The Transcription of Flagella of Enteropathogenic Escherichia coli O127:H6 Is Activated in Response to Environmental and Nutritional Signals

The flagella of enteropathogenic Escherichia coli (EPEC) O127:H6 E2348/69 mediate adherence to host proteins and epithelial cells. What environmental and nutritional signals trigger or down-regulate flagella expression in EPEC are largely unknown. In this study, we analyzed the influence of pH, oxygen tension, cationic and anionic salts (including bile salt), carbon and nitrogen sources, and catecholamines on the expression of the flagellin gene (fliC) of E2348/69. We found that sodium bicarbonate, which has been shown to induce the expression of type III secretion effectors, down-regulated flagella expression, explaining why E2348/69 shows reduced motility and flagellation when growing in Dulbecco’s Minimal Essential Medium (DMEM). Further, growth under a 5% carbon dioxide atmosphere, in DMEM adjusted to pH 8.2, in M9 minimal medium supplemented with 80 mM glucose or sucrose, and in DMEM containing 150 mM sodium chloride, 0.1% sodium deoxycholate, or 30 µM epinephrine significantly enhanced fliC transcription to different levels in comparison to growth in DMEM alone. When EPEC was grown in the presence of HeLa cells or in supernatants of cultured HeLa cells, high levels (4-fold increase) of fliC transcription were detected in comparison to growth in DMEM alone. Our data suggest that nutritional and host signals that EPEC may encounter in the intestinal niche activate fliC expression in order to favor motility and host colonization.


Introduction
Enteropathogenic Escherichia coli (EPEC) is a world-wide cause of acute infantile diarrhea. EPEC attaches to host epithelial cells using a repertoire of adhesins including bundle-forming pilus (BFP), the outer membrane protein intimin, flagella, curli, and E. coli common pilus (ECP) [1][2][3][4]. In the small bowel, EPEC causes attaching and effacing (AE) lesions that lead to the destruction of the intestinal microvilli, opening of tight junctions, and release of water and electrolytes [1]. The gene products responsible for the production of the AE lesions are encoded in a chromosomal pathogenicity island called the locus of enterocyte effacement (LEE) [5].

Bacterial Strains and Construction of the fliC::lacZ Transcriptional Fusion
Transcriptional fusions have been widely employed to analyze gene expression of numerous bacterial virulence factors [35][36][37][38][39]. Here, a transcriptional fusion consisting of the EPEC fliC promoter linked to a promoter-less lacZ reporter gene was constructed to monitor the expression of fliC in the EPEC O127:H6 strain E2348/69 under a variety of growth conditions. Primer FAV3 (5 -CGCGGATCCAAACGGTTAGCAATCGCCTG-3 ), which was derived from the coding region of the EPEC fliC gene (www.sangerinstitute.org (accessed on 11 September 2021)), and FAV5 (5 -TCAAGCTTGGAACTTAAATCCAG-3 ), derived from the coding region of the fliD gene, were used for amplification of the fliC promoter with Pwo polymerase using E2348/69 chromosomal DNA as the template. The amplicon was then ligated into plasmid pCR-Blunt II-TOPO yielding pFAV35. An EcoRIdigested fragment containing the fliC promoter from pFAV35 was ligated into the EcoRI site of the pRS551 vector [40], yielding pFAV36. The pRS551 vector contains the promoterless lacZ reporter gene. The pFAV36 plasmid was transformed into E2348/69, and the transcriptional activity of fliC was monitored by measuring β-galactosidase activity, as previously described [40,41]. In all experiments, E2348/69 carrying the empty vector pRS551 was employed as a negative control.

Culture Conditions and β-Galactosidase Assays
EPEC E2348/69 containing the fliC::lacZ fusion or the empty vector was grown with shaking for 21 h at 37 • C, diluted 1:50 in fresh Luria-Bertani (LB containing 10 g/L NaCl) (Sigma Aldrich, St. Louis, MI, USA) or low-glucose DMEM (Gibco) and M9 minimal medium (25.6 g/L of Na 2 HPO 4 , 6 g/L of KH 2 PO 4 , 1 g/L of NaCl, and 2 g/L of NH 4 Cl) (Gibco). These media were reconstituted with different compounds at various concentrations and used to measure fliC expression after growth of EPEC to an OD 600 of 0.65-0.70 at 37 • C. The pH of the supplemented media was adjusted to pH 7.2 before growth [42]. Compounds of study included divalent cationic salts, ammonium and sodium salts, different carbon and nitrogen sources, and hormones (epinephrine and norepinephrine) used at physiological and non-physiological concentrations. Lysophosphatidic acid (LPA) (Avanti Polar Lipids) was used at 200 mM in LB-grown bacteria after stimulation for 10 min, 1 h, 2 h, or 3 h of incubation at 37 • C, as previously described [43]. The cultures were then diluted 1:5 in Z buffer (0.06 M Na 2 HPO 4 , 0.04 M NaH 2 PO 4 , 0.01 M KCl, 0.001 M MgSO 4 , and 0.05 M β-mercaptoethanol), and the β-galactosidase activity was assayed using the O-nitrophenyl-β-D-galactopyranoside (ONPG) substrate, as previously described [40]. The β-galactosidase experiments were repeated at least three times in triplicate for each condition tested. Basal β-galactosidase activities in LB or DMEM media alone were used as a reference for the analysis.

Adherence to Epithelial Cells and Detection of Flagella by Immunofluorescence
HeLa cells were cultivated at 37 • C under a 5% CO 2 atmosphere in polystyrene 24well plates (Cellstar) containing glass coverslips, as previously described [2]. Briefly, cell monolayers at 80% confluency, were infected from 1 h to 6 h with 10 µL (10 7 bacteria) of a bacterial suspension with an OD 600 of 1.1 (10 9 bacteria/mL) previously grown overnight in LB. The cells were washed with PBS to remove unbound bacteria and then fixed with 2% formalin/PBS for immunofluorescence. Primary rabbit anti-H6 antibodies were added for 1 h at a 1:3000 dilution in 10% normal horse serum. After washing, the cells were incubated for 1 h with secondary anti-rabbit IgG Alexa 488 fluor-conjugated antibodies diluted 1:3000. The cells were washed extensively and mounted in glycerol-PBS and visualized under UV light using a Zeiss Axiolab microscope.

Pre-Conditioned Medium
Monolayers of HeLa cells that had been extensively washed with PBS were incubated in DMEM without antibiotics or fetal bovine serum for 24-48 h. The supernatant from these cell cultures referred as to "pre-conditioned medium" (P-DMEM) was collected and the pH adjusted to 7.4 and filtered through a 0.2 µm membrane. The P-DMEM was supplemented with ampicillin and used to grow E2348/69 containing pRS551 or pFAV36 to quantitate the expression of flagellin.

Motility Assays
Motility assays were performed in 0.3% agar plates containing tryptone media (1% tryptone) and reconstituted with different compounds. Briefly, the agar was spiked with an overnight culture using a needle and incubated at 37 • C. The level of motility was assessed by examining the radius of opacity as a result of bacterial swimming away from the inoculation point after 16 h of incubation as follows: +++: highly motile; ++: moderate motile; +: weak motile; -: non motile.

Statistical Analysis
All data were the averages of at least 3 independent experiments performed in triplicate. GraphPad Prism 9 software (GraphPad, San Diego, CA, USA) was used for statistical differences. One-way ANOVA followed by Tukey's multiple comparison test and an unpaired Student's t test was performed. A p-value ≤ 0.05 was considered statistically significant.

Glucose Activates fliC Expression in EPEC
In the niche of the intestinal tract, there is a gradient of glucose concentration that diminishes from the small to the large intestine. In contrast, the concentration of ammonium, which is generated from protein degradation, increases from the small bowel towards the large bowel [44][45][46]. Thus, we were interested in determining if glucose had any effect on the expression of fliC in EPEC. We began this study by determining the effect of glucose on EPEC fliC expression in M9 minimal media supplemented with 20 mM to 160 mM glucose (M9 + Glu). In parallel assays, we also employed a second hexose, sucrose, as an alternate source of carbon during growth (M9 + Suc). A dose-dependent effect on fliC expression was noted for both sugars with maximum expression reached at 80 mM glucose (p < 0.001) and 40 mM sucrose (p < 0.001) (Figure 1). Sucrose at 20 mM induced a 3-fold increase in fliC expression as compared to 20 mM glucose. At a 40 mM concentration, sucrose induced fliC expression~1.55-fold higher than glucose at 40 mM. fliC expression began to decline when glucose and sucrose were present over 80 mM and 40 mM, respectively. Thus, flagella activation was not glucose-specific since sucrose also induced fliC transcription.

Glucose Activates fliC Expression in EPEC
In the niche of the intestinal tract, there is a gradient of glucose concentration that diminishes from the small to the large intestine. In contrast, the concentration of ammonium, which is generated from protein degradation, increases from the small bowel towards the large bowel [44][45][46]. Thus, we were interested in determining if glucose had any effect on the expression of fliC in EPEC. We began this study by determining the effect of glucose on EPEC fliC expression in M9 minimal media supplemented with 20 mM to 160 mM glucose (M9 + Glu). In parallel assays, we also employed a second hexose, sucrose, as an alternate source of carbon during growth (M9 + Suc). A dose-dependent effect on fliC expression was noted for both sugars with maximum expression reached at 80 mM glucose (p < 0.001) and 40 mM sucrose (p < 0.001) (Figure 1). Sucrose at 20 mM induced a 3-fold increase in fliC expression as compared to 20 mM glucose. At a 40 mM concentration, sucrose induced fliC expression ~1.55-fold higher than glucose at 40 mM. fliC expression began to decline when glucose and sucrose were present over 80 mM and 40 mM, respectively. Thus, flagella activation was not glucose-specific since sucrose also induced fliC transcription.

Role of Sodium Bicarbonate in Flagella Expression
Sodium bicarbonate is a component of DMEM and was reported to stimulate the expression of genes contained in the LEE region of EHEC O157:H7 [47]. EPEC also contains a homolog of the EHEC LEE region [48]. That report led us to determine whether the addition of sodium bicarbonate to M9 minimal medium would affect fliC expression in EPEC. In contrast to the bicarbonate-mediated activation of EHEC LEE-encoded genes previously reported, we found that the presence of this compound reduced, in a doseresponse manner, the expression of fliC transcription in EPEC ( Figure 2). Analysis of the composition of the DMEM employed revealed that bicarbonate was present at 44 mM, a concentration at which we saw significant reduction of fliC transcriptional activation. This result may explain, at least in part, the negative regulation exerted on flagellation and motility previously reported in EPEC E2348/69 growing in DMEM [2].

Role of Sodium Bicarbonate in Flagella Expression
Sodium bicarbonate is a component of DMEM and was reported to stimulate the expression of genes contained in the LEE region of EHEC O157:H7 [47]. EPEC also contains a homolog of the EHEC LEE region [48]. That report led us to determine whether the addition of sodium bicarbonate to M9 minimal medium would affect fliC expression in EPEC. In contrast to the bicarbonate-mediated activation of EHEC LEE-encoded genes previously reported, we found that the presence of this compound reduced, in a doseresponse manner, the expression of fliC transcription in EPEC ( Figure 2). Analysis of the composition of the DMEM employed revealed that bicarbonate was present at 44 mM, a concentration at which we saw significant reduction of fliC transcriptional activation. This result may explain, at least in part, the negative regulation exerted on flagellation and motility previously reported in EPEC E2348/69 growing in DMEM [2].

Modulation of fliC Expression in EPEC by Ammonium
Ammonium was previously shown to influence the production of BFP in the EPEC O111:NM strain B171 [49]. Here, we sought to study the effect of ammonium in fliC expression in EPEC E2348/69 by using three ammonium salts (sulfate, chloride, and oxalate) added at 5 mM, 20 mM, and 100 mM to M9 minimal medium. We did not see any effect of these salts on fliC expression when added to M9 medium (data not shown); thus, we studied their effect in LB-grown bacteria. Among the three salts tested, ammonium oxalate showed the highest levels of fliC transcription (Figure 3). At 5 mM of ammonium salts, a 2.4-fold increase in fliC transcription was recorded compared to growth in LB without ammonium (p < 0.001). We also studied the effect of ammonium in DMEM-grown bacteria; however, the β-galactosidase activity found in DMEM with ammonium sulfate at any given concentration was even less than in DMEM alone (data not shown), suggesting that growth in DMEM with ammonium sulfate had a negative effect on fliC expression.

Modulation of fliC Expression in EPEC by Ammonium
Ammonium was previously shown to influence the production of BFP in the EPEC O111:NM strain B171 [49]. Here, we sought to study the effect of ammonium in fliC expression in EPEC E2348/69 by using three ammonium salts (sulfate, chloride, and oxalate) added at 5 mM, 20 mM, and 100 mM to M9 minimal medium. We did not see any effect of these salts on fliC expression when added to M9 medium (data not shown); thus, we studied their effect in LB-grown bacteria. Among the three salts tested, ammonium oxalate showed the highest levels of fliC transcription (Figure 3). At 5 mM of ammonium salts, a 2.4-fold increase in fliC transcription was recorded compared to growth in LB without ammonium (p < 0.001). We also studied the effect of ammonium in DMEM-grown bacteria; however, the β-galactosidase activity found in DMEM with ammonium sulfate at any given concentration was even less than in DMEM alone (data not shown), suggesting that growth in DMEM with ammonium sulfate had a negative effect on fliC expression.

Modulation of fliC Expression in EPEC by Ammonium
Ammonium was previously shown to influence the production of BFP in the EPEC O111:NM strain B171 [49]. Here, we sought to study the effect of ammonium in fliC expression in EPEC E2348/69 by using three ammonium salts (sulfate, chloride, and oxalate) added at 5 mM, 20 mM, and 100 mM to M9 minimal medium. We did not see any effect of these salts on fliC expression when added to M9 medium (data not shown); thus, we studied their effect in LB-grown bacteria. Among the three salts tested, ammonium oxalate showed the highest levels of fliC transcription (Figure 3). At 5 mM of ammonium salts, a 2.4-fold increase in fliC transcription was recorded compared to growth in LB without ammonium (p < 0.001). We also studied the effect of ammonium in DMEM-grown bacteria; however, the β-galactosidase activity found in DMEM with ammonium sulfate at any given concentration was even less than in DMEM alone (data not shown), suggesting that growth in DMEM with ammonium sulfate had a negative effect on fliC expression.

Effect of Sodium Chloride and Sodium Bisulfite on Flagella Expression
Sodium (chloride and bisulfite) salts were used for the assessment of fliC transcription. NaCl triggered dose-dependent fliC expression when used at 50 mM to 250 mM in LB with a peak at 200 mM, which represents a 3-fold increase compared to LB without NaCl (p < 0.001) ( Figure 4A). Sodium bisulfite was previously shown to induce the synthesis of BFP in EPEC strains [50]. However, here, we found that sodium bisulfite showed an inhibitory effect on flagellin expression ( Figure 4B and Table 1).

Influence of Divalent Cationic Salts on Flagella Expression
Divalent cations are central elements in the integrity of the cell membrane, as well as in many cellular functions. In EPEC, the expression of BFP and T3SS-dependent effector proteins is influenced by the presence of divalent cations (Ca 2+ and Mg 2+ ) in the growth media [49,51]. Growth in DMEM triggers the production of most virulence factors in EPEC [49,51,52]. Therefore, we employed divalent cationic salts MgCl 2 and MnCl 2 at various concentrations (5 mM, 20 mM, and 100 mM) to define their role in fliC transcription when added in parallel to LB and DMEM growth media. When 5 mM MgCl 2 was added to LB, fliC expression was increased 4-fold compared to growth in LB alone (p < 0.001). However, a significant effect on fliC expression was noted only when bacteria were grown in DMEM supplemented with 5 mM of MgCl 2 (p < 0.05) ( Figure 5A,B). Similar results were obtained when magnesium sulfate was employed, suggesting that it is the magnesium cation, and not the sulfate or chloride anions, that triggers fliC expression (data not shown). When 5 mM MnCl 2 was used in LB, increased fliC transcription (1.9-fold, p < 0.05) was observed ( Figure 5B). The presence of calcium chloride did not significantly activate βgalactosidase activity; however, bacterial motility was increased when the salt was added to DMEM (Table 1). Thus, Mg 2+ seems to be the most important of these divalent cations in activating flagella expression. To further support a role for divalent cations in fliC transcription, 5 mM EDTA (chelator of divalent cations) was added to the LB and DMEM growth media. As predicted, a significant reduction in fliC expression was seen with the addition of EDTA to LB (~70%) (p < 0.001) and DMEM (64%) (p < 0.001) ( Figure 5C), stressing the need for divalent cations in flagellin expression.

Influence of Divalent Cationic Salts on Flagella Expression
Divalent cations are central elements in the integrity of the cell membrane, as well as in many cellular functions. In EPEC, the expression of BFP and T3SS-dependent effector proteins is influenced by the presence of divalent cations (Ca 2+ and Mg 2+ ) in the growth media [49,51]. Growth in DMEM triggers the production of most virulence factors in EPEC [49,51,52]. Therefore, we employed divalent cationic salts MgCl2 and MnCl2 at various concentrations (5 mM, 20 mM, and 100 mM) to define their role in fliC transcription when added in parallel to LB and DMEM growth media. When 5 mM MgCl2 was added to LB, fliC expression was increased 4-fold compared to growth in LB alone (p < 0.001). However, a significant effect on fliC expression was noted only when bacteria were grown in DMEM supplemented with 5 mM of MgCl2 (p < 0.05) ( Figure 5A,B). Similar results were obtained when magnesium sulfate was employed, suggesting that it is the magnesium cation, and not the sulfate or chloride anions, that triggers fliC expression (data not shown). When 5 mM MnCl2 was used in LB, increased fliC transcription (1.9-fold, p < 0.05) was observed ( Figure 5B). The presence of calcium chloride did not significantly activate β-galactosidase activity; however, bacterial motility was increased when the salt was added to DMEM (Table 1). Thus, Mg 2+ seems to be the most important of these divalent cations in activating flagella expression. To further support a role for divalent cations in fliC transcription, 5 mM EDTA (chelator of divalent cations) was added to the LB and DMEM growth media. As predicted, a significant reduction in fliC expression was seen with the addition of EDTA to LB (~70%) (p < 0.001) and DMEM (64%) (p < 0.001) ( Figure 5C), stressing the need for divalent cations in flagellin expression.

Host Gut Signals Influence Expression of EPEC fliC
Host signals that prevail at the site of bacterial colonization in the intestine, such as bile acids, hormones, oxygen, carbon dioxide, and pH, determine in part which virulenceassociated products are expressed in vivo. Thus, we sought to evaluate the role of these intestinal cues in fliC expression in vitro after growth in LB and DMEM media. We found that the presence of sodium deoxycholate (bile) in both LB and DMEM was associated with a dose-dependent β-galactosidase activity. At 0.2% sodium deoxycholate, a 3.6-fold and 3-fold increase in transcription were detected in LB and DMEM (p < 0.001), respectively, in comparison to the medium without deoxycholate ( Figure 6A). At this physiological concentration, bile salt may presumptively act as an activator of flagellin synthesis in the small intestine.

The Presence of Epithelial Cells Triggers Flagella Expression in EPEC
We monitored time-dependent (1-6 h) production of flagella by bacteria adhering to cultured HeLa cell monolayers by immunofluorescence using anti-flagella H6 antibodies [2]. At 1 h of infection, only a few bacteria adhered to the epithelial cells. As the time of infection increased, so did the number of bacteria attached to the cell monolayer. Interestingly, within 6 h of infection, the bacteria did not shut off flagella production. Instead, they continued to produce flagella as the adhering bacteria multiplied on the surface of the eukaryotic cells ( Figure 7A). In contrast to these observations, in S. enterica and B. bronchiseptica, the synthesis of flagella was turned off once the bacteria were programmed to penetrate host epithelial cells [53,54]. In addition, transcription of fliC was measured in the presence and absence of epithelial cells. HeLa cells were infected with EPEC for 3 h, In their passage from the stomach to and through the small and large intestines, bacteria experience a pH gradient from low-to-high. Thus, β-galactosidase activity was also measured in EPEC bacteria growing in LB and DMEM media adjusted to a pH range of 6.2 to 8.2 [43]. The pH of the media was measured before and after growth to ensure that the pH did not change by more than 0.05 pH units (data not shown). The optimal pH for activation of EPEC fliC transcription was 8.2 in both LB and DMEM, whereas the β-galactosidase activity was significantly lower at pH 6.2 and pH 7.2 ( Figure 6B).
In the lumen of the large intestine, facultative anaerobic bacteria such as E. coli live under an atmosphere of reduced oxygen tension, which could in turn act as a signal to activate genes required for intestinal colonization. Thus, we studied fliC transcription in bacteria grown with aeration or in a 5% carbon dioxide atmosphere. We found that the β-galactosidase activity was 5-fold increased after growth in both LB (p < 0.001) and in DMEM (p < 0.001) under a 5% carbon dioxide atmosphere compared to growth under aeration. Up to this point, carbon dioxide was the only element, amongst all compounds and conditions tested, that triggered the highest level of fliC expression in bacteria grown in DMEM ( Figure 6C). This result is conceivably significant considering that the oxygen tension in the small intestine, the site of EPEC infection and colonization, is reduced relative to carbon dioxide [3].

The Presence of Epithelial Cells Triggers Flagella Expression in EPEC
We monitored time-dependent (1-6 h) production of flagella by bacteria adhering to cultured HeLa cell monolayers by immunofluorescence using anti-flagella H6 antibodies [2]. At 1 h of infection, only a few bacteria adhered to the epithelial cells. As the time of infection increased, so did the number of bacteria attached to the cell monolayer. Interestingly, within 6 h of infection, the bacteria did not shut off flagella production. Instead, they continued to produce flagella as the adhering bacteria multiplied on the surface of the eukaryotic cells ( Figure 7A). In contrast to these observations, in S. enterica and B. bronchiseptica, the synthesis of flagella was turned off once the bacteria were programmed to penetrate host epithelial cells [53,54]. In addition, transcription of fliC was measured in the presence and absence of epithelial cells. HeLa cells were infected with EPEC for 3 h, after which the transcription of flagella was determined using the β-galactosidase assay. fliC transcriptional activation in E2348/69 was increased 3.5-fold in the presence of HeLa cells compared to growth in DMEM without HeLa cells ( Figure 7B), supporting our hypothesis that eukaryotic cells or a cellular-secreted product trigger flagella expression in EPEC [2].

Role of Endocrine Adrenergic Molecules in EPEC fliC Expression
Catecholamines such as epinephrine and norepinephrine were previously shown to enhance the growth of commensal and pathogenic E. coli and to enhance virulence-associated properties of some E. coli pathogroups [55]. Regarding expression flagella and motility, it was reported that epinephrine and norepinephrine activate the expression of flagellar genes in E. coli O157:H7 [56,57]. In the present study, we tested epinephrine and norepinephrine as potential inducers of fliC transcription in EPEC. Bacteria were grown aerobically in DMEM supplemented with two concentrations (0.3 µ M and 30 µ M) of these molecules. We chose DMEM for these experiments because fliC transcription is considerably reduced in DMEM compared to LB, allowing us to better determine any positive effect on expression. The presence of epinephrine significantly enhanced (p < 0.001) the expression of EPEC flagella at the two different concentrations employed with a 3.6-and 1.9-fold increase, respectively ( Figure 8A). Norepinephrine showed a more discrete increase in fliC transcription at 30 µ M (1.4-fold, p < 0.05), but did not increase the expression of fliC at 0.3 µ M ( Figure 8B).

Role of Endocrine Adrenergic Molecules in EPEC fliC Expression
Catecholamines such as epinephrine and norepinephrine were previously shown to enhance the growth of commensal and pathogenic E. coli and to enhance virulenceassociated properties of some E. coli pathogroups [55]. Regarding expression flagella and motility, it was reported that epinephrine and norepinephrine activate the expression of flagellar genes in E. coli O157:H7 [56,57]. In the present study, we tested epinephrine and norepinephrine as potential inducers of fliC transcription in EPEC. Bacteria were grown aerobically in DMEM supplemented with two concentrations (0.3 µM and 30 µM) of these molecules. We chose DMEM for these experiments because fliC transcription is considerably reduced in DMEM compared to LB, allowing us to better determine any positive effect on expression. The presence of epinephrine significantly enhanced (p < 0.001) the expression of EPEC flagella at the two different concentrations employed with a 3.6-and 1.9-fold increase, respectively ( Figure 8A). Norepinephrine showed a more discrete increase in fliC transcription at 30 µM (1.4-fold, p < 0.05), but did not increase the expression of fliC at 0.3 µM ( Figure 8B).

Host-Cell-Produced Lysophospholipids Do Not Activate fliC Transcription in EPEC
A previous report showed that host-cell-produced lysophospholipids (lysophosphatidic acid (LPA) and lysophosphatidylcholine (LPC)) trigger the synthesis and secretion of flagellin by S. enterica serovar Typhi during infection of intestinal epithelial cells [42]. We were interested in determining if cellular lysophospholipids activate transcription of fliC in EPEC. Commercially available LPA used at 200 mM in LB-grown bacteria showed no apparent effect on the transcription of fliC after stimulation for 10 min, 1 h, 2 h, or 3 h of incubation at 37 • C (data not shown). These observations suggest that different host signals activate fliC in EPEC and pathogenic Salmonella.

Reconstitution of a Bacteriological Medium Optimal for Flagella Expression
Based on the information obtained above, we attempted to formulate a growth medium containing the signals that induced the highest levels of fliC transcription. Since growth in DMEM showed poor induction of fliC expression, we adjusted the pH of this medium to 8.2 and added the key components that triggered flagellin expression: 150 mM sodium chloride and 0.1% sodium deoxycholate sodium. This supplemented medium is referred to as DMEM-S. We also obtained supernatants from cultured HeLa cells grown without antibiotics and filtered them to remove any cellular debris. This medium was labeled as pre-conditioned medium (P-DMEM). For comparison, we also supplemented P-DMEM with NaCl and bile salt to yield P-DMEM-S. Bacteria were incubated in DMEM, DMEM-S, P-DMEM, and P-DMEM-S media under a 5% carbon dioxide atmosphere at 37 • C. Maximal activation of fliC transcription was found in P-DMEM-S with an~3-fold increase in comparison to DMEM alone (Figure 9), indicating that host cell signals play a significant role in fliC expression.
Microorganisms 2022, 10, x FOR PEER REVIEW 12 of 17 Figure 9. Reconstitution of media with fliC activating signals. Sodium chloride at 150 mM and 0.1% sodium deoxycholate were used in DMEM and P-DMEM (see the text for details) adjusted to pH 8.2 and then used to grow bacteria under a 5% carbon dioxide atmosphere. β-galactosidase activity was monitored as before. These data are the mean of at least three experiments performed in triplicate. The negative control for background lacZ activity was E2348/69(pRS551) (white bars). *** p < 0.001.

Motility Assay
The motility of EPEC E2348/69 was assayed in DMEM containing 0.3% agar and supplemented with the compounds listed in Table 1. The results were in strong correlation with the fliC transcription data obtained with the β-galactosidase assay.

Discussion
Production and secretion of virulence factors in many bacterial pathogens are influenced by factors intrinsic to the host, the microenvironment, and the bacteria themselves. For example, the effect of the bacterial growth phase, the presence of carbon and nitrogen sources, salts, bile, hormones, pH, temperature, oxygen, and carbon dioxide may be important to trigger the expression of virulence genes in the host during infection or in extrahost environmental niches [23,[58][59][60][61][62][63]. In EPEC, the environmental and nutritional signals and then used to grow bacteria under a 5% carbon dioxide atmosphere. β-galactosidase activity was monitored as before. These data are the mean of at least three experiments performed in triplicate. The negative control for background lacZ activity was E2348/69(pRS551) (white bars). *** p < 0.001.

Motility Assay
The motility of EPEC E2348/69 was assayed in DMEM containing 0.3% agar and supplemented with the compounds listed in Table 1. The results were in strong correlation with the fliC transcription data obtained with the β-galactosidase assay.

Discussion
Production and secretion of virulence factors in many bacterial pathogens are influenced by factors intrinsic to the host, the microenvironment, and the bacteria themselves. For example, the effect of the bacterial growth phase, the presence of carbon and nitrogen sources, salts, bile, hormones, pH, temperature, oxygen, and carbon dioxide may be important to trigger the expression of virulence genes in the host during infection or in extra-host environmental niches [23,[58][59][60][61][62][63]. In EPEC, the environmental and nutritional signals that modulate the expression of BFP and other virulence factors have been previously reported [49,51,[64][65][66]. Generally, the growth of EPEC in DMEM at 37 • C, with or without epithelial cells, allows the production of BFP and LEE-encoded products and the secretion of T3SS-associated effector molecules. The majority of studies regarding the signals that regulate flagella expression and motility have been conducted in Salmonella and laboratory E. coli strains [27,28,[67][68][69]. Our previously published data suggested that environmental, nutritional, and host signals modulate fliC expression in EPEC, at least in vitro [2]. What cues have an effect on flagella production during colonization of the human gut by EPEC remains largely unknown. Thus, to learn about the chemical and biological nature of the signals that might be present in the intestinal tract and that trigger flagella production, we measured the transcriptional activation of fliC in E2348/69 propagated in a wide range of laboratory growth conditions.
Previously, we showed that the EPEC strain E2348/69 produces flagella upon growth in LB, whereas flagella production and motility are greatly diminished in DMEM. Interestingly, however, the bacteria were able to regain these phenotypes upon contact with cultured epithelial cells replicating in DMEM or when growing in HeLa cell supernatants, suggesting that host cell factors overcome the negative regulation exerted by DMEM components [2]. Some reports have shown that bicarbonate ion stimulates the expression of virulence genes [47,70,71]. We studied fliC transcription in EPEC growing in LB without and with sodium bicarbonate at 22 mM and 44 mM. Converse to the effect that bicarbonate has on the expression of LEE-encoded genes in EHEC, we saw a dose-response negative effect on flagella expression in EPEC. This negative effect may explain why the production of flagella and motility in EPEC are down-regulated when growing in DMEM, which contains 44 mM bicarbonate. In support of this idea, we found that bicarbonate might be such a component, since its addition to LB also inhibited fliC expression.
The influence of different carbohydrates on flagella and motility in E. coli K-12 has been investigated previously. Glucose was shown to inhibit motility at 0.01 M, and it was suggested that this negative effect on flagellation was through catabolic repression [28,68]. More recently, increased expression of fliC in adherent-invasive E. coli growing in glucosesupplemented media was shown [72]. In the present study, we found that the presence of glucose and sucrose in LB at the concentration range of 20 mM to 160 mM increased EPEC flagellin gene expression. These data suggest that the flagella in E. coli K-12 and EPEC are regulated differently by glucose. Positive regulation by glucose may be related to enhanced availability of energy sources. It is well established that there is a high-to-low glucose gradient from the small to the large intestine [45]. Thus, the presence of glucose in the lumen of the small intestine would presumptively be important to initiate the synthesis of flagella, as these appendages are required for the bacteria to swim across the mucus layer that bathes target epithelial sites.
Divalent cations such as Ca 2+ , Mg 2+ , and Mn 2+ are central to the stability and integrity of the cell membrane and are required in many cellular functions. Calcium and magnesium ions regulate virulence gene expression in E. coli, Klebsiella, Yersinia, and Salmonella species, respectively [60,62,[73][74][75]. In EPEC, the expression of BFP and translocated T3SSeffector proteins is also influenced by the presence of Ca 2+ and Mg 2+ salts in the growth media [49,51,[76][77][78]. We demonstrated that the presence of cations such as Mg 2+ and Mn 2+ increased flagellin expression, but did not trigger motility. In favor of this notion, we showed that this effect was reversed by the presence of the chelating agent EDTA. In contrast, Ca 2+ did not activate flagella expression, but favored motility.
The concentration and availability of oxygen and carbon dioxide are relative to the site of the small and large intestines. E. coli is a facultative anaerobic organism and, as such, may or may not need oxygen to grow. It is clear from different studies in the literature that oxygen and carbon dioxide directly affect the production of virulence factors by regulating gene expression or indirectly by altering the environmental pH [51,61,79,80]. In this study, carbon dioxide was one of the most important inducers of fliC expression; this result correlates well with our observations that fliC expression was increased at pH values higher than 7.4.
Bile salts are steroids with detergent properties, which help in fat digestion and absorption through the intestinal wall [81,82]. The most abundant of the bile salts in humans are cholate and deoxycholate, and they are normally conjugated with either glycine or taurine to give glycocholate or taurocholate, respectively. Bile was also shown to be important for the expression of virulence factors in V. cholerae [83][84][85]. However, in Proteus mirabilis and E. coli K-12, 0.1% sodium deoxycholate inhibited flagellation and motility [86]. We found that sodium deoxycholate at a physiological concentration of 0.1% increased fliC expression in EPEC. We speculate that this bile component is important during infection of the small intestine by EPEC.
Earlier studies showed that during infection of intestinal epithelial cells, S. typhi secretes abundant pro-inflammatory flagellins upon activation by the host-cell-produced lysophospholipids (LPA) and LPC) [42]. In our study, we did not see an increase in the transcription of EPEC fliC or in the secretion of flagellin when LPA was used as the activator, suggesting that different host signals activate fliC expression in EPEC and pathogenic Salmonella.
The relationship between stress-related neuroendocrine hormones known as catecholamines, which include adrenaline (epinephrine) and noradrenaline (norepinephrine), and bacterial pathogenesis has been recognized for over half a century [87,88]. Increased levels of norepinephrine in the intestinal lumen play an important role in bacterial pathogenesis [89,90]. At high concentrations, norepinephrine has been shown to enhance the growth of commensal and pathogenic E. coli and to enhance the virulence properties of enterotoxigenic (ETEC) and E. coli O157:H7 [55,56,87,91]. Increased EHEC and adherentinvasive E. coli (AIEC) adherence to porcine colonic mucosa and Caco-2 cells, respectively, were noted after treatment with norepinephrine [20,92]. We found that catecholamines activated fliC expression in EPEC.
It is apparent that the relationship between flagellar and virulence gene regulation is tightly linked and that a complex interplay is occurring at the molecular level to coordinate multiple phenotypes as they are needed in the host niche. The importance of this interplay with regard to flagella regulation is underscored by the attenuation in the virulence of mutants with altered flagella expression. Understanding the environmental and molecular regulators that ultimately determine flagella gene expression and its coordinated regulation with virulence factors can provide insights into the significance of the production of flagella in the niche of the intestinal tract and their role in host colonization.