Ileal lesions of 36.4% of patients with Crohn's disease (CD), an inflammatory bowel disease in humans, are colonized by pathogenic adherent-invasive Escherichia coli (AIEC), and high levels of antibodies directed against E. coli OmpC are present in 37–55% of CD patients. We therefore investigated the expression of OmpC and its role in the interaction of CD-associated adherent-invasive E. coli strain LF82 with intestinal epithelial cells. High osmolarity induced a significant increase in the ability of LF82 bacteria to interact with Intestine-407 cells, which correlates with increased OmpC expression. Deletion of ompC gene markedly decreased the adhesion and invasion levels of the corresponding mutant. A LF82-ΔompR mutant impaired in OmpC and OmpF expression, showed decreased adhesion and invasion, and unlike a K-12-negative OmpR mutant did not express flagella and type 1 pili. Interestingly, the wild-type phenotype was restored when OmpC or OmpF expression was induced in the LF82-ΔompR mutant. Overexpression of RpoE in the LF82-ΔompR isogenic mutant restored a full wild-type phenotype without restoring OmpC expression. Increased expression of RpoE was observed in wild-type strain LF82 at high osmolarity. Hence, the role of OmpC in the AIEC LF82 adhesion and invasion is indirect and involves the σE regulatory pathway.
Electron microscopic examination of AIEC reference strain LF82-infected intestinal epithelial cells revealed a macropinocytosis-like process of entry dependent on actin microfilaments and microtubule recruitment, and characterized by the elongation of membrane extensions, which surround the bacteria at the site of contact between entering bacteria and epithelial cells (Boudeau et al., 1999). Type 1 pili-mediated adherence plays an essential role in the invasive ability of strain LF82 by inducing membrane extensions (Boudeau et al., 2001) and nucleotide sequences of fim genes revealed that strain LF82 produces variant of type 1 pili compared with those of E. coli K-12. We also reported that these type 1 pili variants have to be expressed in the genetic background of strain LF82 in order to promote bacterial uptake because their expression in E. coli strain K-12 is not sufficient to confer invasiveness. Flagella also play a direct role in the adhesion-invasion process of AIEC strain LF82 via motility, and an indirect role in the interaction between bacteria and epithelial cells by downregulating the expression of type 1 pili and of unknown factor(s) involved in invasiveness that remain(s) to be investigated (Barnich et al., 2003). In addition, the lipoprotein NlpI, which is probably located in the inner-membrane, is thought to operate in a regulatory pathway involved in the synthesis of flagella, type 1 pili, and other virulence factors yet to be identified (Barnich et al., 2004). We also found that the YfgL lipoprotein is required to confer invasive ability on AIEC strain LF82 independently of type 1 pilus and flagellum expression, but in correlation with release of outer membrane vesicles in which bacterial effectors are entrapped and contribute to the invasion process (Rolhion et al., 2005).
Recent studies have reported an abnormal reactivity to microbial components, including E. coli outer membrane protein OmpC, in patients with CD. High levels of antibodies directed against E. coli OmpC are present in 37–55% of patients with CD, whereas no more than 5% of non-inflammatory bowel disease individuals express anti-OmpC (Arnott et al., 2004; Beaven and Abreu, 2004; Mow et al., 2004). Reactivity to E. coli OmpC is associated with severe CD characterized by small bowel involvement, frequent disease progression, longer disease duration, and greater need for intestinal surgery (Landers et al., 2002; Mow et al., 2004). Thus, the aim of the present study was to investigate the expression of OmpC and its role in the interaction of CD-associated adherent-invasive E. coli strain LF82 with human intestinal epithelial cells.
Increased adhesion and expression of OmpC in AIEC strain LF82 after bacterial growth at high osmolarity
A significant increase in the number of AIEC LF82 bacteria associated with intestinal epithelial cells was observed when bacteria were cultured in the presence of 20 g l−1 NaCl compared with those grown without NaCl (Fig. 1A). The K-12 strain MG1655 showed lower adhesion levels than those of strain LF82. However, relative values of the number of cell-associated bacteria after growth at increased osmolarities compared with growth in the absence of NaCl indicated similar increases in the adhesion levels for K-12 and AIEC LF82 bacteria (Fig. 1A). Analysis of the expression of OmpA, OmpC and OmpF for AIEC LF82 and MG1655 bacteria grown at various osmolarities was performed by Western immunoblotting using anti-OmpA and anti-OmpC/F antibodies. After growth of LF82 or MG1655 bacteria in the presence of increased concentrations of NaCl from 5 to 20 g l−1, OmpA expression was not affected, that of OmpF decreased, and that of OmpC increased (Fig. 1B). As shown by the determination of the ratios OmpC/OmpA (Fig. 1C), a 1.5-fold increases in OmpC expression was observed for the AIEC strain LF82 when bacteria were grown at high osmolarity compared with bacteria grown in the absence of NaCl. For K-12 strain MG1655, a 1.4-fold increase in OmpC expression was observed under such conditions. Comparison of OmpC expression in AIEC strain LF82 and K-12 strain MG1655 showed a 1.8-fold higher expression in bacteria grown at high osmolarity. The levels of ompF and ompC mRNA were measured by real-time reverse transcription polymerase chain reaction (RT-PCR) after growth of the bacteria at low and high osmolarities, 14.0-fold and 14.7-fold decreases in ompF mRNA levels were observed for strains LF82 and MG1655, respectively, with bacteria grown at high osmolarity compared with low osmolarity (Fig. 1D). This confirmed the results of protein analysis. In addition, 2.0-and 2.3-fold increases in the ompC mRNA levels were observed for strain LF82 and strain MG1655, respectively, with bacteria grown at high osmolarity compared with low osmolarity. Taken together, these results indicate that at high osmolarity OmpC expression is increased in the AIEC strain LF82 and in the E. coli K-12 strain MG1655. Likewise, a significant increase in bacterial adhesion after growth of bacteria at high osmolarity was observed. This suggests that increased OmpC at high osmolarity could play a role in the interaction of bacteria with host cells.
Phenotype of LF82-ΔompC and LF82-ΔompF isogenic mutants
LF82 isogenic mutants with the ompC or the ompF gene deleted were constructed. Quantitative adhesion-invasion assays showed that the LF82-ΔompC mutant was consistently reduced in its abilities to adhere to and to invade Intestine-407 epithelial cells, having a 19.6% ± 8.6% residual adhesion level and a 15.5% ± 5.3% residual invasion level compared with wild-type strain LF82, taken as 100% (Fig. 2A). Transcomplementation experiments were performed with the entire ompC gene cloned into the arabinose inducible expression plasmid vector pBAD24, forming pPBI08. Increased adhesion and invasion levels of the LF82-ΔompC isogenic mutant transcomplemented with cloned ompC were observed with increased concentrations of l-arabinose (Fig. 2A). The arabinose-induced expression of OmpC in the LF82-ΔompC isogenic mutant transcomplemented with cloned ompC was confirmed by Western immunoblotting using anti-OmpC/F antibodies (Fig. 2B). Transcomplementation of the LF82-ΔompC isogenic mutant with the vector alone had no effect on its adhesion and invasion levels. In contrast, the adhesion level of the LF82-ΔompF isogenic mutant was 88.9% ± 6.8% of that of strain LF82 (Fig. 2A). Interestingly, the invasion level of the LF82-ΔompF isogenic mutant was increased, reaching 147.6% ± 29.2% of that of strain LF82 (Fig. 2A), and Western-blot analysis of whole bacteria extracts using antibodies raised against OmpC/F indicated an increased expression of OmpC in the LF82-ΔompF isogenic mutant compared with the wild-type strain (Fig. 2B). Thus, the increased invasion of the LF82-ΔompF mutant can be correlated with increased expression of OmpC in the absence of OmpF.
The decreased adhesion and invasion of LF82-ΔompC isogenic mutant are not due to decreased expression of type 1 pili and flagella
As type 1 pili and flagella have been previously shown to play a key role in the ability of strain LF82 to adhere to and to invade Intestine-407 cells (Boudeau et al., 2001; Barnich et al., 2003), we analysed whether the decreased adhesion and invasion of the LF82-ΔompC isogenic mutant were due to defects of type 1 pili and flagella expression. Electron microscopic examination of negatively stained bacteria indicated that only 7% of LF82-ΔompC bacteria were flagellated compared with 71% for the wild-type strain LF82 (Table 1). Concerning type 1 pili expression, 67% of the LF82-ΔompC bacteria were piliated, compared with 99% for the wild-type strain and we observed that when piliated the LF82-ΔompC bacteria produced only a few type 1 pili in comparison to the wild-type strain (Fig. 3A). These results were confirmed by gold immunolabelling (Fig. 3B) and Western immunoblotting with whole-cell extracts (Fig. 3C) using polyclonal antibodies raised against purified type 1 pili. Interestingly, transcomplementation of the LF82-ΔompC isogenic mutant with the cloned ompC gene (plasmid pPBI08) restored flagella and type 1 pili expression to levels similar to those of the wild-type strain LF82 (Table 1, Fig. 3A–C), while transformation with the vector pBAD24 alone did not affect the phenotype. Thus, we further analysed whether defects in type 1 pili expression in the LF82-ΔompC isogenic mutant could be due to phase variation or to the absence of type 1 pili subunit polymerization on the surface of the bacteria. To check the phase variation, we performed PCRs using two sets of primers specific for the phase-ON and phase-OFF orientations of the invertible element (Schwan et al., 1992). The orientation of the invertible element in the wild-type strain LF82 was mostly in phase-ON (Fig. 3D). In contrast, PCR amplification of the LF82-ΔompC isogenic mutant DNA showed both phase-ON and -OFF orientations of the invertible element. Thus, the decrease in type 1 pili expression in the LF82-ΔompC isogenic mutant is not only correlated with a switch of the invertible element to the phase-OFF orientation.
Table 1. Expression of type 1 pili and flagella in wild-type strain LF82 compared with that in ompC or ompR mutants.
Expression of flagella and type 1 pili was monitored by electron microscopic examination of negatively stained bacteria. For each strain, 100 bacteria were counted and data are the means of three separate experiments.
When piliated, the bacteria expressed very few type 1 pili.
Because the absence of OmpC can perturb outer membrane integrity and therefore the biogenesis of fimbriae, we investigated whether the LF82-ΔompC isogenic mutant could express functional type 1 pili on the bacterial surface. The LF82-ΔompC isogenic mutant was transformed with plasmid pPBI01 harbouring the entire fim operon blocked in the phase-ON orientation to force the bacteria to express these pili. As a consequence of induced type 1 pilus expression, 98% of LF82-ΔompC/pPBI01 bacteria were piliated (Table 1), indicating that the decrease in biogenesis of type 1 pili in the LF82-ΔompC isogenic mutant was not directly related to the absence of OmpC in the outer membrane. To investigate the respective roles of type 1 pili, flagella and OmpC in the ability of strain LF82 to adhere to and to invade intestinal epithelial cells, quantitative adhesion and invasion assays were performed with induced type 1 pili expression and forced contact between bacteria and Intestine-407 epithelial cells. LF82-ΔompC isogenic mutant transformed with cloned fim operon still showed reduced abilities to adhere to and to invade intestinal epithelial cells, having, respectively, a 26.6% ± 4.0% residual adhesion and 15.6% ± 5.5% residual invasion levels compared with wild-type strain LF82 (Fig. 4). As the decrease in adhesion and invasion of this mutant may be also related to the absence of bacterial motility, we added a centrifugation step to establish a close contact between bacteria and epithelial cells. We performed experiments with the LF82-ΔfliC isogenic mutant as a control, which is not able to express flagella and type 1 pili and has a greatly reduced ability to adhere to and to invade intestinal epithelial cells. After centrifugation, while the adhesion level of the LF82-ΔfliC transformed with pPBI01 was fully restored to the level of strain LF82, the adhesion levels of LF82-ΔompC/pPBI01 were not higher than those in assays performed without centrifugation (Fig. 4A). In addition, while a centrifugation step did partially restore the invasion level of the LF82-ΔfliC/pPBI01 to the level of strain LF82/pPBI01, it did not increase the invasion level of LF82-ΔompC/pPBI01 (Fig. 4B). Thus, the decreases in adhesion and invasion levels observed for the LF82-ΔompC isogenic mutant were neither related to decreased type 1 pili expression nor to impaired bacterial motility.
The EnvZ/OmpR two-component system, which controls OmpC expression, is involved in the adherent-invasive phenotype of LF82
To better investigate the role of OmpC in strain LF82, we deleted the ompR gene, as OmpR is known to regulate OmpC and OmpF expression in E. coli strains (Hall and Silhavy, 1981a,b). As shown by Western-blot using anti-OmpC/F antibodies, the LF82-ΔompR isogenic mutant like the K-12 strain MG1655-ΔompR mutant had an absence of OmpC and OmpF expression (Fig. 5A). Transcomplementation of the LF82-ΔompR isogenic mutant with cloned ompR gene (plasmid pPBI09) restored OmpC and OmpF expression to levels similar to those observed in the wild-type strain LF82, and transcomplementation of the wild-type strain LF82 with cloned ompR did not increase the expression of OmpC and OmpF (Fig. 5A), nor adhesion and invasion (Fig. 5B). Quantitative adhesion and invasion assays showed that the LF82-ΔompR isogenic mutant was consistently reduced in its ability to adhere to and to invade Intestine-407 cells, having 15.5% ± 2.7% residual adhesion and 13.0% ± 1.5% residual invasion levels compared with wild-type strain LF82, respectively (Fig. 5B). Transcomplementation of the LF82-ΔompR isogenic mutant with the cloned ompR gene (plasmid pPBI09) restored adhesion and invasion levels to levels similar to those of wild-type strain, while transcomplementation with pBAD24 alone did not restore the wild-type phenotype (Fig. 5B). Deletion of the ompR gene in strain LF82 resulted in a high decrease of type 1 pili expression as shown by gold immunolabelling assays and Western-immunoblotting using anti-type 1 pili antibodies (Fig. 6A and B). Electron microscopic examination of negatively stained bacteria indicated that only 39% of LF82-ΔompR bacteria were piliated compared with 99% for the wild-type strain (Table 1). This effect was specific for strain LF82, because deletion of ompR in K-12 strain MG1655 induced increased type 1 pili expression (Fig. 6A and B). Moreover, LF82-ΔompR isogenic mutant showed a loss of motility on swim agar plates (Fig. 6C) and electron microscopic examination of negatively stained bacteria showed that only 21% of LF82-ΔompR bacteria were flagellated (Table 1). In contrast, and as already reported by others for various E. coli K-12 strains (Shin and Park, 1995; Oshima et al., 2002), the MG1655-ΔompR mutant was more motile than the MG1655 wild-type strain (Fig. 6C). In order to investigate whether the decrease in type 1 pili and flagella expression were involved in the decreased adhesion and invasion of the LF82-ΔompR mutant, we performed experiments with induced type 1 pili expression and forced contact between bacteria and Intestine-407 cells. The LF82-ΔompR transformed with the cloned fim operon (pPBI01) expressed type 1 pili and 90% of LF82-ΔompR/pPBI01 bacteria were piliated (Table 1). After centrifugation, the adhesion and invasion levels of LF82-ΔompR/pPBI01 were higher than those observed in assays performed without centrifugation, however, the adhesion and invasion levels reached only 27.9% ± 5.7% and 27.6% ± 6.8% of those of wild-type strain, respectively (Fig. 5B). Thus, the decreases in adhesion and invasion levels observed for the LF82-ΔompR isogenic mutant were neither related to decreased type 1 pili expression nor to a defect in bacterial motility, in contrast with a LF82-ΔfliC mutant for which, as shown above, induced type 1 pili expression and forced contact between bacteria and epithelial cells fully or partially restored the levels of adhesion and invasion, respectively (Fig. 4).
Induced expression of OmpC or of OmpF is able to restore adhesion, invasion, and flagella and type 1 pili expression in LF82-ΔompR mutant
Transcomplementation of the LF82-ΔompR isogenic mutant with the cloned LF82 ompC gene (plasmid pPBI08) restored type 1 pili expression as shown by gold immunolabelling assays and Western-immunoblotting using polyclonal antibodies raised against purified type 1 pili (Fig. 6A and B). As shown in Table 1, 100% of the LF82-ΔompR/pPBI08 bacteria expressed type 1 pili. Interestingly, the induced expression of the ompC gene in the LF82-ΔompR isogenic mutant also restored flagella expression and motility (Fig. 6C and Table 1). In addition, the induced expression of OmpC in the OmpR-negative mutant restored the ability to adhere to and to invade Intestine-407 cells. The adhesion and invasion levels of LF82-ΔompR transcomplemented with cloned ompC were 87% ± 18.2% and 75.0% ± 15.0%, respectively, of those of strain LF82 (Fig. 7A) and similar results were also observed with polarized T84 intestinal epithelial cells (Fig. 7B). Thus, induced expression of OmpC in the LF82-ΔompR isogenic mutant restored a wild-type phenotype.
The induced expression of the ompF gene cloned in the pBAD33 expression vector (pPBI12) in the LF82-ΔompR isogenic mutant fully restored type 1 pili and flagella expression (Table 1) and the ability to adhere to and to invade (Fig. 7). This indicated that the adherent-invasive phenotype was linked to the expression of outer membrane proteins OmpC or OmpF. However, the role of OmpF in wild-type AIEC LF82 is limited because OmpF is not expressed at high osmolarity and because a LF82-ΔompF isogenic mutant was not affected in its ability to adhere to and to invade intestinal epithelial cells.
The role of OmpC in adhesion and invasion of AIEC LF82 is indirect and involves the σE regulatory pathway
In E. coli, overproduction of outer membrane protein or misfolding proteins in the outer membrane or in the periplasm activate the alternative sigma factor σE or the two-component regulatory system CpxRA (Raivio and Silhavy, 1999). We checked the involvement of these regulatory systems in the restoration of virulence of the LF82-ΔompR isogenic mutant transcomplemented with cloned ompC gene (plasmid pPBI08). After growth of the bacteria in Luria–Bertani (LB) medium in presence of l-arabinose, the cpxR and rpoE mRNA levels were measured by real-time RT-PCR using primers described in Table 2, because it is well known that CpxR and RpoE upregulate their own transcription. The levels of cpxR mRNAs in the LF82-ΔompR isogenic mutant with induced expression of the ompC gene were not different from those of the wild-type strain LF82 or those of the LF82-ΔompR isogenic mutant (Fig. 8A). In addition, as shown in Fig. 8B, the levels of adhesion and invasion of a LF82-ΔcpxR isogenic mutant were not significantly different from those of the wild-type strain LF82. These results indicate that the two-component regulatory system CpxRA is not involved in the adhesion and invasion abilities of AIEC strain LF82. In contrast, the levels of rpoE mRNAs in the LF82-ΔompR isogenic mutant with induced expression of the ompC gene were greater than those of the wild-type strain LF82 or those of the LF82-ΔompR isogenic mutant (Fig. 8A), indicating that the σE-regulatory pathway was activated in the LF82-ΔompR mutant when OmpC was overexpressed. As previously reported, the RpoE factor may be involved in LF82 bacterial viability, since repeated attempts to isolate LF82-ΔrpoE isogenic mutant failed (Bringer et al., 2005). In order to analyse the role of RpoE in the virulence of AIEC strain LF82, we induced in the LF82-ΔompR isogenic mutant increased expression of the rpoE gene cloned in the pBAD24 expression vector, because we speculated that all RpoE overexpressed proteins would not be sequestered to the inner membrane and therefore could activate σE-gene transcription. Interestingly, the increased expression of RpoE in the LF82-ΔompR isogenic mutant obtained with increased concentrations of l-arabinose led to restoration of adhesion and invasion (Fig. 8C), irrespective of the OmpC levels because no OmpC expression was still observed with increased σE expression (Fig. 8D). The adhesion and invasion levels of the LF82-ΔompR isogenic mutant were restored with induced expression of RpoE but not in the presence of similar concentrations of l-arabinose (Fig. 8C). Hence, different genes may be involved in adhesion or in invasion and their expression is variously controlled by RpoE. In addition, we observed that the induced expression of the rpoE gene in the OmpR-negative mutant restored the type 1 pili and flagella expression to levels similar to those observed for the wild-type strain (Table 1). Thus, in the absence of OmpC, overproduction of RpoE in the LF82-ΔompR isogenic mutant fully restored the adherent-invasive phenotype.
Table 2. Oligonucleotides used for PCR and RT-PCR experiments.
High osmolarity activates the σE regulatory pathway
Comparative analysis of the ability of strain LF82 and the related omp isogenic mutants to interact with intestinal epithelial cells was performed after growth of bacteria in LB broth at low or high osmolarity (Fig. 9A). At low osmolarity, the ompR and ompC mutants display low adherence rates and the ompF mutant behaves like the wild-type strain. At high osmolarity, a significant 1.6-fold increase in the number of associated bacteria, similar to that seen in the wild-type strain LF82, was observed for the LF82-ΔompF isogenic mutant. Such increases were not observed for LF82-ΔompR and -ΔompC mutants. Growth of LF82 bacteria at high osmolarity induced, in addition to the increase in ompC mRNA levels as shown in Fig. 1D, a significant (1.7-fold) increase in ompR mRNA levels (Fig. 9B). Similarly the rpoE mRNA level increased 2.8-fold after growth of the LF82 bacteria at high osmolarity (Fig. 9C). In contrast, the degS mRNA levels were similar after growth of the LF82 bacteria at low and high osmolarities (Fig. 9B). The determination of rpoE mRNAs levels in the LF82-ΔompC, -ΔompF and -ΔompR isogenic mutants after growth of bacteria at low and high osmolarities indicated that at low osmolarity the levels of rpoE mRNAs in the various omp mutants were not different from those of the wild-type strain LF82 (Fig. 9C). Thus, decreases in adhesion and invasion levels of the ompR and ompC mutants at low osmolarity can not be explained by a decrease in RpoE expression. This suggests that under low osmolarity conditions, OmpC plays a direct or indirect role in AIEC adhesion and invasion, independently of the RpoE regulatory pathway. In contrast, at high osmolarity, a 3.1-fold significant increase in the rpoE mRNA levels was observed for the LF82-ΔompF mutant, while no increase was observed for LF82-ΔompC and -ΔompR mutants.
The aim of the present study was to investigate the OmpC expression and the role of OmpC in CD-associated adherent-invasive E. coli strain LF82 under conditions of high osmolarity encountered by bacteria in the gastrointestinal tract (Fordtran and Ingelfinger, 1968; Chowdhury et al., 1996) because it has been reported that passage of bacteria from the stomach into the small intestine represents an osmotic upshift that in many food-borne pathogens, serves to trigger the expression of genes that are necessary for survival and colonization (Nikaido and Rosenberg, 1983; Foster and Spector, 1995). Increased concentrations of NaCl increased OmpC expression in AIEC LF82. This was also observed for the non-pathogenic E. coli strain MG1655, confirming previous reports with various K-12 strains, for reviews (Forst and Inouye, 1988; Mizuno and Mizushima, 1990; Pratt et al., 1996). The high OmpC expression in AIEC strain at high osmolarity, together with the abnormal ileal colonization with AIEC bacteria as previously reported (Darfeuille-Michaud et al., 1998), could explain in part why high levels of anti-E. coli OmpC are observed in CD patients (Landers et al., 2002; Mow et al., 2004). Increased bacterial adhesion to intestinal epithelial cells was also observed when AIEC LF82 bacteria were grown at high osmolarity, and therefore we investigated whether OmpC could be involved in the interaction of AIEC LF82 bacteria with host cells.
A LF82-ΔompC mutant showed decreases in its abilities to adhere to and to invade, indicating a possible role of OmpC as an adhesin or an invasin in AIEC bacteria. Conflicting reports exist regarding the role of OmpC in the virulence of pathogenic bacteria. A Shigella flexneriΔompC mutant was considerably impaired in its ability to invade HeLa and Caco-2 intestinal epithelial cells (Bernardini et al., 1993) and mutants of an enterotoxigenic E. coli strain deleted for ompC and ompF, in addition to deletion of aroC, are attenuated in virulence in humans (Turner et al., 2001). However, in Salmonella enterica serovar Typhimurium, no significant difference in binding to intestinal epithelial cells was observed between an OmpC-negative mutant and the wild-type strain (Hara-Kaonga and Pistole, 2004). The decreased adhesion and invasion levels of the OmpC-negative LF82 mutant could result from the decrease in type 1 pili expression, as already reported for LF82 mutants deficient for flagellar biogenesis or lipoprotein NlpI expression (Barnich et al., 2003; 2004). Induced expression of type 1 pili in the LF82-ΔompC isogenic mutant did not restore the ability of the OmpC null mutant to adhere to and to invade intestinal epithelial cells, even after a forced contact between bacteria and host cells. This indicates that OmpC acts independently of type 1 pili and flagella expression, and perhaps has a role per se or regulates the expression of other factors involved in adhesion and invasion processes.
Co-regulated expression of the outer membrane proteins OmpC and OmpF exists in E. coli and involves the two-component OmpR–EnvZ regulatory system (Hall and Silhavy, 1981a,b). We observed that deletion of ompR gene in AIEC strain LF82, as in non-pathogenic E. coli K-12 strain MG1655, induced absence of OmpC and OmpF expression. Interestingly, the phenotypes of the negative OmpR mutants engineered from AIEC strain LF82 or K-12 strain MG1655 were different. The LF82-ΔompR mutant synthesized very small amounts of flagella and type 1 pili, and deletion of ompR in K-12 strain MG1655 resulted in increased motility and type 1 pili expression. This indicates that OmpR functions differently in pathogenic AIEC and commensal bacteria. It has been previously reported that in E. coli K-12 OmpR negatively regulates the master flagellar regulatory operon flhDC (Shin and Park, 1995; Pruss, 1998; Oshima et al., 2002) but in Salmonella enterica serovar Typhimurium inactivation of ompR does not affect flhDC expression (Kutsukake, 1997). The LF82-ΔompR isogenic mutant showed decreased abilities to adhere to and to invade intestinal epithelial cells Intestine-407 and T84. As in results obtained with an OmpC null mutant, induced type 1 pili expression and forced contact between bacteria and intestinal epithelial cells did not restore the ability of the OmpR mutant to adhere to and to invade. Interestingly, the expression of the cloned ompC or ompF gene in the LF82-ΔompR isogenic mutant restored flagella and type 1 pili expression, as well as the ability of the bacteria to adhere to and to invade intestinal epithelial cells. One explanation could be that overexpression of OmpC or OmpF in a ΔompR mutant context could transduce outer membrane or periplasmic stress signals to the cytoplasm. However, in a wild-type strain context and at high osmolarity, we can speculate that the expression of flagella, type 1 pili and interaction with host cells is linked to high OmpC levels because we observed a quasi absence of OmpF and an increase in OmpC. This hypothesis is reinforced by the fact that an OmpF null mutant showed an increased invasion level, together with increased OmpC expression. However, the phenotypes of the OmpC mutant expressing OmpF and the OmpR mutant lacking OmpC with induced overexpression of OmpF were different. A possible explanation is that the induced overexpression of OmpF in the ompR mutant can lead to envelope perturbations and thus to outer membrane or periplasmic stress signals sufficient to promote AIEC adhesion and invasion via RpoE activation.
The Cpx and σE extracytoplasmic stress responses sense and respond to misfolded or overexpressed proteins in the bacterial envelope (Mecsas et al., 1993; Danese et al., 1995; Raina et al., 1995; Raivio and Silhavy, 1999). Overexpression of OmpC in the LF82-ΔompR mutant did not modify the level of cpxR mRNA. As activation of the Cpx regulatory pathway is known to induce cpxRA operon transcription (Raivio and Silhavy, 1999), we hypothesized that the CpxRA regulatory pathway is not involved in the OmpC overexpression-induced responses in AIEC strain LF82. This was confirmed because a CpxR-negative mutant of strain LF82 had a similar phenotype as the wild-type strain. In E. coli K-12, overproduction of the various outer membrane proteins including OmpF and OmpC causes an increase in σE activity (Mecsas et al., 1993; Grigorova et al., 2004). Likewise, the overproduction of OmpC activates the σE pathway, because we observed that induced expression of OmpC in LF82-ΔompR isogenic mutant led to increased rpoE mRNA level. Experiments performed to create a LF82-ΔrpoE isogenic mutant to verify the role of RpoE in AIEC LF82 have failed. Interestingly, induced expression of RpoE increased the abilities of the LF82-ΔompR mutant to interact with host cells and restored flagella and type 1 pili expression. In addition, an increase in rpoE mRNA levels was observed in AIEC strain LF82 after growth of bacteria at high osmolarity together with increases in ompR and ompC mRNA levels. It has been reported that overexpression of OmpC directly activates protease DegS, which then cleaves RseA and thereby activates RpoE regulatory pathway (Ades et al., 1999; Alba et al., 2002; Walsh et al., 2003). No increase in degS mRNA levels were observed when LF82 bacteria were grown at high osmolarity, indicating that activation of the RpoE regulatory pathway in strain LF82 does not need any increase in the DegS protease level. After growth of the bacteria at high osmolarity, no increase in rpoE mRNA levels was observed for LF82 OmpR and OmpC null mutants, as well as no increase in the ability of these mutants to adhere to intestinal epithelial cells. All together this indicates a link between the OmpC level, the RpoE level and bacterial interaction with host cells. This hypothesis is reinforced by the fact that at high osmolarity the behaviour of the LF82 OmpF null mutant, in which we observed an increase in OmpC expression, was similar to that of the wild-type strain LF82.
The σE regulon encodes many pathogen-related functions and this explains why cells lacking σE are defective in pathogenesis. Rhodius et al. performed a promoter prediction model for E. coliσE, which enabled them to predict a total of 89 unique σE-controlled transcription units in E. coli K-12 and eight related genomes (Rhodius et al., 2006). This study and others have shown that in addition to transcribe genes encoding chaperones and proteases targeted to the cell envelope that will degrade overexpressed proteins, σE transcribes genes that ensure the synthesis, assembly and homeostasis of lipopolysaccharide and outer membrane proteins (Dartigalongue et al., 2001; Rezuchova et al., 2003; Rhodius et al., 2006). In addition, the σE regulon encodes multiple functions related to pathogenesis (Humphreys et al., 1999; Cano et al., 2001; Manganelli et al., 2001; Kovacikova and Skorupski, 2002; Testerman et al., 2002; Bang et al., 2005; Rhodius et al., 2006). Interestingly among the genes regulated or predicted to be regulated by σE in E. coli, the yfgL gene is upregulated when RpoE is overexpressed (Onufryk et al., 2005; Rhodius et al., 2006). We previously reported that, in the AIEC strain LF82, YfgL is required for invasive ability in relation to outer membrane vesicle release (Rolhion et al., 2005). In addition, it has been recently reported that in E. coli DH5α a link between outer membrane vesiculation and σE-pathway could exist (McBroom et al., 2006). But among the σE-regulated genes predicted in the various genomes (Rhodius et al., 2006), we found none that could be involved in type 1 pili or flagella expression. Restoration of type 1 pili and flagella expression in LF82-ΔompR isogenic mutant was observed when RpoE was overexpressed. This may indicate that one or several intermediates, whose transcription is σE-dependent, are involved in the regulation of type 1 pili and flagella in AIEC strain LF82. The whole genome of strain LF82 is currently sequenced and in order to identify all AIEC σE-regulated genes, we will subsequently search for such intermediates by performing a promoter prediction model for σE.
In summary, we propose a model for the involvement of OmpC in the interaction of AIEC bacteria with intestinal epithelial cells under conditions of high osmolarity similar to that of the gastrointestinal tract (Fordtran and Ingelfinger, 1968; Chowdhury et al., 1996). At high osmolarity, increased expression of OmpC in AIEC LF82 bacteria and activation of the σE regulatory pathway were observed. This can modulate flagella and/or type 1 pili encoding gene expression but also the expression of genes encoding other yet unidentified virulence factors also involved in AIEC interactions with host cells (Fig. 9D). In addition, activation of the RpoE regulatory pathway can bypass the effect of OmpC, as shown in the LF82 OmpR mutant. Our results rose also the question of whether OmpC is really involved as adhesin or invasin in the interaction of some pathogenic bacteria with host cells as already reported by others, or whether its role is linked to the σE regulatory pathway. We also showed that flagella and/or type 1 pili encoding gene regulation involving the EnvZ/OmpR two-component system is opposite in AIEC strain LF82 and in a K-12 strain. This could indicate that AIEC bacteria have evolved from non-pathogenic to pathogenic bacteria by elaborating intestinal environment adaptation mechanisms for which high osmolarity constitutes key signals to activate the expression of virulence genes.
Bacterial strains, plasmids and cell lines
Strain LF82 was isolated from chronic ileal lesion of a patient with CD and belongs to E. coli serotype O83:H1. It adhered to and strongly invaded Intestine-407, HEp-2 and Caco-2 cells (Boudeau et al., 1999). E. coli strain JM109 was used as host strains for cloning experiments. Bacterial strains and plasmids used in this study are listed in Table 3. Bacteria were grown routinely in LB broth and in M9 minimal media supplemented with 1 mM MgSO4, 0.2% carbon source (glucose or l-arabinose) and 0.00005% thiamine without shaking or on LB agar plates (Institut Pasteur Production) overnight at 37°C. Plasmid vectors pBAD24 and pBAD33 were used for cloning procedures (Guzman et al., 1995). Antibiotics were added to media at the following concentrations: ampicillin (50 μg ml−1), kanamycin (50 μg ml−1) and chloramphenicol (25 μg ml−1). Intestine-407 cells (derived from human intestinal embryonic jejunum and ileum) and T84 (derived from human colon carcinoma) were purchased from Flow Laboratories and cultured according to the manufacturer's protocols.
Table 3. Bacterial strains and plasmids used in this study.
Strain or plasmid
Source or reference
E. coli isolated from an ileal biopsy of a CD patient
pBAD24 harbouring the 1.1kb HindIII–SalI fragment with the entire ompC gene of strain LF82
pBAD24 harbouring the 0.7kb HindIII–SalI fragment with
the entire ompR gene of strain LF82
pBAD24 harbouring the 0.6kb HindIII–SalI fragment with the entire rpoE gene of strain LF82
pBAD33 harbouring the 1.1kb HindIII–SalI fragment with the entire ompF gene of strain LF82
Adhesion and invasion assay
The bacterial invasion was performed using the gentamicin protection assay as described previously (Boudeau et al., 1999). Briefly, Intestine-407 and T84 cells were seeded in 24-well tissue culture plates with 4 × 105 and 2 × 105 cells per well respectively. Monolayers were then infected at a multiplicity of infection of 10 bacteria per cell in 1 ml of the cell culture medium without antibiotics and with heat-inactivated fetal calf serum (FCS) (Biowhittaker Cambrex Compagny, Verviers, Belgium). After a 3 h incubation period at 37°C, monolayers were washed three times in phosphate-buffered saline (PBS, pH 7.2). The epithelial cells were then lysed with 1% Triton X-100 (Euromedex, Mundolsheim, France) in deionized water. Samples were diluted and plated onto Muller-Hinton agar plates to determine the number of colony-forming units (cfu) corresponding to the total number of cell-associated bacteria (adherent and intracellular bacteria). To determine the number of intracellular bacteria, fresh cell culture medium containing 100 μg ml−1 gentamicin was added for 1 h to kill extracellular bacteria. Monolayers were then lysed with 1% Triton X-100, and bacteria were quantified as described above. When needed, the infected monolayers were centrifuged for 10 min at 1000 g before the 3 h infection period.
Bacterial strains were grown overnight at 37°C without agitation on LB broth and 4 μl of the culture were inoculated into the centre of a 0.3% LB agar plates. The plates were incubated at 37°C for 6 h, and motility was assessed qualitatively by examining the circular swim formed by the growing motile bacterial cells.
Transmission electron microscopy
Negative staining. Bacteria were grown overnight in LB broth containing 10 g l−1 NaCl at 37°C without shaking, placed for 1 min on carbon-Formvar copper grids (Electron Microscopy Sciences, Hatfield, England) and negatively stained during 1 min with acid phosphotungstic pH 6.0. Grids were examined with Hitachi H-7650 transmission electron microscope.
Immunolabelling. Gold immunolabelling was performed by the method of Levine (Levine et al., 1984). A drop of bacteria grown overnight in LB broth at 37°C without shaking was placed on carbon-Formvar copper grids. Excess liquid was removed and the grid was placed face down on a suitable dilution of antiserum raised against type 1 pili for 15 min. After 30 washings in wash solution (PBS + 1% bovine serum albumin and 1% Tween 20), the grid was placed on a drop of gold-labelled goat anti-rabbit serum (BB International, Cardiff, UK) for 15 min. After a further thorough washing, the grid was negatively stained with acid phosphotungstic pH 6.0.
Construction of isogenic mutants and transcomplementation assays
Isogenic mutants were generated with a PCR product using the method described by Chaveroche et al. (Chaveroche et al., 2000). Briefly, the strategy was to replace a chromosomal sequence with a selectable antibiotic resistance gene (kanamycin) generated by PCR. This PCR product was generated by using primers with 50-nt extensions that are homologous to regions adjacent to the gene to delete and template E. coli strain carrying a kanamycin resistance gene (Table 2). In addition, strain LF82 was transformed with pKOBEG, a plasmid encoding Red proteins from phage λ under the control of a promoter inducible by l-arabinose. These proteins protect linear DNA from degradation in bacteria. The plasmid was maintained in bacteria at 30°C with 25 μg ml−1 of chloramphenicol and was suicided at 42°C.
Strain LF82/pKOBEG was grown at 30°C with 1 mM l-arabinose to induce Red protein expression. When OD620 reached 0.5, the bacterial culture was incubated for 10 min at 42°C in order to suicide the plasmid. Bacteria were washed three times with 10% glycerol, and PCR products were electroporated. The isogenic mutants were then selected on LB agar containing 50 μg ml−1 kanamycin. The replacement of the gene by the kanamycin resistance cassette in each isogenic mutant was confirmed by PCR.
A 1142, 754 and 615 bp PCR product obtained using, respectively, primers OmpC SalI/OmpC HindIII, OmpR SalI/OmpR HindIII and RpoE SalI/RpoE HindIII containing, respectively, the entire ompC, ompR and rpoE gene were cloned into the pBAD24 vector and designated, respectively, pPBI08, pPBI09 and pPBI10 (Tables 2 and 3). A 1149 bp PCR product obtained using OmpF SalI/OmpF HindIII containing the entire ompF gene was cloned into the pBAD33 vector and designed pPBI12 (Tables 2 and 3). These construct were used to transform the wild-type strain LF82, LF82-ΔompC or LF82-ΔompR isogenic mutant.
Growth of the wild-type strain LF82, LF82-ΔompC, LF82-ΔompF or LF82-ΔompR isogenic mutant was performed at 37°C in the bacterium-cell incubation medium used for adhesion and invasion experiments (Eagle minimal essential medium cell culture medium supplemented with 10% heat-inactivated FCS). The growth curves for wild-type strain LF82 and the three isogenic mutants were similar at all time points (data not shown).
Protein preparation, SDS-PAGE and Western immunoblotting analysis
Expression of OmpC, F and A was analysed with whole-cells extracts. After an overnight incubation at 37°C in LB broth with or without NaCl, bacteria were centrifugated and resuspended in SDS-PAGE loading buffer (2% SDS, 50 mM Tris-HCl pH 6.8, 12.5% glycerol, 400 mM β-mercaptoethanol and 0.01% Bromophenol Blue). For outer membrane preparations, cells were recovered and treated as previously described (Pugsley and Schnaitman, 1979). Protein preparations were heated for 5 min at 95°C and separated by SDS-10% or -12% PAGE in presence or not of urea 6 M. Expression of type 1 pili was analysed with whole-cell extracts. After on overnight incubation at 37°C in LB broth or M9 without shaking, bacteria at the same OD were centrifuged and resuspended in SDS-PAGE loading buffer and heated for 5 min at 95°C. Protein preparations were acidified with HCl, heated for 5 min at 95°C and separated by SDS-12% PAGE. Western immunoblotting was performed according to the procedure of Towbin (Towbin et al., 1979). Proteins were electroblotted onto nitrocellulose membranes (Amersham International) and the membranes were reacted with the rabbit antiserum against OmpC/F (diluted 1:1000), OmpA (diluted 1:10 000), E. coli type 1 pili (diluted 1:750). Immunoreactants were detected using horseradish peroxydase anti-rabbit IgG antibody (diluted 1:10 000), ECL reagents (Amersham International) and autoradiography. OmpC and A were quantified by densitometry (Scion Image).
RNA manipulations, reverse transcription and RT-PCR
Cultures were grown at 37°C in LB with or without NaCl to OD 0.2 at 620 nm and when needed, l-arabinose (0.2%) was added to induce the overexpression of OmpC. Total RNAs were extracted from bacteria at OD 0.4 at 620 nm and treated with DNase (Roche Diagnostics, Manheim, Germany) to remove any contaminating genomic DNA. The RNAs were reverse transcribed and amplified using specific primers to cpxR, rpoE, ompR, ompC, ompF and degS mRNAs or 16S rRNA (Table 2). Amplification of a single expected PCR product was confirmed by electrophoresis on a 2% agarose gel. RT-PCR was performed using a Light Cycler (Roche Diagnostic), and quantification of the cpxR, rpoE ompR, ompC, ompF and degS mRNA levels or 16S rRNA (as a control) was performed using RNA master SYBR Green I (Roche Diagnostic) with 0.5 μg of total RNA.
For analysis of the significance of differences in adhesion and invasion levels, Student's t-test was used. All experiments were repeated at least three times. A P-value less than or equal to 0.05 was considered statistically significant.
This study was supported by the Ministère de la Recherche et de la Technologie (EA3844), by the INRA (USC 2018) and by grants from the Association F. Aupetit (AFA), and Institut de Recherche des Maladies de l'Appareil Digestif (IRMAD, Laboratoire Astra France). We are grateful to Jean Michel Betton (CNRS URA 2185, Institut Pasteur, Paris, France) for his interest in our work and helpful discussions. We thank Roland Lloubès (CNRS UPR 9027, Institut de Biologie Structurale et Microbiologie, Marseille, France) for OmpC/F antibodies, Nico Nouwen (Departement of Molecular Microbiology, University of Groningen, Haren, the Netherlands) for OmpA antibodies and Maryvonne Moulin-Schouleur (Pathogénie Bactérienne, UR86, INRA, Nouzilly, France) for E. coli type 1 pili antibodies. We also thank Christelle Drégneaux and Claire Szczepaniak (CICS, Université d'Auvergne, Clermont-Ferrand, France) for technical assistance with electron microscopy.