Research in Veterinary Science 92 (2012) 18–23

Contents lists available at ScienceDirect

Research in Veterinary Science

journal homepage: www.elsevier .com/locate / rvsc

Determination of flagellar types by PCR-RFLP analysis of enteropathogenicEscherichia coli (EPEC) and Shiga toxin-producing E. coli (STEC) strains isolatedfrom animals in São Paulo, Brazil

Claudia de Oliveira Ayala a, Ana Carolina Ramos Moreno b, Marina Baquerizo Martinez b,Ylanna Kelner Burgos a, Antonio Fernando Pestana de Castro a,⇑, Silvia Yumi Bando c

a Departamento de Microbiologia, Instituto de Ciências Biomédicas, Universidade de São Paulo, Av Professor Lineu Prestes, 1374, Dip. Bact. (London 69-70), 05508-000 São Paulo,SP, Brazilb Departamento de Análises Clínicas e Toxicológicas, Faculdade de Ciências Farmacêuticas, Universidade de São Paulo, Brazilc Departamento de Pediatria, Faculdade de Medicina, Universidade de São Paulo, Brazil

a r t i c l e i n f o

Article history:Received 27 May 2010Accepted 25 October 2010

Keywords:Domestic animalsEscherichia coliMolecular typingH antigenFlagellinPCR-RFLP

0034-5288/$ - see front matter � 2010 Elsevier Ltd. Adoi:10.1016/j.rvsc.2010.10.025

⇑ Corresponding author. Tel.: +55 11 3091 7298; faE-mail addresses: apestana@usp.br, apestana@usp.

a b s t r a c t

This study evaluated the polymerase chain reaction- restriction fragment length polymorphism (PCR-RFLP) analysis of fliC for typing flagella antigen (H) of Shiga toxin-producing Escherichia coli (STEC) andenteropathogenic E. coli (EPEC) strains isolated from different animals. The molecular typing of the H typewas efficient in the determination of 93 (85%) strains. Two nonmotile (H-) E. coli strains showed a PCR-RFLP electrophoretic profile that did not match known H type patterns. The fliC nucleotide sequence ofstrains B2N and 4a revealed a nucleotide substitution at the restriction site and a nucleotide insertionthat generated a stop codon, respectively. The results of this study showed that PCR-RFLP analysis of fliCis faster, less laborious and as efficient for the determination of H type E. coli isolated from animals, com-pared to serotyping and that it is useful in determining H type in nonmotile strains and strains expressingnon-reactive H antigens. Moreover, the fliC sequence of strain B2N suggests that we could have found anew flagellin antigen type.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction E. coli. The variability of the H antigen is found within the flagellar

Typical enteropathogenic Escherichia coli (tEPEC) strains arenormally associated with diarrhoea in children under one (Nataroand Kaper, 1998; Trabulsi et al., 2002). However, strains of thispathotype are nowadays less often the cause of diarrhoea through-out the world, and atypical EPEC (aEPEC) strains are overwhelm-ingly responsible for EPEC-mediated diarrhoeal infections inchildren and adults (Franzolin et al., 2005). In addition, differentanimals can be carriers of aEPEC, though not all of them belongto the human serotypes (Trabulsi et al., 2002).

Another important pathotype of E. coli is Shiga toxin-producingEscherichia coli (STEC) strains. These strains cause diarrhoea andare also associated with hemorrhagic colitis (HC) and haemolyticuremic syndrome (HUS). STEC strains are characterized by the pro-duction of two powerful toxins similar to the Shigella dysenteriaetoxin: the Shiga toxins 1 and 2 (Stx1 and Stx2), encoded byprophages within their chromosomes.

The combination of O and H antigens defines the serotype andcan be very useful for the determination of different pathotypes of

ll rights reserved.

x: +55 11 3091 7354.br (A.F. Pestana de Castro).

filament (Iino et al., 1988). The flagellar antigen of E. coli corre-sponding to the filament is formed by the polymerization of a singleprotein subunit, flagellin, encoded by the fliC gene (Macnab, 1992).

Diarrheagenic E. coli strains can be analysed for H type by thepolymerase chain reaction-restriction fragment length polymor-phism (PCR-RFLP) method. Fields et al. (1997) showed that STECstrains of the O157:H- serotype had the same restriction patternas STEC O157:H7 strains. Botelho et al. (2003) also demonstratedthat PCR-RFLP is a satisfactory typing method for H type determi-nation in nonmotile E. coli strains.

2. Material and methods

2.1. Strains studied

E. coli isolates (n = 109 strains) originating from cattle (Leomilet al., 2003, 2005; Aidar-Ugrinovich et al., 2007), rabbits (Penteadoet al., 2002), monkeys (Carvalho et al., 2003), sheep (Vettoratoet al., 2003) and dogs (Nakazato et al., 2004) were used. Themajority of these strains (85 strains) were previously classified asEPEC (57 strains) or STEC (28 strains) based upon virulencemarkers.

Table 1Strains used in this study, diarrheagenic pathotypes, serotype determined by serology and fliC gene restriction analysis.

Strain Serotyping- byseroagglutination

PCR foreae

PCR forstx1/2

fliC fragmentsize (bp)

Restriction fragments (bp) DECpathotype

H antigen determinedby PCR-RFLP

151MB-13a O7:H10 � � 1324 519/321/303 – O7:H10Mora-1a O132:H51 � � 1940 350/304/257/238/194/140/135/99 – O132:H51Mora-5a O132:H51 � � 1955 352/307/259/241/198/140/132/101 – O132:H516 CP(8)-1a O10:H51 � +/� NAi – STEC –Ficha-4a O127:H40 + � 1577 291/247/154 EPEC O127:H40129MB-2a O123:H2 + +/+ NA – STEC –Elga(5)-1a O123:H2 � +/� NA – STEC –Florinda-19a O80:H26 � � 1473 523/310/134 – O80:H2673CC-11a O86:H36 � � 1939 684/549/276/200/139/98 – O86:H3665CC-3a O123:H11 + � 1553 586/272/173/148/113 EPEC O123:H1134CC-3a O123:H11 + � 1551 531/506/346/207/127 EPEC O123:H11001HV-2a O8:H21 � �/+ 1613 374/313/291/271/169/148 STEC O8:H21001HV-5a O116:H20 � � 1831 378/299/276/202/178 – O116:H20002HV-5a O111:H19 + +/� NA – STEC –16FB-12a O55:H19 � +/� 1869 614/480/280/243/189 STEC O55:H19151MB-3a O175:H49 � �/+ 1568 432/331/309/274/209 STEC O175:H4992MB(4)-1a ONT:H8 � +/� NA – STEC –21FC-19a O7:H7 � �/+ 1579 531/506/346/207/127 STEC O7:H7N61-12a O22:H16 � +/+ NA - STEC –43CC-2a O123:HNT + � 1947 656/529/413 EPEC O123:H447J(15)-1a O119:H- � � NA – – –132MB(9)-1a O119:H- + +/+ 1594 591/434/382/127 STEC O119:H2Beleza-2a O111:H- + +/� 1520 729/320/285/152 STEC O111:H8Beleza-4a O111:H- + +/� 1553 742/322/287/153 STEC O111:H8166b O4:H1 � � 1852 752/439/294/173 – O4:H1167b O4:H1 � � 1758 766/443/304/169 – O4:H1143b,* O128:H2 + � 1429 625/460/347 EPEC O128:H2144b O128:H2 + � NA – EPEC –128b,* O132:H2 + � 1662 605/441/337 EPEC O132:H2130b,* O132:H2 + � 1490 597/433/336 EPEC O132:H2133b,* O132:H2 + � 1502 612/445/347 EPEC O132:H2237b O1:H45 � � 1738 439/388/314/215/132 – O1:H45129b O103:H19 � � 1758 690/466/303/266/200 – O103:H19Lb O103:H19 � � 1852 672/442/285/248/203 – O103:H19183b O103:H19 � � 1863 655/434/279/245/199 – O103:H19113b,* O110:H6 + � 1702 560/448/355/229/150 EPEC O110:H6119b,* O110:H6 + � 1649 529/426/331/217/152 EPEC O110:H6236b,* O153:H7 + � 1737 596/533/363/239/136 EPEC O153:H7238b O153:H7 + � 1737 589/540/362/239/137 EPEC O153:H7243b O153:H7 + � 1798 593/547/380/251/139 EPEC O153:H7160b O20:H28 � � 1747 439/348/206/171 – O20:H2821b O128:NR � � 1525 64/447/351 – O128:H238b NT:NR � � 1747 429/340/203/171 – NT:H28117b NT:NR � � 1636 749/329/285/143 – NT:H8145b NT:NR � � 1772 610/545/411 – NT:H1412b O126:H- � � NA – – –101b O41:H- � � 1393 466/343/304/237/169 – O41:H20102b O41:H- � � 1533 402/333/305/236/161 – O41:H20C13-13c ONT:H8 � +/+ 1430 712/318/287/154 STEC ONT:H8RB15-16c O174:H8 � +/+ 1467 716/316/286/153 STEC O174:H8124-3c O75:H8 � +/+ 1556 700/315/285/147 STEC O75:H8C8-8c O87:H16 � �/+ 1646 377/330/290/146 STEC O87:H16220-2c ONT:H16 � �/+ 1486 380/336/299/157 STEC ONT:H16243-18c O146:H21 � +/+ 1556 374/316/285/220/147 STEC O146:H21CA2-10c O128:H2 � +/� 1496 563/406/312/125 STEC O128:H2CA3-14c O128:H2 � +/� 1439 565/415/323/129 STEC O128:H2IT3-1c ONT:H14 � +/� 1627 617/555/413 STEC ONT:H14RB18-8c O88:H25 � � 1271 946/312 – O88:H25RB18-6c O88:H25 � � 1215 946/311 – O88:H25RB24-8c O110:H28 � � 1712 413/330/196/163 – O110:H28RB7-4c ONT:H19 � +/� 1892 609/578/273/240 STEC ONT:H19RB6-2c O178:H7 + � 1627 533/489/349/237/134 EPEC O178:H7RB6-3c O178:H7 + � 1837 536/498/349/229/125 EPEC O178:H7RB31-3c O153:H7 + � 1683 536/496/349/238/130 EPEC O153:H7RB31-5c O153:H7 + - 1837 525/489/342/227/125 EPEC O153:H7RB20-1c O5:H- � +/� 1892 625/595/272/237 STEC O5:H19123-4c O5:H- � +/� 1903 625/587/272/235 STEC O5:H19210-7c O75:H- � +/+ 1497 718/310/279/145 STEC O75:H8

(continued on next page)

Claudia de Oliveira Ayala et al. / Research in Veterinary Science 92 (2012) 18–23 19

Table 1 (continued)

Strain Serotyping- byseroagglutination

PCR foreae

PCR forstx1/2

fliC fragmentsize (bp)

Restriction fragments (bp) DECpathotype

H antigen determinedby PCR-RFLP

RB24-14c O75:H- � +/+ 1517 718/314/280/145 STEC O75:H8234-7c O16:H- � +/+ 1298 549/324/304 STEC O16:H101d* O128:H2 + � 1514 577/413/317/124 EPEC O128:H22d O49:H46 + � 1608 461/316/292/242/190 EPEC O49:H463d,* O127:H40 + � 1477 298/262/148 EPEC O127:H4012a,d,* ONT:H40 + � 1466 285/260/156 EPEC ONT:H404bd OR:H34 + � 1583 651/542/413 EPEC OR:H345d O167:H9 + � 1767 1008/303/276/163 EPEC O167:H912bd,* O26:H7 + � 1752 532/495/358/221 EPEC O26:H77ad,* O132:H31 + � 1565 530/427/378/217/152 EPEC O132:H317bd,* O132:H31 + � 1476 538/401/389/209/153 EPEC O132:H318d,* O132:H31 + � 1628 572/421/398/233 EPEC O132:H316d O49:H6 + � 1619 521/419/325/218/149 EPEC O49:H613ad,* O167:H6 + � 1603 520/417/332/224/150 EPEC O167:H613bd,* O167:H6 + � 1578 512/417/331/223/152 EPEC O167:H615ad O142:H6 + � 1619 522/408/333/220/145 EPEC O142:H615bd O142:H6 + � 1621 522/418/328/216/150 EPEC O142:H616a,d,* O142:H6 + � 1614 518/409/330/222/151 EPEC O142:H616bd,* O142:H6 + � 1595 521/415/326/222/149 EPEC O142:H617ad O139:H14 + � 1590 631/568/416 EPEC O139:H1417bd O139:H14 + � 1611 626/565/415 EPEC O139:H144a,d O33:H- + � 1636 720/579/439 EPEC O33:H34008e O98:H28 + � 1753 437/334/198/171 EPEC O98:H28HE10e ONT:H40 + � 1558 299/272/156 EPEC ONT:H40HE8e O111:H25 + � NA – EPEC –HE9e O111:H25 + 8 NA – EPEC –BIO2e O157:H16 + � NA – EPEC –BIO4e O157:H16 + � NA – EPEC –HE4e O11:H16 + � 1673 390/346/311/301/165/149 EPEC O11:H16BIO12e,* O119:H2 + � 1584 612/444/338 EPEC O119:H2SPA14e,* O119:H2 + � 1636 610/443/345 EPEC O119:H2B1e O55:H7 + � 1736 591/540/366/240/137 EPEC O55:H7B4e O125:H49 + � 1743 432/331/305/269/201/131 EPEC O125:H49C32e O4:H6 + � 1646 520/421/325/223/151 EPEC O4:H6S6e,* O142:H6 + � 1830 525/425/345/227/153 EPEC O142:H6C27- col. 3e O167:H6 + � 1830 531/426/331/217/152 EPEC O167:H6C27- col. 1e O25:H8 + � NA – EPEC –QSF4e O88:H- + � 1278 1278 EPEC O88:H5/O88:H55HE13e O15:H- + � 1339 757/402/196 EPEC O15:H52B17e O156:H- + - 1309 1016/339 EPEC O156:H25B2Ne ONT:H- + � 1331 829/470/156 EPEC ONT:H-

fH-, nonmotile strain; gNR, non-reactive flagellar antigen; hNT, O undetermined by serotyping.a strains isolated from cattle.b strains isolated from rabbits.c strains isolated from sheep.d strains isolated from monkeys.e strains isolated from dogs.f Means that strains with H- are nonmotile strains.g NR means that although harboring the H antigen (H-) it does not reacto with any of the known H antisera.h NT means that the strain did not reacted with the O antisera available.i NA, No amplification.

* strains isolated from animals with diarrhoea; strains with no asterisk were isolated from animals without diarrhoea.

20 Claudia de Oliveira Ayala et al. / Research in Veterinary Science 92 (2012) 18–23

2.2. PCR-RFLP of the flagellin gene

A single colony of each strain was cultivated in 3 mL TSB med-ium (Tryptone Soya Broth, OXOID�) and incubated for 18 h at 37 �Cwith constant shaking (200 rpm). An aliquot of 1 mL of this culturewas subjected to DNA extraction using the Easy DNA Kit (Invitro-gen™) according to the manufacturer’s instructions. After quantifi-cation, DNA was diluted to a final concentration of 10 ng/lL andused for amplification of the fliC gene. PCR was carried out as de-scribed by Fields et al. (1997). The amplicons obtained were di-gested with endonuclease RsaI as previously described by RamosMoreno et al. (2006). The restriction products were separated byhorizontal electrophoresis, and the migration patterns were ana-lysed by PhotoCapt MW Program, version 10.01, for Windows.

2.3. Sequence analysis of flagellin gene

Two PCR amplification products of fliC were sequenced as de-scribed below. The PCR products were purified using PureLinkTM

PCR Purification Kit (InvitrogenTM) and cloned into pCR�2.1-TOPOplasmid vector, according to the manufacturer’s instructions(TOPO TA Cloning� Kit for Sequencing, InvitrogenTM). The chemicaltransformation and the selection of tranformants were performedaccording to the manufacturer’s instructions.

A single colony was grown in 3 mL TSB medium containingkanamycin for extraction of plasmid DNA. The sequencing reac-tions were performed with the BigDye� Terminator v3.1 CycleSequencing Kit (Applied Biosystems), primers M13F and M13Rand 10 ng plasmid DNA. The sequencing products were electropho-

Fig. 1. RFLP of fliC gene of bovine strains (lanes 3 and 4), rabbit strains (lanes 2 and6) and ovine strains (lane 5). Lane 1: Ladder, 100 bp; Lane 2: strain 21 (O128:NR,identified as O128:H2); Lane 3: strain 132MB9-1 (ONT:H-, identified as ONT:H2);Lane 4: strain Beleza 2 (O111:H-, identified as O111:H8); Lane 5: strain 123-4(O5:H-, identified as O5:H19); Lane 6: strain 101 (O41:H-, identified as O41:H20);Lane 7: Ladder, 100 bp.

Fig. 2. RFLP of fliC gene of two strains isolated from different animals showing adifferent pattern when compared to 56 H types. Lane 1: Ladder 100 bp; Lane 2: dogstrain B2N (ONT:H-); Lane 3: ovine strain 234-7 (O16:H-, identified as O16:H10);Lane 4: monkey strain 4a (O33:H-); Lane 5: Ladder 100 bp.

Claudia de Oliveira Ayala et al. / Research in Veterinary Science 92 (2012) 18–23 21

resed in a Model 3130xl Genetic Analyzer (Applied Biosystems).The sequences were assembled and edited in the SeqMan Pro modeof the LaserGene version 7.1.0 (DNASTAR, Inc. Madison, USA) soft-ware package.

Fig. 3. Nucleotide and inferred amino acids sequence alignment of fliC of strains 4a atermination codon and the arrow indicates a nucleotide insertion in the fliC sequence o

All sequences obtained were aligned and analysed with theBasic Local Alignment Search Tool (BLAST) program, available atthe National Center for Biotechnology Information website (http://blast.ncbi.nlm.nih.gov/Blast.cgi).

3. Results

3.1. Restriction analysis of flagellin gene

PCR amplicons were obtained for 95 strains (87%) studied here-in. The fragment sizes of the amplification products varied between1215 and 1955 bp. This range is similar for strains isolated from dif-ferent animal species. The restriction profiles obtained for 95strains were compared with those previously obtained for 56known H types (Ramos Moreno et al., 2006). Among the 109 strains,73 had their H type confirmed by PCR-RFLP. Twenty strains thatwere classified as H- (nonmotile) or HND (not-determined withany known H antisera) had their flagellar type identified only byPCR-RFLP (Table 1). Fig. 1 shows the restriction profiles of five Htype strains that were not characterized by serotyping, but weretyped by RFLP profiles and corresponded to H2, H8, H19 and H20(Fig. 1). Only two nonmotile strains, one isolated from dog (strainsB2 N) and another from monkey (strain 4a), displayed restrictionprofiles that did not have a corresponding profile when comparedto 56 H type profiles (Fig. 2). Therefore, the PCR products of fliC weresequenced. All STEC strains were isolated from non-diarrheic ani-mals. Conversely, 21 EPEC strains were isolated from animals withdiarrhoea or at least loose stools (Table 1). These strains includedthree from dogs, seven from rabbits and 11 from marmosets.

3.2. Analysis of nucleotide sequence of flagellin gene

The partial nucleotide sequence of flagellin gene obtained foreach strain was deposited in the GenBank database. The analysisof the nucleotide sequence of strain 4a (accession numberGU784813) showed 98.6% similarity with the fliC gene of E. coliserotype H34 (GenBank accession number AY250016).The se-quence alignment revealed the insertion of one nucleotide (cyto-sine) at position 22 of the sequence analyzed. This eventcorresponded to the loss of one restriction site and generated apremature stop codon (Fig. 3).

The nucleotide sequence of fliC obtained for strain B2N (acces-sion number GU784812) showed 76.6% similarity with the flagellargene of E. coli serotype H52 (GenBank accession numberAY250028.1). The sequence alignment revealed that the N-termi-nal and C-terminal ends are conserved but that the central regionshows nucleotide substitutions, deletions and insertions. However,the inferred amino acid sequence showed 80% identity with flagel-lin H52.

4. Discussion

The serotypes of the E. coli strains isolated from the animalsstudied are known to cause diarrhoea in humans and animals.Moura et al. (2009) showed that EPEC strains isolated from human

nd E. coli serotype H34 (GenBank accession n. AY250016). Asterisk indicates thef strain 4a.

22 Claudia de Oliveira Ayala et al. / Research in Veterinary Science 92 (2012) 18–23

and animals and of the same serotype belonged to a single clonalgroup. In addition, Bettelheim et al. (2000) isolated STEC strainsfrom sheep, belonging to three serotypes that have been associatedwith human diseases. These studies demonstrated the importanceof serotyping E. coli strains isolated from different animals, becauseanimals could serve as reservoirs of human pathogenic E. coliserotypes.

The classical method used for H antigen determination is sero-logical agglutination (SA). However, this technique is time con-suming and is performed only in reference laboratories.Moreover, it is based on cross-reactions between antigens ex-pressed on the surface of the strains studied, which may result infalse-positive results.

Many strains of E. coli have O or H antigens, or both, that do notreact with the respective available antisera (Gyles, 2006). There-fore, flagellin genes have been used for the detection and identifi-cation of bacteria. The PCR-RFLP technique employed to obtainelectrophoretic profiles of the fliC gene has been shown to be veryuseful for the determination of H types of E. coli (Fields et al., 1997;Botelho et al., 2003; Amhaz et al., 2004; Beutin et al., 2004; RamosMoreno et al., 2006).

In this study, we evaluated the efficiency of this method todetermine the H types of EPEC and STEC strains isolated from dif-ferent animals, including nonmotile strains. Comparing our resultswith those in the literature (Fields et al., 1997; Zhang et al., 2000;Botelho et al., 2003; Ramos Moreno et al., 2006), we observed anaccuracy in the restriction patterns of the different fliC alleles,independent of the diarrheagenic pathotype of E. coli.

Nonmotile strains generally possess the structural gene of fla-gellin, but are not able to produce functional flagella (Macnab,1992). Here, we confirmed that 21 nonmotile strains show fliCamplification products, with the exception of two strains, and 17strains had restriction profiles corresponding to 10 different Htypes.

Twelve motile strains did not amplify fliC. This fact may be dueto inadequate homology of the primers. Wang et al. (2003) showedthat the prevention of flagellin PCR amplification was a conse-quence of nucleotide insertion and deletion in the primer position.

Two strains analyzed here did not show a restriction profilewith known H types. Botelho et al. (2003) also observed that eightstrains displayed unknown patterns and suggested that these ser-otypes are rarely isolated or even non-existent in Brazil. Moreover,Winstanley and Morgan (1997) reported that the variability inRFLP patterns must be interpreted with caution, because mutationsof a single nucleotide may be responsible for such differences. Hor-izontal transfer of the fliC gene may contribute to the diversity ofthe flagellar serotypes due to recombination among strains ofE. coli. The host’s immune system is the main factor in the selectionand continuous evolution of the flagellar antigens (Reid et al.,1999).

For these reasons, we performed the nucleotide sequencing offliC of strains 4a and B2N. The nucleotide sequence obtained forstrain 4a showed high identity with the fliC gene which encodesa flagellin corresponding to H34. However, the sequence showeda nucleotide substitution at one RsaI restriction site (position1609), thus confirming the discrepant profile obtained for theH34 type. In addition, this sequence also has a nucleotide insertionat position 22 and generated a premature stop codon. This resultindicates the translation of a truncated protein. The expression ofdefective flagellin could explain the failure to identify H antigenby serotyping.

The sequencing of fliC of strain B2N showed similarity with thegene coding for the flagellin typed as H52. The sequence showsconserved C-terminal and N-terminal regions, but a variablecentral region. The N-terminal and C-terminal portions of the fla-gellin are essential for the polymerization and secretion of flagellin

molecules and these regions are highly conserved (Hyman andTrachtenberg, 1991). On the other hand, the variability in the cen-tral region has been found in different pathogenic E coli serotypes.There is considerable evidence supporting the notion that lateralgene transfer and recombination of foreign DNA are major sourcesof variation, creating new fliC alleles and generating antigenicdiversity (Reid et al., 1999). Alternative alterations of the flagellarphenotype of Escherichia coli strain P12b, the standard type for fla-gellar antigen H17, possessing a new non-fliC flagellin gene flnAand a possible loss of the original flagellar phenotype and genotypein the course of subculturing on semisolid media has been recentlyreported (Ratiner et al., 2010). Briefly, these authors concluded thatthe original strain of E. coli P12b possesses two distinct flagellingenes, fliCH4 (cryptic) and flnAH17 (expressed). H4 and H17 variantscan be distinguished by means of PCR using common fliC primers.According to the authors, H4 cells arising spontaneously in the ori-ginal culture of P12b can fully replace H17 cells in the course ofsubculturing this strain on some semisolid media in motility tubes,thus indicating the possible loss of the original (H17) P12b strainduring laboratory maintenance. The unexpected diversity betweenB12N and the H52 fliC genes could be explained by a similar phe-nomenon. Conversely, one cannot exclude the possibility that theB12N fliC gene sequence analyzed here probably corresponds to anew flagellar antigen. In both cases, additional studies arerequired.

In conclusion, this study demonstrated that the PCR-RFLP isuseful for typing H types of EPEC and STEC strains isolated fromanimals. Moreover, this molecular technique is faster than tradi-tional serotyping in the identification of E. coli H antigens, takingonly two days to perform, when compared to the classical serotyp-ing method which takes seven or more days and is efficient in typ-ing nonmotile strains. In addition, some tiny or significantnucleotide modification can be detected as it was also demon-strated for the fliC sequence of strain 4a that was equal to the se-quence of H34, but showed a nucleotide substitution at one RsaIrestriction site (position 1609), thus confirming the discrepant pro-file obtained for H34 type. In addition, this sequence also has anucleotide insertion at position 22 and a consequent prematurestop codon. This result indicates the translation of a truncated pro-tein. The expression of defective flagellin could explain the failureto identify H antigen by serotyping.

Acknowledgement

FAPESP, CNPq.

References

Aidar-Ugrinovich, L., Blanco, J., Blanco, M., Blanco, J.E., Leomil, L., Dahbi, G., Mora, A.,Onuma, D.L., Silveira, W.D., Pestana de Castro, A.F., 2007. Serotypes, virulencegenes, and intimin types of Shiga toxin-producing Escherichia coli (STEC) andenteropathogenic E Coli (EPEC) isolated from calves in São Paulo, Brazil.International Journal of Food Microbiology 115, 297–306.

Amhaz, J.M.K., Andrade, A., Bando, S.Y., Tanaka, T.L., Moreira-Filho, C.A., Martinez,M.B., 2004. Molecular typing and phylogenetic analysis of EnteroinvasiveEscherichia coli using the fliC gene sequence. FEMS Microbiology Letters 235,259–264.

Bettelheim, K.A., Bensink, J.C., Tambunam, H.S., 2000. Serotypes of Verocytotoxin-producing (Shiga toxin-producing) Escherichia coli isolated from healthy sheep.Comparative Immunology, Microbiology and Infectious Diseases 23, 1–7.

Beutin, L., Krause, G., Zimmermann, S., Kaulfuss, S., Gleier, K., 2004. Characterizationof Shiga toxin-producing Escherichia coli strains isolated from human patients inGermany over a 3-year period. Journal of Clinical Microbiology 42, 1099–1108.

Botelho, B.A., Bando, S.Y., Trabulsi, L.R., Moreira-Filho, C.A., 2003. Identification ofEPEC and non-EPEC serotypes in the EPEC O serogroups by PCR-RFLP analysis ofthe fliC gene. Journal of Microbiology Methods 54, 87–93.

Carvalho, V.M., Gyles, C.L., Ziebell, K., Ribeiro, M.A., Catão-Dias, J.L., Sinhorini, I.L.,Otman, J., Keller, R., Trabulsi, L.R., Pestana de Castro, A.F., 2003. Characterizationof monkey Enteropathogenic Escherichia coli (EPEC) and human typical andatypical EPEC serotype isolates from neotropical nonhuman primates. Journal ofClinical Microbiology 41, 1225–1234.

Claudia de Oliveira Ayala et al. / Research in Veterinary Science 92 (2012) 18–23 23

Fields, P.I., Blom, K., Hughes, H.J., Helsel, L.O., Feng, P., Swaminathan, B., 1997.Molecular characterization of the gene encoding H antigen in Escherichia coliand a development of a PCR-restriction fragment length polymorphism test foridentification of E. coli O157:H7 and O157:NM. Journal of Clinical Microbiology35, 1066–1070.

Franzolin, M.R., Alves, R.C., Keller, R., Gomes, T.A., Beutin, L., Barreto, M.L., Milroy, C.,Strina, A., Ribeiro, H., Trabulsi, L.R., 2005. Prevalence of diarrheagenicEscherichia coli in children with diarrhea in Salvador, Bahia, Brazil. Memóriasdo Instituto Oswaldo Cruz 100, 359–363.

Gyles, C.L., 2006. Shiga toxin-producing Escherichia coli: An overview. Journal ofAnimal Science 85, E45–E62.

Hyman, H.C., Trachtenberg, S., 1991. Point mutations that lock Salmonellatyphimurium flagellar filaments in the straight right-handed and left-handedforms and their relation to filament superhelicity. Molecular Biology 220, 79–88.

Iino, T., Komeda, Y., Kutsukake, K., Macnab, R.M., Matsumura, P., Parkinson, J.S.,Simon, M.I., Yamaguchi, S., 1988. New unified nomenclature for the flagellargenes of Escherichia coli and Salmonella typhimurium. Microbiology Reviews 52,533–535.

Leomil, L., Aidar-Ugrinovich, L., Guth, B.E.C., Irino, K., Vettorato, M.P., Onuma, D.L.,Pestana de Castro, A.F., 2003. Frequency of Shiga toxin-producing Escherichiacoli (STEC) isolates among diarrheic and non-diarrheic calves in Brazil.Veterinary Microbiology 97, 103–109.

Leomil, L., Pestana de Castro, A.F., Krause, G., Schmidt, H., Beutin, L., 2005.Characterization of two major groups of diarrheagenic Escherichia coli O26strains which are globally spread in human patients and domestic animals ofdifferent species. FEMS Microbiology Letters 249, 335–342.

Macnab, R.M., 1992. Genetics and biogenesis of bacterial flagella. Annual ReviewGenetics 26, 129–156.

Moura, R.A., Sircili, M.P., Leomil, L., Matté, M.H., Trabulsi, L.R., Elias, W.P., Irino, K.,Pestana de Castro, A.F., 2009. Clonal relationship among atypicalenteropathogenic Escherichia coli strains isolated from different animalspecies and humans. Applied Environmental Microbiology 75, 7399–7409.

Nakazato, G., Gyles, C., Ziebell, K., Keller, R., Trabulsi, L.R., Gomes, T.A., Irino, K.,da Silveira, W.D., Pestana de Castro, A.F., 2004. Attaching and effacing

Escherichia coli isolated from dogs in Brazil: characteristics and serotypicrelationship to human Enteropathogenic E. Coli (EPEC). VeterinaryMicrobiology 101, 269–277.

Nataro, J.P., Kaper, J.B., 1998. Diarrheagenic Escherichia coli. Clinical MicrobiologicalReviews 11, 142–201.

Penteado, A.S., Aidar-Ugrinovich, L., Blanco, J., Blanco, M., Blanco, J.E., Mora, A.,Andrade, J.R., Correa, S.S., Pestana de Castro, A.F., 2002. Serobiotypes andvirulence genes of Escherichia coli strains isolated from diarrheic and healthyrabbits in Brazil. Veterinary Microbiology 89, 41–51.

Ramos Moreno, A.C., Guth, B.E.C., Martinez, M.B., 2006. Can the fliC PRC-RestrictionLength Polymorphism Technique replace classic serotyping methods forcharacterization of the H antigen of Enterotoxigenic Escherichia coli strains?Journal of Clinical Microbiology 44, 1453–1458.

Ratiner, Y.A., Sihvonen, L.M., Liu, Y., Wang, L., Siitonen, A., 2010. Alteration offlagellar phenotype of Escherichia coli strain P12b, the standard type strain forflagella antigen H17, possessing a new non-fliC flagellin gene flnA, and possibleloss of original flagellar phenotype and genotype in the course of subculturingthrough semisolid media. Archives Microbiology 192, 267–278.

Reid, S.D., Selander, R.K., Whittam, T.S., 1999. Sequence diversity of flagellin (fliC)alleles in pathogenic Escherichia coli. Journal Bacteriology 181, 153–160.

Trabulsi, L.R., Keller, R., Gomes, T.A.T., 2002. Typical and atypical EnteropathogenicEscherichia coli. Emerging Infectious Diseases 8, 508–513.

Vettorato, M.P., Leomil, L., Guth, B.E.C., Irino, K., Pestana de Castro, A.F., 2003.Properties of Shiga toxin-producing Escherichia coli (STEC) isolates from sheepin the State of São Paulo, Brazil. Veterinary Microbiology 95, 103–109.

Wang, L., Rothemund, D., Curd, H., Reeves, P.R., 2003. Species-wide variation in theEscherichia coli flagellin. Journal Bacteriology 185, 2936–2943.

Winstanley, C., Morgan, J.A.W., 1997. The bacterial flagellin gene as a biomarker fordetection, population genetics and epidemiological analysis. Microbiology 143,3071–3084.

Zhang, W., Bielaszewska, M., Bockemuhl, J., Scheutz, F., Karch, H., 2000. Molecularanalysis of H antigens reveals that human diarrheagenic Escherichia coli O26strains that carry the eae gene belong to the H11 clonal complex. Journal ofClinical Microbiolology 38, 2989–2993.