OUP user menu

Lack of expression of bundle-forming pili in some clinical isolates of enteropathogenic Escherichia coli (EPEC) is due to a conserved large deletion in the bfp operon

Mafalda R. Bortolini , Luis R. Trabulsi , Rogéria Keller , Gad Frankel , Vanessa Sperandio
DOI: http://dx.doi.org/10.1111/j.1574-6968.1999.tb08723.x 169-174 First published online: 1 October 1999

Abstract

Enteropathogenic Escherichia coli (EPEC) produces a plasmid-encoded type IV pilus, called the bundle-forming pilus (BFP), involved in the formation of the localized adhesion onto epithelial cells. In this study, we demonstrate that clinical isolates of serotypes O128ab:H2 and O119:H2 contain a ca. 13-kb deletion in the bfp operon, resulting in a lack of expression of these pili. An IS sequence with homology to the IS66 of Agrobacterium tumefaciens replaced the deleted bfp genes. These results suggest that the bfp operon was deleted through a transpositional event and that other adherence factors may mediate attachment of these bacteria to the host cells.

Keywords
  • Enteropathogenic Escherichia coli
  • Bundle-forming pilus
  • Localized adherence
  • Type IV pilus

1 Introduction

Enteropathogenic Escherichia coli (EPEC) is one of the main causes of infant diarrhea in children in developing countries [1,2]. EPEC adheres to epithelial cells forming microcolonies, a pattern called localized adherence (LA) [3]. In addition, it produces a distinct histopathological lesion on the intestinal mucosa referred to as the attaching and effacing (AE) lesion [4], which is characterized by effacement of microvilli and intimate adherence of the bacterium to the apical surface of the epithelial cells [4].

The LA phenotype is dependent on the presence of an EPEC adherence factor plasmid (EAF) [5]. This plasmid harbors a 14-gene operon which encodes a type IV pilus, called the bundle-forming pilus (BFP) [6,7]. A subset of three genes, also encoded on the EAF plasmid, called the plasmid encoded regulator (perABC) genes, is involved in the transcriptional regulation of virulence genes [8], including bfp [9].

The AE lesion formation requires the products of several chromosomal genes encoded on a pathogenicity island of 35 kb called the locus of enterocyte effacement (LEE) [10]. The LEE encodes a type III secretion system [11], an outer membrane protein adhesin, termed intimin [12], and several secreted proteins including EspA, EspB and EspD [10]. Two chromosomal loci (selC and pheU) were identified as the LEE integration sites in different EPEC serotypes [13]. However, some EPEC strains may contain the LEE inserted in a third, unidentified integration site [13].

In this study, we report the genetic analysis of EPEC clinical isolates with regard to their virulence traits.

2 Materials and methods

2.1 Bacterial strains

We studied eight EPEC strains belonging to serotypes O128ab:H2 and O119:H2, isolated from children with diarrhea between 1960 and 1993 (Table 1). These isolates were of particular interest because they do not produce BFP and a typical LA pattern on epithelial cells, although they hybridize with the bfpA probe [1416]. These strains fall within the same phylogenetic group as previously determined by multilocus enzyme electrophoresis, ERIC-typing and Ribotyping [1416]. Prototypic EPEC strains E2348/69 (O127:H6) [17] and B171 (O111:NM) [18] have been described before.

View this table:
Table 1

Study of the LEE region and of the virulence genes in the EAF plasmid

StrainsSerotypeProbesProbes and PCRPCR
LEE-ABCDbfpABfpBBfpD–FBfpI–LperABCBfpA-IS66selCpheU
MB18O128ab:H2+++++
MB20O128ab:H2+++++
MB21O128ab:H2+++++
MB72O119:H2+++++
MB79O119:H2+++++
MB80O119:H2+++++
MB109O119:H2+++++
MB112O119:H2+++++
E2348/69O127:H6+++++++
JPN15O127:H6++
B171O111:NM++++++ntnt
nt=not tested.
  • JPN15 is a plasmid-cured derivative from E2348/69.

2.2 PCR and probing for per and bfp plasmid genes

The perABC genes were detected by PCR, using primers PerA-F/PerA-R, PerB1/PerB2 and PerC1/PerC2 (all annealing temperatures were 50°C) (Table 2), which produced amplicons of 824, 329 and 177 bp, respectively. The bfpB gene was assayed by PCR using primers BfpB1/BfpB2 (annealing at 56°C) (Table 2). Primers designed to anneal to the bfpA gene (BfpAB2) and the IS66-like element (DEL12) are described in Table 2 (annealing at 56°C). Primers designed to map the deletion in the bfp operon: bfpDbfpF, expected amplicon size 2690 bp, and bfpIbfpL, expected amplicon size 1751 bp, are also described in Table 2 (annealing at 57°C). All reactions were performed with 200 ng of purified genomic DNA template, 100 ng of each primer, 200 µM deoxynucleoside triphosphates, 1 U Taq DNA polymerase and 1.5 mM MgCl2 in Taq polymerase buffer (Life technologies, Gaithersburg, MD, USA). After denaturing the template at 94°C for 10 min, a total of 30 cycles were performed at 94°C for 1 min, annealing for 1 min and 72°C for 2 min.

View this table:
Table 2

Primers used in PCR

PrimersSequence
PerA-F5′-ATGCTTACATCTAAAAAAGAAATG-3′
PerA-R5′-TTATGGCAATGTTCCTTGTGT-3′
PerB15′-GTAGTTTTTGATGGATGTATG-3′
PerB25′-TCACCGCCTCCTTCCATCGT-3′
PerC15′-GACGAGCTGCAGATGCCTTGG-3′
PerC25′-GCCCCATTTTCTTATATGCCT-3′
BfpB15′-CGCCAGAAGCCTTGAGATCA-3′
BfpB25′-GAACAGTGCAACAGGCGGA-3′
BfpAB25′-GATTGAATCTGCAATGGTGC-3′
DEL125′-TGATGACACGCCTGTCCCGG-3′
BfpD5′-CTTATCAGGCTGCTCGTATAG-3′
BfpF5′-CAAGCTCACGTGCATCCATC-3′
BfpI5′-TCACTGATTGAAGCGTCATTAG-3′
BfpL5′-GCTCGTCCGACAGAATACTC-3′

The perABC probe is a 3.5-kb EcoRI fragment of pCVD450 [9]. The probes for bfpB, bfpDbfpF and bfpIbfpL were the products of the amplification of these genes from the prototype EPEC strain E2348/69. All strains were hybridized by a colony blot procedure under high stringency conditions [19].

2.3 PCR and probing for the LEE region

All strains were hybridized by a colony blot procedure with LEE-A, LEE-B, LEE-C and LEE-D probes as described by McDaniel et al. [10]. After electrophoresis in 0.7% agarose gels, the fragments were purified with a Wizard PCR Preps DNA Purification System kit (Promega), labelled with [α-32P]dATP by random priming using a Random Primed DNA labelling kit (Boehringer Mannheim) according to the manufacturer's instructions and hybridized under high stringency conditions [19].

Insertion of the LEE in selC was assayed by PCR as described by McDaniel et al. [10] and in pheU as described by Sperandio et al. [13].

3 Results and discussion

Expression of BFP requires the function of at least 14 genes which comprise the complete bfp operon (bfpABCUDEFPHIJKLM) (Fig. 1) and are present in the EAF plasmid [20,21]. It has also been shown that the transcriptional activation of bfp is dependent on the perABC (bfpTVW) genes [8,9].

Figure 1

(A) Schematic organization of the bfp operon and the per genes in EPEC prototype strains E2349/69 and B171. (B) Schematic representation of the interrupted bfpA (*) gene and the IS66 insertion sequence that replaced the remaining of the bfp operon in strain MB20 (O128ab:H2) and the per genes. The small arrows represent the position of the primers described in Table 2. The interrupted bar means that the distance between the bfp operon and per is unknown.

We studied clinical isolates of EPEC serotypes O128ab:H2 and O119:H2, which were previously found to hybridize to the bfpA probe, but they did not produce the BFP fimbriae and only adhered to epithelial cells after 6 h of infection. To understand the reason why these strains were unable to produce BFP, we performed a genetic analysis on them to determine whether all the bfp and per genes were present. The perABC genes were detected by PCR and probe hybridizations in all of these strains (Table 1 and Fig. 1), indicating that the per regulon is present. In contrast, the bfpB gene probe did not hybridize with any of the O128:H2 or the O119:H2 strains, suggesting that this gene was either missing or shared a low sequence homology with bfpB from E2348/69. In order to clarify this, we cloned a 2.1-kb BamHI/SalI DNA fragment, that hybridized with the bfpA gene probe, in pBluescript II SK (Stratagene) from one of the strains, MB20 (O128ab:H2), generating plasmid pMB11. The DNA sequence of pMB11 was obtained by DNA sequencing using the Ready Reaction Dye Dideoxy Terminator Cycle Sequencing kit in an Automatic Sequencer (Applied Biosystems 373A). This sequence revealed that at least part of the bfp operon was deleted (GenBank accession number AF119170), since only the first 262 bp of the bfpA gene were detected (Fig. 1 and GenBank accession number AF119170). Consistent with this, no BFP was detected by Western blotting in any of the eight strains (data not shown). Downstream of the truncated bfpA, we identified an open reading frame (ORF) of 1343 bp in the opposite orientation that runs into it and which shared 61% sequence similarity with ORFs 2 and 3 of IS66 (Fig. 1 and GenBank accession number AF119170) from the Ti plasmid of Agrobacterium tumefaciens (GenBank accession number M10204).

To determine if the deletion in the bfp operon is common to all the strains included in this study, we have used primers designed, according to the DNA sequence obtained from strain MB20 (O128ab:H2), to anneal to bfpA and the IS66-like element. This PCR produced amplicons of 1053 bp with the eight studied strains while no amplicon was produced with the control wild-type EPEC strains (E2348/69 and B171), the EAF plasmid-cured derivative of E2348/69 (strain JPN15) (Table 1) or the laboratory strain DH5α (data not shown). The products of amplification from each of the O128 and O119 strains were then purified and digested with PstI. This enzyme had only one restriction site along the amplicon of strain MB20 and the digestion of amplicons of all the eight strains produced two DNA fragments (188 and 865 bp) for each strain (data not shown), suggesting that the EAF plasmids of these O128ab:H2 and O119:H2 strains have a similar deletion in the bfp operon.

The extent of the deletion in the bfp operon was measured by PCR reactions with two additional sets of primers, designed to amplify the last genes of the operon: bfpDbfpF and bfpIbfpL (Table 2 and Fig. 1). Using these primers for PCR yielded no product for any of the O128:H2 and O119:H2 strains (Table 1), while the expected PCR products were detected using E2348/69. This result was confirmed by hybridizing these eight strains with the probes bfpD–F and bfpI–L (Table 1), which were amplified by PCR using DNA from strain E2348/69. These results show that a deletion of the entire bfp operon occurred in all the O128:H2 and O119:H2 strains.

To fully characterize these strains, we also looked for the presence of the LEE region by probe hybridizations and PCR. All strains hybridized with the LEE probes A, B, C and D [11] (summarized in Table 1). The LEE integration site in these strains was determined by PCR as previously described [10,13]. The integrity of the selC locus, determined as described by McDaniel et al. [10], revealed that while no PCR product was detected in the prototype E2348/69 EPEC strain (in which the LEE was inserted into selC) [10], this locus was not disrupted in either of the O128:H2 or the O119:H2 strains (Table 1). In contrast, using PCR primers and conditions as described by Sperandio et al. [13], we found that the pheU gene was disrupted in all the O128:H2 and O119:H2 strains, while it was intact in strains E2348/69 (Table 1) and DH5α (data not shown). The insertion of the LEE region in the pheU gene was first reported in a Shiga-toxin producing E. coli strain of serotype O26:H[22], which is in a phylogenetic cluster classified as EHEC2 [23]. Strains O128ab:H2 and O119:H2 have been classified in a phylogenetic cluster known as EPEC2, which is closely related to EHEC2, and both seemed to have evolved from the same ancestral strain [23]. Curiously, the EPEC2 cluster is more related to EHEC2 than to EPEC1 and our finding that the LEE in the O128 and O119 strains was inserted into the pheU locus is consistent with this idea.

In conclusion, since we identified a DNA sequence downstream of bfpA in the EAF plasmid of these strains that was similar to IS66 of Ti plasmid of A. tumefaciens, we propose that a transpositional event could have been responsible for the deletion in the bfp operon of these strains. These strains have been previously grouped in the same phylogenetic cluster by multilocus enzyme electrophoresis, ERIC-typing and Ribotyping [1416], suggesting that this deletion in the bfp operon, which is conserved in all of them, occurred in the ancestral strain. The LEE region of these strains is also inserted in the same chromosomal locus (pheU), reinforcing the idea that these strains share a common ancestor.

Hicks and collaborators [24] recently showed that BFP was not the initial adherence factor of EPEC to intestinal epithelial cells but was involved in the formation of tridimensional colonies, implying that BFP may not be involved in the adherence of the bacteria to the epithelial cells but rather in tethering bacteria together. Here, we report that some bacterial isolates from diarrhea share a deletion in the bfp operon. The studied strains were isolated from sporadic cases of diarrhea, not from epidemic outbreaks, so even though BFP may not be essential for virulence in these strains, it probably does enhance it in epidemic strains.

Acknowledgments

We thank Dr James B. Kaper from the University of Maryland and Dr Timothy McDaniel from Stanford University for the four LEE probes, Jussara dos Santos da Silveira and Viviane Furlan Lozano for their help with the bfpB hybridizations, Dr Jorge A. Girón from the Universidad Autonoma de Puebla and Dr Mohamed A. Karmali for revising this manuscript. This work was supported by Grants FAPESP 96/4148-5, FINEP/MCT/PRONEX 41.96.0881.00. Vanessa Sperandio is a Pew fellow.

References

  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
View Abstract