Abstract
In this study, we characterized the genetic background of various nleA variants in 106 Shiga toxin-producing Escherichia coli (STEC) and enteropathogenic Escherichia coli (EPEC) strains. The flanking regions of eight nleA variants were analyzed by DNA sequencing and compared with the corresponding regions of five previously described NleA-encoding prophages. The analyzed nleA variants were all located downstream of the DNA region responsible for phage morphogenesis. In particular, the type III effector genes avrA, ospB, nleH, and nleG and IS elements were detected in the neighborhood of nleA. The structure of the eight analyzed regions flanking nleA primarily resembled the corresponding region of the NleA4795-encoding prophage BP-4795. Using PCR, the gene order flanking 13 nleA variants in strains of different serogroups was compared to the respective regions in reference strains. The analyses showed that strains which harbor prophages with conserved flanking regions of a particular nleA variant predominantly occurred, and IS elements were additionally detected in these regions. We were able to mobilize nleA by transduction in 20% of strains determined, which comprised in particular EPEC strains harboring an nleA variant, the gene encoding the protein known as “EspI-like.” Plaque hybridization was used to identify phages that harbor the genes stx and nleA. However, only two strains harbored variant nleA4795 in the genome of an Stx1 prophage.
INTRODUCTION
Enteropathogenic Escherichia coli (EPEC) and Shiga toxin-producing E. coli (STEC) are ubiquitously occurring pathogens causing human diseases ranging from watery diarrhea to hemolytic-uremic syndrome (HUS) (19). A type III secretion system functions as an important virulence determinant of EPEC and many STEC strains for survival under environmental conditions in the intestine (8). By translocation of a broad spectrum of type III effector proteins, these bacteria are able to modulate numerous signaling pathways in eukaryotic host cells (8, 28, 29).
One of the most frequently occurring type III effector genes in STEC and EPEC is nleA (also known as espI) (4, 5). NleA/EspI could be assigned as a virulence determinant in a mouse model (18). Translocated NleA colocalizes with Golgi markers in eukaryotic cells (6, 9). One reported function of NleA is the alteration of protein traffic in host cells by inhibition of the COPII-dependent vesicular transport from the endoplasmic reticulum to the Golgi apparatus (14). Furthermore, the involvement of NleA in the disruption of tight junctions was described previously (27). In addition, 15 putative eukaryotic binding partners of NleA were identified, but until the present study further effects on host cells by the interaction of NleA with most of these proteins were unknown (15).
Whereas several type III effector genes seem to be conserved, nleA shows multiple allelic variations. So far, 15 variants of nleA have been described for STEC and EPEC (5, 7). These variants were designated nleA4795, the EspI-like gene, z6024, and nleA1 to nleA12, and the deduced amino acid sequences of the variants reveal identities between 71% and 96%. Variants nleA6 and nleA8 were further classified into the subvariants nleA6-1 to nleA6-3, as well as nleA8-1 and nleA8-2. STEC and EPEC harbor either one or two nleA alleles (5, 7).
In E. coli strains, type III effector genes occur in three types of exchangeable effector gene loci: pathogenicity islands, lambdoid prophages, and non-phage-associated loci (28). Phages seem to be particularly important vehicles for the spread of type III effector genes (26, 28). Thereby, type III effector genes are additional genes with distinct transcription control sequences in the phage genome and can be transcribed autonomously from the rest of the phage genome (12). Lambdoid prophages encoding type III effectors always carry more than one effector gene, which are exclusively located downstream of the tail fiber genes (21, 28).
According to available STEC and EPEC genome sequences, nleA is predominantly located in prophage genomes which seem to be cryptic (11, 13, 20, 21, 22). There is only one report of an nleA variant contained on a plasmid in an atypical O55:H7 EPEC strain (30). Furthermore, there occur some inducible NleA-encoding phages which could spread this gene by horizontal gene transfer (7). In a similar manner, the Cif-encoding prophage of EPEC O103:H2 strain E22 with the variant EspI-like gene produces infectious phages in vitro (16). Shiga toxin (Stx)-converting bacteriophages rarely harbor type III effector genes, in particular alleles of nleG (21). However, variant nleA4795 is colocated with stx1 in the genome of the inducible prophage BP-4795 (6). The flanking regions of nleA consist primarily of type III effector genes and IS elements. However, only the flanking regions of the nleA variants z6024 in O26 and O157 STEC strains and nleA8-2 in EPEC strain E2348/69 are identical to some extent. In contrast, comparable regions of prophages harboring nleA variants, such as the EspI-like gene, nleA8-1, and nleA4795, show low levels of gene similarity both among themselves and in comparison with phages carrying z6024 or nleA8-2 (6, 13, 16, 21, 22).
The aim of the present study was to determine the genetic background of a number of nleA variants. We wanted to analyze whether nleA variants are located together with certain effector genes in genetic cassettes which could be transferred between E. coli strains, and we were interested in the genetic structure of such cassettes. Finally, the frequency of inducible Stx-converting phages additionally harboring an nleA variant was analyzed.
MATERIALS AND METHODS
Bacterial strains.
The 106 STEC and EPEC strains used in this study were described previously (7). Important phenotypic and genotypic characteristics of the strains are listed in Table 1 and in Table S4 in the supplemental material. E. coli K-12 strain C600 was used for cloning experiments and as a negative control. E. coli K-12 strains harboring either plasmid pK18 (23) or pACYC184 (3) were used in transduction experiments. E. coli strains were routinely grown in LB medium or on LB agar plates at 37°C. Antibiotics were added to final concentrations of 20 μg/ml for chloramphenicol and 50 μg/ml for kanamycin.
Table 1.
stx subtypes and nleA variants of 106 pathogenic E. coli strains, characteristics of regions flanking nleA, and transduced genetic traitsa
| Serotypeb (no. of strains) | stx subtype | nleA variant | Flanking region of nleAc | Plaquesd | Transduced traits (hybridization)e |
|---|---|---|---|---|---|
| O15:H− (1) | z6024 | z6024 (O157, nleF-espM = O26) | − | ||
| O26:H11 (1) | 2 | 8-1 | ND | ++ | stx2 |
| O26:H11 (1) | 2 | 8-1 | 8-1 (O26) | ++ | stx2 (H) |
| O26:H11 (1) | 1 | z6024/8-1 | ND | Lysis | |
| O26:H− (1) | 1 | z6024/8-1 | ND | + | stx1 |
| O26:H11 (1) | 1 | z6024/8-1 | z6024 (nleG-espM = O26)/8-1 (O26) | Lysis | |
| O26:H− (1) | 1 | z6024/8-1 | z6024 (nleG-espM = O26)/8-1 (O26) | + | nleA**, stx1 (H) |
| O26:H11 (1) | 1 | z6024/8-1 | z6024 (O26)/8-1 (nleG-yfaS = O26) | (+) | nleA**, stx1 |
| O26:H− (1) | 2 | z6024/8-1 | ND | Lysis | stx2 |
| O26:H− (1) | 2 | z6024/8-1 | z6024 (nleG-espM = O26)/8-1 (O26) | Lysis | |
| O26:H11/− (2) | 2 | z6024/8-1 | z6024 (nleG-espM = O26)/8-1 (O26) | ++ | stx2 |
| O26:H− (1) | 1/2 | z6024/8-1 | z6024 (O26)/8-1 (nleG-yfaS = O26) | Lysis | stx2 |
| O49:H2 (2) | espI-like | espI-like | + | espI-like (1 × H) | |
| O49:H2 (1) | espI-like | espI-like (cif-nleA: +1.2 kb) | − | ||
| O49:H10 (1) | z6024 | z6024 (O157, nleF-espM = O26) | Lysis | ||
| O49:H10/− (2) | z6024/8-1 | z6024 (O157, nleA-nleH: +1.4 kb; nleF-espM = O26)/8-1 (?) | Lysis | ||
| O49:H10 (1) | 2 | 2 (nleA-yehV: −0.4 kb) | ++ | (H) | |
| O49:H35 (1) | 2 | 2 (nleA-yehV: −1.2 kb) | (+) | ||
| O49:H− (1) | 10 | 10 | Lysis | nleA10 | |
| O84:H4/28 (2) | 1 | nleA4795 | nleA4795 | (+) | nleA4795, stx1 (H) |
| O84:H28 (1) | 1 | nleA4795 | nleA4795 | ND | ND |
| O84:H− (1) | 1 | 7/8-2 | 7/8-2 (dst nleA ?; ust nleA +0.4 kb) | + | stx1 |
| O84:H2/−/nt (3) | 1 | 7/8-2 | 7/8-2 (dst nleA ?; ust nleA +0.4 kb) | − | |
| O84:H− (1) | 1 | 7/8-2 | 7/8-2 (dst nleA ?; ust nleA +0.4 kb) | ND | ND |
| O84:H− (1) | 8-1/8-2 | 8-1 (?)/8-2 | Lysis | ||
| O84:Hnt (1) | 1 | 9 | 9 | + | nleA9**, stx1 (H) |
| O84:H2 (1) | 1 | 9/8-2 | 9/8-2 (dst nleA ?; ust nleA +0.4 kb) | − | nleA**, stx1** |
| O84:Hnt (1) | 1 | 9/8-2 | 9/8-2 (dst nleA ?; ust nleA +0.4 kb) | ND | ND |
| O84:Hnt (1) | 1 | 9/8-2 | 9/8-2 (ND) | ND | ND |
| O103:H2 (1) | espI-like | espI-like | (+) | espI-like | |
| O103:H2 (1) | espI-like/11 | espI-like (lom-nleB: +1.5 kb)/11 | + | (H) | |
| O103:H2 (1) | 1 | espI-like | ND | (+) | stx1 |
| O103:H2 (1) | 1 | espI-like | espI-like | (+) | stx1 |
| O103:H2 (1) | 1 | espI-like | espI-like | − | |
| O103:H2 (1) | 1 | espI-like | espI-like (cif-nleA: +1.2 kb) | Lysis | |
| O103:H2 (2) | 1 | espI-like | espI-like (nleB-nleC: +1.2 kb) | ++ | (1 × H) |
| O103:H11 (1) | 1 | 8-1 | 8-1: O26 | Lysis | stx1** |
| O103:H2 (1) | 2 | espI-like | espI-like (nleH-cif: +1.2 kb) | + | stx2 (H) |
| O103:H2 (1) | 2 | espI-like | espI-like (nleH-cif: +1.2 kb) | − | |
| O111:H25 (1) | z6024 | z6024 (O157, nleA-nleH: +1.4 kb; nleF-espM = O26) | − | ||
| O111:H2/− (2) | espI-like | espI-like (nleB-ybhB = O103) | ++ | ||
| O111:H8/− (4) | 1 | 8-1 | 8-1 (O111, ant-nleG: +1.4 kb) | Lysis | |
| O111:H− (1) | 1/2 | 8-1 | 8-1 (O111, ant-nleG: +1.4 kb) | Lysis | |
| O111:H− (1) | 1/2 | 8-1 | 8-1 (O111, ant-nleG: +1.4 kb) | Lysis | stx2 |
| O111:H− (1) | 1/2 | 8-1 | 8-1 (O111) | Lysis | stx2 |
| O111:H− (2) | 1/2 | 8-1 | ND | Lysis | stx2 |
| O118:H5 (1) | z6024 | z6024 (O157, nleF-espM = O26) | − | ||
| O125:H− (1) | 8-2 | 8-2 (+1–2 kb at various sites) | − | ||
| O127:H6 (1) | 8-2 | 8-2 | + | ||
| O128:H2 (1) | espI-like | ND | − | ||
| O128:H2 (2) | espI-like | espI-like | − | ||
| O128:H2 (1) | espI-like | espI-like | + | ||
| O128:H2 (1) | espI-like | espI-like | Lysis | espI-like | |
| O128:H2 (1) | espI-like | espI-like (no PBS of Var-Seq1) | − | espI-like | |
| O128:H2 (1) | espI-like | ND | + | espI-like | |
| O128:H2 (1) | 2f | espI-like | espI-like (nleB-nleA = O103; nleA-ybhB: +1.2 kb) | − | stx2f |
| O128:B12 (1) | 2f | espI-like | espI-like (nleB-nleA = O103; nleA-ybhB: +1.2 kb) | Lysis | |
| O128:H2 (1) | 2f | espI-like | espI-like (nleC-ybhB = O103; nleB-nleC: +1.2 kb) | Lysis | |
| O145:H25 (1) | 2 | z6024 | z6024 (dst nleA ?; nleF-espM: −1.2 kb) | + | stx2 (H) |
| O145:H− (1) | 2 | 3 | 3 | Lysis | stx2** |
| O145:H28 (1) | 1 | 5 | 5 | Lysis | |
| O145:H28 (1) | 1 | 5 | 5 | Lysis | nleA5**, stx1** |
| O145:H28 (1) | 2 | 6-1 | 6-1 | ++ | nleA6-1*, stx2 (H) |
| O145:H28 (1) | 11 | 11 | Lysis | ||
| O145:H− (1) | 1/2 | 2 | 2 | ND | ND |
| O145:H− (1) | 1 | 2 | 2 | − | stx1 |
| O145:H− (1) | 1 | 2 | 2 | + | stx1 (H) |
| O145:H− (1) | 2 | 2 | 2 | ND | ND |
| O145:H− (1) | 2 | 2 | 2 (nleA-yehV: −0.4 kb) | ++ | stx2 (H) |
| O145:H− (1) | 2 | 2 (nleA-yehV: −1.2 kb) | Lysis | ||
| O156:H8 (2) | z6024 | z6024 (O157, nleF-espM = O26) | Lysis | ||
| O156:H25 (1) | 3/8-2 | 3 (nleA-yehV: +0.4 kb)/8-2 (dst nleA ?; ust nleA +1.3 kb) | Lysis | nleA3** | |
| O156:H25 (1) | 1 | 3/8-2 | 3 (lom-nleA: −3.8 kb; nleA-yehV: +0.4 kb)/8-2 (dst nleA ?; ust nleA +1.3 kb) | + | nleA**, stx1 (H) |
| O156:H25 (1) | 1 | 3/8-2 | 3 (lom-nleA: −3.4 kb; nleA-yehV: +0.4 kb)/8-2 (dst nleA ?; ust nleA +1.3 kb) | + | nleA**, stx1 (H: stx1) |
| O156:H25 (2) | 1 | 6-1/8-2 | 6-1 (lom-nleA: +2.1 kb)/8-2 (dst nleA ?; ust nleA +1.3 kb) | − | |
| O156:H25 (1) | 6-1/8-2 | 6-1/8-2 (ND) | Lysis | nleA** | |
| O156:H25 (1) | 6-2/8-2 | 6-2/8-2 (dst nleA ?; ust nleA +1.3 kb) | Lysis | nleA6-2* | |
| O157:H7 (1) | z6024 | z6024 (no nleH or nleF) | − | ||
| O157:H− (1) | z6024 | z6024 (O157, nleF-espM = O26) | + | (H) | |
| O157:H− (1) | 2 | z6024 | z6024 (O157, nleF-espM = O26) | ++ | stx2 |
| O157:H7/− (3) | 2 | z6024 | ND | ++ | stx2 |
| O157:H7/− (2) | 2 | z6024 | z6024 (O157) | ++ | stx2 (H) |
| O157:H7 (1) | 1/2 | z6024 | z6024 (O157) | ++ | stx1/2 (H: stx1**, stx2) |
| O157:H7/− (2) | 1/2 | z6024 | z6024 (O157) | ++ | stx1/2 (H: stx2) |
| O157:H7 (1) | 1/2 | z6024 | z6024 (no nleH or nleF) | Lysis | stx1/2 |
| O157:H7/− (2) | 1/2 | z6024 | ND | ++ | stx1/2 |
Abbreviations: ND, not detected; dst, downstream; ust, upstream; PBS, primer binding site; espI-like, espI-like gene.
Different serotypes are indicated and separated by a slash.
Flanking regions have been determined by a comparison of all obtained PCR products with those of the respective reference strains. Discrepancies from the reference strains are indicated in parentheses.
−, no plaques; (+), <50 plaques with 1 ml phage lysate used in plaque test; +, multiple plaques; ++, >1,000 plaques and partial lysis.
*, few transductants; **, no single transductants, several hybridized positive plaques; (H), result of plaque hybridization identical to result of transduction.
Preparation of phage DNA fragments for sequencing.
For sequencing of the flanking regions of variants nleA6-2 and nleA10, phage lysates were prepared as described earlier (7). Subsequently, DNA of these lysates was prepared as described previously (24). Finally, the phage DNA was dissolved in 100 μl of distilled water, restricted with EcoRI overnight, and analyzed by Southern blot hybridization (7), and nleA-positive DNA fragments were cloned in the vector pBluescriptIIKS+ (Stratagene).
DNA fragments containing variants nleA2, nleA3, nleA5, nleA7, nleA9, and nleA11 were amplified by PCR from genomic DNA of the respective strains with high-fidelity PCR enzyme mix (Fermentas) according to the manufacturer's recommendations for synthesis of PCR fragments of >3 kb. Specific parameters for this PCR were an annealing temperature of 52°C and an elongation time of 10 min. The primers used are listed in Table S1 in the supplemental material. The respective PCR products were cloned using the CloneJet PCR cloning kit (Fermentas).
Sequencing and sequence analysis.
DNA sequencing of both strands was performed with the CEQ 8000 genetic analysis system (Beckman Coulter) using the CEQ Dye Terminator cycle sequencing (DTCS) quick start kit (Beckman Coulter) following the manufacturer's recommendations, or PCR products were sent to and sequenced by the Sequence Laboratory (Seqlab; Göttingen, Germany). The initial sequence information for clones with variants nleA6-2 and nleA10 was obtained with primers universal-forward (5′-ACGACGTTGTAAAACGACGGCCAG-3′) and universal-reverse (5′-TTCACACAGGAAACAGCTATGACC-3′), that for clones with variants nleA2, nleA3, nleA5, and nleA11 was obtained with primers of the CloneJet PCR cloning kit (Fermentas), and that for PCR products of variants nleA7 and nleA9 was obtained with the amplification primers lom-for, yehV-for, VarA-Seq, and V83-rev2. These sequences were used to create customized primers. Each nucleotide was determined at least twice.
The sequences obtained from the raw data were edited and analyzed with BioEdit (10). Searches for open reading frames (ORFs) and predictions of translation start positions were performed with EditSeq and GeneQuest (DNASTAR Lasergene 9.0) and online with GeneMark.hmm (17). Searches for homologous DNA and protein sequences were conducted with BLASTN and BLASTX programs (National Center for Biotechnology Information, NCBI, Bethesda, MD).
PCR techniques.
Amplification of fragments of nleA-flanking regions for verification of specific gene orders was performed in total volumes of 25 μl containing 250 ng of genomic DNA. Genomic DNA was prepared either as described earlier (7) or using the QIAamp DNA mini kit (Qiagen) following the manufacturer's protocol. Furthermore, the PCR mixes consisted of 200 μM each deoxynucleoside triphosphate, 15 pmol of each primer, 1× Taq polymerase buffer, and 1.25 U of Taq DNA polymerase and were adjusted to the final volume with sterile water. PCR programs were used as described previously (5), and specific parameters for annealing temperatures and elongation times are listed in Table S2 in the supplemental material.
Preparation of phage lysates for transduction and plaque assay.
An overnight culture of each E. coli strain was used to inoculate 50 ml of LB broth containing 5 mM MgSO4, following an incubation at 37°C with vigorous shaking until an optical density at 600 nm (OD600) of 0.6 to 0.8 was reached. After the culture was adjusted with 0.5 μg/ml norfloxacin (12.5 mg/ml in glacial acetic acid) and 5 mM MgSO4 was added, the flask with the bacterial suspension was incubated at 37°C and 180 rpm. After 4 h, 1.5 ml of the culture was transferred in an Eppendorf tube and centrifuged at 13,000 rpm for 5 min and 4°C. The supernatant was filtered through a 0.22-μm filter and used for transduction experiments and plaque assays.
Transduction and plaque assay.
For transduction, 100 μl and 200 μl or 10-fold to 1,000-fold dilutions in LB medium of the obtained phage lysate were mixed with a 100-μl bacterial culture of E. coli C600/pACYC184 or C600/pK18 according to the antibiotic resistance of the donor strain (OD600, ∼0.8) and 2 μl of 1 M CaCl2. The cultures were incubated and handled further as described previously (7). For PCR analysis, grown bacterial colonies were either picked or rinsed from the agar plates with 2 ml of a 0.9% NaCl solution. Undiluted and 1:10-diluted aliquots of these bacterial suspensions were used as templates for PCR for the detection of nleA, eae, and the relevant stx variant when required (see Table S3 in the supplemental material).
For the plaque assay, 100 μl and 1 ml of the obtained phage lysate were mixed with a 100-μl bacterial culture of E. coli C600 or C600/pACYC184 (OD600, ∼0.8) and 20 μl of 1 M CaCl2. The assay was performed as described earlier (25).
Plaque hybridization.
For plaque hybridization, 1 to 3 nylon membrane discs (Biodyne A; Pall) were placed for 2 min on precooled agar plates with well-defined plaques from the plaque assay. Afterwards, 1 μl of an nleA or stx PCR product was spotted on the membrane discs as a positive control for the probe. Plaque hybridization was performed with the digoxigenin (DIG) DNA labeling and detection kit (Roche Diagnostics) following the manufacturer's protocol.
DNA probes were generated by PCR amplification (see Table S3 in the supplemental material). The PCR products were purified with the QIAEX II gel extraction kit (Qiagen), and an aliquot of 5 μl was then reamplified, purified with the QIAquick PCR purification kit (Qiagen), and labeled using the DIG DNA labeling and detection kit (Roche Diagnostics).
Nucleotide sequence accession numbers.
The sequences of the regions flanking nleA of various prophages have been deposited in the GenBank database library under the continuous accession numbers JN159726 to JN159733.
RESULTS
Characterization of the regions flanking various nleA variants.
In order to characterize the flanking regions of the nleA variants nleA1 to nleA7 and nleA9 to nleA11, prophages were induced from the corresponding host strains, and DNA was prepared and restricted from the obtained phage lysates. DNA fragments carrying nleA variants and flanking sequences were ligated in the vector pBluescriptIIKS+ for sequencing. This method was successful for NleA-encoding phages in EPEC O156 strain CB8745 (nleA6-2) and EPEC O49 strain CB7690 (nleA10). To obtain template DNA for DNA sequencing of the other variants, the NleA-encoding regions between the genes lom and yehV were amplified by PCR. For strains LTEC94460 and CB7197, harboring variants nleA1 and nleA4, respectively, no PCR products appropriate for DNA sequencing were obtained, and therefore these two variants were excluded from this study. The PCR products of the flanking regions of six nleA variants were ligated in vector pJET1.2/blunt for sequencing. Clones were obtained for variants nleA2, nleA3, nleA5, and nleA11 after amplification with O145 strains 2074/97, 4557/99, CB6242, and DG264/4 as templates. Because cloning amplified fragments from STEC O84 strains CB6403 (nleA7) and 03-00175 (nleA9) was not successful, these PCR products were used directly for sequencing.
DNA regions ranging from 6,194 bp up to 10,269 bp from eight bacteriophage genomes were sequenced. Comparable to known sequences of NleA-encoding phages, the characterized phage genome regions contained lom, tail fiber genes, and also hypothetical genes and the type III effector genes nleA and avrA in combination with ospB, nleH, or nleG. Some of these genes are presumably not functional due to frameshift mutations or the insertion of transposase genes (Fig. 1). Similar to the NleA4795-encoding prophage BP-4795, many of the sequenced regions contain IS elements between lom and nleA. Furthermore, the regions flanking the 3′ and the 5′ ends of the nleA variants studied here showed different characteristics. Whereas the region from the 5′ end of nleA to the end of the respective phage genome was similar to that of prophage BP-4795 in all characterized phage fragments, the region between lom and the 3′ end of nleA was different (Fig. 1). Upstream of the 5′ end of nleA, all determined phage regions included a gene for a hypothetical protein and the type III effector gene avrA. Only some phages carried the gene dinI. Only in phage CP-2074 of E. coli strain 2074/97 was an IS element integrated in avrA (Fig. 1). The highest similarity of the region between gene lom and the 3′ end of nleA to NleA4795-encoding phage BP-4795 was present in the bacteriophages CP-4557, CP-6242, and BP-8745, harboring variants nleA3, nleA5, and nleA6-2. In comparison to BP-4795, they lack one IS629 element but also carry the gene of an unknown protein in putative ISEc8 (ORF79 of BP-4795) and the type III effector gene ospB, also termed ibe (2), followed by an IS629 element. In contrast to phage BP-4795, these three phages harbor additional genes between the IS629 element and nleA. The truncated type III effector gene avrA and the gene of a hypothetical protein are located in this region (Fig. 1). Furthermore, searches for related protein sequences revealed similarities to a truncated putative ferredoxin in this region (data not shown). In contrast to the NleA6-2-encoding phage, the bacteriophages encoding variant NleA3 and NleA5 carry another IS element (Fig. 1). Moreover, the sequenced region between lom and the end of the phage genome of the NleA6-1-encoding phage in strain CB4973 resembles those of the NleA6-2-encoding phage of strain CB8745, with the exception of 31 bp that were different. Twenty-five of these different base pairs are located in the gene nleA (data not shown). The NleA10-encoding prophage BP-7690 also possesses two IS elements similar to the NleA3- and NleA5-encoding phages, but in the region between lom and the 3′ end of nleA10 is located only the type III effector gene nleH1-1, which is disrupted by an IS element in this prophage (Fig. 1). Similar to the sequence of the NleA10-encoding phage BP-7690, the sequences of the NleA2- and NleA7-encoding phages CP-2074 and CP-6403 carry truncated effector gene nleH1-1 followed by a transposase gene but no complete IS elements in the region between the genes lom and the 3′ end of nleA (Fig. 1). Transposase genes and complete IS elements were absent only from the sequenced fragments of the NleA9- and NleA11-encoding phages. In contrast to the other determined sequences, the region between lom and the 3′ end of nleA comprises only a variant of the type III effector gene nleG in these two phages (Fig. 1). Thereby, NleA9-encoding phage CP-00175 harbors nleG2-1, and NleA11-encoding phage CP-264 carries variant nleG2-2 or nleG2-3, similar to prophages harboring nleA variant z6024, the EspI-like gene, or nleA8-1 as well as nleG variant nleG2-1, nleG2-2, or nleG5 adjacent to the 3′ end of nleA. Further substantial similarities regarding the occurrence of type III effector genes between the determined phage fragments and already-known NleA-encoding phages of E. coli strains E22, E2348/69, and EDL933, which harbor four to six different effector genes, are not observable (Fig. 1).
Fig. 1.
Schematic representation of the flanking regions of nleA between the gene lom and the end of the respective phage genomes in different NleA-encoding prophages. Sequences of the reference strains are available under the accession numbers AJ556162 (prophage BP-4795 of E. coli O84:H4 strain 4795/97), AAJV01000058 (E. coli O103:H2 strain E22), AP010953 (E. coli O26:H11 strain 11368), NC_011601 (E. coli O127:H6 strain E2348/69), and NC_002655 (E. coli O157:H7 strain EDL933). A light gray background depicts phage core genome, light gray arrows depict type III effector genes, dark gray arrows depict phage genes, and black arrows depict transposase genes; asterisks depict pseudogenes.
Thus, the flanking regions of nleA in eight determined NleA-encoding phage genome fragments showed similarities mainly to the corresponding region of the NleA4795-encoding prophage BP-4795. Besides IS elements, the type III effector genes avrA, ospB, nleH, and nleG, contained on both strands, were detected in different phage sequences at particularly high levels.
Comparison of the genetic structures of the regions carrying type III effector genes in various NleA-encoding prophages.
By use of PCR, the genetic structures of the regions flanking nleA in 92 E. coli strains of various serogroups were compared with the sequenced phage fragment containing the respective nleA variant of 13 different analyzed bacteriophage regions (Table 1; also see Table S4 in the supplemental material). Some PCR detection, like that for lom-nleA, failed to amplify fragments for some strains used as controls because of the selected elongation time. In some PCR analyses, products were amplified from the strains used as negative controls because these strains harbor some of the determined genes in the genome of nleA-negative prophages or due to unspecific primer binding (see Table S4). For the various nleA alleles, the number of isolates analyzed ranged from 2 isolates, for rarely occurring variants such as nleA5 or nleA11, to 25 strains, for frequently occurring variants such as z6024. PCR products differing from the expected size were not analyzed in detail. The characterization of the flanking regions showed predominantly that those of the determined strains resembled those of the respective reference strain or else that additional DNA was integrated at distinct positions of the regions flanking several nleA variants. Various PCR fragments were approximately 1.2 kb larger than the expected size, indicating the presence of IS elements.
All amplified fragments of the regions flanking the nleA variants nleA4795, nleA5, nleA7, nleA9, and nleA11 were identical to those of the respective reference strains. These five rarely occurring nleA variants were present in two to five strains of predominantly one serogroup (Table 1; also see Table S4 in the supplemental material). The flanking regions of the sequenced variants nleA2, nleA3, and nleA6 did not exhibit such a homogenous pattern. While three out of the eight strains harboring nleA2 resembled the reference strain 2074/97, four isolates revealed partly truncated PCR fragments. The differences were located in the region between nleA and yehV and were due to an IS element integrated in the gene avrA of reference strain 2074/97 (Fig. 1; also see Table S4). In each case, two strains lacked either 1.2 kb or only 0.4 kb and therefore presumably the IS element or parts of the IS element or adjacent DNA. In contrast, the five nleA6-positive strains harbored identical regions between nleA and yehV, but two strains carried 2.1 kb of additional DNA in the region between lom and nleA compared to reference strains CB4973 and CB8745 (Table 1; also see Table S4). The 6.6-kb region between lom and nleA of one out of the four determined nleA3-positive isolates was identical to that of the reference strain, while two strains lacked 3.4 to 3.8 kb of this region and possibly the type III effector genes ospB and avrA. Furthermore, in contrast to reference strain 4557/99, the other three determined nleA3-positive strains seemed to possess the gene dinI, like strains with variant nleA5 (Fig. 1 and Table 1; also see Table S4). Moreover, all phages mentioned above, such as bacteriophage BP-4795 of E. coli O84 strain 4795/97 (6), appear to be integrated in the gene yehV (see Table S4).
Furthermore, the already-described flanking regions of the frequently occurring nleA variants nleA8-1, nleA8-2, z6024, and the EspI-like gene were determined in strains of four to eight different serogroups. Some isolates appeared to lack a primer binding site; aside from this, various divergences in these regions occurred in comparison to the respective reference strains. For the determined 19 nleA8-1-positive strains, the amplified PCR products in seven strains, predominantly of serogroup O26, corresponded to the expected fragments of the nleA8-1 flanking regions in the sequenced genome of E. coli O26 strain 11368 (21). Only one O111 isolate seemed to be identical in terms of the flanking region of nleA8-1 in the sequenced genome of E. coli O111 strain 11128 (21), whereas the other six O111 isolates carried an additional IS element between genes ant and nleG5, which was confirmed by sequencing. For the 16 nleA8-2-positive strains determined, only the PCR analyses with reference strain E2348/69 and one O84 isolate resulted in all expected fragments. The 25 z6024-positive strains determined by PCR included two O49 isolates, determined in a previous study to have two alleles of nleA8-1 (7), which yet appear to harbor besides nleA8-1 one z6024 allele. The seven analyzed z6024-positive O26 strains resembled reference strain 11368 (21), with the exception that the gene order between the core phage genome and the additional acquired type III effector genes could not be identified in most cases. Four O157 strains were identical to O157 reference strain EDL933. PCR analyses of the regions flanking the EspI-like gene in 22 isolates revealed that only eight strains of various serogroups resembled the O103 reference strain E22 (Table 1; also see Table S4 in the supplemental material). Moreover, the following variations of the flanking regions occurred. For several strains positive for nleA8-1 or the EspI-like gene, it was not possible to identify the gene order between the core phage genome and the acquired type III effector genes, but apart from that, the flanking regions of nleA in these isolates resembled those of the respective reference strains. Likewise, the gene order between the core phage genome and nleA8-2 could not be identified in 13 isolates of serogroups O84 and O156. In comparison to reference strain E2348/69, the seven O84 strains possessed 0.4 kb of additional DNA between nleA and nleH, whereas the six O156 isolates harbored 1.3 kb of additional DNA between nleH and espO. One nleA8-2-positive isolate and 10 EspI-like gene-positive strains of different serogroups carried 1.2 to 2 kb of additional DNA at various positions, and 3 z6024-positive isolates harbored 1.3 kb of additional DNA between genes nleA and nleH; also, 1 z6024-positive isolate with an unknown gene order downstream of nleA lacked 1.2 kb upstream of this gene. Two O157 strains lacked nleH and nleF or parts of these genes, and 10 non-O26 z6024-positive strains lacked 3.5 kb; therefore, genes associated with an IS element in the region between nleF and espM, similar to what was seen for O26 strains. Furthermore, the order of the genes flanking nleA8-1 in three isolates of serogroup O49 and O84 strains could not be identified, because most of the PCR analyses failed to amplify fragments (Table 1; also see Table S4).
The investigated strains harbored either prophages with conserved flanking regions of a particular nleA variant or additional DNA, presumably IS elements, which were integrated in these regions.
Mobilization of different nleA variants.
The possibility of mobilizing the region harboring nleA was determined for 100 pathogenic E. coli strains by transduction experiments and subsequent PCR as well as for selected isolates with plaque hybridization experiments. In this collection of 100 strains, 14 isolates, which were already determined with transduction in an earlier study (7), were included as controls or for closer examination.
With the conditions used for the transduction experiments and by subsequent PCR, in 20 E. coli isolates nleA variants could be mobilized and transferred to an E. coli laboratory strain (Table 1; also see Table S4 in the supplemental material). By determination of 10 single colonies for 11 donor strains, 20 to 100% of transductants were detected to be stable for at least three passages on LB agar plates. These strains comprised eight EPEC isolates of serogroups O49, O103, O128, and O156, which predominantly harbored the nleA variant EspI-like gene, and three STEC strains. Whereas the numerous transductants of the two O84 STEC strains were positive for nleA4795 and stx1, the few transductants obtained for the O145 STEC strain were positive for either nleA6-1 or stx2. For nine donor strains, no single nleA transductants could be detected by the analysis of 10 selected colonies; however, after complete agar plates were rinsed, PCR products of nleA were obtained. These nine donor strains comprised two EPEC O156 strains either with alleles nleA6-1 and nleA8-2 or with alleles nleA3 and nleA8-2, as well as seven stx1-positive STEC strains of serogroups O26, O84, O145, and O156, harboring variants nleA3, nleA5, nleA9, and z6024 partly in combination with an nleA8 allele (Table 1; also see Table S4). By sequencing of PCR products obtained for two donor strains harboring nleA3 and nleA6-2, respectively, in combination with nleA8-2, the transferred nleA variants were assigned as nleA3 and nleA6-2. Furthermore, the mobilizations of Stx-converting bacteriophages were determined; apart from stx2f, subtypes of stx1 and stx2 were not analyzed. For 51% of the determined stx1-positive donor strains and 85% of the stx2-positive donor strains, the respective stx subtypes could be transduced to an E. coli laboratory strain with the conditions used in these experiments (Table 1; also see Table S4).
Besides transduction, 22 of the 100 strains were also determined in plaque hybridization experiments to analyze the combined occurrence of nleA and stx in one prophage genome. These strains comprised 7 STEC isolates and 1 EPEC isolate which were able to transfer nleA by transduction, as well as 11 STEC and 3 stx-negative strains not able to transfer nleA as controls for the verification of the transduction experiments. By plaque hybridization, only two nleA4795-positive O84 isolates which harbor nleA in the genome of an inducible Stx1-converting bacteriophage could be identified, and one of these isolates was E. coli strain 4795/97, harboring the already-characterized bacteriophage BP-4795 (6). The plaque hybridization with the other five STEC donor strains with transferable nleA resulted predominantly in plaques positive for stx1 or stx2 and a few plaques positive for nleA. Thereby, the plaques were positive for either stx or nleA. With the four stx-negative strains, the results of the transduction experiments were verified, with one strain being positive for nleA and three strains negative by plaque hybridization. The 11 STEC strains analyzed as controls by plaque hybridization comprised two stx1-positive strains, six stx2-positive strains, and three stx1/2-positive strains. For the stx1-positive and the stx2-positive donor strains, the results of transduction experiments were verified. The three stx1/2-positive O157 donor strains hybridized clearly with the stx2 probe but not or to a minor degree with the stx1 probe (Table 1; also see Table S4 in the supplemental material).
Thus, under the conditions used, a few strains, in particular EPEC strains harboring the nleA variant EspI-like gene, were able to transfer nleA. Furthermore, only variant nleA4795 could be demonstrated to be located in the genome of a Stx prophage.
DISCUSSION
The analysis of the flanking regions of 13 nleA variants indicates the existence of four different variations of these regions with regard to their gene composition. The NleA-encoding prophages which are integrated either between the genes ybhC and ybhB or in the gene yfaT (16, 21) with variants of the EspI-like gene or nleA8-1 show minor similarities to the corresponding regions of other NleA-encoding phages. In contrast, the flanking region of nleA8-2 is probably a truncated version of the respective region of Z6024-encoding phages (Fig. 1), and these two NleA-encoding phages are integrated in the gene ompW (11, 13, 22). NleA-encoding prophages integrated in the gene yehV, which are positive for nleA4795, nleA2, nleA3, nleA5-nleA7, and nleA9-nleA11, possess slightly different flanking regions of this gene, predominantly due to the insertion or deletion of additional DNA and different type III effector genes downstream of nleA. On the one hand, very similar nleA variants with identities of 97%, such as nleA3 and nleA5, or nleA7 and nleA10, show identical gene compositions of their flanking regions with regard to the occurring type III effector genes. Interestingly, on the other hand, some nleA variants with identical compositions of type III effector genes in their flanking regions exhibit lower identities to each other than to nleA variants possessing another type III effector gene in the region downstream of nleA. For example, nleA9 and nleA11 share 87% identity, while their sequenced flanking regions between lom and the end of the phage show over 95% identity. In contrast, nleA9 and nleA11 show respective identities of 92% and 95% to nleA3 and nleA5, which harbor the genes ospB and avrA instead of nleG. Therefore, it is not possible to draw a conclusion from a specific nleA variant to the gene composition of the corresponding flanking region.
It was shown that the occurrence of a specific nleA variant is closely serogroup associated (7). Moreover, 52% of strains examined in this study were already characterized by multilocus sequence typing (MLST) (5). According to the MLST characterization, there are certain nleA variants, such as nleA9 or the EspI-like gene, which occur exclusively in closely related strains that can belong to various serogroups. Such strains are often concentrated in certain evolutionary lineages of EPEC and STEC. In contrast, several nleA variants, such as nleA8-2 and z6024, appear in a number of distantly related strains. The distribution pattern of NleA-encoding phages with different flanking regions of various nleA variants raises the question about their evolution. This pattern suggests the possibility that bacteriophages with the integration site in yehV could originate from one common ancestor. This could also be true for phages containing the nleA variants nleA8-2 and z6024. As suggested by Ogura et al. (21), such ancestors could have undergone diverse genomic rearrangements, such as integration or excision of IS elements potentially linked with certain type III effector genes, resulting in the development of the existing variety of NleA-encoding phages. In addition, strains with very similar flanking regions of certain nleA variants, such as nleA3, nleA5, and nleA6, could harbor largely identical prophages which accumulated diverse mutations only in the gene nleA. Bacteriophages with dissimilar flanking regions of nleA could have evolved independently. The combined occurrence of two NleA-encoding prophages with dissimilar gene compositions downstream of the morphogenesis region and different integration sites in several pathogenic E. coli strains could support this hypothesis. Furthermore, as speculated by Loukiadis et al. (16), tandem integrations of two prophages in the same chromosomal site might have occurred, with the subsequent deletion of one of these prophage genomes resulting in the generation of newly arranged flanking regions of nleA. So, the gene nleA alone or in combination with nleH or nleG could have been integrated in different previously nleA-negative phage genomes. Whereas the transfer en bloc is supposable for several type III effector genes, such as nleB and nleE or nleH and cif (4, 5, 16), nleA does not seem to be a part of a type III effector gene cassette, even though it is present in combination with diverse nleG2 variants in different NleA-encoding prophages.
Because we previously demonstrated the transmission of four nleA variants by horizontal gene transfer, without determination whether these variants were contained in the genomes of Stx phages (7), the analysis of the mobilization of NleA-encoding phages was extended. In contrast to Stx-converting prophages, in particular Stx2 phages, which are often intact and can be spread by horizontal gene transfer, NleA-encoding bacteriophages seem to be predominantly immobilized. However, it could not be ruled out that more NleA-encoding phages can be mobilized under conditions different from those used in this study, for example by induction with different antibiotics or by preparation of more-concentrated phage lysates. No distinct transductants positive for nleA were obtained for 20% and 70% of inducible EPEC and STEC donor strains, respectively. The reason for this observation is presumably the existence of inter-prophage interactions, as described by Asadulghani et al. (1). By inter-prophage interactions, chimeras of two prophages can emerge. If the two bacteriophages use very similar prophage proteins necessary for the lytic life style, this mechanism mostly complements defects in the morphogenetic functions of a defective prophage (1). The detection of stx1 transfer by transduction but not or to a minor degree by plaque hybridization with stx1/2-positive O157 strains is also in agreement with results for E. coli O157 strain Sakai (RIMD050995) obtained by Asadulghani et al. (1), who showed on a small scale the transfer of the defective Stx1 phage to an E. coli laboratory strain. Alternatively, several intact prophages may integrate foreign DNA-like type III effector genes during recombination events in the lytic life cycle. However, none of the minor mobilized nleA variants could be detected in combination with stx in the transduction and plaque hybridization experiments. Only variant nleA4795 is located in the genome of an intact Stx1-converting phage, as already described (6).
The results of our study have shown that the gene nleA is always linked to varying numbers of other type III effector genes, in particular avrA, ospB, nleH, and nleG. Until the present study, it was not clear what selection advantage can be obtained for strains harboring NleA-encoding prophages. The low mobility of nleA demonstrates that in most isolates analyzed, a stable structure of virulence factors has been achieved. Future research is needed to analyze the existence of synergistic functions of type III effectors of NleA-encoding prophages in detail and to elucidate the role of these type III effectors in the ecology of pathogenic E. coli strains.
Supplementary Material
ACKNOWLEDGMENT
This work was supported by grant Schm 1360/3-1 from the Deutsche Forschungsgemeinschaft (DFG).
Footnotes
Supplemental material for this article may be found at http://aem.asm.org/.
Published ahead of print on 14 October 2011.
REFERENCES
- 1. Asadulghani M., et al. 2009. The defective prophage pool of Escherichia coli O157: prophage-prophage interactions potentiate horizontal transfer of virulence determinants. PLoS Pathog. 5:e1000408. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Buss C., Müller D., Rüter C., Heusipp G., Schmidt M. A. 2009. Identification and characterization of Ibe, a novel type III effector protein of A/E pathogens targeting human IQGAP1. Cell. Microbiol. 11:661–677 [DOI] [PubMed] [Google Scholar]
- 3. Chang A. C., Cohen S. N. 1978. Construction and characterization of amplifiable multicopy DNA cloning vehicles derived from the P15A cryptic miniplasmid. J. Bacteriol. 134:1141–1156 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Coombes B. K., et al. 2008. Molecular analysis as an aid to assess the public health risk of non-O157 Shiga toxin-producing Escherichia coli strains. Appl. Environ. Microbiol. 74:2153–2160 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Creuzburg K., et al. 2011. Evolutionary analysis and distribution of type III effector genes in pathogenic Escherichia coli from human, animal and food sources. Environ. Microbiol. 13:439–452 [DOI] [PubMed] [Google Scholar]
- 6. Creuzburg K., et al. 2005. The Shiga toxin 1-converting bacteriophage BP-4795 encodes an NleA-like type III effector protein. J. Bacteriol. 187:8494–8498 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Creuzburg K., Schmidt H. 2007. Molecular characterization and distribution of genes encoding members of the type III effector NleA family among pathogenic Escherichia coli strains. J. Clin. Microbiol. 45:2498–2507 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Garmendia J., Frankel G., Crepin V. F. 2005. Enteropathogenic and enterohemorrhagic Escherichia coli infections: translocation, translocation, translocation. Infect. Immun. 73:2573–2585 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Gruenheid S., et al. 2004. Identification and characterization of NleA, a non-LEE-encoded type III translocated virulence factor of enterohaemorrhagic Escherichia coli O157:H7. Mol. Microbiol. 51:1233–1249 [DOI] [PubMed] [Google Scholar]
- 10. Hall T. A. 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symp. Ser. 41:95–98 [Google Scholar]
- 11. Hayashi T., et al. 2001. Complete genome sequence of enterohemorrhagic Escherichia coli O157:H7 and genomic comparison with a laboratory strain K-12. DNA Res. 8:11–22 [DOI] [PubMed] [Google Scholar]
- 12. Hendrix R. W., Lawrence J. G., Hatfull G. F., Casjens S. 2000. The origins and ongoing evolution of viruses. Trends Microbiol. 8:504–508 [DOI] [PubMed] [Google Scholar]
- 13. Iguchi A., et al. 2009. Complete genome sequence and comparative genome analysis of enteropathogenic Escherichia coli O127:H6 strain E2348/69. J. Bacteriol. 191:347–354 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Kim J., et al. 2007. The bacterial virulence factor NleA inhibits cellular protein secretion by disrupting mammalian COPII function. Cell Host Microbe 2:160–171 [DOI] [PubMed] [Google Scholar]
- 15. Lee S. F., et al. 2008. A C-terminal class I PDZ binding motif of EspI/NleA modulates the virulence of attaching and effacing Escherichia coli and Citrobacter rodentium. Cell. Microbiol. 10:499–513 [DOI] [PubMed] [Google Scholar]
- 16. Loukiadis E., et al. 2008. Distribution, functional expression, and genetic organization of Cif, a phage-encoded type III-secreted effector from enteropathogenic and enterohemorrhagic Escherichia coli. J. Bacteriol. 190:275–285 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Lukashin A. V., Borodovsky M. 1998. GeneMark.hmm: new solutions for gene finding. Nucleic Acids Res. 26:1107–1115 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Mundy R., et al. 2004. Identification of a novel Citrobacter rodentium type III secreted protein, EspI, and roles of this and other secreted proteins in infection. Infect. Immun. 72:2288–2302 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Nataro J. P., Kaper J. B. 1998. Diarrheagenic Escherichia coli. Clin. Microbiol. Rev. 11:142–201 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Ogura Y., et al. 2008. Systematic identification and sequence analysis of the genomic islands of the enteropathogenic Escherichia coli strain B171-8 by the combined use of whole-genome PCR scanning and fosmid mapping. J. Bacteriol. 190:6948–6960 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Ogura Y., et al. 2009. Comparative genomics reveal the mechanism of the parallel evolution of O157 and non-O157 enterohemorrhagic Escherichia coli. Proc. Natl. Acad. Sci. U. S. A. 106:17939–17944 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Perna N. T., et al. 2001. Genome sequence of enterohaemorrhagic Escherichia coli O157:H7. Nature 409:529–533 [DOI] [PubMed] [Google Scholar]
- 23. Pridmore R. D. 1987. New and versatile cloning vectors with kanamycin-resistance marker. Gene 56:309–312 [DOI] [PubMed] [Google Scholar]
- 24. Recktenwald J., Schmidt H. 2002. The nucleotide sequence of Shiga toxin (Stx) 2e-encoding phage φP27 is not related to other Stx phage genomes, but the modular genetic structure is conserved. Infect. Immun. 70:1896–1908 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Schmidt H., Bielaszewska M., Karch H. 1999. Transduction of enteric Escherichia coli isolates with a derivative of Shiga toxin 2-encoding bacteriophage φ538 isolated from Escherichia coli O157:H7. Appl. Environ. Microbiol. 65:3855–3861 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Stavrinides J., McCann H. C., Guttman D. S. 2008. Host-pathogen interplay and the evolution of bacterial effectors. Cell. Microbiol. 10:285–292 [DOI] [PubMed] [Google Scholar]
- 27. Thanabalasuriar A., et al. 2010. The bacterial virulence factor NleA is required for the disruption of intestinal tight junctions by enteropathogenic E. coli. Cell. Microbiol. 12:31–41 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Tobe T., et al. 2006. An extensive repertoire of type III secretion effectors in Escherichia coli O157 and the role of lambdoid phages in their dissemination. Proc. Natl. Acad. Sci. U. S. A. 103:14941–14946 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Wong A. R., et al. 2011. Enteropathogenic and enterohaemorrhagic Escherichia coli: even more subversive elements. Mol. Microbiol. 80:1420–1438 [DOI] [PubMed] [Google Scholar]
- 30. Zhou Z., et al. 2010. Derivation of Escherichia coli O157:H7 from its O55:H7 precursor. PLoS One 5:e8700. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.