OUP user menu

Single multiplex assay to identify simultaneously enteropathogenic, enteroaggregative, enterotoxigenic, enteroinvasive and Shiga toxin-producing Escherichia coli strains in Brazilian children

Katia R. S. Aranda, Sandra H. Fabbricotti, Ulysses Fagundes-Neto, Isabel C. A. Scaletsky
DOI: http://dx.doi.org/10.1111/j.1574-6968.2006.00580.x 145-150 First published online: 1 February 2007


A multiplex PCR to differentiate typical and atypical enteropathogenic Escherichia coli (EPEC), enteroaggregative E. coli (EAEC), enterotoxigenic (ETEC), enteroinvasive E. coli (EIEC) and Shiga toxin-producing E. coli (STEC) strains was developed and evaluated. The targets selected for each group were eae and bfpA for EPEC, aggR for EAEC, elt and est for ETEC, ipaH for EIEC and stx for STEC isolates. This PCR was specific and sensitive for rapid detection of target isolates in stools. Among 79 children with acute diarrhea, this technique identified 13 (16.4%) with atypical EPEC, four (5%) with EAEC, three (3.8%) with typical EPEC, one (1.3%) with ETEC and one (1.3%) with EIEC.

  • multiplex PCR
  • Escherichia coli
  • diarrhea


Diarrhea remains an important public health problem for children of low socio-economic status in Brazil (Trabulsi et al., 1985). In an etiologic study of acute diarrhea in children less than 2 years of age from different regions of Brazil, Scaletsky (2002a, b) showed that diarrheagenic Escherichia coli organisms have an important role as a cause of enteric diseases. However, these pathogens are probably underestimated due to inappropriate diagnostic methods in clinical practice.

Diarrheagenic E. coli can be divided into five main categories on the basis of distinct epidemiological and clinical features, specific virulence determinants and an association with certain serotypes (Nataro & Kaper, 1998). The most commonly reported diarrheagenic E. coli strains in Brazilian children are enteropathogenic E. coli (EPEC) and enteroaggregative E. coli (EAEC). EPEC harbor the ‘locus of enterocyte effacement’ (LEE) pathogenicity island, which encodes factors responsible for the attaching and effacing (A/E) phenotype on host enterocytes (Jerse et al., 1990). These EPEC strains can also harbor the EPEC adherence plasmid (EAF) comprising the cluster of genes encoding the bundle-forming pilus (BFP) (Girón et al., 1993). EPEC strains with the EAF plasmid are classified as ‘typical EPEC’, whereas EPEC strains that do not possess the EAF plasmid are classified as ‘atypical EPEC’ (Kaper, 1996). EAEC are characterized by an aggregative adherence (AA) pattern on cultured epithelial cells, and produce fimbrial colonization factors called aggregative adherence factors (AAFs). The three other categories seem to be less prevalent: enterotoxigenic E. coli (ETEC), which produces the heat labile (LT) and/or heat-stable (ST) enterotoxins; enteroinvasive E. coli (EIEC) invades the colonic epithelium; and Shiga toxin-producing E. coli (STEC), which produces Shiga toxins 1 and 2 (Stx 1 and Stx 2) and in some strains the presence of the LEE region (Nataro & Kaper, 1998).

Identification of diarrheagenic E. coli strains requires that these organisms be differentiated from nonpathogenic members of the normal flora. Serotypic markers correlate, sometimes very closely, with specific categories of diarrheagenic E. coli; however, these markers are rarely sufficient in themselves to identify strains reliably as diarrheagenic. Thus, the detection of diarrheagenic E. coli has focused increasingly on the identification of certain characteristics that themselves determine the virulence of these organisms. This identification process may include HEp-2 cell adherence, DNA hybridization and PCR assays to detect the presence of specific virulence traits or the genes encoding these traits. The first two types of assays require special expertise, use cell culture and radioactive material and are time-consuming.

Considering the importance of diarrheagenic E. coli as a cause of childhood diarrhea in Brazil, we previously designed two multiplex PCRs to detect EPEC, ETEC, EIEC, STEC and EAEC, in stool samples (Aranda et al., 2004). In order to simplify diagnosis, we set up a single multiplex PCR assay by combining seven primer pairs to detect these types of E. coli strains simultaneously in a single reaction. The present study was undertaken to evaluate the application of this new assay to categorized pathogenic E. coli isolates and determine their distribution among children with and without diarrhea.

Materials and methods

Clinical specimens

From 1 January to 31 March 2005, all children under 5 years of age with acute diarrhea who were brought to the emergency room of Hospital São Paulo, Monday through Friday, were enrolled in the study. Every fecal specimen was examined by standard methods for the presence of Shigella spp. Salmonella spp., Yersinia enterocolitica, Campylobacter spp. and rotavirus. Four separate lactose-fermenting colonies and two nonlactose-fermenting colonies from each child were selected from MacConkey agar plates for testing. A total of 277 E. coli isolates were obtained from the 79 children. In addition, 210 E. coli isolates from 60 children without diarrhea were tested as controls. In total, 487 E. coli isolates were individually tested by the multiplex PCR assay and also screened by colony hybridization with specific DNA probes designed to detect EPEC (eae and bfpA probes), EAEC (CVD432 probe), ETEC (LT and ST probes), EIEC (Inv probe) and STEC (stx1 and stx2 probes) (Nataro & Kaper, 1998). These probes were labeled with [α-32P]dCTP, and colony hybridization assays were performed as described previously (Scaletsky et al., 2002a).

Bacterial strains

Diarrheagenic E. coli reference strains used as positive controls in the PCR assays included the EPEC E2348/69 (eae, bfpA), EAEC O42 (aggR), ETEC H10407 (elt, est), EIEC EDL1284 (ipaH) and EHEC EDL931 (eae, stx1, stx2) strains. The nonpathogenic E. coli K12 DH5αstrain was used as a negative control. We used 330 additional reference strains obtained from our laboratory collection to evaluate the multiplex PCR assay. These included 110 nondiarrheagenic E. coli isolates and 220 diarrheagenic E. coli isolates, divided into 50 EAEC, 50 atypical EPEC, 50 typical EPEC, 50 ETEC and 20 EIEC. We also included 115 E. coli and 17 Shigella species strains isolated from 150 diarrheic patients in a previous study (Aranda et al., 2004).

Preparation of DNA templates for PCR

All strains examined by PCR were grown on MacConkey agar plates at 37°C. DNA was extracted from bacteria by resuspending one bacterial colony in 50 µL of sterile water, boiling the suspension for 5 min and centrifuging it at 10 000 g for 1 min. The supernatant was then used as the DNA template for PCR.

Development of multiplex PCR assay

The DNA templates were subjected to multiplex PCR with specific primers for the detection of the following virulence markers: eae (structural gene for intimin of EPEC and EHEC), bfpA (structural gene for the BFP of typical EPEC), aggR (transcriptional activator for the AAFs of EAEC), elt and est (enterotoxins of ETEC), ipaH (invasion plasmid antigen H found in EIEC and Shigella), and stx (Shiga toxins 1, 2 and variants). Primers for stx and aggR were previously described by Toma (2003), and those for eae, bfpA, elt, est and ipaH were described by Aranda (2004). DNA sequences and sizes of PCR products are shown in Table 1.

View this table:
Table 1

PCR primers used in this study

Primer designationPrimers (5′–3′)Target geneAmplicon size (pb)Primer conc. (pmol)

The multiplex PCR assay was carried out with a 50-µL of reaction mixture containing 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl2, 2 mM of each deoxynucleoside triphosphate, 1.5 U of AccuPrime Taq DNA polymerase (Invitrogen), 2 µL of the DNA template and the PCR primers. The optimal concentration of each primer pair in the reaction mixture was determined empirically. Each primer pair concentration was varied independently until the PCR products exhibited equal intensities on 2% agarose gels when a DNA mixture of five prototype E. coli strains was used as the PCR template (Fig. 1). The concentration for each primer pair in the final reaction is given in Table 1. The PCR mixtures were then subjected to the following cycling conditions: 50°C (2 min, 1 cycle), 95°C (5 min, 1 cycle), 40 cycles of 95°C (1 min), 50°C (1 min), 72°C (1 min) and 72°C (7 min, 1 cycle) in a thermal cycler (model system 2400; Perkin-Elmer Corporation, Norwalk, Conn.). PCR products (10 µL) were visualized after electrophoresis in 2% agarose gels in Tris-borate-EDTA buffer and ethidium bromide staining, and the amplicons were identified based only on the size of the amplified product. In all further experiments, the DNA mixture from the five diarrheagenic reference strains from pure cultures served as the positive control, while E. coli K12 DH5α was the negative control.

Figure 1

Sensitivity of multiplex PCR of EPEC E2348/69 (eae, bfpA) from a spiked stool sample. Lane M, DNA molecular size markers (100-bp ladder); lane 1, nonspiked stool sample; lanes 2–7, dilutions from 102 to 107 CFU mL−1; lane 8, EPEC E2348/69 (eae, bfpA) and EIEC EDL1284; lane 9, positive control (DNA mix from the five prototype Escherichia coli: EPEC E2348/69 (eae, bfpA), EAEC O42 (aggR), ETEC H10407 (elt, est), EIEC EDL1284 (ipaH) and EHEC EDL931 (stx1, stx2, eae).

In order to determine the limit detection of the multiplex PCR, stool samples negative for diarrheagenic E. coli were spiked with a phosphate-buffered saline (PBS) suspension of reference diarrheagenic E. coli strains in serial 10-fold dilutions to give 102–108 CFU mL−1. Each serial dilution of the spiked stool sample was spread on MacConkey agar plates at 37°C and colonies were subjected to multiplex PCR. The sensitivity of the assay was defined as the lowest concentration of diarrheagenic E. coli that yielded positive results for each dilution. There may be a possibility that two diarrheagenic E. coli strains are in the same sample, we therefore mixed each pair of two reference strains and carried out the multiplex PCR to detect these E. coli strains.


Sensitivity of multiplex PCR

The sensitivity of the diagnostic multiplex PCR assay was determined from the number of diarrheagenic E. coli cells (in CFU per milliliter) spiked into each milliliter of stool sample that could be detected by this method. Repeat experiments confirmed the limit of detection of diarrheagenic E. coli was c. 103 CFU mL−1 of stool suspension. Figure 1 shows the limit of detection of typical EPEC E2348/69 (eae, bfpA) from a spiked stool sample (the results of other diarrheagenic E. coli strains are not shown), and that the multiplex PCR could detect two diarrheagenic E. coli (typical EPEC and EIEC) strains in a spiked stool sample.

Specificity of multiplex PCR

The specificity of multiplex PCR was tested with 330 additional reference strains. The strains included 220 diarrheagenic E. coli isolates and 110 nondiarrheagenic E. coli isolates. The multiplex PCR showed positive results for the diarrheagenic E. coli strains and negative results for the nondiarrheagenic E. coli strains (Table 2).

View this table:
Table 2

Results of multiplex PCR with 330 verified Escherichia coli strains

E. coli strain and geneNo. of strainsNumber of positives by multiplex PCR
Diarrheagenic E. coli (n=220)
Nondiarrheagenic E. coli1100

Validation of the multiplex PCR with clinical isolates

To demonstrate the utility of the multiplex PCR assay, 115 E. coli and 17 Shigella strains isolated from diarrheic patients were subjected to the multiplex PCR, and the results were compared with those obtained by monoplex PCRs (Table 3; Fig. 2). The comparison of the analysis of 65 pathogenic E. coli comprising typical and atypical EPEC, EAEC, EIEC and STEC by multiplex PCR and monoplex PCRs targeting eae, bfpA, aggR, ipaH and stx as single genes demonstrated that both assays yielded the same virulence genes. Moreover, eight of 17 Shigella strains carrying the ipaH gene were also detected by the multiplex PCR. The 50 nonpathogenic E. coli strains and the nine Shigella ipaH-negative tested negative in the multiplex PCR.

View this table:
Table 3

Results of monoplex and multiplex PCR assays with Escherichia coli and Shigella strains obtained from clinical isolates

Number of positive strains by PCR
StrainNumber of strainsGene(s)multiplexmonoplex
Typical EPEC20eae, bfpA2020
Atypical EPEC21eae2121
STEC3eae, stx133
E. coli5000
Shigella spp900
Figure 2

Multiplex PCR of clinical Escherichia coli isolates and stool samples. Lane M, DNA molecular size markers (100-bp ladder); lane 1, positive control (DNA mix from the five prototype E. coli): lane 2, EPEC HSP 7-1 (eae); lane 3, EPEC HSP 43-1 (eae and bfpA); lane 4, EIEC MA 245/5 (ipaH); lane 5, EAEC MA233-1 (aggR); lane 6, Shigella flexneri (ipaH); lane 7, negative control (E. coli DH5α); lanes 8 and 9, atypical and typical EPEC isolated from children with diarrhea.

Validation of multiplex PCR with stool samples from children with and without diarrhea

A total of 24 diarrheagenic E. coli strains were isolated from 79 stool samples from children with diarrhea and from 60 samples from children without diarrhea. The prevalence of diarrheagenic E. coli in both groups was significantly different (P<0.003). The PCR assays detected 13 (16.4%) atypical EPEC isolates (eae PCR positive), three (3.8%) typical EPEC isolates (eae and bfpA PCR positive), four (5.1%) EAEC isolates (aggR PCR positive) and one (1.3%) EIEC isolate (ipaH PCR positive) and one (1.3%) ETEC isolate (elt PCR positive) from the group with diarrhea. The prevalences of atypical EPEC and EAEC were similar (1.7%) in the healthy group. No STEC strains were isolated from any of the children examined. There was agreement between results of the PCR multiplex and DNA hybridization assays for all strains (Table 4). EPEC and EAEC strains detected by multiplex PCR were serotyped by an agglutination test using a commercial antiserum (PROBAC do Brasil, São Paulo). The typical EPEC strains detected belonged to serogroups O111, O127 and O142. One atypical EPEC belonged to serogroup O26 and one EAEC was of serogroup O126; both strains were isolated from children with diarrhea.

View this table:
Table 4

Diarrheagenic isolates of Escherichia coli from children with diarrhea or without diarrhea

Number of tested strains/number of positive strains
ChildAge (years)DiarrheaE. coli groupGene(s) and/or probePCRDNA probeSerogroup
44YesTypical EPECeae, bfpA2/22/2O111
62YesAtypical EPECeae3/33/3ND
81YesAtypical EPECeae3/33/3ND
142YesTypical EPECeae, bfpA3/33/3O127
172YesEAECaggR, pCVD4323/23/2ND
203YesEAECaggR, pCVD4323/23/2ND
223YesAtypical EPECeae4/34/3ND
252YesAtypical EPECeae2/12/1ND
322YesAtypical EPECeae2/22/2ND
334YesEAECaggR, pCVD4323/33/3O126
342NoEAECaggR, pCVD4324/44/4ND
394YesAtypical EPECeae3/23/2O26
421YesAtypical EPECeae4/34/3ND
453NoAtypical EPECeae3/23/2ND
492YesAtypical EPECeae3/23/2ND
571YesEAECaggR, pCVD4323/23/2ND
581YesAtypical EPECeae2/22/2ND
595YesEIECipaH, pInv2/22/2ND
623YesTypical EPECeae, bfpA3/33/3O142
651YesAtypical EPECeae3/23/2ND
691YesAtypical EPECeae3/33/3ND
702YesAtypical EPECeae3/33/3ND
721YesAtypical EPECeae4/24/2ND
  • ND, serogroup not determined.

No Shigella spp., Salmonella spp., Yersinia enterocolitica, Campylobacter spp. or rotavirus were isolated from the diarrhea or the control group.


Numerous multiplex PCRs methods have been developed for the identification of E. coli pathotypes (Pass et al., 2000; Toma et al., 2003; Kimata et al., 2005; Müller et al., 2006). However, most of the assays harbor limitations in terms of the number of targeted genes, specificity, the resolution of amplified fragments in agarose electrophoresis, nonspecific amplification and the inability to differentiate all categories of diarrheagenic E. coli strains. Recently, Vu Nguyen (2005) reported a multiplex PCR assay to detect eight genes for identification of EPEC, EAEC, ETEC, EIEC and STEC. However, it appears that a differentiation between typical and atypical EPEC and ETEC-LT and EIEC might not be straightforward. As a result of the similarity of the sizes of the DNA fragments amplified, from eae and bfpA (376 and 367 bp, respectively) and from elt and ial (322 and 320 bp, respectively), it has been necessary to perform PCRs with specific primers after the multiplex PCR in order to verify the result.

Therefore, in the present study, a novel multiplex PCR assay has been developed that allows the simultaneous detection of virulence genes from typical and atypical EPEC (eae and bfpA), ETEC (elt and est), EIEC (ipaH), EHEC (stx1, stx2 and variants), and EAEC (aggR) in a single reaction. In order to identify these virulence genes, we must compare the sizes of the DNA fragments amplified with those of the positive controls or DNA molecular markers.

It was estimated that the limit of detection of diarrheagenic E. coli by the multiplex PCR was c. 103 CFU mL−1 in a stool sample. It was also shown that the presence of two types of diarrheagenic E. coli in a stool sample could be detected by multiplex PCR.

The specificity of the multiplex PCR assay was demonstrated using several reference strains, as well as with well-characterized clinical isolates. It showed positive results for the diarrheagenic E. coli strains tested and negative results for all nondiarrheagenic E. coli strains. There was complete agreement between the results of single and multiplex PCRs for all clinical isolates tested, indicating the high degree of specificity of the assay.

Using the multiplex PCR assay, we detected pathogenic E. coli (typical and atypical EPEC, EAEC, ETEC, and EIEC) in the fecal samples of 22 (27.8%) children with acute diarrhea, whereas only two (0.3%) asymptomatic children were found to harbor atypical EPEC and EAEC strains. The same results were obtained by colony DNA hybridization with specific DNA probes.

In the field study, atypical EPEC was the most commonly isolated category of diarrheagenic E. coli and significantly associated with diarrhea (P<0.005). Furthermore, our findings demonstrate the previously recognized importance of atypical EPEC as a cause of childhood diarrhea in the field study and in other cities (Scaletsky et al., 1999; Gomes et al., 2004; Rodrigues et al., 2004). Moreover, our results also support the evidence from recent prospective case–control studies showing a reduced etiological role for typical EPEC in cases of diarrhea in children (Gomes et al., 2004; Rodrigues et al., 2004). Indeed, this pathogenic category was detected only in three children.

We diagnosed only four EAEC infections by PCR using aggR-derived primers. This primer set had previously been found to be very sensitive (Toma et al., 2003). One ETEC-LT and one EIEC strain were detected, but they were less common categories of diarrheagenic E. coli found in Brazil. We did not diagnose any STEC infection by the multiplex PCR and believe that STEC strains are relatively rare in Brazil (Vaz et al., 2004).

In conclusion, the multiplex PCR presented in this paper offers a practical possibility for rapid identification of diarrheagenic E. coli in a single reaction tube and could be used in a routine diagnostic laboratory.


This work was supported by Fundação de Amparo a Pesquisa do Estado de São Paulo (FAPESP) and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq).


  • Editor: Rob Delahay


View Abstract