Zbl. Bakt. 277,170-178 (1992) © Gustav Fischer Verlag, StuttgartlNew York

A Sensitive Method for the Detection of Enterotoxigenic Escherichia coli by the Polymerase Chain Reaction Using Multiple Primer Pairs AKIO ABEl, HIROMI OBATA 2, SHIGERU MATSUSHITA 2, SUMIO YAMADAZ, YASUO KUDOH 2, AROON BANGTRAKULNONTH 3 , ORN-ANONG RATCHTRACHENCHAT 3 , and HIROFUMI DANBARA 1 * 1 Department

of Bacteriology, The Kitasato Institute, 5-9-1 Shirokane, Minato-ku, Tokyo 108, Japan, 2Department of Microbiology, Tokyo Metropolitan Research Laboratory of Public Health, 3-24-1, Hyakunin-cho, Shinjuku-ku, Tokyo 169, Japan, and 3 Division of Clinical Pathology, Department of Medical Science, Ministry of Public Health, Bangkok, Thailand

With 5 Figures· Received August 16, 1991 . Revision received January 2, 1992 . Accepted February 12, 1992

Summary In this study, a polymerase chain reaction (PCR) method has been developed for the detection of enterotoxigenic Escherichia coli (ETEC). Three different sets of oligonucleotide primers synthesized were used to amplify the enterotoxin genes of heat-labile (LTh) and heat-stable (STIa and STIb) enterotoxins of ETEC. These primers amplified a 627, 240, or 169 base pair (bp) DNA fragment from LTh, STIa and STIb gene, respectively, of the reference ETEC strains. The addition of RNase A (10 ug/ml) to the PCR reaction solution diminished nonspecific amplification of DNA fragments other than the enterotoxin genes. Five types of ETEC strains corresponding to the LTh, STIa, STIb, LTh-STIa, or LTh-STIb genotypes were distinguished by a single procedure of PCR using the mixture of the three sets of primers. PCR, hybridization, and conventional methods were subjected to one hundred stool specimens from diarrheal patients. It was found that PCR was the most sensitive method among them. These results suggested that PCR with triple primer pairs would be useful for the laboratory diagnosis of ETEC in the stool specimens.

Zusammenfassung In der vorliegenden Untersuchung wurde ein die Polymerase-Ketten-Reaktion (PCR) nutzendes Verfahren fur die Bestimmung enterotoxigener Escherichia coli (ETEC) entwickelt. Es wurden drei verschiedene Satze synthetisierter Oligonucleotid-Primer verwendet, urn die " Corresponding author

Detection of ETEC by PCR Method

171

Enterotoxin-Gene hitzelabiler (LTh ) und hitzestabiler (S11a und S11b) ETEC-Enterotoxine zu verstarken, Diese Primer verstarkten [eweils ein aus 627, 240 oder 169 Basenpaaren bestehendes DNS-Fragment der LTh- bzw. S11a- und S11b-Gene der ETEC-Referenzstiimme. Eine Zugabe von RN ase A (10 ug/rnl) zu der PCR-Reaktionslosung verhinderte die unspezifische Verstiirkung anderer als der Enterotoxin-Gene. Bei Anwendung eines einzigen PCR-Verfahren s unter Verwendun g eines Gemisches der drei Sarze von Primern lielSen sich funf Typen von ETEC-Stammen unterscheiden, die den Genotypen LTh, S11a, S11b, LThS11a und LTh-S11b entsprachen. 100 Stuhlproben von Dur chfallpatienten wurd en mirtels PCR, Hybridisierung und konventionellen Methoden untersucht. Es ergab sich, daIS die PCR das empfindlichste Verfahren darstellte. Die Ergebnisse deuten darauf hin, daIS die PCR mit dreifachen Primer-Paaren fU r die Labordiagnose von ETEC in Stuhlproben von Nutzen sein durfte.

Introduction Enterotoxigenic Escherichia coli (ETEC) causes diarrhea in humans and animals. ETEC produces two types of ent erotoxin, heat-labile enterotoxin (LT) and heat-stable ent erotoxin (ST) . LT type h (LT h) of ETEC is functionally, stru ctura lly and genetically closely related to the chole ra toxin (CT ) of Vibrio cholerae 01. ST is classified into STI and ST U, and STI is subtyped as ST type Ia (STIa) and type Ib (STIb). The mo st frequent entero toxins found in hum an ET EC are LTh , STIa, and STIb. Th e pr oduction of ente ro toxi ns of ETEC or V. cho lerae 01 is conventiona lly tested by animal models or immunological meth od s. The ligated rabbit int estina l loop test (7) and the reversed pa ssive agglutina tion test (9) are used to detect CT and LT. ST is detected by the suckling mou se test (4). A DNA-DNA hybridization method using gene probes has been develop ed for th e detection of ETEC (1, 2, 5, 6, 8, 11 ). A simple method is required for the lab or atory diagnosis of the causa tive agents of infectious diseases. The polymerase cha in reaction (PCR) meth od (15, 16 ) ha s been applied in the detection of microor gani sms such as Salmonella typ him urium (18), Chlamydia trachomatis (14), ETEC (13) and My cobacterium tub erculosis (19). In th e pre sent paper, we describe a sensitive PCR meth od to detect ET EC producing LTh , STIa, and STIb in th e feces o f travellers with diarrhea.

Materials and Methods

Bacterial strains. E. coli C600 strain (our laboratory strain) was used for the propagation of recombinant plasmids. Wild-type E. coli producing LTh, S11a, S11b, LTh-S11a, and LThSTlb were used as the authentic reference strains of ETEC reported previously (3). Media, chemicals, and enzymes. PAB (antibiotic medium 3, Difco Lab., Detroit, Mich. ) supplemented with 10 ug/ml thymidine, PAB agar , and DHL agar (Eiken Kagaku, Tokyo, Japan) were used. A therrnoresistanr DN A polymerase (Taq DNA polymerase, The PerkinElmer Corp., Norwalk, CT, USA) restriction enzymes (Tokyobo, Osaka, Japan), and the Ee L gene detection system on Hyperfilm ECL (Amersham, Buckinghamshire, U.K.) were pur chased commercially. Synthesis of oligonucleotide LTh, STla, and STlb primers. Oligonucleotide LTh, STla, and S11b primers were synthesized based on the nucleotide sequences reported by Yamamoto et al. (2 1), So et al. (20), and Moseley et al. (12), respectively. They were synthesized by automated DNA synthesizer (Cyclone DNA synthesizer, Bioserch, Inc., CA, USA), and purified by 16% polyacrylamide gel electrophoresis. The base sequence and position of primers wirh respect to the reported sequences are shown in Table 1.

172

A.Abe et al.

Detection of ETEC by PCR. The template DNA for PCR was prepared by boiling from the authentic reference ETEC strains and diarrheal stool specimens as follows. A single bacterial colony of the reference ETEC strains was grown on PAB agar and suspended in 200 I-ll of distilled water. For the stool specimens,S mg of diarrheal feces was inoculated into 3 ml of PAB and grown overnight at 37°C. Bacteria harvested from 200 I-ll of the overnight culture were resuspended in 200 I-ll of distilled water. These suspensions were boiled for 5 min and centrifuged at 12000 rpm for 5 min. A 38 I-ll portion of the supernatant was used as the source of template DNA for PCR as reported by Saiki et al. (15, 16). Briefly, 38 I-ll of the template DNA was mixed with 11 I-ll of reaction buffer at a final concentration of 20 mM Tris-HCI (pH 8.0), 1.5 mM MgCh, 25 mM KCI, 0.05% Tween 20, 100 ug/ml bovine serum albumin, 50 I-lM each of dATP, dGTP, dCTP, and dTTP, 0.21-lM each of the six primers and 10 ug/ml RNase A. The mixture was heated at 98°C for 5 min to denature the DNA and rapidly cooled to room temperature. Thereafter, 1 I-ll of Taq DNA polymerase (2.5 units) was added to the mixture. The mixture was overlayed by 40 I-ll of mineral oil. The first step of annealing and extension was at 58°C for 2 min, and at 72 °C for 3 min, respectively. The subsequent 29 cycles of the denaturation, annealing and extension were performed at 94°C for 50 sec, at 58 °C for 1 min, and at 72 °C for 2 min, respectively. Southern blot hybridization of PCR products. DNA amplified by PCR was electrophoresed on a 2.0% agarose gel and transferred to Hybond N + membrane (Amersham, Buckinghamshire, U.K.) by capillary blotting in the presence of 0.4 M NaOH (14). The membrane was washed with a solution containing 0.3 M NaCl and 0.03 M sodium citrate (pH 7.0, 2 x SSC solution), and DNA was cross-linked by UV light using UV stratalinker 1800 (Stratagene Co., La Jolla, CA, USA). Plasmid pKAD008 (1) was digested with Xba I and Eco RI and the resulting DNA fragment of 676 bp, 236 pb, and 173 bp were conjugated with horseradish peroxidase and used as the probes for LTh, STIa, and STIb enterotoxins, respectively. Chemiluminescence generated after hybridization was detected by the ECL gene detection system on Hyperfilm (Amersham) as previously reported (1). Detection of ETEC by conventional and hybridization methods. Conventionally, enterotoxin production was assayed for 5 bacterial colonies of E. coli from the diarrheal stool specimens grown on DHL agar. LTh production was tested by reversed passive agglutination (9) with the Serotoxin LT Kit (Eiken Kagaku, Tokyo, Japan), and ST production by the fluid accumulation in the intestinal loop of suckling mice (4). For stool hybridization,S ul of stool suspension in PAB (1 g of diarrheal feces/ml) layered on nitrocellulose filter was used as the target. A 1300 bp DNA fragment prepared from pKAD008 by digestion with XbaI

Table 1. Oligonucleotide primers used for PCR Primers

DNA fragments amplified (bp)

Sequence"

Position b

LTh

627

STla

240

STlb

169

(+)ATATATGGATGGTATCGTGTT (-)TCCTTCATCCTTTCAATGGC (+)CCGTGAAACAACATGACG (-)ACATCCAGCACAGGCAGGATT (+)TTCACCTTTCCCTCAGGATG (-)GCACCCGGTACAAGCAGGATT

424-444 253-272 -27--10 193 - 213 45- 64 193 ~ 213

a

b

Oligonucleotide primers were synthesized for coding (+) and noncoding (-) strands of LTh, STIa, and STlb genes. Position" 1" corresponds to the adenine base of first ATG codon of open reading frame of the enterotoxin genes. The base positions of (+) and (-) strands of LTh gene signify the number of bases from the first ATG of open reading frame of subunits A and B, respectively.

Detection of ETEC by PCR Method

173

was used as LTh-STla-STIb gene probe. Stool hybridization was carried out for the detection of LTh, STIa, or STIb gene of ETEC in the diarrheal stool specimens according to the method previously reported (1).

Results

Establishment of optimal PCR condition using reference ETE C strains The optimal PCR condition was dete rmined using the LTh , STIa, and STIb primers synt hesized. The template DNA was prep ared from the reference ETEC strains. The temperature was optimized at 58 °C for annealing of the primers with DNA, 94 °C for denaturation of the pr imers and template DNA, and 72 "C for the extension of DNA synt hesis by Taq DNA pol ymerase. Th e addition of RNase A at a final concentration of 10 ug/rnl diminished non specific template DNA fragments amplified from other sequences than L Th, STla, and STIb genes (Fig. 1).

M

2

3

4

+

+

+

+

5

6

7

8

1632 517 396 344 298 221 154

- LTh (627) - STla (240) - STlb (169)

Fig. 1. Effect of RNase A on PCR. The mixtur e of LTh, STla, and STIb primers was used for PCR. PCR was performed with or without RNase A (10 ug/ml) in the reaction solution using template DNA isolated from the reference ETEC strains producing LTh (lanes 1, 5), STIa (lanes 2, 6), and STIb (lanes 3, 7). Lanes 4 and 8 represent E. coli C600 strain producing no enterotox in. Lane M is a size marker in base pairs obtained from pBR322 by H inf I digestion. Symbols (+ ) or (-) indicates that RNase A was present or absent, respectively, in the reaction solution of PCR.

Th ese primers amplified DNA fragments of 627, 240, or 169 bp from template DNAs of the reference ETEC stra ins producing LTh, STIa, or STlb enterotoxin, respectively, as shown in Fig. 1 (lanes 5, 6, 7). Amplified DNA fragments were analyzed for their specificity by Southern blot hybr idization. The 670, 240 , or 169 bp fragment hybridiz ed specifically with the LTh , STIa, or STIb DN A probes, respectively, as shown in Fig. 2, demonstrating that these fragments were amplified from the correspo nding enterotoxin genes. The mixture of the three sets of LTh , ST Ia, and STIb primers in one PCR reacti on solution was used to ident ify the reference ETEC strai ns. As show n in Fig.3, the identified geno types were exactly the same as the enterotoxin types of the reference strains (STIa; lane s 1-3, LTh-STla; lanes 4-6, STlb ; lanes 7-9, LTh-STlb; lanes 10-12). Th ese results indic ated that five types of ETEC producing LTh , STIa, STlb ,

174

A.Abe er al.

1

2

3

-627 -240 -169

Fig. 2. Southern blot hybridization analysis of PCR products. PCR products with RNase A obtained in Fig. 1 were hybridized with respective DNA probes conjugated with horseradish peroxidase, and hybridization signals of chemiluminescence were detected by the ECL detection system (1). Lanes 1, 2, and 3 are PCR products from the reference ETEC strains producing LTh, STla, and STIb, respectively.

M -

1632

J 517 ../ 396 - 344

~ 298

221 154

Fig. 3. Detection of reference ETEC srrains by PCR. The mixtur e of LTh, STla, and STIb primers was used for PCR against template DNA prepared from the reference ETEC strains produ cing STla (lanes 1-3), LTh-STla (lanes 4-6), STlb (lanes 7- 9), and LTh-STlb (lanes 10-12). Lanes M represent a size marker in base pairs of pBR322 digested with Hinf I.

LTh-S'Tla, and LTh-STlb were detectable by a single PCR procedure using the mixture of thr ee sets of enterotoxin primers.

Detection of ETE C strains in the diarrheal stool specimen s peR meth od using the thr ee sets o f pr imers was applied to detect ETEC str ains in the diarrheal stool specimens. Fig. 4 sho ws the results of PCR fo r a repre sentative set of 12 samples. Th ese sa mples were sho wn to cont ain ETEC with enterotoxin geno types of LTh (lanes 1- 3, 6), STla (lane 9), STIb (lane 8), LTh-STIb (lanes 7, 10-11 ). Three samples (lanes 4, 5, 12) did not contai n ETEC. These result s showe d that PCR with a mixture of th e three sets of LTh, STIa, and STIb primers were ab le to detect ETEC in the dia rrhea l stoo l samples.

Detection of ETEC by PCR Method

M

2

3

4

5

6

7

8

9

10 11

175

12

1632 517 396 344

293

221 154 Fig. 4. Detection of ETEC in the diarrheal stool specimens by PCR. The mixture of LTh, STIa, and STlb primers was used for PCR against template DNA prepared from the stool specimens of diarrheal patients. The stool specimens were found to contain ETEC with genotypes LTh (lanes 1-3, 6), STIa (lane 9), STIb (lane 8), LTh-STIb (lanes 7, 10, 11). Lanes 4,5, and 12 contained no ETEC. Lane M represents a size marker in base pairs of pBR322 digested with Hinf 1. Note that DNA fragments of PCR products from LTh gene (627 bp) show a faint band in lanes 7 and 11.

Comparison of PCR, hybridization, and conventional methods to detect ETEC in the diarrheal stool specimens A total of one hundred stool specimens was collected from diarrheal patients (travellers). The detectability of ETEC in the stool specimens by PCR, hybridization, and conventional methods was compared in Table 2. Group 1 (42 samples) and group 4 (51 samples) were examined and classified as ETEC positive and negative stool specimens, respectively. Thus, a 93% agreement was obtained among the three methods. Group 2 (2 samples) was examined for ETEC-positivity by PCR and hybridization, but for negativity by the conventional method. Group 3 (5 samples) was examined for ETEC-positivity only by PCR. In summary, 49, 44, and 42 samples were determined to be ETEC-positive stool specimens by PCR, hybridization, and conventional methods, respectively, demonstrating that peR method was the most sensitive one.

Table 2. Detection of ETEC in stool specimens from 100 diarrheal patients Methods Groups

PCR

Hybridization"

1 2 .1

4

+ + +

+ +

Conven- No. of tional b specimens

+

42 2 5 51

(+) and (-) indicate ETEC positive and negative, respectively. a

b

The stool specimens were suspended in PAB and a portion of 5 ul layered onto nitrocellulose filter was used as the target for a trivalent LTh-STIa-STlb DNA probe (1). Five colonies from each stool specimens were subjected to the suckling mouse (4) and reversed passive agglutination tests (9) for the detection of ST and LT toxins, respectively.

176

A.Abe et al.

Discussion PCR method developed in the present study is most sensitive. This was demonstrated by the fact that 49,44, and 42 stool specimens were determined to be ETEC-positive by PCR, hybridization, and conventional methods, respectively. When 5 bacterial colonies were subjected to the enterotoxin assays by the conventional method, the stool specimens of group 2 were determined to be ETEC-negative samples. However, an ETEC colony was detected among 45 additional colonies in one of the 2 samples of the group 2 stool specimens. These results suggested that a small number of ETEC cells was contained in these stool specimens. The stool specimens of group 3 were determined to be ETEC-positive only by the PCR method, but negative samples by both stool hybridization and conventional methods. For stool hybridization, 5 mg of feces was layered onto nitrocellulose filter. Our previous report showed that the detection limit is 100 original living ETEC cells per 5 mg of feces by stool hybridization (1). These results, therefore, suggest that a stool specimen containing 100 original living ETEC cells per 5 mg feces is considered to be ETEC-positive using the overnight culture of in vivo amplification step at 3rc for preparation of template DNA for PCR. Subtyping of ST-producing ETEC into STIa and STIb is impossible by the conventional method using suckling mouse assay. Our previous report showed that this subtyping was also difficult when using the hybridization method because only a weak hybridization signal can be obtained between probe and ST gene (1, 8). However, the discrimination of ST producers into STIa and STIb was easily done by our PCR method developed in this study. Moreover, ETEC not only from pure cultures but also those contained in the feces were classified into five genotypes, LTh, STIa, STIb, LTh-STIa, and LTh-STIb, by a single step of PCR and agarose gel electrophoresis. For example, the 42 samples of group 1 stool specimens were characterized to be carriers of ETEC with genotypes LTh (4 samples), STIa (9 samples), STIb (12 samples), LTh-STIa (6 samples), and LTh-STIb (11 samples). Our PCR method with triple primer pairs detected Vibrio cholerae 01 producing cholera toxin (CT). As shown in Fig. 5, a 627 bp DNA fragment was amplified from V. cholerae 01 serotypes Ogawa (NIH-41) and Inaba (569B, NIH-35), and ETEC-producing LTh. The cleavage by HindIII of the amplified 627 bp fragment from the LTh

M

2

3

4

5

6

+ + 7 8

+ + + + 9 10 11 12

1632\... 517

396~

344-

298~

220r 154

Fig. 5. Detection of V. cholerae 01 strains by PCR. The mixture of LTh, STIa, and STIb primers was used for the PCR against template DNA prepared from V. cholerae 01 serotypes Ogawa (NIH41, lanes 4,10), Inaba (569B,lanes 5,11), and Inaba (NIH35, lanes 6,12). Lanes 1-3, and 7-9 were from the reference ETEC strains producing LTh. Lane M represents a size marker in base pairs of pBR322 digestedwith Hinf I. Symbol (+) or (-) indicates with or without Hind III digestion, respectively.

Detection of ETEC by PCR Method

177

gene, but not from the CT gene, into 481 and 146 bp fragments distinguished ETEC from V. cholerae 01 strains. We found that RNase A (10 ug/mll in the PCR reaction solution remarkably reduced nonspecific amplifications resulting in a clear-cut agarose gel electrophoresis pattern. Three types of distinct DNA fragments with 627, 240, and 169 bp corresponding to LTh, STIa, and STIb genes, respectively, were amplified from the reference ETEC strains and also from the stool specimens. Therefore, the genotypes of ETEC contained in the stool specimens could be judged simply by the estimation of the fragment length amplified. One should be cautious to avoid contamination of bacteria or template DNA because PCR is sensitive. Irradiation of containers or pipets by UV light may efficiently reduce the risk of contamination with DNA. Although the reagent and equipment are expensive, we believe that the PCR method is useful particularly for a rapid detection of ETEC. For a more rapid detection of ETEC, direct PCR from the stool specimens of patients will be useful. This is currently under development in our laboratory.

References 1. Abe, A., K. Komase, A. Bangtrakulnonth, a.-A. Ratchtrachenchat, K. Kawahara, and H. Danbara: Trivalent hear-labile- and heat-stable-enterotoxin probe conjugated with horseradish peroxidase for detection of enterotoxigenic Escherichia coli by hybridization. J. Clin. Microbiol. 28 (1990) 2616-2620 2. Danbara, H.: Identification of enterotoxigenic Escherichia coli by colony hybridization using biotinylated LTIh, STIa and STIb enterotoxin probes. In: Gene Probes for Bacteria (A.]. L. Macario and E. C. de Macario, eds.), pp. 167-178. Academic Press, Inc., San Diego/California (1990) 3. Danbara, H., K. Komase, H. Arita, H. Abe, and M. Yoshikawa: Molecular analysis of enterotoxin plasmids of enterotoxigenic Escherichia coli of 14 different a serotypes. Infect. Immun. 56 (1988) 1513-1517 4. Dean, A. c., Y.-c. Ching, R. G. Williams, and L. B. Harden: Test for Escherichia coli enterotoxin using infant mice: Application in a study of diarrhea in children in Honolulu. J. Infect. Dis. 125 (1972) 407-411 5. Echeverria, P. and D. N. Taylor: A comparative study of enterotoxin gene probes and tests for toxin production to detect enterotoxigenic Escherichia coli. J. Infect. Dis. 153 (1986) 255-260 6. Echeverria, P., D. N. Taylor,]. Seriioatana, and C. Moe: Comparative study of synthetic oligonucleotide and cloned polynucleotide enterotoxin gene probes to identify enterotoxigenic Escherichia coli. J. Clin. Microbiol. 25 (1987) 106-109 7. Garbach, S. L., ]. G. Banwell, B. D. Chatterjee, B. Jacobs, and R. B. Sack: Acute undifferentiated human diarrhea in the tropics. I. Alterations in intestinal microflora. J. Clin. Invest. 50 (1971) 881-889 8. Kirii, Y., H. Danbara, K. Komase, H. Arita, and M. Yoshikawa: Detection of enterotoxigenic Escherichia coli by colony hybridization with biotinylated enterotoxin probes. J. Clin. Microbiol. 25 (1987) 1962-1965 9. Kudoh, Y., S. Yamada, S. Matsusita, K. Ohta, M. Thuno, T. Muraaka, N. Obtomo, and M. Ohashi: Detection of heat-labile enterotoxin of Escherichia coli by reversed passive hemagglutination test with specific immunoglobulin against cholera toxin. In: Proceeding of the 14th Joint Conference, US-Japan Cooperative Medical Science Program, Cholera Panel (K. Takeya and Y. Zinnaka, eds.), pp. 266-273. Toho University, Tokyo (1979) 12 Zbl. Bakt. 277/2

178

A. Abe er al.

10. Mekalanos, j. j., D. j. Swart z, G. D. N. Pearson, N. Harfold, F. Groyne, and M. de Wilde : Cholera toxin genes: nucleotide sequence, deletion analysis and vaccine development. Nature (Lond.) 306 (1983) 551-55 7 11. Moseley, S. L., 1. Huq, A . R. M. A . Alim, M. So, M. Samadpour-Motalebi, and S. Falkow: Detection of enterotoxigenic Escherichia coli by DNA colony hybridization. J. Infect. Dis. 142 (1980) 892-898 12. Moseley, S. L., j. W. Hardy, M. 1. Hug, P. Echeverria, and S. Falkow: Isolation and nucleotide sequence determination of a gene encoding a heat-stable enterotoxin of Escherichia coli. Infect. Immun. 39 (1983) 1167-1174 13. Olive , D. M. : Detection of enterotoxigenic Escherichia coli after polymerase chain reaction amplification with a thermostable DNA polymerase. J. Clin. Microbiol. 27 (1989) 261-265 14. 0stergaard, L., S. Birkeland, and G. Christiansen: Use of polymerase chain reaction for detection of Chlamydia trachomatis. j, Clin. Microbiol. 28 (1990) 1254-1260 15. Saiki, R. K., D. H. Gelfand, S. Stoffel, s. j. Scharf, R. Higuchi, G. T. Horn, K. B. Mullis, and H. A. Erlich: Primer directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239 (1988) 487-491 16. Saiki, R. K., S. Scharf, F. Faloona, K. B. Mullis, G. T. Horn, H. A. Erlich, and N. Arnheim: Enzymatic amplification of beta-globin genomic sequences and restriction site analysis for diagnosis of sickle cell anemia. Science 230 (1985) 1350-1354 17. Sambrook, j., E. F. Fritsch, and T. Maniatis: Molecular Cloning, pp. 9.45-9.46. Cold Spring Harbor Laboratory, Cold Spring HarborlNew York (1989) 18. Shyamala, V. and G. F.-L. Ames: Amplification of bacterial genomic DNA by the polymerase chain reaction and direct sequencing after asymmetric amplification : application to the study of periplasmic permeases. J. Bact. 171 (1989) 1602-1608 19.5jobring, U., M. Mecklenburg, A. B. Andersen, and H. Miorner : Polymerase chain reaction for detection of My cobacterium tuberculosis . J. Clin. Microbiol. 28 (1990) 2200--2204 20. So, M. and B. j. McCarthy : Nucleotide sequence of the bacterial tran sposon Tn1681 encoding a heat-stable (ST) toxin and its identification in enterotoxigenic Escherichia coli strains. Proc. Nat!. Acad. Sci. USA 77 (1980) 4011-4015 21. Yamamoto, T., T. Tamura , and T. Yokota: Primary structure of heat-labile enterotoxin produced by Escherichia coli pathogenic for humans. j. BioI. Chern. 259 (1984) 5037- 5044

Dr. Hirofumi Danbara, Department of Bacteriology, The Kitasato Institute, 5-9-1 Shirokane, Minato-ku, Tok yo 108, japan, present address: Department of Microbiology, School of Pharmaceutial Sciences, Kitasaro University, 5-9-1 Shirokane, Minato-ku, Tokyo 108, japan