The Cloned Locus of Enterocyte Effacement from EnterohemorrhagicEscherichia coli O157:H7 Is Unable To Confer the Attaching and Effacing Phenotype upon E. coliK-12

Cloning and mapping the EHEC LEE.The EHEC O157:H7 LEE was cloned from strain 85-170, which is a Δstx variant of strain 84-289. Both 85-170 and 84-289 secrete higher levels of Esp proteins than do EDL933 and other EHEC strains (7), and we have observed that they are also more active in the fluorescent actin stain (FAS) assay. Prior to cloning, the chromosomal locus was mapped by Southern blotting on digested 85-170 chromosomal DNA and probing with the four EPEC LEE fragments identified as LEE A through D by McDaniel et al. (14). The resulting restriction map (Fig. 1) indicated that the LEE of 85-170 was indistinguishable from that of O157:H7 strain EDL933 and similar to that of EPEC. The 85-170 LEE was cloned following the procedure of McDaniel and Kaper (15). Chromosomal DNA was partially digested with PspAI, an isoschizomer of SmaI that generates cohesive ends, ligated into the cosmid vector pCVD551, and transfected into HB101. Colonies were probed with radiolabeled LEE probes A to D. One cosmid, pJAY1512, contained a 37-kb DNA insert that hybridized to all four EPEC LEE probes. Several partial LEE clones were also obtained. pJAY1512 was mapped with restriction enzymes and LEE probes A through D (Fig. 1). This process confirmed and extended the 85-170 LEE map obtained from chromosomal Southern blotting. We found that the cosmid clone pJAY1512 did not contain the LEE-associated prophage found at the extreme end of the EHEC LEE. In addition to pJAY1512, we also obtained the LEE clone pLEE_O157 (18) from F. Blattner. pLEE_O157 was derived from the chromosome of strain EDL933, a wild-type EHEC O157:H7 strain, by a FLP recombinase-based method (19) and has been entirely sequenced (18). It contains the LEE prophage as well as genes flanking the LEE (18).

Characterization of the cloned EHEC LEE functions.Both pJAY1512 and pLEE_O157 in their respective E. coli K-12 host strains (Table 1) were examined in the FAS assay for A/E lesion formation (11). The FAS assay involves incubation of bacteria on a HEp-2 cell tissue culture monolayer for 3 to 6 h during which time the bacteria form A/E lesions on the surfaces of infected cells. The addition of phalloidin-fluorescein isothiocyanate conjugate, which binds to actin, allows visualization of actin condensed below the bacteria in the A/E lesion. FAS test results were scored blindly by at least one other independent observer and were repeated at least three times. When the cloned EHEC LEE was tested in different K-12 genetic backgrounds, including HB101, W3110, and DH5α, none of the resulting strains was able to form A/E lesions (Table 1) or secrete Esp proteins (results not shown). This finding contrasts with that for the cloned EPEC LEE (Table 1).

We then examined whether the failure of HB101(pJAY1512) to elicit A/E lesions was due to an initial failure to adhere. HB101(pJAY1512) was transformed with plasmids encoding either the afimbrial adhesin AFA-I (13) or pMAR7 (1), which encodes the bundle-forming pili of EPEC as well as other factors, notably the Per transcriptional activator of many EPEC virulence genes (5). The variants expressing AFA-I or bundle-forming pili were much more adherent to HEp-2 cells in the standard adhesion assay (1) (data not shown) but were unable to form A/E lesions.

Clearly, the cloned EHEC O157:H7 LEE is phenotypically different from the cloned EPEC O127:H6 LEE, and this difference is reflected in their sequence divergence. While many genes are highly conserved, other genes diverge quite markedly (18) (Fig. 1). These highly divergent genes may be responsible for the inability of the EHEC LEE to form A/E lesions when it is cloned into E. coli K-12. We hypothesized that complementation of pJAY1512 with plasmids containing regions of the EPEC LEE could restore the A/E phenotype. The ability of cloned EPEC LEE genes to complement mutated EHEC LEE genes in an EHEC background has previously been shown for escN(7), ler (3), and eae(21). Indeed, the cloned EPEC eae gene was able to restore A/E lesion formation to an EHEC Δeae mutant and the resultant complemented strain colonized the intestine in a manner characteristic of EPEC rather than of EHEC (21).

To test for complementation of pJAY1512 with cloned EPEC genes, plasmids containing regions of the EPEC O127:H6 LEE were isolated to ensure that entire operons within hypervariable regions were represented (Fig. 1 and Table 1). These plasmids were transformed into HB101(pJAY1512) and examined by the FAS assay. None of the resulting complemented strains exhibited FAS (Table 1). As a control, we transformed pJAY1512 into CVD451, an escN mutant of EHEC O157:H7 that is unable to form A/E lesions or to secrete Esp proteins (7). The resultant strain, CVD451(pJAY1512), was positive in the 6-h FAS test, indicating that the EHEC LEE clone could rescue the chromosomal escN mutation in an EHEC background.

In summary, our data clearly suggest that the genetic differences between EPEC and EHEC LEE elements are reflected phenotypically when the LEE is cloned into an E. coli K-12 background. While the cloned EPEC LEE was sufficient for A/E lesion formation and Esp secretion in an E. coli K-12 host (15), a similarly cloned LEE from EHEC was inactive in all these assays. This result was confirmed with an independent EHEC LEE clone, pLEE_O157. This clone, which was derived from a different O157:H7 strain in a different laboratory by a different methodology, was also negative for the A/E phenotype as shown by the FAS test. The inability of the EHEC LEE clone to mediate A/E lesion formation was not due to an obvious defect in the cloning of the LEE or the inability of the LEE in the native O157:H7 strain to mediate A/E lesions. The EHEC LEE diverges from its EPEC counterpart in several genes, and these genes are a likely source of phenotypic differences between both the wild-type organisms and their cloned LEEs (18). However, complementation of pJAY1512 with subclones representing the entire EPEC LEE was unable to confer A/E lesion activity. Our data also exclude the possibility that failure to form A/E lesions was due to failure to initially associate with the cell, as transformation of HB101(pJAY1512) with plasmids encoding adhesins resulted in increased adhesion but not the A/E phenotype.

These data raise the question of whether only the LEE of EPEC O127:H6 strain E2348/69 is sufficient for A/E lesion formation in an E. coli K-12 background and would imply that the EHEC LEE requires a factor in trans that is not present in E. coliK-12. Recent work (3, 16, 20) has shown that EPEC and EHEC have differences in regulation and also suggest that they may require different factors in trans, a topic which is currently being investigated in our laboratory.