Roles of the Tol-Pal system in the Type III secretion system and flagella-mediated virulence in enterohemorrhagic Escherichia coli

The Tol-Pal system is a protein complex that is highly conserved in many gram-negative bacteria. We show here that the Tol-Pal system is associated with the enteric pathogenesis of enterohemorrhagic E. coli (EHEC). Deletion of tolB, which is required for the Tol-Pal system decreased motility, secretion of the Type III secretion system proteins EspA/B, and the ability of bacteria to adhere to and to form attaching and effacing (A/E) lesions in host cells, but the expression level of LEE genes, including espA/B that encode Type III secretion system proteins were not affected. The Citrobacter rodentium, tolB mutant, that is traditionally used to estimate Type III secretion system associated virulence in mice did not cause lethality in mice while it induced anti-bacterial immunity. We also found that the pal mutant, which lacks activity of the Tol-Pal system, exhibited lower motility and EspA/B secretion than the wild-type parent. These combined results indicate that the Tol-Pal system contributes to the virulence of EHEC associated with the Type III secretion system and flagellar activity for infection at enteric sites. This finding provides evidence that the Tol-Pal system may be an effective target for the treatment of infectious diseases caused by pathogenic E. coli.

Scientific RepoRtS | (2020) 10:15173 | https://doi.org/10.1038/s41598-020-72412-w www.nature.com/scientificreports/ some phenotypes including increased susceptibility to bile salts and population of filamentous morphology 13,15,16 . Some of tol-pal genes are also involved in bacterial pathogenesis, such as survival of Salmonella enterica serovar Typhimurium within macrophages, cytotoxicity of Erwinia chrysanthemi in plant cells, and pustule formation in humans by Haemophilus ducreyi 15,[17][18][19][20] . Previously, we found that mutants in tol-pal genes of uropathogenic E. coli (UPEC) exhibited decreased bacterial internalization in urinary tract cells and impaired motility, which reduced bacterial colonization within the urinary tract of mice 21 . Thus, the Tol-Pal system contributes to the pathogenicity of UPEC in the urinary tract. We aim to obtain further insights into roles of the Tol-Pal system in pathogenesis of E. coli. In this study we focused on characterizing roles of the Tol-Pal system in pathogenesis of EHEC, an enteropathogenic E. coli strain that cause infectious diseases at enteric sites, associated with Type III secretion system, Shiga toxin and flagella-mediated motility.

Results
Deletion of the tolB gene decreases bacterial motility, levels of EspB and EspA, the Type III secretion proteins. To test if the Tol-Pal system of pathogenic E. coli is involved in the pathogenesis at enteric sites, we used EHEC O157:H7 as a typical strain, which causes infectious disease at enteric sites. We constructed an in-frame deletion mutant of the tolB gene, which is a member of tol-pal genes, that lacks the Tol-Pal system. Similar to UPEC, the tolB mutant in EHEC exhibited a reduced motility on semi-solid agar compared with the wild-type parent, and the introduction of pTH18krtolB, a heterologous tolB expression plasmid, increased its motility up to the level of the wild-type parent (Fig. 1a). We observed that the tolB mutant was also less motile than the wild-type parent in broth (see Supplementary Video online). Similar to the tolB mutant in UPEC, the tolB mutant in EHEC produced defective flagella (Fig. 1b). We also examined the level of flagellin, encoded by fliC. Because we were unable to detect FliC even in wild-type EHEC by western-blotting with a commercial FliC antibody, we constructed EHEC strains carrying pTH18krfliC-VSVG, which produces the recombinant VSVG-tagged FliC protein under an innate fliC promoter. As predicted, we detected FliC-VSVG in both the cell lysate and secreted fractions from wild-type culture (Fig. 1c) Supplementary Fig. 1). However, FliC-VSVG in those from the tolB mutant culture was undetectable. Thus, deletion of tolB decreases FliC level, and leads to reduction of flagellar production and motility.
The tol-pal mutants in some bacteria including Vibrio cholerae and Erwinia chrysanthemi exhibited an impaired cell division, and formed extensive filaments during growth 15,16 . However, we did not observe this phenotype in the tolB mutant of EHEC. We examined the cell morphology of the tolB mutant with the wildtype parent on microscopy. The tolB mutant exhibited a rod-shaped morphology similar to the wild-type parent (Fig. 2a). We also compared CFU and OD 600 values between these strains to assess population of filaments. These values of the tolB mutant were moderately lower than those of the wild-type parent, while CFU/OD 600 values of the tolB mutant and the wild-type parent were similar (Fig. 2b,c). Type III secretion system proteins are a subset of typical proteins that are involved in the pathogenicity of EHEC. EspB is a member of these proteins, and it is required for the translocation of other effector proteins into host epithelial cells, and this protein also has an effector activity [22][23][24] . Since most of EspB produced by EHEC is secreted into culture media, we compared the EspB level in culture supernatants from the tolB mutant with that from the wild-type parent by western-blotting with an EspB antiserum. The EspB level in the tolB mutant was lower than that in the wild-type parent (Fig. 3a) (Supplementary Fig. 2). In contrast, no apparent difference in the levels of intracellular EspB extracted from whole cell fractions in the wild-type parent and tolB mutant was observed ( Fig. 3a) (Supplementary Fig. 2). We repeated this assay in DMEM. Although LB medium is commonly used for culturing EHEC, Type III secretion system proteins of EHEC are fully activated in expression when grown with DMEM. Levels of both extracellular and intracellular EspB in wild-type EHEC grown with DMEM were higher than those grown with LB. Similar to LB medium, deletion of tolB decreased the level of extracellular EspB, but not intracellular EspB (Fig. 3a) (Supplementary Fig. 2). We confirmed that decreased EspB secretion by tolB deletion was restored when the pTH18krtolB plasmid for heterologous tolB expression was introduced ( Fig. 3a) (Supplementary Fig. 2). In addition to EspB, we examined the secretion of EspA, other protein which is secreted via the Type III secretion system. We found that the level of EspA in the tolB mutant was also lower than that in the wild-type parent, and the heterologous expression of tolB in this mutant elevated EspA up to the wild-type level (Fig. 3b) (Supplementary Fig. 3). We also measured the transcript level of genes that encode Type III secretion system proteins by qPCR. These gene clusters consist of five operons (LEE1-LEE5), and the espB gene is transcribed with the espA gene in LEE4 10 . As target genes for qPCR analyses, we selected ler, espA, tir, escJ and escV from each operon. These transcript levels in the tolB mutant were moderately higher than those in the wild-type parent, although the reason for this remains unknown (Fig. 3c). These combined results suggest that the tolB gene contributes to the secretion, but not the expression of EspB and EspA.
We also estimated the activity of Shiga toxins as the other subset of proteins that is required for EHEC virulence, in the wild-type parent and its tolB mutant by latex agglutination assays. No significant difference in agglutination titers of both Stx1 and Stx2 was observed between the wild-type and tolB mutant (Fig. 4a). We also found that transcript levels of stx1 and stx2 in the tolB mutant were similar to the wild-type parent (Fig. 4b).
The tolB gene product forms protein complexes with Pal, TolA, TolQ, and TolR, proteins, to form the Tol-Pal system. Any members are indispensable for the activity of the Tol-Pal system. We determined whether decreased motility and EspA/B secretion are associated with a defect in the Tol-Pal system. We constructed the pal deletion mutant and analyzed motility and EspA/B secretion. The pal mutant was less motile than the wild-type parent ( Fig. 1d) (see Supplementary Video online) Similar to the tolB mutant, the pal mutant produced fewer flagella than the wild-type (Fig. 1c,e) ( Supplementary Fig. 1). Western-blotting analysis with an EspB antiserum showed a lower level of EspB secretion in the pal mutant than that in the wild-type parent while the levels of intracellular  Fig. 2). We also found that deletion of pal decreased the level of EspA secretion, but not Stx1 and Stx2 (Figs. 3b, 4a) ( Supplementary  Fig. 3). Thus, the pal mutant behaves similarly to the tolB mutant. These results suggest that the activity of the Tol-Pal system contributes to EHEC motility and virulence. www.nature.com/scientificreports/ the tolB mutant exhibits reduced adhesion to epithelial cells, and forms A/E lesions with low efficiency. EHEC initially adheres to host epithelial cells, and then the bacteria secrete effector and translocator proteins including EspB followed by the establishment of A/E lesions 4,7 . To test if the tolB gene also contributes to the initial adhesion process, we compared the ability of the tolB mutant to adhere to HeLa cells with that of the wild-type parent. Although EHEC adheres to intestinal epithelial cells in vivo, it also adheres to HeLa cells, and its Type III secretion system proteins, including EspB target the cells in vitro 4,25 . Therefore, HeLa cells are conveniently used to evaluate the in vitro ability of EHEC to adhere to host epithelial cells and determine the activity of the Type III secretion system. The tolB mutant exhibited an approximately three-fold lower rate of adhesion than the wild-type parent (Fig. 5a). We also observed that the adhesion rate of the tolB mutant was comparable to the wild-type level following the introduction of pTH18krtolB (Fig. 5a).
To test whether the tolB mutant has a reduced ability to form A/E lesions in epithelial cells compared to the wild-type, HeLa cells were infected with the wild-type parent or the tolB mutant, then actin accumulation with bacterial cells in HeLa cells were investigated. We found that the tolB mutant produced fewer A/E lesions than  www.nature.com/scientificreports/ the wild-type parent (Fig. 5b). These results indicate that the tolB gene is required for optimal adhesion and A/E lesion formation in epithelial cells.
the tolB mutant of C. rodentium exhibits low virulence in mice. To characterize the role of the tolB gene in the Type III secretion system-associated with EHEC pathogenesis in vivo, we used a murine intestinal infection model with C. rodentium. Since EHEC is a human-specific pathogen, it does not cause typical symptoms of the disease in mice that reflect those observed in human infections. C. rodentium is a natural pathogen of mice, and it carries LEE genes but not genes that encode Shiga toxins 26 . For this reason, C. rodentium is frequently used to evaluate the Type III secretion system associated virulence in mice. We constructed the tolB deletion mutant from the C. rodentium DBS100 strain that is highly virulent in C3H/HeJ mice. When the mice were administrated with the wild-type parent, all mice died within 8 days following a marked decrease in body weight on the previous day ( Fig. 6a,b). In contrast, the mice infected with the tolB mutant survived for at least 21 days. A moderate decline in body weight was observed in these mice until the twelfth day post-infection, but their body weight increased thereafter, which implies the recovery from infection. We also characterized intestinal symptoms in mice at 7 days post-infection. The colons of mice infected with wild-type C. rodentium exhibited a significantly shorter in the length than those with MG1655, the non-pathogenic K-12 strain or mice of non-infection group (Fig. 6c). Diarrhea-like stool inside of the colon and some swelling were observed in mice infected with wild-type C. rodentium, suggesting that the wild-type C. rodentium causes inflammation of the colons (Fig. 6d). On the other hand, mice infected the tolB mutant exhibited relatively short colons compared to the control mice, but still significantly longer than mice infected with the wild-type (Fig. 6c). Similar to the control mice, mice infected the tolB mutant exhibited no apparent swelling on the colon and also showed normal stool inside of the colon (Fig. 6d). These combined results indicate that the tolB gene of C. rodentium is necessary for optimal virulence in mice infected at the enteric site. However, the tolB mutant may be more susceptible to acid compared with the wild-type parent, which may attribute the low virulence of this mutant to a low ability of gastric transit under acidic conditions after oral infection. To verify this possibility, we tested the bacterial survival of the DBS100 parent and tolB mutant after incubation in an acidified medium (pH 3.0). No significant difference in the survival rates between the parent and tolB mutant was observed (Fig. 6e). Therefore, the low virulence of the tolB mutant is not solely due to a high susceptibility to acid. It is known that tol-pal mutants are highly susceptible to bile salts 13,15 , therefore deletion of tol-pal genes may reduce fitness cost in intestinal tracts, which may explain low virulence of the tolB mutant. We found that the tolB mutant of C. rodentium DBS100 exhibited lower MICs of bile salts, such as sodium cholate and sodium deoxycholate than the wild-type parent ( Table 1). response against high virulent wild-type C. rodentium, in mice infected with the tolB mutant, we tested the production of anti-bacterial antibodies and cytokines elicited by activation of T cells. Initially, we found that mice infected with the tolB mutant produced a significant amount of anti-bacterial IgG2a, but not IgG1 and IgA antibodies at 7 days post-infection compared to control mice and mice infected with the non-pathogenic K12 strain (Fig. 7a-c). To test an ability of T cell response, we determined levels of Th1 cytokine IFN-γ and Th17 cytokine IL-17 produced from spleen cells after in vitro re-stimulation with an extract from C. rodentium DBS100. We observed that a significant amount of IL-17 production was elicited in spleen cells from mice infected with the tolB mutant (Fig. 7d). These combined results indicate that tolB mutant bacteria have the ability to elicit the immune response against wild-type virulent strain.

Discussion
The Tol-Pal system was first characterized in E. coli, and it consists of two protein complexes formed by TolA, TolB, TolQ, TolR and Pal 13,14 . The subset of tol-pal genes that encode these proteins is also found in many species of other gram-negative bacteria. It has been shown that these Tol-Pal proteins are involved in the pathogenesis 15,[17][18][19][20] . In this study, we suggest that the Tol-Pal system contributes to EHEC virulence associated with the Type III secretion system and flagellum because the tolB and pal deletion mutants that lack the activity of the Tol-Pal system exhibited decreased functional flagella production, secretion of the Type III secretion system proteins, EspA and EspB required for the formation of A/E lesions, and ability to adhere to and to form A/E lesions in host epithelial cells. We also showed that the tolB mutant of C. rodentium did not cause lethality in mice, whereas all mice infected with the parent strain died within 8 days post-infection. In addition, unlike mice infected with the wild-type C. rodentium, mice infected with the tolB mutant did not exhibit symptoms of inflammation and diarrhea-like stool in the colon. This supports the results of in vitro experiments using EHEC. C. rodentium is traditionally used to evaluate virulence associated with the Type III secretion system in the intestine of mice. The pathogenesis of this bacterium reflects that of EHEC associated with the Type III secretion system in mice because C. rodentium produces a subset of Type III secretion system proteins that share high sequence similarity with that in EHEC, but it dose not produce Shiga toxins 26 . The tolB mutant exhibited a higher susceptible to bile salts than the wild-type parent. This may be also involved in attenuated virulence of the tolB mutant in mice. Results of antibody and T cell response suggest that the tolB mutant has the ability to elicit the immune response, including activation of Th17 cells, against wild-type virulent strain although this induction is still moderate. Activation of Th17 cells is proposed to be important for protection from some enteric-pathogens including C. rodentium for mice 27 , which implies that the tolB mutant give a clue to develop the low virulent live vaccine against EHEC without affecting the potential of eliciting the protective immunity. Type III secretion system proteins are the major proteins responsible for virulence in the human intestine because these proteins enable EHEC to induce A/E lesions in enteric epithelial cells, and the induction of A/E lesions leads to severe haemorrhagic diarrhea 4 . During infection, EHEC initially adheres to intestinal epithelial cells, on which it produces a subset of Type III secretion system proteins, and then injects these effector proteins into host cells via transport machinery 28 . Therefore, bacterial adhesion to host cells is an initial key step that enables EHEC to induce virulence associated with the Type III secretion system followed by the establishment of A/E lesions in host intestinal cells. When EHEC adheres to cells, it uses proteins localized in the outer membrane including fimbria, adhesin proteins and flagella 29,30 . The tolB and pal mutants produced defective flagella in addition to decreased EspB secretion. Expression of the flagellum and Type III secretion system proteins is inversely regulated because the expression of fliC which encodes flagellin is repressed when the production of Type III secretion system proteins is induced at later stages after the initial attachment via flagellin 31 . Our results suggest that the activity of the Tol-Pal system in EHEC may regulate both the initial attachment and induction of A/E lesions in host epithelial cells.
The tol-pal mutants that lack activity of the Tol-Pal system exhibited defects in some components localized on the outer-membrane, such as lipopolysaccharide and OmpA protein 32 . In this study, we also found that secretion of EspA/B, the translocator proteins secreted via the Type III secretion system, was decreased following the deletion of tolB and pal genes. However, the intracellular EspA/B level and transcript levels of some genes in LEE operons, including espA and espB were not decreased. We speculate that the activity of the Tol-Pal system may also be involved in the assembly of the transport protein complex for the Type III secretion system, and thus, the secretion ability of effector and translocator proteins including EspA/B in tolB and pal mutants may be defective, which leads to the reduced secretion of these proteins compared with the wild-type parent. In the transport protein complex, EscC is a major protein required to form the "outer ring" embedded in the outer membrane, www.nature.com/scientificreports/ and it is associated with both EscD, an inner membrane protein and the EscF needle structure protein [33][34][35][36] . The disturbance in the outer membrane by deletion of tol-pal genes may impair the precise localization and stabilization of the outer ring, which presumably leads to disassembly of the transport protein complex. However, some proteins localized in the outer membrane, such as fimbrial protein complexes, remain functional in the tol-pal mutants despite the absence of activity of the Tol-Pal system 21 . It would be interesting to gain insight into which proteins are affected by the Tol-Pal system in addition to Type III transport proteins and how these proteins are associated with it. In addition to the results of our previous study with UPEC, we showed here that the Tol-Pal system also contributes to the virulence of EHEC, which causes an infectious disease at enteric sites. This finding may enable us to make the idea that attenuates bacterial virulence by targeting the Tol-Pal system as an option for chemotherapies to treat infections caused by pathogenic E. coli.

Materials and methods
Bacterial strains, host cells and culture conditions. The bacterial strains and plasmids used in this study are listed in Table 2. Bacteria were grown in Luria-Bertani (LB) medium unless otherwise indicated. The optical density at 600 nm (OD 600 ) was measured as an indicator of cell growth. The antibiotics chloramphenicol (45 μg/mL) and kanamycin (50 μg/mL) were added to the growth media for marker selection and maintaining plasmids. HeLa cells, the cervical cancer cells were cultured in Dulbecco's Modified Eagle Medium containing 10% HyClone FetalClone III serum (HyClone Laboratories, Inc., Logan, UT, United States) at 37 °C and in an atmosphere of 5% CO 2 .
Cloning and mutant construction. In-frame gene deletions were produced by sequence overlap extension polymerase chain reaction (PCR), as described previously 37 , using the primer pairs delta1/delta2 and delta3/  Table 3). The upstream flanking DNA comprised 450 bp and the first two amino acid codons of tolB, and the first three amino acid codons of pal. The downstream flanking DNA included, the last amino acid codon (CTG) of tolB in EHEC, the last six amino acid codons of tolB in C. rodentim, the last five amino acid codons of pal, a stop codon, and 450 bp of DNA. These deletion constructs were ligated into the temperature-sensitive plasmid pKO3 37 and introduced into EHEC or the C. rodentium strain. We selected sucrose-resistant/chloramphenicol-sensitive colonies at 30 °C. We also constructed pTH18krtolB. The tolB gene and its 200 bp upstream were cloned into the low-copy number plasmid pTH18kr. To construct a C-terminal vesicular stomatitis virus glycoprotein (VSVG)-tagged FliC expression plasmid, pTH18krfliC-VSVG, the DNA containing the fliC coding region and its 300-bp upstream region was amplified with pTHfliC-F and pTHfliC-VSVG-R primers, and ligated into HindIII and BamHI sites in pTH18kr. This DNA fragment contains a cis-regulatory element for fliC gene expression. In addition, although the pTH18kr vector has a lac promoter sequence upstream of the HindIII and BamHI sites, the introduced fliC gene is oriented in a reverse direction to the lac promoter. Therefore, we expected that the resulting bacteria construct would produce FliC as a C-terminal VSVG-tagged protein from its native promoter. All constructs were confirmed by DNA sequencing. Motility assays. LB medium containing 0.3% agar was spotted with 2 μL of bacteria grown for 18 h at 37 °C in LB medium. Bacterial motility was evaluated by measuring the motility diameters after 8 h at 37 °C under an atmosphere of 5% CO 2 . We also monitored bacterial motility in liquid cultures. Bacteria were grown to the late

Shiga toxins assay.
To test the activity of Shiga toxins (Stx1 and Stx2), we used latex agglutination reagents (Denka Seiken Co. Ltd, Tokyo, Japan). EHEC strains were grown at 37 °C with shaking to the early stationary phase in Muller-Hinton medium, and separated by centrifugation and filtration. These culture supernatants were serially diluted in 96-well round bottom plates containing phosphate-buffered saline (PBS) and an equal volume of the latex suspension sensitized with Stx1 or Stx2 antibody was then added. After incubation for 14 h at 4 °C, titers were determined. The titers are presented as the reciprocal of the dilution of the last well before agglutinations were observed.

RNA extraction and quantitative real-time PCR analyses.
Bacteria were grown to the late logarithmic growth phase at 37 °C in LB medium. Total RNA extraction and cDNA synthesis were performed using a SV Total-RNA Isolation System (Promega Corp., Madison, WI, United States) and ReverTra Ace qPCR RT Manster Mix (TOYOBO Co. Ltd, Osaka, Japan), respectively, following the manufacturer's instructions. Real-time PCR were carried out as described previously 21 . Primers were listed in Table 4.
Bacterial adhesion to host epithelial cells. Bacterial adhesion to HeLa cells were assessed as s described previously with slight modifications 21 . Bacteria were inoculated into HeLa cells in 24-well plates. After incubation for 4 h, the total number of bacteria was determined from a first set of duplicate wells. The number of adhered bacteria was determined by counting the number of bacteria present in a second set of wells after washing five times with PBS + (PBS containing 0.5 mM MgCl 2 and 1 mM CaCl 2 ). Numbers of adhered bacterial cells are represented as relative colony forming units (CFUs) by their ratios (%) to total cell CFUs.

Assessment of A/E lesion formation.
Bacteria were inoculated with an MOI of 100 bacteria per host cell into cultured HeLa cells on glass coverslips in a 6-well plate and incubated at 37 °C for 4 h, then the medium was replaced by a fresh medium. The culture was continued for another 2 h. The coverslips were washed with PBS+. Rhodamine phalloidin (Life Technologies, Carlsbad, CA, United States) was used to visualize actin, and Hoe-chst33342 (Dojindo Laboratories, Tokyo, Japan) was used to stain bacteria and HeLa nuclei. Fluorescent images were acquired in the DAPI and rhodamine phalloidin laser units on an Olympus FV1200 IX81 microscope using a 100 × objective and captured with a CCD camera.

C. rodentium infections in mice.
Four week old female C3H/HeJ mice were obtained from CLEA Japan (Tokyo, Japan). C. rodentium DBS100 or its tolB mutant was grown overnight in LB medium with shaking (180 rpm) at 37 °C. The bacterial cells were harvested, and resuspended in fresh LB medium at a concentration of 1 × 10 9 CFU/mL, and 200 μL of the bacterial suspension (2 × 10 8 CFU) was orally administrated. As a control group, 200 μL of bacteria-free LB broth and/or culture resuspension from MG1655, a non-pathogenic K-12 strain of E. coli, were inoculated into mice. To measure survival rates and body weight of mice, the mice were monitored daily for 21 days. To characterize intestinal symptoms, mice were euthanized 7 days post-infection and the colons were aseptically removed.
Acid survival assays. Bacteria  preparation of C. rodentium antibodies. C. rodentium DBS100 was grown overnight in LB medium with shaking (180 rpm) at 37 °C. Bacterial cells were harvested, then lysed by freeze-thaw and sonication. Bacterial extracts were prepared in PBS with a concentration of 50 μg/mL to detect serum IgGs and fecal IgA by ELISA according to manufacturer's protocol (Bethyl Laboratories, Montgomery, TX, United States) 43 . To detect fecal IgA, fecal extracts were prepared from fresh feces 43 .
In vitro anti-bacterial T cell response. Spleen cells were isolated from mice infected with bacteria and control mice, and 1 × 10 6 cells were re-suspended in 0.2 mL of RPMI1640 medium and seeded in each well of 96-well culture plates 44 . Cells were stimulated with bacterial antigens from 1 × 10 6 cells of C. rodentium DBS100 killed by freeze-thaw, and incubated for 48 h. IFN-γ and IL-17 in culture supernatants were assayed by using ELISA kits in accordance with manufacturer's protocols (eBioscience product from Thermo Fisher Scientific, Waltham, MA, United States).
Approval for experiments. All experimental protocols were approved by the Gunma University Gene Recombination Experiment Safety Committee for all gene recombination and bacteria culture studies (The approval number: 16-002) and by the Committee of Experimental Animal Research of Gunma University for all animal studies (The approval number: 19-094). All experiments were performed in accordance with these committee's guidelines and regulations.

Data availability
All data generated or analysed during this study are included in this published article.