Novel Avian Pathogenic Escherichia coli Genes Responsible for Adhesion to Chicken and Human Cell Lines

Avian pathogenic E. coli is not only pathogenic for commercial poultry but can also cause foodborne infections in humans utilizing the same attachment and virulence mechanisms. Our aim was to identify genes of avian pathogenic E. coli involved in adhesion to chicken and human cells in order to understand the colonization and pathogenesis of these bacteria. In contrast to the recent studies based on genotypic and bioinformatics data, we have used a combination of phenotypic and genotypic approaches for identification of novel genes contributing to adhesion in chicken and human cell lines. Identification of adhesion factors remains important, as antibodies elicited against such factors have shown potential to block colonization and ultimately prevent disease as prophylactic vaccines. Therefore, the data will augment the understanding of disease pathogenesis and ultimately in designing strategies against the infections.

ABSTRACT

Avian pathogenic Escherichia coli (APEC) is a major bacterial pathogen of commercial poultry contributing to extensive economic losses and contamination of the food chain. One of the initial steps in bacterial infection and successful colonization of the host is adhesion to the host cells. A random transposon mutant library (n = 1,300) of APEC IMT 5155 was screened phenotypically for adhesion to chicken (CHIC-8E11) and human (LoVo) intestinal epithelial cell lines. The detection and quantification of adherent bacteria were performed by a modified APEC-specific antibody staining assay using fluorescence microscopy coupled to automated VideoScan technology. Eleven mutants were found to have significantly altered adhesion to the cell lines examined. Mutated genes in these 11 “adhesion-altered mutants” were identified by arbitrary PCR and DNA sequencing. The genes were amplified from wild-type APEC IMT 5155, cloned, and transformed into the respective adhesion-altered mutants, and complementation was determined in adhesion assays. Here, we report contributions of the fdtA, rluD, yjhB, ecpR, and fdeC genes of APEC in adhesion to chicken and human intestinal cell lines. Identification of the roles of these genes in APEC pathogenesis will contribute to prevention and control of APEC infections.

IMPORTANCE Avian pathogenic E. coli is not only pathogenic for commercial poultry but can also cause foodborne infections in humans utilizing the same attachment and virulence mechanisms. Our aim was to identify genes of avian pathogenic E. coli involved in adhesion to chicken and human cells in order to understand the colonization and pathogenesis of these bacteria. In contrast to the recent studies based on genotypic and bioinformatics data, we have used a combination of phenotypic and genotypic approaches for identification of novel genes contributing to adhesion in chicken and human cell lines. Identification of adhesion factors remains important, as antibodies elicited against such factors have shown potential to block colonization and ultimately prevent disease as prophylactic vaccines. Therefore, the data will augment the understanding of disease pathogenesis and ultimately in designing strategies against the infections.

INTRODUCTION

Avian pathogenic Escherichia coli (APEC) is of prime importance for the poultry industry because of its etiological role in embryonic death, septicemia, salpingitis, omphalitis, and chronic respiratory infections and is responsible for significant economic losses in poultry production worldwide (1). Clinically important serogroups of APEC include O2 and O78, which are associated with outbreaks of colibacillosis in poultry and responsible for 80% of poultry disease cases worldwide (2). APEC also has considerable zoonotic potential, as human consumption of contaminated poultry may cause infections, and human intestine can serve as an APEC reservoir (3). Studies comparing extraintestinal pathogenic E. coli (ExPEC) strains from human and avian sources have reported APEC as a probable reservoir of virulence-associated genes for certain E. coli subgroups, indicating APEC as potential zoonotic agents (4, 5). Pathogenicity tests of the APEC isolate IMT5155 in four animal models showed their virulence and suggested their zoonotic potential (6). Pathogenicity of bacteria is a complex process encompassing several interactive steps between host cells and pathogens. Adhesion to the host cell is one of the primary steps in bacterial strategies for successful colonization and progress of infection (7, 8). Adhesion is also important for colonization of the host intestine, which is the reservoir of APEC and other ExPEC. The role of flagella as adhesion proteins was first characterized in pathogenic E. coli (9). Mutations in flagellar genes (fliC) resulted in reduced virulence in a chicken infection model (10) and reduced adhesion to rectal epithelial cells. The complementation of the mutation restored the binding, confirming the role of flagella in host cell adhesion (11). Adhesion mediated by pili or fimbriae, mainly type 1 and P fimbriae, is widespread in Gram-negative bacteria colonizing lungs, air sacs, and internal organs of chickens; however, the lack of expression of P fimbriae in the chicken trachea has suggested a role in later stages of infection (12). A number of other fimbriae, including long-polar fimbriae (Lpf1 and Lpf2), E. coli common pilus (ECP [Mat]), F9, E. coli laminin-binding fimbriae (ELF), sorbitol-fermenting fimbriae protein (Sfp), hemorrhagic E. coli pilus (HCP), etc., have also been reported to play a role in host cell adhesion (13). Complete genome sequencing of an APEC strain (APEC O1:K1:H7) has shown that the function of a large number of genes is still unknown, and the role of many genes in host cell adhesion and virulence remains hypothetical (14, 15).

Attachment of APEC has been studied by spot inoculation of chicken meat with bacterial suspensions followed by quantification of the recovered bacteria or adhesion of APEC to human cell lines studied by lysis of the exposed cell monolayer followed by enumeration of bacteria in serial dilutions on agar media (16). Such approaches for detection and quantification of host-pathogen interactions are indirect and time-consuming. The VideoScan technology involves staining of bacteria with fluorescence and then automated rapid counting of the fluorescent bacteria efficiently (17). The automation and direct visibility of the bacteria make this technology preferable for screening over counting CFU and measuring the whole-well fluorescence, respectively. We applied this direct approach for screening and identification of novel APEC genes contributing to host cell adhesion. A previously reported transposon mutant library of a wild-type APEC strain, IMT 5155 (18), was used in this study to screen mutants for adhesion to chicken and human cell lines in order to identify novel APEC genes and their possible roles in host colonization, host tropism, and zoonotic potential to strengthen the understanding of APEC infections.

RESULTS

Screening of transposon mutant library and identification of potential adhesion genes.

A VideoScan module was developed using antibody staining, which allowed the automated enumeration of bacteria adhering to each of the cell lines and screening of the transposon mutant library of APEC strain IMT 5155 (Fig. 1).

FIG 1.

Rapid screening of transposon mutants using O2-antibody staining method. (A) CHIC-8E11 cells as a negative control in the DAPI channel. (B) Wild-type APEC IMT 5155. (C) Transposon mutant with low adhesion. (D) Transposon mutant with high adhesion.

Using this methodology, out of 1,300 random transposon mutants of APEC IMT 5155 screened for adhesion on CHIC-8E11 and LoVo cell lines, 11 were identified as “adhesion-altered mutants” for one or both cell lines (Fig. 2). No impairment in the growth rate of the adhesion-altered mutants in comparison to the wild-type strain was observed. Among the 11 adhesion-altered mutants, 6 showed reductions in adherent bacteria per millimeter squared of monolayer, whereas 3 mutants showed higher numbers of bacteria adherent to both CHIC-8E11 and LoVo cells than the wild-type APEC IMT5155 strain. One mutant showed higher adhesion to CHIC-8E11 cells only, and one mutant was found to show higher adherence to LoVo cells only, suggesting host-specific adhesion. To rule out any limitations of the antibody staining due to a mutation in the antibody-binding epitope, the adhesion-altered property of these mutants was also confirmed independently using a fluorescent in situ hybridization (FISH) staining method (Fig. S1). Sequencing of arbitrary PCR amplicons revealed the identity of the mutated genes in adhesion-altered mutants potentially contributing to adhesion to CHIC-8E11 and LoVo cell lines (Table 1).

FIG 2.

Adhesion of APEC wild-type strain IMT5155 (average value normalized to 100) and its selected transposon mutants to CHIC-8E11 (red bars) and LoVo (blue bars) intestinal cells. The data from the initial screening of the transposon library are shown as median values and median absolute deviation (MAD) of two separate experiments in triplicate.

TABLE 1.

Genes identified in this study

Mutant name	Gene name/locus	Complete name	(Possible) role/function
B6b	wzx	O-antigen flippase/O50 family O-antigen flippase	O-antigen synthesis
F10e	fliP	Flagellar biosynthesis protein FliP	Flagellar synthesis/assembly
D12k	flgK	Flagellar hook-associated protein FlgK	Flagellar synthesis/assembly
H6n	fdtA	TDP-4-oxo-6-deoxy-alpha-d-glucose-3,4-oxoisomerase	O-antigen synthesis
C7q	rfbD	dTDP-4-dehydrorhamnose reductase	O-antigen synthesis
E4r	AJB35136	Group 1 glycosyl transferase	Probably O-antigen synthesis
A7r	fdeC	Intimin-like adhesin FdeC (factor adherence E. coli)	Adhesin
D2q	ecpR	Helix-turn-helix transcriptional regulator/putative fimbrial transcriptional regulator	ECP transcriptional regulator
A10k	nhaA	Na+/H+ antiporter NhaA	Na+/H+ antiporter
C4s	yjhB	MFS transporter	Probably sialic acid transport
D2t	rluD	23S rRNA pseudouridine synthase D	Pseudouridine synthesis

Motility assays.

The results of the motility assay of adhesion-altered mutants in comparison to WT are shown in Fig. 3. Two mutations which affected flagellar synthesis (fliP, flgK) and one mutation which partially affected O-antigen synthesis (fdtA) showed more than 30% reduction in motility compared to the wild type. The wzx, rfbD, rluD, and group 1 glycosyl transferase (GenBank accession no. AJB35136) mutants showed 4% to 18% lower motility in comparison to the wild type.

FIG 3.

Motility assay of APEC wild-type strain IMT5155 and selected transposon mutants. The data are shown as median values and MAD of three separate experiments. SAEC5148, nonmotile strain used as a negative control.

Complementation of intact potential genes from wild-type APEC IMT 5155.

PCR amplification of the intact genes of the putative adhesion-altered mutants identified in the screening was performed from wild-type APEC IMT 5155, followed by cloning into the pACYC177 vector. The accurate cloning of intact genes with promoters and respective open reading frames was verified by sequencing. The respective complementing plasmids were then introduced into the particular adhesion-altered mutants. All adhesion-altered mutants were confirmed for their susceptibility to ampicillin prior to complementation, and also no effect of complementation on bacterial growth was observed in “complemented adhesion-altered mutants.” Adhesion assay screening of the adhesion-altered mutants and complemented adhesion-altered mutants showed complementation effects for 5 of the genes identified in the original screen (fdtA, ecpR, yjhB, rluD, and fdeC). Transformation of adhesion-altered mutants with the empty vector plasmid alone showed no significant effects on adhesion. The complementation of 2 genes (fdtA and ecpR) resulted in adhesion comparable to that of the wild-type strain, proving their important role in adhesion. The complementation with cloned genes for yjhB, rluD, and fdeC resulted in only moderate curing of adhesion, suggesting a contributory role in bacterial adhesion (as shown in Fig. 4).

FIG 4.

Adhesion of APEC wild-type strain IMT5155, selected transposon mutants, and transposon mutants transformed with either empty vector pACYC177 (Neg, green bar) or pACYC177 plasmids bearing wild-type versions of transposon mutated genes (Comp, red bars) to CHIC-8E11 and LoVo intestinal epithelial cells. The data shown are median values and MAD of three separate experiments in sextuplicates. Wells not incubated with bacteria were used as a negative control.

In prevalence studies, we tested 10 isolates each of 8 different E. coli pathotypes randomly selected from our culture stocks. The gene fdtA was detected in 5 human fecal E. coli (HFEC), 3 APEC, 1 atypical enteropathogenic E. coli (aEPEC), and 1 uropathogenic E. coli (UPEC) isolate, whereas this gene was not detected in the 4 other pathotypes (Fig. 5). The gene ecpR was detected in all isolates except 1 isolate each of aEPEC, APEC, enteroaggregative E. coli (EAEC), and enterotoxigenic E. coli (ETEC). The rluD gene was detected in all isolates except 1 sepsis-associated pathogenic E. coli (SePEC) isolate. The gene yjhB was detected in all isolates from UPEC, ETEC, HFEC, and SePEC but not detected in 1, 2, 3, and 6 isolates of avian fecal E. coli (AFEC), EAEC, aEPEC, and APEC pathotypes, respectively. flgK was detected in all isolates except 1 AFEC and 4 SePEC isolates. The gene wzx was detected in 4 HFEC isolates only. The gene encoding group 1 glycosyl transferase (GenBank accession no. AJB35136) was detected in 1 UPEC and 6 HFEC isolates only. The rfbD gene was not detected in 5 EAEC isolates, 3 AFEC isolates, and 1 isolate each of aEPEC, ETEC, and SePEC. The genes fdeC, nhaA, and fliP were detected in all tested isolates.

FIG 5.

Prevalences of the selected genes in various E. coli pathotypes. Gene prevalence is shown for genes (on the x axis) in E. coli pathotypes (y axis). Grayscale indicates color intensity change with the prevalence of gene of interest. Each rectangle corresponds to the prevalence of a gene of interest in particular pathotype. Pathotypes used include uropathogenic E. coli (UPEC), human intestinal commensal (fecal) E. coli (HFEC), enterotoxigenic E. coli (ETEC), enteropathogenic E. coli (EPEC), sepsis-associated human E. coli isolated from sepsis patients (SAEC), enteroaggregative E. coli (EAEC), avian pathogenic E. coli (APEC), chicken intestinal commensal (avian fecal) E. coli (AFEC), and atypical enteropathogenic E. coli (aEPEC).

DISCUSSION

E. coli is one of the most common inhabitants of the gastrointestinal tract of both animals and humans. Among pathogenic types of E. coli, APEC is a major bacterial pathogen of poultry with considerable zoonotic importance. APEC can potentially serve as a virulence gene reservoir for EHEC and demonstrate attachment to both human and chicken epithelial cells, indicating their transmission potential to humans (16). Investigations on APEC pathogenic mechanisms (19, 20) have previously identified adhesins, toxins, iron uptake systems, and resistance to host serum as major virulence factors. Previous studies have generally determined the attachment of E. coli by allowing them to attach to the host cells followed by detachment of cell monolayers and determination of the adherent bacteria in serial dilutions (21). This indirect and time-consuming approach has also been used for APEC (16). Here, we have developed a specific, rapid, and high-throughput method to detect and quantify the APEC adherence to human and chicken intestinal cell lines. We used the method to phenotypically screen a random transposon mutant library of 1,300 mutants and identified 11 adhesion-altered mutants for chicken and human intestinal cell lines. Complementation of adhesion-altered mutants with the respective cloned genes of the wild-type strain served as an additional validation of our screening method.

Four out of the 11 genes, wzx (encodes O-antigen flippase), rfbD (encodes dTDP-4-dehydrorhamnose reductase), a gene with no name assigned (GenBank accession no. AJB35136, encodes group 1 glycosyl transferase), and fdtA (encodes TDP-4-oxo-6-deoxy-alpha-d-glucose-3,4-oxoisomerase) are involved in lipopolysaccharide (LPS) biosynthesis (22–24). The corresponding transposon mutants (B6b, C7q, E4r, and H6n, respectively) showed greatly decreased ability to adhere to chicken and human cell lines (Fig. 2), and complementation of the H6n mutant with the cloned, intact fdtA gene recovered its ability to adhere to both cell lines (Fig. 4). Contribution of the O-antigen to adhesion and intestinal colonization has been studied in various E. coli pathotypes, but, to our knowledge, there have been no studies on APEC strains. Loss of the bacterial O-antigen is known to have pleiotropic effects on cell adhesion and host colonization (25). Loss of O-antigen was shown to affect bacterial adhesion, suggesting the reduced adhesion of the wzx, rfbD, and AJB35136 mutants is probably due to reduced expression of bacterial adhesins or lack of motility (26). The adhesion-altered mutants with transposon insertions in these genes showed reduced motility, but there are likely other factors that also contribute to decreased adhesion to LoVo and CHIC-8E11 cells. Other reports have indicated that the loss of O-antigen increases adhesion to human HEp-2 and HeLa cells, an effect which may be associated with different effects on the expression and exposition of adhesins at the bacterial cell surface (27, 28). The fdtA, as part of the O-antigen gene cluster, is reported to have isomerase activity and is involved in biosynthesis of 3-acetamido-3,6-dideoxy-d-galactose (d-Fuc3Nac), and its deletion should be responsible for lack of d-Fuc3Nac on the O-antigen side chain, with the main chain intact (24). It has been shown that F-type lectins expressed on the surface of mammalian cells are responsible for binding with molecules containing fucose residues (29, 30). The lack of the surface-exposed d-Fuc3Nac in LPS could be responsible for lowered adhesion to epithelial cells due to the fact that there is no interaction between LPS and lectin expressed on the surface of the host cells. Alternatively, lack of fdtA enzyme has influence on general sugar metabolism in bacteria, which affects other bacterial functions. Our results for the motility assays conducted on the fdtA adhesion-altered mutant (H6n; Fig. 3) would be consistent with the latter suggestion. We suggest that fdtA is essential for adhesion in APEC, and the changes in the O-antigen structure most likely influence adhesion to host cells in vivo and contribute to successful APEC colonization as well, but more work needs to be done.

Lowered motility is likely to affect adhesion to cell lines. It was therefore not surprising that transposon insertions in fliP, encoding a flagellar biosynthesis protein (31), and flgK, encoding a flagellar hook-associated protein (32), had lower motility and adhesion to cell lines than the wild-type strain. We have shown that in the cases of fdtA, wzx, rfbD, AJB35136, fliP, and flgK, decreased motility is associated with lower adhesion to cells. It might be assumed that increased motility could show the opposite effect. Consistent with these results, a recent study also showed that a mutation in nhaA gene encoding the Na⁺/H⁺ antiporter in E. coli K-12 strain had increased motility and FliC protein expression, which corroborates our findings (33).

The gene fdeC encodes an intimin-like adhesin, FdeC, containing an N-terminal β-barrel and a C-terminal extracellular domain and which has been suggested to have a potential role in UPEC colonization during uropathogenesis in a mouse model (34) and EHEC colonization of the terminal rectum of cattle (35). Surprisingly, we observed increased adhesion to both cell lines with the transposon mutant of the fdeC gene. These results might be explained by compensatory effects of an alternative adhesin resulting from regulatory crosstalk between adhesins. Such interactions have been shown previously between the two adhesin gene clusters fim and pap in UPEC (36). In other studies, an E. coli mutant incapable of expressing either T1F or P fimbriae compensated by synthesizing F1C fimbriae (37). We detected the fdeC gene in all tested isolates of different pathotypes (Fig. 5); however, the complementation of fdeC in the adhesion-altered mutants did not fully complement their adhesion defects in the CHIC-8E11 and LoVo cell lines (Fig. 4).

An insertion mutation of the ecpR gene in wild-type APEC IMT 5155 enhanced its adhesion only to the CHIC-8E11 cells (Fig. 2), and complementation with the cloned gene reduced adhesion in both cell lines (Fig. 4). This observation indicates a potential regulatory role in APEC-host cell adhesion. The E. coli common pilus (ECP) has been identified as an adherence factor of commensal and pathogenic E. coli for epithelial colonization (38). The ecpR gene encodes the ECP regulator that is a part of the ECP gene cluster (39). The ecpR-coded EcpR (MatA) protein has also been proposed as a regulatory protein based on its putative LuxR-like C-terminal helix-turn-helix (HTH) DNA-binding motif (39, 40). This gene product was shown to be a positive regulator of ECP that might control expression of other fimbrial clusters, as the transposon mutant of ecpR showed increased binding only to the CHIC-8E11 cell line. Previously, ECP has been shown to contribute to adhesion of APEC, EHEC, and commensal bacteria to HEp-2 and HeLa cells (38, 41). In our study, we suggest that ECP might also influence adhesion to chicken intestinal epithelium or the lack of ECP might result in expression of other fimbrial clusters. Further work will be required to clarify the role of transposon insertion in ecpR gene in the adhesion of APEC IMT 5155 to CHIC-8E11 cells.

The rluD gene encodes the ribosomal large subunit pseudouridine synthase D and is responsible for pseudouridine modifications in E. coli. These modifications not only play a vital role in bridging the two ribosomal subunits and stabilizing the ribosome but also may play an important role in the maturation of 50S ribosomal subunits (42). An rluD deletion mutant of the E. coli K-12 strain has been shown to have decreased motility (43). The increased adherence to LoVo cells seen in this study may therefore be attributed in part to more rapid settling of the bacteria with lowered motility (Fig. 3). The mutation in the rluD gene in the adhesion-altered mutant (D2t) from the transposon mutant library showed enhanced adhesion to LoVo cells only (Fig. 2), whereas its complemented counterpart showed a substantial reduction in its adhesion ability to both cell lines (Fig. 4).

Finally, a strain with transposon insertion into yjhB, which encodes a putative sialic acid transporter, showed improved adherence to both cell lines, suggesting an effect on the sialic acid regulation of type 1 fimbriae expression (44). Sialic acids are found nearly exclusively in animals and play an important role in regulation of immune response. Bacteria can use sialic acids both as a carbon source and as a “coat” providing protection from the immune response (45). Moreover, it has been found that sialic acid can regulate the expression of type 1 fimbriae in E. coli (46). Further research will be required to fully elucidate the role of this gene in pathogenesis of APEC.

In summary, we report the identification of 11 genes affecting adhesion of APEC to chicken and human intestinal epithelial cell lines. In the future, the site-directed mutagenesis to create specific mutants with desired adhesion effects and in vivo experiments can further elucidate the gene functions and validate our hypotheses. We found that most of the identified genes (eight) affected bacteria motility. Moreover, we found a contribution of ECP as adhesins contributing to adhesion to chicken intestinal cells, while its role in APEC pathogenesis has been reported earlier (41). We also report the role of the fdtA gene in APEC adhesion to chicken and human intestinal cell lines. It should be noted that although the cloned, intact genes did not fully complement the adhesion defects of particular adhesion-altered mutants, their role in adhesion cannot be ruled out, as bacteria are known to use multiple strategies in various alternate ways to adhere or colonize the host (47, 48). Identification of adhesion factors remains important, as antibodies elicited against adhesins such as fimH and Gal-Gal pili have shown potential to block bacterial colonization and prevent the disease as prophylactic vaccines (49, 50). This study therefore contributes to a better understanding of APEC pathogenesis, and future identification of the receptors in the host against these bacterial adhesion factors may also contribute to blocking the initial steps of infection, which ultimately can help in disease prevention and control.

MATERIALS AND METHODS

Bacterial strains.

The wild-type clinical isolate APEC IMT 5155 and its mutant library containing 1,300 random transposon mutants (with kanamycin resistance cassette) were obtained from the Free University of Berlin, Germany (18), and recovered on LB agar plates containing kanamycin at 50 μg/ml. The wild-type APEC IMT 5155 (WT) was grown on ChromAgar plates. The WT was used as a positive control in motility assays, while a nonmotile, sepsis-associated E. coli 5148 isolate was used as a negative control (51).

Cell lines and cell culture.

The human intestinal LoVo cell line (ATCC CCL-229) was from our laboratorýs cell line stocks, and the chicken intestinal cell line, clone CHIC-8E11 (MicroMol GmbH) was obtained from Karsten Tedin (Freie Universität Berlin, Germany). The cell lines were grown at 37°C with 5% CO₂ and passaged at a confluence of 80 to 90% according to standard protocols. The cell culture medium used for LoVo and CHIC-8E11 consisted of Dulbecco modified Eagle medium (DMEM)/Ham’s F12 salts (Millipore) supplemented with 10% bovine serum (Millipore), 2 mM l-glutamine, and 100 IU/100 μg per ml penicillin/streptomycin (Millipore). For E. coli adhesion assays, cell lines were seeded in 96-well plates (Nunclon).

Screening of transposon mutant library for adhesion.

Cells grown to near confluence in 96-well plates were washed twice with 200 μl of phosphate-buffered saline (PBS), and 100 μl of cell culture media without antibiotics was added. Wild-type APEC IMT 5155 and each of the transposon mutants was grown in 1 ml of LB broth at 37°C with shaking at 180 rpm in 96-deep well plates for 16 h. Before every adhesion assay, the optical density at 600 nm (OD₆₀₀) of each overnight culture was measured, and accordingly, the bacterial cultures were diluted in cell culture media without antibiotics. The same amount of each bacterial culture was added to the monolayers in triplicate with a multiplicity of infection of bacteria:host cells of 100 (approximately 3 × 10⁸ CFU/ml). Six wells without added bacteria served as negative controls, and six wells were inoculated with the wild-type APEC IMT 5155 strain as a positive control. The bacteria were allowed to adhere to the monolayer for 3 h at 37°C with 5% CO₂. The medium containing nonadherent bacteria in wells was discarded, and the wells were washed 3 times with 100 μl of 1× PBS.

A specific antibody staining method was optimized for detection and quantification of bacteria adherent to monolayers of cell lines. As the wild-type strain (APEC IMT 5155) belonged to serogroup O2, we used anti-E. coli O2 serogroup antibodies (Sifin, Germany). Cells/bacteria were fixed with 50 μl of 4% PFA (paraformaldehyde) in 1× PBS at 4°C for 1 h. The wells were washed 3 times with 100 μl of 1× PBS, and the plates were kept overnight or stored at 4°C. For antibody staining, the wells were blocked with 100 μl of blocking buffer (1× PBS containing 0.5% bovine serum albumin [BSA]) at room temperature (RT) for 5 min and then incubated with 50 μl of a 1:50 dilution of anti-coli O2 antibodies in blocking buffer and placed at 37°C for 30 min. The wells were then washed 3 times with 100 μl of blocking buffer following by staining with 50 μl of a 1:200 dilution of goat anti-rabbit IgG labeled with Alexa Fluor 647 in blocking buffer and kept at RT for 30 min in the dark. The wells were then washed 3 times with 100 μl of 1× PBS. The wells were counterstained with 50 μl of DAPI (4′,6-diamidino-2-phenylindole) for 30 s at RT for nuclear staining to validate the cell monolayer confluence and then washed 3 times with 100 μl of 1× PBS.

The plates were visualized by a fluorescence microscope coupled with an automated VideoScan technology with a module developed for counting antibody-stained bacteria (17). Briefly, an image of the well in the DAPI fluorescence channel was captured to ensure proper monolayers, and then the fluorescent bacteria in the Cy5 channel were photographed as 6 fields per well and counted by the automated software to obtain the results as the number of adherent bacteria per millimeter squared of monolayer. The average number of adherent wild-type strain IMT 5155 was normalized to 100, and transposon mutants showing either less than or equal to one-third (less than 33 bacteria per mm²) or at least 3 times higher (more than 300 bacteria per mm²) adhesion to the host cells were designated adhesion-altered mutants. The screening of all 1,300 mutants in triplicate was performed in two independent experiments, and the adhesion-altered mutants were again reexamined for adhesion in sextuplicates in 3 independent experiments to verify the particular adhesion-altered mutants. In order to rule out the effect of transposon mutation on bacterial growth, the growths of wild-type APEC IMT 5155 and the adhesion-altered mutants were studied in shake flask at 37°C with 180 rpm for 11 h by measuring the OD₆₀₀ after every hour.

FISH method.

A FISH staining method optimized at Brandenburg University of Technology, Germany (52), was applied to all of the adhesion-altered mutants to confirm their adhesion-altered property as an alternative antibody- and LPS-independent method for staining and quantification of adherent bacteria on chicken and human cell lines. The adhesion assay of overnight growth of adhesion-altered mutants was performed on chicken and human cell lines as mentioned above until the fixation step with PFA. The plates were washed three times with distilled water. The cells were dehydrated with 50 μl of 95% ethanol for 5 min, dried, and stored at 4°C until performing FISH. Forty microliters of FISH probe EUB338 Atto647N with a final concentration of 5 ng/μl were added to each well, and the plates were incubated at 46°C for 1 h in a humid chamber (52). The plates were washed once with washing buffer and incubated at 48°C for 10 min in washing buffer. Nuclei were stained with DAPI (50 μg/ml in distilled water) and washed once with water, and the plates were dried at room temperature for VideoScan analysis.

Identification of transposon insertion sites in adhesion-altered mutants.

The adhesion-altered mutants were subjected to arbitrary PCR using AmpliTaq Gold polymerase to locate the transposon cassette insertion site in these mutants and to identify the mutated genes. One primer was complementary to the kanamycin resistance gene in the transposon cassette (P9; Table S1 in the supplemental material), whereas the second primer used was arbitrary (Arbi5). The products of the first-round PCR were subjected to a second round/nested PCR using one arbitrary primer (Arbi2, as homologous to the 5′ sequence of Arbi5) and a transposon I terminus-specific primer (P6) as described earlier (18). The genomic DNA of wild-type APEC, and a reaction without template, served as negative controls. The amplicons were purified and Sanger sequenced commercially (LGC’s Agowa Genomics, Berlin, Germany). The obtained DNA sequences were analyzed in NCBI public databases using BLASTn, and the genes potentially contributing to adhesion were noted.

Cloning of intact potential genes from wild-type APEC IMT 5155 into the pACYC177.

The genes identified as affected in the adhesion-altered mutants were amplified from the wild-type APEC IMT 5155 strain by PCR using Phusion polymerase (Thermo Scientific) with primers designed for cloning in expression vector pACYC177 (Table S1). PCR products were purified by PCR purification kit (Qiagen), while the pACYC177 plasmid was isolated using the Midiprep kit (Qiagen). DNA concentrations were determined with a Colibri microvolume spectrometer (Titertek-Berthold). Purified PCR products were restriction digested with Fast Digest (FD) BamHI/HindIII enzyme (Thermo Scientific), ligated into pACYC177 previously digested with BamHI/HindIII using T4 DNA ligase (Thermo Scientific), and subsequently transformed by heat shock into E. coli XL1-Blue chemocompetent cells (53). Transformants were screened using a rapid colony screening protocol (54). Positive colonies were processed for plasmid isolation. Plasmids were purified and digested with BamHI/HindIII, and plasmids which showed restriction fragments of the expected sizes were subjected to Sanger sequencing using sequencing primers in Table S1. Positive clones harboring plasmids with cloned genes showing correct promoter, and open reading frames were stored at −20°C.

Complementation of adhesion-altered mutants with intact genes.

For adhesion complementation studies, electrocompetent cells of each of the 11 adhesion-altered mutants were prepared according to prior protocols with minor modifications (55). Briefly, each mutant was grown in 100 ml of LB media to reach an OD₆₀₀ of ∼0.6 and put immediately into an ice bath for 5 min before centrifugation for 10 min at 5,000 × g at 4°C. The pellet was washed once with 50 ml of sterile precooled (4°C) water and twice with 25 ml precooled 10% glycerol (in water) and finally suspended in 0.2 ml of 10% glycerol and frozen in 50-μl aliquots. For transformation, 100 ng of plasmid DNA was added to a particular mutant’s electrocompetent cells, incubated on ice for 1 min, transferred to a cuvette (4-mm gap), and electroporated at a voltage of 2.5 kV, a capacity of 25 μF, and resistance of 200 Ω. The bacteria were also transformed with vector plasmid alone (without insert) as a negative control. The bacterial colonies on agar plates were designated “adhesion-altered mutants complemented with gene of interest” and “adhesion-altered mutants complemented with empty plasmid only.” To exclude the effect of complementation on bacterial growth, the growths of wild-type APEC IMT 5155, adhesion-altered mutants, and the adhesion-altered mutants with complemented gene of interest were studied in a shake flask at 37°C with 180 rpm as mentioned above. The wild-type APEC IMT5155 and adhesion-altered mutants with or without complementing plasmids were again examined for their adhesion characteristics on CHIC-8E11 and LoVo cell monolayers in 96-well plates as described above. The adhesion of each mutant was determined in 6 wells in three independent experiments, and the number of bacteria adherent per millimeter squared of monolayers was detected by automated fluorescence microscopy as above. The average number of adherent wild-type bacteria in each of the 6 wells in 3 experiments was normalized to 100 and compared for statistical significance with similarly treated adherent bacteria of adhesion-altered mutants with and without complementation with potential genes to see the impact of plasmid complementation.

Motility assays.

Analysis of literature associated with transposon-affected genes in adhesion-altered mutants suggested that some of these genes might also have altered motility. Therefore, we tested selected strains with transposon mutations in genes nhaA, fdtA, flgK, fliP, rluD, wzx, rfbD, a group 1 glycosyl transferase encoded in E. coli APEC IMT5155 (AJB35136), and wild-type APEC IMT5155 for motility, while strain SAEC5148 was used as a negative control. Each bacterium was grown in 1-ml LB medium for 16 h at 37°C with 180-rpm shaking. Each isolate was inoculated into the center of a motility agar plate (LB containing 0.25% agar), and the plates were incubated at 37°C for 16 h. The diameter of the diffusion zone from the point of inoculation was noted. A diameter of > 5 mm was considered positive motility, whereas ≤ 5 mm diameter was considered negative motility (51). Three independent experiments were performed for each isolate.

Prevalence of potential adhesion genes.

The 11 identified potential adhesion genes were used for prevalence studies in 8 other pathotypes of E. coli. Primers were designed (Table S1), and conditions were optimized for amplification of the gene fragments from 10 isolates of each of the pathotypes. The pathotypes included atypical enteropathogenic E. coli (aEPEC), uropathogenic E. coli (UPEC), avian pathogenic E. coli (APEC), enterotoxigenic E. coli (ETEC), enteroaggregative E. coli (EAEC), human fecal E. coli (HFEC), avian fecal E. coli (AFEC), and sepsis-associated pathogenic E. coli (SePEC).