OUP user menu

A gene cluster closely related to type II secretion pathway operons of Gram-negative bacteria is located on the large plasmid of enterohemorrhagic Escherichia coli O157 strains

Herbert Schmidt, Beatrix Henkel, Helge Karch
DOI: http://dx.doi.org/10.1111/j.1574-6968.1997.tb10299.x 265-272 First published online: 1 March 1997


Analysis of 14.162 kb of DNA derived from plasmid pO157 of enterohemorrhagic Escherichia coli (EHEC) O157:H7 strain EDL933, extending in the 5′ direction of the recently described EHEC-hly operon, revealed 13 open reading frames (ORF) which showed great similarities to genes of members of the type II pathway secretion systems of Gram-negative bacteria. We named the ORFs etpC to etpO for Embedded Image HEC Embedded Image ype II secretion Embedded Image athway. In addition, an IS911-like insertion element was found to separate the etp genes from the EHEC-hlyC gene. Hybridization experiments with a specific etp probe and various categories of enteric E. coli pathotypes revealed that the etp gene cluster occurred in all 30 EHEC strains of serogroup O157 (100%) tested and is distributed sporadically among other EHEC serogroups (60%). In addition, the etp genes were rarely detected in STEC isolated from bovine feces (10%). Moreover, it was found not to occur in enteropathogenic E. coli, enteroaggregative E. coli, enterotoxigenic E. coli and enteroinvasive E. coli. The results obtained with the etp probe were confirmed by a PCR approach to specifically detect an internal fragment of the etpD gene.

  • Enterohemorrhagic Escherichia coli
  • Plasmid pO157
  • EHEC type II secretion pathway

1 Introduction

Enterohemorrhagic Escherichia coli (EHEC) are a subgroup of Shiga toxin-producing E. coli (STEC) which is defined by its ability to cause bloody diarrhea and hemolytic uremic syndrome (HUS). EHEC are involved in sporadic cases and outbreaks of HUS and bloody diarrhea world-wide [1]. The occurrence of a large plasmid of approximately 90 kb (pO157) in addition to the expression of Shiga toxins [2], the heat-stable enterotoxin EAST1 [3] and intiminO157[4], is characteristic for this organism [5]. Plasmid pO157 of EHEC O157:H7 strain EDL933 has recently been mapped by restriction analysis [6]. Molecular studies revealed that this plasmid encodes the EHEC hemolysin, which belongs to the RTX family exoproteins [7]. The EHEC hemolysin operon is present in almost all EHEC of serogroup O157 and more sporadically in other EHEC serotypes [7, 8]. A further pO157-encoded protein, KatP, acts as a bifunctional catalase-peroxidase and is mainly located in the periplasm of EHEC strains [9]. At present it is not clear if these proteins play a direct role in EHEC-caused diseases or if they act rather as accessory factors for the survival, spread or maintenance of EHEC O157:H7 strains.

Several species of Gram-negative bacteria secrete extracellular proteins using a two-step general secretory pathway (GSP), which is termed ‘type II secretion pathway’ [10]. It begins with a Sec-dependent, signal peptide-mediated translocation across the cytoplasmic membrane [11]. Further transport across the outer membrane involves components of the main terminal branch (MTB) of the GSP. The MTB of the GSP is employed by a wide variety of pathogenic Gram-negative bacteria for extracellular protein export [11], examples of which are the pul genes of Klebsiella oxytoca and K. pneumoniae[12], the out genes of Erwinia chrysanthemi and E. carotovora and the exe operon of Aeromonas hydrophila[13, 14]. One of the most intensively studied systems is the pullulanase secretion pathway, in which at least 14 genes and proteins are involved [11, 15, 16].

In this study, we analyzed a cluster of open reading frames (ORFs) located in the 5′ direction of the EHEC-hly operon which encode 13 putative proteins closely related to the main terminal branch of type II secretion pathways of Gram-negative bacteria and analyzed their distribution among pathogenic E. coli strains.

2 Materials and methods

2.1 Bacterial strains and plasmids

The EHEC O157:H7 strain EDL933 was isolated from a patient with hemorrhagic colitis and harbors a 60 MDa plasmid designated pO157 as well as one smaller plasmid [17]. Recombinant plasmids pE12 and pB12 carrying fragments derived from pO157 were described recently [6]. E. coli HB101 (F, mcrBmrr, hsdS20(rBmB), recA13, leuB6, ara-14, proA2lacY1, galK2, xyl-5, mtl-1, rpsL20, supE44, λ) was used as host for recombinant plasmids pB12 and pE12 [6]. The wild-type E. coli strains used in this study were isolated in the last 2 years from patients with hemolytic uremic syndrome and diarrhea throughout Germany. The bovine strains were isolated from cattle throughout Southern Germany. Shiga toxin-producing E. coli strains were detected by PCR with stx1- and stx2-specific primers [18] and serotyped with the method described by Bockemühl et al. [19]. The enteropathogenic (EPEC), enteroaggregative (EAggEC), enterotoxigenic (ETEC) and enteroinvasive (EIEC) strains were from our strain collection.

2.2 Preparation of plasmids

Plasmids of STEC wild-type strains were prepared according to the method of Birnboim and Doly [20]. Briefly, single colonies were grown overnight in 3 ml L-broth with moderate shaking and subjected to the above mentioned procedure. After ethanol precipitation and washing, plasmid DNA was dissolved in 30 μl of sterile water.

2.3 Southern blot hybridization

DNA was transferred from agarose gels to nylon membranes according to standard methods [21]. For hybridization assays, the Nonradioactive DNA Labeling and Detection Kit (Boehringer GmbH, Mannheim, Germany) was used according to the manufacturer's instructions. The specific washing step was carried out at 55°C in a solution consisting of 0.03×SSC (4.5 mM NaCl, 0.45 mM sodium citrate) and 0.1% SDS to effect a stringency of 90%. The etp probe was prepared by restriction of pE12 with EcoRI and subsequent purification of the 11.9 kb fragment from agarose gels with the Prep-a-Gene kit (BioRad, Munich). The probe was labeled randomly as described earlier [22].

2.4 PCR

Oligonucleotides D1 (5′-cgt cag gag gat gtt cag-3′) and D13R (5′-cga ctg cac ctg ttc ctg att a-3′) were designed to amplify a 1062 bp internal fragment of the etpD gene covering positions 571–1632 of the etpD ORF. 50 ng of plasmid DNA was subjected to PCR. Amplification was carried out in a total volume of 50 μl containing each deoxynucleosidetriphosphate at 200 μM, 30 pmol of each primer, 5 μl 10-fold concentrated polymerase synthesis buffer containing 100 μM MgCl2 and 2.0 U of AmpliTaq-DNA polymerase (Perkin-Elmer, Applied Biosystems Weiterstadt, Germany). The DNA was denatured at 94°C for 30 s, annealed at 52°C for 60 s, and then extended for 70 s at 72°C. After 30 cycles were completed, a final extension step of 5 min at 72°C was conducted.

2.5 DNA sequencing

Nucleotide sequencing was carried out with universal and reverse primers for pUC/M13 vectors and customized primers as described previously [7]. Nucleotide sequence analyses and the searches for homologue DNA sequences in the EMBL and Genbank database libraries were performed with the program package HUSAR (Heidelberg Unix Sequence Analysis Resources, German Cancer Research Center, Heidelberg) and the Lasergene software package (DNASTAR, Madison, WI, USA).

2.6 Nucleotide sequence accession number

The nucleotide sequence of the 14.162 bp fragment derived from pO157 will appear in the EMBL/GenBank Nucleotide Sequence Data Library under accession number Y09824.

3 Results and discussion

3.1 Nucleotide sequence analysis

The high sequence similarity between the EHEC-hly operon and the α-hemolysin operon of E. coli prompted us to analyze the proximal and distal 5′ region of the EHEC-hly operon to identify potential regions necessary for regulation of gene expression. In order to achieve this, nucleotide sequence analysis was performed with plasmids pB12 and pE12 [6] carrying overlapping restriction fragments derived from plasmid pO157. Plasmid pB12 was recently shown to carry the EHEC-hly operon and was used in this study to analyze the yet uncharacterized 5′ end. More distal regions were analyzed using plasmid pE12. The inserts of both plasmids covered the region from 6.8 kb to 28.8 kb in the pO157 map [6]. Nucleotide sequencing was conducted by subcloning EcoRI-BamHI fragments from pE12 in vector pK18 and subsequently by the use of universal and customized sequencing primers. Since the analysis of the proximal 5′ region revealed sequences highly homologous to the pul genes of K. pneumoniae, we decided to sequence the whole insert of plasmid pE12.

A total sequence of 14.162 bp was revealed and this was analyzed for ORFs. Criteria for the search for ORFs were a translational start codon (atg) and a translational stop codon (tga, taa, tag). A total of 15 potential ORFs were found. The search for homologous DNA sequences in the EMBL database libraries showed that 13 ORFs exhibited sequence similarities with the pullulanase operons of K. pneumoniae and K. oxytoca, the out operons of E. chrysanthemi and E. carotovora and other members of the type II secretion pathway family of Gram-negative bacteria EHECEmbedded Image ype II Embedded Image athway. The positions and sizes of the ORFs and the sequence similarity to the pul genes are listed in Table 1.

View this table:

Sizes and positions of ORFs belonging to the etp gene cluster located on plasmid pO157 of EHEC O157:H7 strain EDL933 and sequence similarity to the pul genes of Klebsiella pneumoniae

Designation of ORFPosition in sequence (bp)aSize of ORF (bp)% Sequence similarityb to the respective pul genes of Klebsiella pneumoniae
etpC .349–1.22487645.5
  • aThe first postion represents the first nucleotide of the translational start codon, the second position the last base of the translational stop codon.bThe sequence similarity was determined with the Lipman-Pearson Algorithm using a k-tuple of 2, a gap penalty of 3 and a window size of 20.

PulD of K. pneumoniae and related proteins such as pIV [23], YscC of Y. enterocolitica[24], OutD of E. chrysanthemi[25], XpsD of Xanthomonas campestris[26] or MxiD of Shigella flexneri[27] belong to a superfamily of proteins involved in various secretion pathways of Gram-negative bacteria [28]. These proteins show a modular organization with two major domains, a dissimilar N-terminal domain and highly conserved C-terminal domain separated by a short central region of approximately 30–100 amino acid residues. Four highly conserved regions (A–D) could be distinguished within the C-terminal domain, one of them containing the common motif (V,I)PXL(S,G)XIPXXGXLF. From the N-terminal domains the PulD family can be divided into at least two subgroups, correlating with distinct secretion pathways. In order to clarify if the putative EtpD protein also belongs to this family we performed an alignment of EtpD, PulD and OutD (Fig. 2). All four regions proposed by Genin and Boucher are present including the 14 amino acid motif (V,I)PXL(S,G)XIPXXGXLF in region D. Therefore, EtpD could be shown to be a further highly conserved member of this family.

Figure 2

Alignment of a region within the C-terminal domain of EtpD of E. coli O157:H7 strain EDL933 (upper line), PulD of Klebsiella pneumoniae (middle line) and OutD of Erwinia carotovora (lower line). Amino acids that are identical are underlined. Four blocks sharing a nearly perfect sequence match are indicated (A, B, C, D). The boxes depict highly conserved amino acid residues. The alignment starts at positions 228 (EtpD), 228 (PulD) and 281 (OutD) of the respective amino acid sequences.

The DNA sequence at positions 12071–13323 (1250 bp) is highly related to IS911 of Shigella dysenteria[29]. Alignment with the Lipman-Pearson algorithm using a k-tuple of 3 revealed a sequence similarity of both IS elements of 91.3%. The IS911 relative of plasmid pO157 carries two imperfect terminal inverted repeats of 27 bp and at least two potential ORFs. The location of the IS911 relative, between the EHEC-hlyC gene which starts at position 13866 and the etpO gene, which ends at position 11965, separates these two gene clusters from each other. Homology searches in the EMBL database library of the DNA sequences between the etpO and the E-hlyC gene revealed no significant homologies to the upstream regulatory regions present in other RTX operons such as hlyR[30]. The recently described 35 bp OPS (operon polarity suppressor) sequence spanning the element ggcggtag which is described to occur upstream of various bacterial operons [31] was found at positions 11184–11191. However, this was placed inside the ORF of the putative etpN gene. The genetic organization of the entire analyzed DNA fragment is shown in Fig. 1.

Figure 1

Structure of the 14.162 bp region analyzed in this study. The arrows indicate the length and direction of the ORFs belonging to the etp gene cluster. Restriction enzyme sites are indicated above. The last ORF at the 3′ end which appears truncated, represents the 5′ end of the EHEC-hlyC gene. Above the figure, clones pE12 and pB12 are depicted.

3.2 Distribution of the etp gene cluster among enteric E. coli

The discovery of a gene cluster presumably encoding a type II secretion system in EHEC strain EDL933 prompted us to investigate its occurrence in (a) EHEC of serogroup O157, (b) non-O157 EHEC, (c) Shiga toxin producing E. coli (STEC) isolated from bovine feces and (d) other enteric E. coli pathogroups (EAggEC, EPEC, EIEC and ETEC). The results achieved for all tested STEC E. coli by hybridization with the 11.9 kb etp gene probe and by PCR applied with etpD-specific primers D1 and D13R are listed in Table 2. Representative hybridization results are shown in Fig. 3. The etp genes occurred obligatorily (100%) in EHEC strains of serogroup O157 and more sporadically in non-O157 EHEC (60%). Only 10% of the STEC isolates from bovine feces carried the etp operon. Moreover, none of the other E. coli enteric pathogroups (EPEC, EAggEC, EIEC, ETEC) was found to carry the etp gene cluster (data not listed in Table 1).

View this table:

Results of hybridization with a 11.9 kb etp gene probe and by PCR with etpD-specific primers D1 and D13R applied to Shiga toxin-producing E. coli from various sources

Groups of E. coliNumber of strains testedSerotypesToxin genotypesNumber of etp-positive strains bySource
EHEC30O157:H7; Hstx 2, stx1/stx23030aisolates from patients with HUS or diarrhea
Non-O157 EHEC30O103:H2;stx1, stx2, stx1/stx21818aisolates from patients with HUS or diarrhea
O26:H11; H;
O111:H; H8
STEC30O55:H18; O98:H8;stx 1, stx233aisolates from bovine feces
O110:H; O8:H21
  • aThe strains which are positive by D1/D13R PCR are the same as the strains which are positive by hybridization.

Figure 3

Alignment of a region within the C-terminal domain of EtpD of E. coli O157:H7 strain EDL933 (upper line), PulD of Klebsiella pneumoniae (middle line) and OutD of Erwinia carotovora (lower line). Amino acids that are identical are underlined. Four blocks sharing a nearly perfect sequence match are indicated (A, B, C, D). The boxes depict highly conserved amino acid residues. The alignment starts at positions 228 (EtpD), 228 (PulD) and 281 (OutD) of the respective amino acid sequences.

The occurrence of these genes among Shiga toxin producing E. coli is therefore restricted. It appears to be closely associated with Stx production in strains of serogroup O157, but only sporadically in non-O157 strains. In addition, we found no correlation between the occurrence of this gene cluster and the disease caused by the respective infectious organism. Restriction analyses of EHEC plasmids indicate that the plasmid population of these strains is heterogeneous. Moreover, the occurrence of the IS element shows evidence that the etp gene cluster, as well as the hly operon, is possibly part of gene cassettes interchangeable between EHEC plasmids.

Since its first description in 1983 [32], the large plasmid (pO157) of EHEC of serotype O157:H7 has been implicated in the pathogenesis of EHEC-associated diseases. However, no conclusive results have been obtained in animal experiments or cell culture experiments on the role of pO157 [17, 3335]. Only one study showed a statistically significant linkage between the occurrence of the plasmid encoded EHEC hemolysin in STEC O111 strains and the development of HUS in patients infected with this organism [8]. To date, only molecular studies provide reliable information on the function of plasmid pO157, which is present in almost all O157:H7 strains. Substantial information has come from our group on the plasmid map [6], the EHEC hemolysin [7, 36], the KatP catalase-peroxidase [9] and the etp gene cluster (this study). Hybridization studies with the plasmid pO157-specific gene probes for the above mentioned genes indicate that the large plasmids of EHEC strains are more heterogeneous than was thought, presumably carrying gene cassettes interchangeable among strains (data not shown). Our present study presented a new pO157 located gene cluster highly homologous to type II transport systems of Gram-negative bacteria. To date, we have no information on the type of protein transported or, moreover, if the system is functional. However, the etp gene cluster is present in important EHEC serogroups.

Previous investigations and the results of this study have shown that the EHEC-hly, katP and etp genes occur virtually obligatorily in EHEC O157:H7 strains, whereas the distribution of these genes in non-O157 EHEC is more sporadic. In these strains, different combinations of the genes were found (data not shown). In conclusion, hybridization studies with gene probes specific for genes located on plasmid pO157 provide substantial information on the plasmid populations among EHEC strains and are powerful methods for subtyping of EHEC plasmids and strains. Much epidemiological work has to be done to improve our understanding of the distribution of plasmid encoded genes among various EHEC serotypes of animal and human origin.


This work was supported by Grant Ka 717/2-3 from the Deutsche Forschungsgemeinschaft.


View Abstract