![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Previous Article | Next Article 
Journal of Clinical Microbiology, August 2007, p. 2498-2507, Vol. 45, No. 8
0095-1137/07/$08.00+0 doi:10.1128/JCM.00038-07
Copyright © 2007, American Society for Microbiology. All Rights Reserved.
Institute of Food Science and Biotechnology, Department of Food Microbiology, University of Hohenheim, 70599 Stuttgart, Germany
Received 5 January 2007/ Returned for modification 24 February 2007/ Accepted 23 May 2007
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
It has been shown that additional effector proteins encoded by genes outside the LEE in cryptic or intact prophages are translocated by the LEE-encoded T3SS. The majority of these effectors have been identified by a proteomics approach with the mouse A/E pathogen Citrobacter rodentium (8) as well as by using bioinformatics, proteomics, and translocation assay approaches with the E. coli O157:H7 strain RIMD 0509952 (37). This group of non-LEE-encoded effectors also includes Cif (24), NleA/EspI (16, 27), TccP/EspFU (3, 15), EspJ (7), NleB (20), and EspK (38). The cycle-inhibiting factor Cif blocks the cell cycle at the G2/M-phase transition and is involved in the formation of stress fibers (24). The Tir cytoskeleton coupling protein TccP/EspFU binds N-WASP and leads to Nck-independent actin polymerization (3, 15). EspJ may play a role in host survival and pathogen transmission (7). NleB is probably a virulence determinant (20), whereas EspK could be involved in intestinal colonization (38).
The non-LEE-encoded effector NleA/EspI of C. rodentium shows 81% identity at the amino acid level to the protein Z6024, encoded by phage CP-933P in E. coli O157:H7 strain EDL933 (30), and 78% and 76% identity, respectively, to NleA4795, which is encoded by the Stx1-converting prophage BP-4795 of E. coli O84:H4 strain 4795/97 (6), and the EspI-like protein, encoded together with Cif by a prophage in the genome of the rabbit EPEC O103:H2 strain E22 (24). The non-LEE-encoded effectors NleA and NleA4795 localize close to the Golgi apparatus of HeLa cells (6, 16). Moreover, experiments with a mouse model showed that NleA/EspI is necessary for virulence (16, 27), but the function of this effector protein is still unknown. Mundy et al. (26) examined 232 EPEC and 93 EHEC strains for the presence of espI using colony hybridizations. They could detect espI in 53% of the LEE-positive EPEC strains tested. In eae-positive EHEC isolates, it was found more frequently. Consequently, 37 of 43 (86%) LEE-positive EHEC strains contained espI, and the authors assumed a correlation between the presence of espI and certain intimin subtypes in EPEC strains. However, it was not possible to define such an association for the occurrence of espI and a specific intimin type in EHEC strains. Furthermore, they could detect espI more commonly in strains from patients suffering from a more severe disease (26).
The aim of the present study was to determine the distribution of nleA4795 and related variants among pathogenic intestinal E. coli strains. Furthermore, we were interested in a possible association and correlation of the presence of nleA variants with serotypes and eae types.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
, 
, 
, and 
was performed as described previously (33, 40). The PCR protocols for the specific detection of other eae subtypes have been modified by increasing the annealing temperature to 67°C (for ß and 
), 52°C/62°C (for 
), and 58°C (for 
) (29, 34, 40). The amplification of nleA variants for subsequent restriction was performed in a total volume of 70 µl. Other PCRs were carried out in a total volume of 50 µl containing 5 to 6 µl of bacterial suspension and containing one to three single colonies in 150 mM NaCl, 200 µM of each deoxynucleoside triphosphate, 30 pmol of each primer, 1x Taq polymerase buffer, and 3 U of Taq DNA polymerase (Genaxxon Bioscience). PCR started with an initial denaturation for 5 min at 94°C, followed by 30 cycles of amplification, consisting of denaturation for 30 s at 94°C, annealing for 60 s at a specific temperature (Table 1), and elongation at 72°C for a certain time course (Table 1). After the last cycle, a final elongation step of 5 min at 72°C completed the reaction.
|
For DNA sequencing, nleA variants were amplified with primer pairs VarA-for/VarA-rev and VarA-for/VarA-rev2. To separate mixed PCR products, the respective product was extracted either from a gel using the QIAquick gel extraction kit (QIAGEN) or by being cloned into the pCR2.1-TOPO vector using the TOPO TA cloning kit (Invitrogen) according to the manufacturers' instructions. DNA sequencing was performed with the CEQ 8000 genetic analysis system (Beckman Coulter) using the CEQ dye terminator cycle sequencing quick-start kit (Beckman Coulter) by following the manufacturer's recommendations. The sequences obtained from the raw data were edited and analyzed with BioEdit (17).
Preparation of genomic DNA. For preparation of genomic DNA, 2 ml of an overnight culture was centrifuged for 10 min at 6,000 rpm, resuspended in 1 ml 0.9% NaCl solution, and centrifuged again for 5 min at 13,000 rpm in a bench-top centrifuge. The bacterial pellet was resuspended in 400 µl STET buffer (233 mM sucrose, 50 mM Tris-HCl [pH 8.0], 20 mM EDTA [pH 8.0], 5% [vol/vol] Triton X-100, and 1.8 mg/ml lysozyme), incubated for 5 min at room temperature, heated for 1 min to 100°C, and immediately cooled on ice for 2 min. After this procedure, 30 µl of 10% (wt/vol) sodium dodecyl sulfate and 3 µl proteinase K (20 mg/ml) (Carl Roth) were added, and the solution was incubated for 1 h at 56°C. After the addition of 3 µl RNase A (100 mg/ml) (Sigma-Aldrich) and 200 µl distilled water, the mixture was incubated for another 30 min at 37°C. Finally, after extraction with an equal volume of phenol-chloroform-isoamyl alcohol (25:24:1), the liquid phase was transferred into a fresh tube. The genomic DNA was precipitated with 2.5 volumes of ethanol and a 1/10 volume of sodium acetate (pH 7.2) for 30 min at –20°C, followed by centrifugation for 30 min at 13,000 rpm and 4°C. The pellet was washed with 70% ethanol and dissolved in 100 µl distilled water.
Southern blot hybridization.
Ten micrograms of genomic DNA was digested overnight with MluI (Fermentas), separated on a 0.8% agarose gel, and transferred to a nylon membrane by vacuum blotting (32). Southern blot hybridization was performed with the DIG DNA labeling and detection kit (Roche Diagnostics) by following the manufacturer's recommendations. The hybridization was conducted at 68°C overnight using an nleA4795 probe. This probe was generated from E. coli strain 4795/97 by PCR with the primer pair V83-for2/V83-rev3 after purification of two gel extracts with the QIAEX II gel extraction kit (QIAGEN) and was labeled with the Klenow fragment of the DNA polymerase included in the DIG DNA labeling and detection kit (Roche Diagnostics). For identification of the size of the obtained DNA fragments, a GeneRuler 1-kb DNA ladder and 
-mix marker 19 (Fermentas) were used.
Phage transduction. An overnight culture of the respective E. coli strain was used to inoculate 200 ml of Luria-Bertani (LB) broth containing 1 ml 1 M CaCl2, followed by an incubation with vigorous shaking until an optical density at 600 nm of 0.8 was obtained. After adjusting the culture with 0.05 to 0.25 µg/ml norfloxacin (25 mg/ml in glacial acetic acid) and addition of 1 ml 1 M CaCl2, the flask with the bacterial suspension was incubated overnight. The phage particles were separated from the cell debris by centrifugation (7,500 x g, 30 min, 4°C), followed by filtration through a funnel filter (Whatman). To remove bacterial nucleic acids, DNase I and RNase A (Sigma-Aldrich) were added to final concentrations of 0.5 µg/ml each. After incubation at 37°C for 45 min, sodium chloride was added to a final concentration of 5.8% (wt/vol) and was dissolved, and the solution was incubated on ice for 1 h. After a centrifugation step (7,500 x g, 10 min, 4°C), the phage particles were precipitated by adjusting the supernatant to 10% (wt/vol) polyethylene glycol 6000. After polyethylene glycol 6000 was dissolved at room temperature, the mixture was incubated on ice for 1 h. Phage particles were harvested by centrifugation (9,500 x g, 30 min, 4°C). The resulting phage pellet was dried at room temperature and dissolved in 1 ml of SM buffer (100 mM NaCl, 8 mM MgSO4·7H2O, 50 mM Tris-HCl [pH 8], and 0.01% [wt/vol] gelatin).
For transduction, 100 µl of different dilutions of the obtained phage lysate was mixed with 100 µl bacterial culture of E. coli C600/pk18 (optical density at 600 nm, 
0.8) and 2 µl 1 M CaCl2. The culture was incubated for 4 h at 37°C without shaking. Afterwards, 4 ml LB medium, 20 µl 1 M CaCl2, and 50 µg/ml kanamycin were added, followed by an incubation for 48 h at 37°C and shaking at 180 rpm. The culture then was centrifuged for 30 min at 3,500 x g and at 4°C, and the pellet was plated on LB agar containing 50 µg/ml kanamycin.
Nucleotide sequence accession numbers. The coding sequences of the variants nleA1 to nleA11 have been deposited in the GenBank database under continuous accession numbers AM421995 to AM422007.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The presence of nleA was determined by PCR with primer V83-for2 in combination with either V83-rev2 or V83-rev3, which is complementary to conserved regions of this gene. Primer V83-rev3 was constructed because it was not possible to amplify a PCR product from all strains with the primer V83-rev2. Total DNA of PCR-negative strains was hybridized with an nleA4795 probe to exclude the possibility that negative PCR results were due to variations in primer-binding sites. Only one PCR-negative rabbit O49:H2 strain was detected with the nleA4795 probe. Members of the nleA gene family were detected in 150 out of 170 strains, and 148 of these isolates carried one of the eae alleles ß, , , , , , and (Table 2). Interestingly, we were able to detect nleA in two eae-negative human isolates with serotypes O103:H2 and O111:H–. nleA genes were absent in two eae -positive human strains of serogroup O145. Moreover, nleA was not detectable in the EPEC strain E2348/69, which expresses intimin 
, in two eae-positive strains of serogroup O49, and in three human eae-positive O156:H8 isolates (Table 2). nleA members were present in the two food isolates but were absent from four eae-negative animal isolates and one eae-positive animal isolate. Moreover, they were found in 119 out of the 135 human isolates, whereas the 15 nleA-negative isolates included 8 eae-positive and 8 eae-negative E. coli strains. In addition, we were able to detect nleA genes in all 48 HUS isolates as well as in 40 out of 48 human strains associated with diarrheal disease (Table 2).
Restriction analyses of the nleA variants. The PCR products obtained with primer pairs V83-for2/V83-rev2 and V83-for2/V83-rev3 did not have the same length. Their sizes ranged from 959 bp to 1,168 bp (Table 1). Two restriction enzymes were chosen for differentiation because of the similarity of nleA4795 and Z6024. Separate restrictions of all PCR products with BseNI (Fig. 1A) and PstI (Fig. 1B) showed 11 different patterns among the 150 PCR products, in addition to the already known variants nleA4795, Z6024, and the espI-like gene of E. coli. The expected restriction pattern for nleA/espI of C. rodentium was not observed. After restriction with BseNI, PCR products obtained with strains 5721/96 and 0917/99, depicted in lanes 4 and 5, respectively, showed the same pattern (Fig. 1), whereas different patterns were obtained by using PstI. Furthermore, PCR products of strains 5721/96, 3439/00, CF11201, and CB6389 in lanes 4, 8, 9, and 10 as well as those of strains PT272 and CB8745 in lanes 6 and 11 share the same PstI restriction pattern (Fig. 1), in contrast to their varying BseNI patterns. Moreover, the addition of molecular weights of BseNI and PstI restriction fragments of strain 1247/95 in lane 13 revealed molecular weights that were approximately double the weights of the other strains. Based on this observation and the arrangement of the restriction fragments, we concluded that this PCR product is a mixture of the two nleA variants shown in lane 2 and lane 8 (Fig. 1).
|
|
DNA fragments were amplified by using primer VarA-for either in combination with VarA-rev, which binds in the region downstream of nleA, or in combination with VarA-rev2, which binds at the 3' end of the gene. For DNA sequencing of PCR products of isolates with two copies of nleA, the following strategies were used. If the PCR product appeared as a single band, it could be cloned directly into the pCR2.1-TOPO vector of the TOPO TA cloning kit (Invitrogen). If two PCR products were amplified, the PCR product with the expected size was extracted from a gel prior to sequencing.
DNA sequencing resulted in the identification of 11 new nleA variants, termed nleA1 to nleA11, besides the three known variants of pathogenic E. coli (Fig. 3). We defined an open reading frame (ORF) with a cutoff value of less than 97% sequence identity at the deduced amino acid level as an individual variant of nleA. By sequencing, two nleA variants were identified that shared a restriction pattern that was the same as that of the espI-like gene. To distinguish variants nleA1, nleA2, and the espI-like gene, PCR products with the primer pair V83-for2/V83-rev2 from all concerned strains were restricted either by Bpu1102I or by CseI. Moreover, many members of variant nleA8-1 (see below) differ in 1 bp in the recognition site of BseNI, resulting in different restriction patterns. This is shown in lanes 4 and 8 in Fig. 1. To distinguish this variant, nleA8-1, from nleA3 and nleA4, which had the restriction pattern shown in lane 4 (Fig. 1), the restriction enzymes BclI, NheI, and SphI were used.
|
Most of the strains of serogroups O26, O84, and O156, as well as three O49:NM isolates from pigs, probably originating from the same farm, and one O103:H2 strain, harbored two copies of the gene nleA (Table 2). Moreover, one copy of nleA was disrupted in isolates of serogroups O26, O84, and O156. Sequencing of variant Z6024 of serogroup O26 revealed the deletion of 1 bp, resulting in a truncated, possibly nonfunctional putative protein. This was also the case for nleA8-1 of the O84:H– isolate that harbored variants nleA8-1 and nleA8-2. Other strains of serogroup O84 harboring nleA7 and nleA8-2 or nleA9 and nleA8-2 showed a disrupted ORF of variant nleA8-2 because of the insertion of the insertion sequence (IS) element ISEc8 651 bp upstream of the 3' end of the gene. In contrast, 117 bp of the 5' end of variant nleA8-2, present in serogroup O156, was missing. The inserted sequence resembled the region upstream of nleA4795 of the prophage BP-4795 and, to some extent, an ISEc8 element. Therefore, 179 bp that was in BP-4795 was missing from this sequence. Furthermore, the first 34 bp of the espI-like gene of one O128:H2 pigeon isolate was deleted due to insertion of an IS element, in contrast to other strains of serotype O128:H2 that harbored a complete ORF of the espI-like gene. Variant nleA4 also may encode a truncated protein because of the insertion of 5 bp located 78 bp downstream of the 5' end of the gene.
Most of the isolates of serogroups associated with severe human disease harbored variant Z6024, nleA8-1, or the espI-like gene (Table 2). Moreover, only one or three different variants of nleA could be detected from the serogroups O26, O157, O103, and O111. In contrast, serogroup O145 appeared to be heterogeneous, with six different variants. Whereas serotype O145:NM was nearly uniform, four variants of nleA were detected in the six O145:H28 isolates examined. This also was the case for serogroups O49, O84, and O156. These serogroups harbored five to six different variants, whereas nleA-positive O128 strains possessed only the espI-like gene. Variants Z6024, nleA8-1, and the espI-like gene occurred most frequently, followed by variants nleA8-2 and nleA2. In contrast, the other nine variants were detected only in one to six isolates (Table 2). Furthermore, most of the eae subtypes ß, , , , and , detected in a larger number of isolates, were associated with four to eight different nleA variants (Table 2). Because of the rare occurrence of most of the nleA variants as well as the association of the espI-like gene, nleA8-1, and Z6024 with at least two different eae subtypes, a direct correlation between a certain variant of nleA and a specific eae subtype could not be defined.
Transduction of nleA variants. Previously, it was shown that nleA4795 and the espI-like gene are located in the genome of prophages, which are fully inducible to produce phage particle progeny, whereas Z6024 is located on a cryptic prophage (6, 24, 30). In order to determine whether the newly discovered variants are within intact phages that are able to spread the nleA-encoded T3SS effector by horizontal gene transfer, transduction experiments were carried out. A collection of 24 pathogenic E. coli strains harboring different nleA variants were chosen as donor strains for these experiments (Table 3). E. coli C600/pK18 was used as the recipient to prevent the growth of single wild-type cells, which were not detached during preparation of phage lysates. The 11 isolates carrying stx2 or stx2f that were examined carried inducible Stx2-converting prophages, which were transducible in E. coli C600/pK18 (Table 3). All strains harboring stx2 belonged to serogroups associated with severe human disease, and most of them originated from patients with HUS. Only one strain, the O145:H28 isolate CB4973, also possessed the variant nleA6-1 in the genome of an inducible prophage. Furthermore, the three isolates 4795/97, 01-08612, and CB6389 of serogroup O84, as well as the O49:NM strain CB7690, obviously harbored functional prophages carrying an nleA variant, which could be transduced in the E. coli K-12 strain C600/pK18 (data not shown). Strains CB7690, 4795/97, and 01-08612 carried variant nleA10 or nleA4795 in the genome of an inducible bacteriophage. Strain CB6389 possessed an intact prophage harboring variant nleA9 that is not disrupted by an IS element, and the similar variant nleA8-2 was located in the genome of this isolate as well. Moreover, each of these three O84 strains exhibited an intact Stx1-converting prophage. This also was the case for the O84:H2 isolate CB7197. However, it was not possible to transduce the variant nleA4 of this isolate in the E. coli strain C600/pK18. No bacteriophages harboring stx1 or a variant of nleA and originating from the other eight stx1-positive or stx-negative strains were detectable after transduction in C600/pK18. Thus, we were able to demonstrate the transduction of five nleA-carrying phages with the method described above. However, we cannot exclude completely the possibility that the other 19 analyzed E. coli strains also harbor inducible nleA phages that could not be detected in this assay.
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
DNA sequencing revealed 15 gene variants of different lengths. Interestingly, a region of 4 to 51 deduced amino acids approximately located in the middle of the encoded deduced protein was lacking in some variants. This region, characterized by the predominant occurrence of alanine, serine, and threonine, includes a putative transmembrane helix. Therefore, in variants of nleA with a deletion of 36 to 51 amino acids, this helix is missing. Thus, these variants exhibit only one putative transmembrane helix, whereas other members of the nleA gene family possess two helices. At present, the role of the number of helices is unknown. Possibly, there are differences in the location or function of the deduced proteins inside eukaryotic cells.
The intimin subtypes seem to be responsible for different host tissue tropisms in the intestine (9). Therefore, they may contribute to the severity of symptoms during a disease caused by EHEC or EPEC. Such an association also was shown for Stx variants. Stx2 often causes more severe disease than those caused by Stx1, whereas differences appear among the heterogeneous members of the Stx2 group (2, 12). Until now, no significant correlation could be determined between the occurrence of a certain nleA variant and the appearance of a specific eae type, stx type, or pathotype. The distribution of different members of the nleA gene family seems to be associated more closely with the serogroup of E. coli strains. Thus, with the exception of E. coli O145 strains, the serogroups associated with more severe disease in humans predominantly contain one or two different nleA variants, whereas less important serogroups contain a larger number of variants (Table 2). The majority of O26, O103, O111, and O157 strains harbor Z6024, the espI-like gene, or nleA8-1. On the other hand, strains of serogroups O49, O84, and O156 harbor a variety of members of the nleA gene family. Therefore, these strains may be depicting a pool for genetic rearrangements.
Whereas Z6024 is harbored by the cryptic prophage CP-933P (30), nleA4795 and the espI-like gene are carried by inducible bacteriophages (6, 24). Transduction experiments also revealed the location of nleA6-1, nleA9, and nleA10 in the genome of inducible phages. T3SS effector protein-encoding genes often are present at one end of bacteriophages, presumably a result of incorrect excision during the lytic life cycle. Moreover, the transduction of bacteriophages that carry a variant of the gene nleA to a laboratory E. coli strain raises the possibility that NleA-converting bacteriophages can be spread by horizontal gene transfer. Some strains harbored two copies of mostly different nleA variants. In the majority of the concerned isolates, one of these copies was deleted by an IS element, or some base pairs were missing. This perhaps results from the ability of bacteriophages harboring a member of the nleA gene family to infect the E. coli strain when parts of another lysogenic phage in the genome of this E. coli strain were destroyed by genetic rearrangements. These results indicate a major role for bacteriophages in the distribution of the members of the nleA gene family.
Recently, it was shown that nleA4795 is harbored together with stx1 on a single bacteriophage (6). Whether this is the case for other bacteriophages, such as BP-01-08612, BP-CB6389, and BP-CB4973, remains to be elucidated.
Interestingly, the nleA variants investigated in this study and those investigated by other authors are linked to phage DNA. This fits with the concept that the specific characteristics of pathogenic E. coli strains are located on mobile genetic elements. To elucidate the role of the phage-encoded type III effectors in more detail, further research is needed.
|
ACKNOWLEDGMENTS |
|---|
This work was supported by grant Schm1360/1-3 from the Deutsche Forschungsgemeinschaft.
|
FOOTNOTES |
|---|
Published ahead of print on 6 June 2007.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Antimicrob. Agents Chemother. | Clin. Microbiol. Rev. |
|---|---|
| Clin. Vaccine Immunol. | ALL ASM JOURNALS |
|---|
