native CT when the toxicity was analysed by using Y1 adrenal cells in vitro. When high (1 microg) and low (0.1 microg) doses of these CT mutants, together with high (1 microg) and low (0.1 microg) doses of influenza A/PR/8/34 virus (H1N1) vaccine, respectively, were administered intranasally into BALB/c mice in a two dose regimen (twice, 4 weeks apart), they produced both anti-PR8 hemagglutinin (HA) IgA and IgG antibody (Ab) responses roughly in a dose-dependent manner. The relatively low level of anti-HA Ab responses, induced by the low dose CT mutants, were enough to provide complete protection against the homologous virus infection. Under these vaccination conditions, no anti-CTB IgE Ab responses were induced. The mutant CT112 K, which showed a relatively high adjuvant activity, the lowest toxicity and relatively high yields in a bacterial culture, seems to be the most effective and safest adjuvant for nasal influenza vaccine among those examined. The low dose of CT derivatives or vaccine used in the mouse model (0.1 microg/20 g mouse) corresponded to 100 microg/20 kg, the estimated dose per person. A tentative plan for safety standards for human use of CT (or LT) derivatives as an adjuvant of nasal influenza vaccine is discussed.

Mutants of Escherichia coli heat-labile enterotoxin as an adjuvant for nasal influenza vaccine.

The effectiveness and safety of known mutants of Escherichia coli heat-labile enterotoxin (LT) as an adjuvant for nasal influenza vaccine were examined. Six mutants, called LT7K (Arg to Lys), LT61F (Ser to Phe), LT112K (Glu to Lys), LT118E (Gly to Glu), LT146E (Arg to Glu) and LT192G (Arg to Gly) were constructed by the replacement of one amino acid at one position of the A1 subunit to another using site-directed mutagenesis. All mutants were confirmed to be less toxic than wild-type LT when analyzed using Y-1 adrenal cells in vitro. When influenza vaccine was administered intranasally with LT7K and LT192G, BALB/c mice developed high levels of serum and local antibodies to the HA molecules. The adjuvant activity of these mutant LTs corresponded to that of wild-type LT when 1 microgram of these mutant LTs (or wild-type LT) was coadministered with the vaccine. From the point of view of safety, LT7K was considered to be the most potent mucosal adjuvant and was examined in more detail. The adjuvant activity of the mutant was lowered more rapidly with a decrease in dose than was that of wild-type LT. The low level of adjuvant of a relatively small amount of LT7K was heightened by adding LTB to the mutant LT. These results suggest that LT7K supplemented with LTB can be used as a less toxic, effective adjuvant for nasal influenza vaccine.

A comparison of nucleotide sequences of measles virus L genes derived from wild-type viruses and SSPE brain tissues.

The nucleotide sequences of the large protein (L) gene derived from two wild-type measles viruses (MV) and two SSPE brain-derived viruses have been determined. All sequences have single large open reading frames encoding 2183 amino acid residues. The deduced L proteins are well conserved and the proposed functional domains which have been identified for rhabdo- and paramyxoviruses are completely conserved in all strains. The degree of variability of L proteins is the lowest of all structural proteins of MV, reflecting its role in virus reproduction and persistence. Biased hypermutation was not observed in the L genes derived from SSPE brain tissue. None of the nucleotide changes can be associated with the attenuated phenotype of the Edmonston vaccine viruses.

Escherichia coli heat-labile enterotoxin B subunits supplemented with a trace amount of the holotoxin as an adjuvant for nasal influenza vaccine.

Escherichia coli heat-labile enterotoxin B subunit (LTB) (2 micrograms), supplemented with a trace amount of the holotoxin (LT) (0.02-20 ng), was examined for the adjuvant effect on antibody (Ab) responses against influenza inactivated haemagglutinin (HA) vaccine in Balb/c mice. Each mouse received a primary intranasal (i.n.) inoculation with the vaccine (1.5 micrograms), prepared from PR8 (H1N1) virus, together with LT-containing LTB and in 4 weeks a second i.n. inoculation of the vaccine alone. The inoculation of the vaccine with the LT-containing LTB induced significantly high primary and secondary anti-HA IgA and IgG Ab responses in the nasal wash and the serum, while the vaccine with LTB or less than 2 ng of LT induced little response. The synergistic adjuvant effect was maximal in the concentration of LTB supplemented with 0.2-2 ng of LT. Under these conditions, the augmented IgA and IgG Ab responses, which are cross-protective to PR8 HA molecules, provided complete cross-protection against PR8 virus challenge in mice immunized with heterologous vaccine within the same subtype. These results suggest that LTB containing a trace amount of LT can be used as a potent adjuvant for nasal vaccination of humans against influenza.

Virulence region of plasmid pNL2001 of Salmonella enteritidis.

The virulence region of the Salmonella enteritidis 55 kb plasmid pNL2001 was identified by Tn1-insertion mutagenesis, DNA hybridization studies, and Western blot analysis of proteins encoded in the virulence region of the plasmid. DNA hybridization studies showed that the pNL2001 plasmid contained a 6.4 kb SalI-EcoRI fragment homologous to the 6.4 kb SalI-EcoRI Salmonella plasmid virulence (spv) region of the S. choleraesuis 50 kb plasmid (pKDSC50). One of the 247 Tn1-insertion mutants of S. enteritidis, designated strain TA19, showed a reduced mouse lethality, and the Tn1-insertion of strain TA19 was located within this homologous 6.4 kb region, suggesting that the 6.4 kb SalI-EcoRI fragment of pNL2001 contained the spv region. Two contiguous SalI-EcoRI fragment were cloned into the expression vectors. By the 6.4 kb SalI-EcoRI fragment were cloned into the expression vectors. By Western blot analysis using four Spv peptide antisera, each specific for individual proteins encoded in the spvR, spvA, spvB and spvC genes of pKDSC50, four proteins encoded in the 6.4 kb SalI-EcoRI fragment of pNL2001 were identified. Protein SpvR with an apparent molecular mass of 32 kDa was produced from the 2.3 kb SalI-EcoRI fragment, and protein SpvA, SpvB and SpvC with apparent molecular masses of 32, 70 and 29 kDa, respectively, were produced from the 4.1 kb EcoRI-EcoRI fragment. From the 4.1 kb EcoRI::Tn1 fragment of the TA19 plasmid, proteins SpvA and SpvB were expressed, but not SpvC. It was therefore suggested that the spvC gene may contribute to the expression of virulence of S. enteritidis. Furthermore, the nucleotide sequence of the 6.4 kb SalI-EcoRI fragment encoding these four proteins was determined. Four open reading frames which encoded the four proteins with deduced molecular masses of 33,906, 28,200, 65,349 and 27,646 Da were detected. Deduced amino acid sequences of each protein showed a high degree of identity to corresponding sequences in the virulence region of S. choleraesuis, S. dublin and S. typhimurium virulence plasmids. Therefore, we confirmed that the virulence plasmids of Salmonellae including S. enteritidis share the highly conserved region responsible for virulence.

Synergistic action of cholera toxin B subunit (and Escherichia coli heat-labile toxin B subunit) and a trace amount of cholera whole toxin as an adjuvant for nasal influenza vaccine.

Cholera toxin B subunit (CTB) and Escherichia coli heat-labile toxin (LTB) (2 micrograms), each supplemented with a trace amount of cholera toxin (CT) (0.02-20 ng), were examined for the adjuvant effect on antibody (Ab) response against influenza inactivated HA (haemagglutinin) vaccine in Balb/c mice. Each mouse received a primary intranasal (i.n.) inoculation of the vaccine (1.5 micrograms) and the CT-containing CTB and in 4 weeks a second i.n. inoculation of the vaccine alone. The primary inoculation of the vaccine with CTB alone did not induce either anti-HA IgA or IgG Ab response, or haemagglutination-inhibition Ab responses in the serum. The vaccine with less than 2 ng of CT also failed to induce Ab response. On the other hand, the vaccine with CT-containing CTB induced a high Ab response, which increased depending on the CT dose. Moreover, the second vaccine induced a response more than ten times higher than the primary one and the response increased depending on the CT dose. Similar enhancement was found in the local anti-HA IgA Ab response in the nasal wash. Such synergistic effects were observed also between LTB and CT. The amount of Ab produced by the synergism was considered to be enough to protect against virus infection. These results suggest that CTB (or LTB) containing a trace amount of CT (about 0.1%) can be used practically as a potent adjuvant for nasal vaccination of humans against influenza.

Complete nucleotide sequence of the phosphoprotein of the Yamagata-1 strain of a defective subacute sclerosing panencephalitis (SSPE) virus.

The complete nucleotide sequence of the phosphoprotein (P) gene of the Yamagata-1 strain of a defective subacute sclerosing panencephalitis (SSPE) virus was determined. Comparison with the P gene of the Edmonston strain of measles virus (MV) revealed 44 differences of which 23 nucleotides substitutions were identical with those revealed between other SSPE viruses and MV (Cattaneo et al. (1989) Virology 173, 415-425). The consensus sequence of the G insertion site was completely conserved, whereas mRNAs with one or three non-templated G residue insertions were found in addition to the mRNA of the exact genome copy. As a result of the frameshift downstream of the site of G insertion, the cysteine-rich V protein was predicted from the one G-inserted mRNA besides the P and C proteins predicted from the genome-copied mRNA.

Molecular analysis of virus-producing and non-producing clones derived from a defective SSPE virus Yamagata-1 strain.

Two virus clones were isolated from a defective SSPE virus, the Yamagata-1 strain, and designated as the YA and YF clones. The YA clone-infected cells produced neither cell-free virus nor cell-associated virus, whereas the YF clone-infected cells produced both cell-associated and cell-free virus. No difference of epitopes on structural proteins was observed between these two clones. Both clones had hemadsorption activity. Quantitation of structural protein by Western dot blots showed relatively a lower amount of M protein in the YA-infected cells than that in the YF-infected cells. The ratio, P plus M dicistronic/M monocistronic mRNA, in the YA-infected cells was about twice that in the YF-infected cells. Sequence analysis of cDNA corresponding to P plus M dicistronic mRNA revealed that the deduced M protein of the YF virus was smaller than that of the YA virus by five amino acids from the carboxy terminal. These results suggest that abundant production of P plus M dicistronic mRNA is responsible for the reduced amount of M protein in the non-productive YA clone.

Trivalent heat-labile- and heat-stable-enterotoxin probe conjugated with horseradish peroxidase for detection of enterotoxigenic Escherichia coli by hybridization.

A 1,268-bp polynucleotide probe for heat-labile and heat-stable enterotoxins (LTh, STIa, STIb) was conjugated with horseradish peroxidase (HRP). The HRP-conjugated trivalent probe was applied to the detection of enterotoxigenic Escherichia coli (ETEC) by colony and stool hybridizations. The binding of the probe to its targets was assayed by the addition of HRP substrates hydrogen peroxide and luminol in the presence of an enhancer, and the chemiluminescence was recorded by exposure to X-ray film. Slot blot hybridization demonstrated that the HRP-conjugated trivalent probe specifically hybridized with the DNA isolated from ETEC strains. The trivalent probe also specifically identified bacterial colonies of ETEC that produced LTh, STIa, STIb, LTh-STIa, or LTh-STIb. Treatment of targets with sodium dodecyl sulfate and proteinase K remarkably reduced nonspecific hybridization to DNAs of non-ETEC strains. Furthermore, this probe was able to detect stool specimens seeded with 10(2) original ETEC cells per 5 mg of feces. These results suggest that the HRP-conjugated trivalent probe is a candidate for use in the clinical laboratory to detect ETEC.

Molecular analysis of structural protein genes of the Yamagata-1 strain of defective subacute sclerosing panencephalitis virus. III. Nucleotide sequence of the hemagglutinin gene.

The full-length cDNA corresponding to the mRNA for the hemagglutinin (H) protein of the Yamagata-1 strain of the subacute sclerosing panencephalitis (SSPE) virus was cloned and the nucleotide sequence was determined. The mRNA corresponding to the H protein was composed of 1952 nucleotides and contained a single large open reading frame, which encoded 620 amino acids with a predicted molecular weight of 69,723. This cDNA clone expressed the H protein in Cos 7 cells, and the transfected cells showed hemadsorption. The nucleotide and amino-acid sequence homology with the Edmonston strain of MV were 98.0% and 96.6%, respectively. The deduced amino acid sequence had a single hydrophobic domain near the N-terminus that was long enough to serve as an anchor in the membrane. Five potential glycosylation sites were found on the H protein at identical positions as in the H protein of MV. Cysteine and proline were located at almost identical positions as those of the H protein of MV. In addition, monoclonal antibody study revealed that three epitopes, including the domains that were involved in the biological activities of the H protein of MV, were conserved on the Yamagata-1 strain. These results suggested that the H protein of the Yamagata-1 strain of defective SSPE virus is structurally and functionally similar to that of the Edmonston strain of MV.

Molecular analysis of structural protein genes of the Yamagata-1 strain of defective subacute sclerosing panencephalitis virus. I. Nucleotide sequence of the nucleoprotein gene.

The complete nucleotide sequence of cloned cDNAs corresponding to the full-length mRNA encoding the NP protein of the Yamagata-1 strain of subacute sclerosing panencephalitis (SSPE) virus was determined. The gene is composed of 1683 nucleotides and contains a single large open reading frame, which is capable of encoding 525 amino acids with a molecular weight of 58,399. Comparison of the nucleotide and predicted amino acid sequences with those of the Edmonston strain of measles virus (MV) showed that the gene and the protein were highly conserved. However, the antigenic sites on the NP protein of the Yamagata-1 strain were found to be changed by an epitope analysis using monoclonal antibodies against the NP protein of MV. Only 1 of 4 monoclonal antibodies reacted with the NP protein of SSPE virus, and the other three antibodies did not. Almost identical changes in nucleotides and amino acids were found to occur in the NP gene of the Yamagata-1 strain when compared with the IP-3-Ca strain of another SSPE virus. In addition, the deduced secondary structure of the NP protein of the IP-3-Ca strain was similar to that of the Yamagata-1 strain, but differed from the MV. These results suggest that the NP proteins of SSPE viruses have a common property that is different from MV.

Molecular analysis of structural protein genes of the Yamagata-1 strain of defective subacute sclerosing panencephalitis virus. IV. Nucleotide sequence of the fusion gene.

The full-length cDNA corresponding to the mRNA of the fusion (F) protein of the Yamagata-1 strain of subacute sclerosing panencephalitis (SSPE) virus was cloned, and its complete nucleotide sequence was determined. The F gene was composed of 2369 nucleotides and contained a single large coding region, which is located between two noncoding regions. The 5'-terminal noncoding region consisted of 584 nucleotides comprising 44.9% cytosine, and had several inverted repetitious sequences. The 3'-terminal noncoding region had a relatively low homology of 91.7% with the MV. The coding region was expanded for nucleotides 585-2189, which encoded 534 amino acids with a molecular weight of 57,963. The homology of the amino acid sequence of the F protein between the MV and SSPE virus was 96.27%, and the positions of cysteine and proline were almost identical in the two viruses. The functional domains of SSPE-virus F protein closely resembled those of MV F protein, including the cleavage site, a signal sequence, the fusion-related stretch, the transmembrane region, and four potential glycosylation sites. Four antigenic epitopes on the MV F protein were also conserved on the SSPE-virus F protein. However, deletion of one nucleotide (position 2155) of the SSPE virus was found when compared with the MV, and shifted the coding frame, causing the substitutions of 27 C-terminal amino acids of the MV F protein with 11 different residues. The variations of the C-terminal region of the F protein were observed with two other SSPE viruses, suggesting that this may be a common property of SSPE virus that differs from MV.

Molecular analysis of enterotoxin plasmids of enterotoxigenic Escherichia coli of 14 different O serotypes.

A total of 104 isolates of enterotoxigenic Escherichia coli derived from diarrheal patients from more than 10 countries were examined for serotype and toxigenicity. The transferability and molecular structure of the enterotoxin plasmids from each isolate were also examined. Enterotoxin plasmids from serotypes such as O6, O25, O27, O126, O128, and O159, which are frequently associated with E. coli diarrhea (classical strains) generally did not transfer by conjugation from clinical isolates, whereas those from serotypes such as O7, O17, O80, O98, O139, O150, and O153, which are rarely associated with diarrhea (rare strains) transferred almost always from the clinical isolates by conjugation. Analyses of enterotoxin plasmids by restriction endonucleases and DNA-DNA hybridization with the enterotoxin probes revealed that the strains with the same O serotype and toxigenicity carry closely related enterotoxin plasmids. These results suggest that classical strains resulted from the dissemination of ancestral clones which received enterotoxin plasmids long ago, while the rare strains acquired the enterotoxin plasmids recently by conjugation and have not yet been spread to the same degree as the ancestral clones.

Detection of enterotoxigenic Escherichia coli by colony hybridization with biotinylated enterotoxin probes.

By using biotinylated enterotoxin DNA probes, a method to detect enterotoxigenic Escherichia coli by colony hybridization was developed. The treatment of colonies on nitrocellulose membrane filters with proteinase K and Triton X-100 was essential for obtaining the specific hybridization. A total of 200 E. coli strains isolated from travelers with diarrhea were tested for colony hybridization by using a probe encoding heat-labile toxin (LT) type h. All strains (86 of 86) that produced LT, but none of the non-LT producers, hybridized with 32P-labeled and biotinylated LT type h probes. A total of 36 strains chosen randomly from the 200 isolates were tested for colony hybridization by using heat-stable enterotoxin (ST) probes. All but two strains that hybridized with the 32P-labeled ST type Ia probe also hybridized with the biotinylated ST type Ia probe. All strains that hybridized with the 32P-labeled ST type Ib probe also hybridized with the biotinylated ST type Ib probe. Thus, almost all E. coli strains tested were judged to be the same by colony hybridization with biotinylated or 32P-labeled enterotoxin probes. These results demonstrate that the biotinylated enterotoxin probes are useful in the diagnosis of enterotoxigenic E. coli strains by colony hybridization.

Analysis of the plasmids of Escherichia coli O148:H28 from travellers with diarrhea.

98 Escherichia coli strains of serotype O148:H28 isolated from diarrheal patients from 10 Asian countries and Mexico at Osaka Airport Quarantine were analyzed for enterotoxigenicity and plasmid profile. They were classified into three groups. The first group contained 44 strains that were non-enterotoxigenic and carried 3.9 kb and 50 kb non-enterotoxin plasmids. The second group contained 9 strains that produced LT and ST. They carried a 45 kb enterotoxin plasmid, and 4.6 kb and 9.2 kb non-enterotoxin plasmids. The third group contained 45 strains that produced ST. They carried a 40 kb enterotoxin plasmid, and non-enterotoxin plasmids other than 3.9 kb, 4.6 kb, 9.2 kb and 50 kb. Southern blot hybridization demonstrated that all the non-enterotoxin or enterotoxin plasmids carried by the strains of the same group were identical or similar. These results suggested that the 98 E. coli strains with O148:H28 serotype were derived from three clones, and that the individual strains among each group were derived from a single clonal strain.