Bull. Vet. Inst. Pulawy 47, 363-375, 2003

COMPARISON OF ARBITRARY PRIMER (AP) PCRAND PULSED-FIELD GEL ELECTROPHORESIS METHODS

FOR GENOTYPIC DIFFERENTIATIONOF ESCHERICHIA COLI O157 STRAINS

MARCIN WEINER AND JACEK OSEK

Department of Microbiology,National Veterinary Research Institute,

24-100 Pulawy, Polande-mail: josek@piwet.pulawy.pl

Received for publication July 03, 2003.

The genetic relatedness of 12 Escherichia coli O157 strains isolated from cattle andhumans was examined by methods based on arbitrary primer amplification of DNA sequences(AP-PCR) and by pulsed-field gel electrophoresis (PFGE) of chromosomal DNA digested withXbaI or NotI endonucleases. The usefulness of the methods for molecular differentiation of theseisolates was evaluated using computer program analysis of the obtained DNA profiles withmanually made corrections as well as with the Simpson’s index of diversity. The resultsindicated that PFGE with the XbaI restriction endonuclease, possessing the index of diversity D= 1.00, was the most powerful tool to reveal genetic relationship among E. coli O157 strainsisolated from different sources.

Key words: E. coli O157, AP-PCR, PFGE, genotypic differentiation.

Escherichia coli O157 strains are an important group of food-borne pathogenscausing severe human diseases, including bloody diarrhoea (haemorrhagic colitis -HC), haemolytic uremic syndrome (HUS) and thrombotic thrombocytopenic purpura(TTP) (1, 9, 16). E. coli O157 possess several factors implicated in the pathogenesis ofthe diseases, including Shiga toxins (Stx1 and Stx2) and the pathogenicity island calledLEE, containing the eaeA gene encoding the intimin protein, involved in the intimateattachment of bacteria to enterocytes (14, 27). Moreover, the 60-MDa plasmid,carrying the ehly gene encoding enterohaemolysin has been found in nearly all E. coliO157 strains and has been suspected to play a role in pathogenicity of infections (16,27, 29).

Ruminants, especially cattle, are known to be the most important reservoir ofE. coli O157 strains which are carried in their gastrointestinal tract and transmitted tohumans through food contaminated with fecal material (1, 16). Moreover, transmissionof E. coli O157 by person-to-person contact, drinking of raw milk, water or eatingvegetables and fruits has also been described (1).

Because of the clonal nature of E. coli O157 strains, highly sensitivemolecular biology-based subtyping methods are needed to differentiate and analysestrains recovered from the same or different sources (7, 34). Several assays have beenused for genetic investigation of E. coli O157 isolates. These tests are based either on

364

restriction fragment length polymorphism (RFLP) and pulsed-field gel electrophoresis(PFGE) (2, 5, 13, 15) or amplification of DNA sequences with PCR (AP-PCR), i.e.Random Amplified Polymorphic DNA (RAPD-PCR), Enterobacterial RepetitiveIntergenic Consensus (ERIC-PCR), BOX-PCR and sequences located between therepetitive IS3 genomic element (IS-PCR) (4, 6, 11, 25, 32, 35).

The purpose of this study was to compare selected PCR-based methods withPFGE test for the determination of molecular relationship of E. coli O157 strainsisolated from different sources.

Material and Methods

Bacterial strains. The E. coli strains used in this study are presented in Table 1.

Table 1Characteristics of E. coli O157 strains used in the study

Virulence marker geneE. coli strain Origin* Serotype

stx1 stx2 eaeA ehly344 C O157:H7 + + + +408 C O157:H7 + - + +409 C O157:H7 - + + +414 C O157:H7 - + + +420 C O157:H7 - + + +422 H O157:H- + + + +424 H O157:H- - + + +441 H O157:H7 - + + +442 H O157:H7 - + + -443 H O157:H7 + + + -451 H O157:H7 + + + +454 H O157:H7 + + + +

*E. coli strains were isolated from cattle (C) or from humans (H)

The strains were obtained from the E. coli Reference Laboratory in Dessau,Germany (numbers 344, 414, 441, 442, 443, 451), the National Institute of Hygiene inWarsaw, Poland (numbers 408, 422, 424), the Department of Food Hygiene, Universityof Warmia and Mazury, Olsztyn, Poland (number 420), and from the own straincollection (numbers 409, 454). The animal strains (one isolate from one animal) wererecovered from 3-5-month-old cattle as described previously (22). No animal contactsbetween the Polish and German cattle farms were noted. Furthermore, the humanstrains were isolated from persons who had no contacts with the animals tested in thestudy. The E. coli isolates were previously analysed for the stx1, stx2, ehly, eaeAvirulence factor genes, using single or multiplex PCR tests (18, 19).

Preparation of genomic DNA for PCR tests. Isolation and purification ofgenomic DNA was performed from E. coli bacteria grown in Luria Bertani (LB) brothusing AquaPure Genomic DNA Isolation Kit (Bio-Rad, Hercules, USA). The purityand concentration of the DNA preparations were measured spectrophotometrically at260 and 280 nm.

AP-PCR amplification conditions. All PCR primer sequences andconcentrations used in the study are shown in Table 2.

Table 2PCR primers and amplification conditions used in the study

PCRtechnique Primer code Primer sequence

Primerconcentration

(µM)Amplification conditions Reference

RAPD 1254 CCGCAGCCAA 0.594ºC 5 min,35X (36ºC 1 min, 72ºC 2 min, 94ºC 1 min),72ºC 10 min

24, 32

BOX BOXA1R CTACGGCAAGGCGACGCTGACG 0.594ºC 5 min,35X (40ºC 1 min, 72ºC 2 min, 94ºC 1 min),72ºC 10 min

6

ISIS3A

IS3B

CACTTAGCCGCGTGTCC

CGCTTAGGCCTGTGTCCA0.5

94ºC 5 min,35X (36ºC 1 min, 72ºC 2 min, 94ºC 1 min),72ºC 10 min

31

ERICERIC1R

ERIC2

ATGTAAGCTCCTGGGATTCAC

AAGTAAGTGACTGGGGTGAGCG0.2

94ºC 5 min,35X (52ºC 1 min, 72ºC 2 min, 94ºC 1 min),72ºC 10 min

11, 35

365

366

∑=

−−

−=S

jnn

NND jj

1)1(

)1(1

1

PCR reactions were performed in 50-µl reaction mixture consisting of E. coligenomic DNA (50 ng for RAPD-PCR, 10 ng for ERIC-, BOX- and IS-PCR,respectively), 1X Taq buffer (10 mM Tris-HCl [pH 8.8], 50 mM KCl, and 0.8%Nonidet P40), MgCl2 in final concentration (mM) as follows: 2.0 (RAPD-PCR), 3.0(BOX- and IS-PCR), and 5.0 (ERIC-PCR), 2 U of the Taq thermostable DNApolymerase (Fermentas, Vilnius, Lithuania), 200 µM of each dATP, dCTP, dGTP anddTTP (Fermentas), and DNase-, RNase-free, deionized water (ICN Biomedicals, CostaMesa, USA). DNA amplification was carried out in a thermal cycler (PTC-100, MJResearch, Watertown, USA), using the program conditions as described in Table 2.The PCR products were visualized by standard gel electrophoresis in a 2% agarose gel(Type I, Low EEO, Sigma, St. Louis, USA) stained with ethidium bromide (5 µg/ml)for 1 min. The gels were photographed under UV light using the Gel Doc 2000documentation system (Bio-Rad, Hercules, USA). All AP-PCR analyses wereperformed several times with slightly different amplification conditions until the bandprofiles were stable and highly reproducible. The fingerprints showed in this studybased on at least three independent amplifications which provided the same AP-PCRpictures.

Pulsed-field gel electrophoresis (PFGE). Bacterial strains were grown in 10ml of LB broth at 37ºC for 18 h with 150 rpm. After centrifugation at 10 000×g for 5min at 4ºC, the bacterial pellet was washed with 10 mM Tris-HCl and resuspended to afinal concentration of 2×109 bacterial cells/ml. Equal volumes of the bacterialsuspension and 2% agarose (low melting point; Sigma), pre-warmed at 55ºC, weremixed and immediately added into block formers (Bio-Rad). The cell lysis and DNAdigestion were carried out essentially as described by Watanabe et al. (33). Tworestriction endonucleases were used: XbaI (40 U) or NotI (50 U) (Fermentas),according to the manufacturer’s instruction. PFGE was performed with the CHEF DRII system (Bio-Rad) in 0.5X TBE at 14ºC for 23 h of a constant voltage of 200 V andthe pulse time from 5 to 40 s. After electrophoresis, the gels were stained with ethidiumbromide (5 µg/ml) for 25 min, rinsed in distilled water for 5 min and photographedXQGHU�89�OLJKW�XVLQJ�WKH�*HO�'RF������GRFXPHQWDWLRQ�V\VWHP��$����NE� ��3)*(�'1$marker (Sigma) was included into each agarose run.

Data analysis. Gel images were scanned and analysed using the MolecularAnalyst Fingerprinting Plus Software (Bio-Rad). Dendrograms were created from amatrix of band matching using the Dice coefficient and the unweighted pair groupmethod with arithmetic mean (UPGMA) analysis. Some DNA bands, mistakenlyidentified as positive by the computer software were deleted. Moreover, some PCRamplicons not recognized automatically by the computer were manually added for theanalysis. The isolates sharing less then 90% of the similarity were regarded as different(30). The discriminating ability of the tests used was determined by the Simpson’sindex of diversity according to the following equation:

where N is the total number of strains in the sample population, s is the total number oftypes being described, and nj is the number of strains belonging to the jth type (12).

367

Results

Arbitrary Primer PCR analyses. RAPD-PCR amplification performed withthe primer 1254 generated 5 to 8 weaker or stronger DNA bands with each of the E.coli isolates. The UPGMA analysis of the fingerprints revealed that 12 strains testedcould be classified into 4 clonal groups (Fig. 1). The RAPD-PCR method possessed thediscriminatory index D = 0.79.

344 408 409 414 420 422 424 441 442 443 451 454

Fig. 1. Agarose gel electrophoresis showing RAPD-PCR fingerprintings and thedendrogram outlining the clonal relationship of 12 E. coli O157 strains performed withUPGMA analysis and Dice’s coefficient.

100908070

408414420451409422443441344424442454

368

The genetic relationship among E. coli strains tested with BOX-PCRdetermined with the single BOXA1R primer resulted in 8 to 10 DNA amplicons. Thestrains were then classified into 7 clonal types with genetic similarity less than 90%(Fig. 2). The equation of Simpson’s index of diversity showed that BOX-PCR used inthis study had the discrimination D = 0.77.

344 408 409 414 420 422 424 441 442 443 451 454

Fig. 2. Agarose gel electrophoresis showing BOX-PCR fingerprintings and thedendrogram outlining the clonal relationship of 12 E. coli O157 strains performed withUPGMA analysis and Dice’s coefficient.

1009080706050

344409424442451422441443414420454408

369

The genetic analysis of the E. coli O157 isolates determined by IS-PCRresulted in 3 to 10 DNA amplicons that could be clustered into 9 different clonalgroups (Fig. 3). The index of diversity was determined as D = 0.91.

344 408 409 414 420 422 424 441 442 443 451 454

Fig. 3. Agarose gel electrophoresis showing IS-PCR fingerprintings and thedendrogram outlining the clonal relationship of 12 E. coli O157 strains performed withUPGMA analysis and Dice’s coefficient.

100908070605040

422424442443420451441454409414344408

370

Further investigation of the genetic relationships among 12 E. coli O157strains tested performed with ERIC-PCR resulted in 10 different fingerprints which hadthe degree of similarity less than 90% (Fig. 4). In this test the Simpson’s index wascalculated as 0.97.

344 408 409 414 420 422 424 441 442 443 451 454

Fig. 4. Agarose gel electrophoresis showing ERIC-PCR fingerprintings and thedendrogram outlining the clonal relationship of 12 E. coli O157 strains performed withUPGMA analysis and Dice’s coefficient.

100908070605040

344414409408441443442420424422451454

371

PFGE analysis. Using the XbaI restriction endonuclease, 12 closely relatedPFGE profiles were observed among 12 E. coli O157 strains isolated from humans andcattle (Fig. 5).

M 344 408 409 414 420 422 M 424 441 442 443 451 454 M

Fig. 5. PFGE patterns of XbaI-digested DNA of E. coli O157 strains and thedendrogram outlining the clonal relationship of these isolates generated using UPGMAanalysis and Dice’s coefficient. Lane M - molecular VL]H�VWDQGDUG�����NE� ��'1$��

100908070605040

443454441451344408414424420409422442

372

UPGMA analysis with Dice coefficient and Simpson’s index of diversityshowed that PFGE analysis performed with XbaI had the highest discrimination of D =1.00. With the NotI endonuclease, DNA of only 9 E. coli O157 strains was digestedwhereas 3 isolates (i.e. 408, 443, 454) could not be differentiated with this enzyme.Moreover, digestion with NotI generated only two DNA band types with a very lowindex of diversity D = 0.22 (data not shown).

Discussion

Several methods have been used for genetic differentiation of E. coli O157strains. Arbitrary amplification (AP-PCR) of repeated DNA sequences in E. coli strainshas been reported as a useful method for genetic characterization of these bacteria.There are several variants of this technique, such as RAPD-PCR, ERIC-PCR, BOX-PCR, and IS-PCR. It has been shown that all of them enable the generation of DNAfingerprintings which discriminate E. coli bacterial strains of the same serotype (4, 6,11, 21, 23, 25, 32, 35).

In the present study four arbitrary amplification techniques for typing of E.coli O157 isolates were compared. The results indicate that among them, ERIC-PCRtest had the highest discriminatory capacity as calculated by Simpson’s index ofdiversity (0.97). On the other hand, Osek (20) previously showed that the DNAfingerprinting profiles received with BOX and ERIC primers were more effective forgrouping of E. coli strains isolated from pigs with diarrhoea than those obtained withRAPD and IS primers. In another study of Osek (21) performed with E. coli O157:H7-

isolates recovered from pigs, the genetic relatedness of the strains was evaluated withthe BOX-PCR method. The results revealed that the isolates obtained from differentfarms and possessing the same virulence marker genes were closely related or identicalas tested with the UPGMA analysis. However, the E. coli O157 strains testedpreviously by Osek (21) were not the same as analysed in the present study.

One of the major advantages of the AP-PCR technique is that it is fast and canbe used without previous knowledge of the nucleotide sequence of the bacterial targetDNA. On the other hand, this method has been criticized due to lack of reproducibility(3, 10, 26). However, under the PCR conditions and with the oligonucleotide primersused in the present study, the stable fingerprints were obtained.

PFGE is widely used as a molecular subtyping method of E. coli O157 strains,due to its high discriminatory power and good reproducibility. This method is based onthe analysis of restriction endonuclease-digested genomic DNA fragments separated bypulsed-field gel electrophoresis (PFGE) (2, 5, 15, 17, 19, 28). The UPGMA analysiswith Dice coefficient and Simpson’s index of diversity showed that PFGE with XbaIdigestion possessed the index of discrimination D = 1.00 and therefore it was the mostpowerful tool to reveal genetic differences among E. coli O157 strains analysed in thepresent study that were isolated from cattle and humans. Davis et al. (5) found that inthe absence of epidemiological data, similarity coefficients from a single-enzymePFGE are poor measures of relatedness of E. coli O157 isolates and even six or morerestriction endonucleases are needed to provide a reasonable estimate using thisdiscriminating method. Moreover, a poor correlation between similarities derived fromthe PFGE and randomly amplified DNA tests for E. coli O157 was also observed (28).Other investigators have found that combination of results obtained with several typing

373

methods, including AP-PCR and PGFE, improve the discriminating power of themolecular differentiation tests used (8, 13).

In conclusion, the results obtained in the present study demonstrated thatPFGE performed with the XbaI restriction endonuclease is a simple, powerfulmolecular tool to reveal genetic differences and clonal relationship among E. coli O157strains of different origin and shows discriminating advantage over the four other AP-PCR methods used.

Acknowledgments: This paper was granted by the Open Society InstituteZUG Foundation, Project Code: 17013 (Log-in ID: 40004053).

References

1. Armstrong G.L., Hollingsworth J., Morris J.G.: Emerging foodborne pathogens:Escherichia coli O157:H7 as a model of entry of a new pathogen into the foodsupply of the developed world. Epidemiol. Rev., 1996, 18, 29-51.

2. Bohm H., Karch H.: DNA fingerprinting of Escherichia coli O157:H7 strains bypulsed-field gel electrophoresis. J. Clin. Microbiol., 1992, 30, 2169-2172.

3. Caetano-Anolles G.: Amplifying DNA with arbitrary oligonucleotide primers.PCR Methods Applic., 1993, 3, 85-94.

4. Cave H., Bingen E., Elion J., Denamur E.: Differentiation of Escherichia colistrains using randomly amplified polymorphic DNA analysis. Res. Microbiol.,1994, 145, 141-150.

5. Davis M., Hancock D.D., Besser T., Call D.R.: Evaluation of pulsed-fieldelectrophoresis as a tool for determining the degree of genetic relatednessbetween strains of Escherichia coli O157: H7. J. Clin. Microbiol., 2003, 41,1843-1849.

6. Dombek P.E., Johnson L.A.K., Zimmerley S.T., Sadowsky M.J.: Use ofrepetitive DNA sequences and the PCR to differentiate Escherichia coli isolatesfrom human and animal sources. Appl. Environ. Microbiol., 2000, 66, 2572-2577.

7. Feng P., Lampel K.A., Karch H., Whittam T.S.: Genotypic and phenotypicchanges in the emergence of Escherichia coli O157:H7. J. Infect. Dis., 1998, 177,1750-1753.

8. Griff K., Karch H., Schneider C., Daschner F.D., Beutin L., Cheasty T., Smith H.,Rowe B., Dierich M.P., Allerberger F.: Comparative study of five differenttechniques for epidemiological typing of Escherichia coli O157. Diagn.Microbiol. Infect. Dis., 1998, 32, 165-176.

9. Griffin P.M., Tauxe R.V.: The epidemiology of infections caused by Escherichiacoli O157:H7, other enterohemorrhagic E. coli, and the associated hemolyticuremic syndrome. Epidemiol. Rev., 1991, 13, 60-98.

10. Hopkins K.L., Hilton A.C.: Optimization of random amplification ofpolymorphic DNA analysis for molecular subtyping of Escherichia coli O157.Lett. Appl. Microbiol., 2001, 32, 126-130.

11. Hulton C.S.J., Higgins C.F., Sharp P.M.: ERIC sequences: a novel family ofrepetitive elements in the genomes of Escherichia coli, Salmonella typhimuriumand other enterobacteria. Mol. Microbiol., 1991, 5, 825-834.

374

12. Hunter P.R., Gaston M.A.: Numerical index of the discriminatory ability oftyping systems: an application of Simpson’s index of diversity. J. Clin.Microbiol., 1988, 26, 2465-2466.

13. Izumiya H., Terjima J., Wada A., Inagaki Y., Itoh K., Tamura K., Watanabe H.:Molecular typing of enterohemorrhagic Escherichia coli O157:H7 isolates inJapan by using pulsed-field electrophoresis. J. Clin. Microbiol., 1997, 35, 1675-1680.

14. Jerse A.E., Yu J., Tall B.D., Kaper J.B.: A genetic locus of enteropathogenicEscherichia coli necessary for the production of attaching and effacing lesions ontissue culture cells. Proc. Natl. Acad. Sci. USA, 1990, 87, 7839-7843.

15. Meng J., Zhao S., Zhao T., Doyle M.P.: Molecular characterisation ofEscherichia coli O157:H7 isolated by pulsed-field electrophoresis and plasmidDNA analysis. J. Med. Microbiol., 1995, 42, 256-263.

16. Nataro J.P., Kaper J.B.: Diarrheagenic Escherichia coli. Clin. Microbiol. Rev.,1998, 11, 142-201.

17. Osek J.: Clonal analysis of Escherichia coli strains isolated from pigs with post-weaning diarrhea by pulsed-field gel electrophoresis. FEMS Microbiol. Lett.,2000, 186, 327-331.

18. Osek J.: Genetic identification of virulence markers in Shiga toxin-producingEscherichia coli (STEC) strains. Bull. Vet. Inst. Pulawy, 2001, 45, 37-42.

19. Osek J.: Molecular characterisation of Shiga toxin-producing Escherichia coliO157 strains isolated in Poland. Res. Vet. Sci., 2001, 70, 175-177.

20. Osek J.: A comparision of methods based on PCR amplification of repeated DNAsequences (AP-PCR) for genotypic analysis of Escherichia coli strains. Med.Wet., 2002, 58, 361-364.

21. Osek J.: Genetic relatedness of Escherichia coli O157:H- strains isolated frompigs determined by random amplification of BOX DNA sequences. Bull. Vet.Inst. Pulawy, 2002, 46, 3-10.

22. Osek J., Gallien P., Protz D.: Characterization of shiga-producing Escherichiacoli strains isolated from calves in Poland. Comp. Immunol. Microbiol. Infect.Dis., 2000, 23, 267-276.

23. Osek J., Gallien P., Timm M.: Molecular characterization and clonal analysis ofstrains of Escherichia coli O157 group isolated from different sources. Med.Wet., 2002, 58, 590-593.

24. Pacheco A.B.F., Guth B.E.C., De Almeida D.F., Ferreira L.C.S.: Characterizationof enterotoxigenic Escherichia coli by random amplification of polymorphicDNA. Res. Microbiol., 1996, 147, 175-182.

25. Pacheco A.B.F., Guth B.E.C., Soares K.C.C., Nishimura L., De Almeida D.F.,Ferreira L.C.S.: Random amplification of polymorphic DNA reveals serotype-specific clonal clusters among enterotoxigenic Escherichia coli strains isolatedfrom humans. J. Clin. Microbiol., 1997, 35, 1521-1525.

26. Parker J.D., Rabinovitch P.S., Burmer G.C.: Targeted gene walking polymerasechain reaction. Nucl. Acids Res., 1991, 19, 3055-3060.

27. Paton J.C., Paton A.W.: Pathogenesis and diagnosis of Shiga toxin-producingEscherichia coli infections. Clin. Microbiol. Rev., 1998, 11, 450-479.

28. Radu S., Ling O.W., Rusul G., Karim M.I.A., Nischibuchi M.: Detection ofEscherichia coli O157:H7 by multiplex PCR and their characterization byplasmid profiling, antimicrobial resistance, RAPD and PFGE analyses. J.Microbiol. Meth., 2001, 46, 131-139.

375

29. Schmidt H., Beutin L., Karch H.: Molecular analysis of the plasmid encodedhaemolysin of Escherichia coli O157:H7 strain EDL933. Infect. Immun., 1995,63, 1055-1061.

30. Tenover F.C., Arbeit R.D., Goering R.V., Mickelsen P.A., Murray B.E., PersingD.H., Swaminathan B.: Interpreting chromosomal DNA restriction patternsproduced by pulsed-field electrophoresis: criteria for bacterial strain typing. J.Clin. Microbiol., 1995, 33, 2233-2239.

31. Thompson C.J., Daly C., Barrett T.J., Getchell J.P., Gilchrist M.J.R., LoeffelholzM.J.: Insertion element IS3-based PCR method for subtyping Escherichia coliO157:H7. J. Clin. Microbiol., 1998, 36, 1180-1184.

32. Versalovic J., Koeuth T., Lupski J.R.: Distribution of repetitive DNA sequencesin Eubacteria and application to fingerprinting of bacterial genomes. NucleicAcids Res., 1991, 19, 6823-6831.

33. Watanabe H., Terajima J., Izumiya H., Iyoda S.: Molecular typing methods forSTEC. Meth. Mol. Med., 2002, 73, 55-65.

34. Whittam T.S., Wachsmuth I.K., Wilson R.A.: Genetic evidence of clonal descentof Escherichia coli O157:H7 associated with hemorrhagic colitis and hemolyticuremic syndrome. J. Infect. Dis., 1998, 157, 1124-1133.

35. Woods C.R., Versalovic J., Koeuth T., Lupski J.R.: Whole-cell repetitivesequence-based polymerase chain reaction allows rapid assessment of clonalrelationships of bacterial isolates. J. Clin. Microbiol., 1993, 31, 1927-1931.