The novel E. coli cell division protein, YtfB, plays a role in eukaryotic cell adhesion.

Characterisation of protein function based solely on homology searches may overlook functions under specific environmental conditions, or the possibility of a protein having multiple roles. In this study we investigated the role of YtfB, a protein originally identified in a genome-wide screen to cause inhibition of cell division, and has demonstrated to localise to the Escherichia coli division site with some degree of glycan specificity. Interestingly, YtfB also shows homology to the virulence factor OapA from Haemophilus influenzae, which is important for adherence to epithelial cells, indicating the potential of additional function(s) for YtfB. Here we show that E. coli YtfB binds to N'acetylglucosamine and mannobiose glycans with high affinity. The loss of ytfB results in a reduction in the ability of the uropathogenic E. coli strain UTI89 to adhere to human kidney cells, but not to bladder cells, suggesting a specific role in the initial adherence stage of ascending urinary tract infections. Taken together, our results suggest a role for YtfB in adhesion to specific eukaryotic cells, which may be additional, or complementary, to its role in cell division. This study highlights the importance of understanding the possible multiple functions of proteins based on homology, which may be specific to different environmental conditions.

No difference in LPS profile and lysozyme sensitivity of UTI89 and UTI89 ΔytfB A. Lipopolysaccharide (LPS) was extracted from overnight cultures of UTI89 (1) and UTI89 ΔytfB (2), and increasing concentrations were separated using a 4-12% SDS gel before being silver stained. No difference in profile was observed between wild type and ΔytfB mutant. Protein standards are shown in kDa.
B. Sensitivity of UTI89 and UTI89 ΔytfB to lysozyme was measured using a sold medium assay (i) or in liquid (ii) and showed no difference in sensitivity between wild type and mutant. For the solid assay, 0.5 mg/ml lysozyme was spotted onto agar containing a bacterial suspension and zones of lysis were recorded. Lysozyme sensitivity was normalised to that of UTI89. For the liquid assay, bacterial suspensions were incubated with 10 mg/ml lysozyme with shaking at 37 °C and a reduction in optical density was recorded over time. Experiments were performed with at least two biological replicates and error bars represent the standard error of mean. YtfB is not required for in vitro bladder cell adhesion, invasion or intracellular growth.
Statically grown UTI89 and UTI89ΔytfB incubated with a monolayer of PD07i bladder epithelial cells and adherence (A), invasion (1h intracellular growth) (B) or intracellular growth (24h intracellular growth) (C) measured. The data are averages of at least three biological replicates, and error bars represent SEM. P values were determined using an unpaired student T test. Adherence is displayed as a percentage compared to wild type UTI89, whose adherence, invasion or growth is represented as a dashed line. YtfB is not required in vivo using competitive UTI and CAUTI infections.
The mice were infected with either an ascending UTI (A) or catheter-associated UTI (B) with a 1:1 ratio of E. coli strains UTI89: UTI89ytfB::kan (labelled as UTI89 in the figure) or UTI89∆ytfB: UTI89ytfB::kan (labelled as UTI89∆ytfB in the figure) at 10 7 colony forming units per inoculum, harvested at 24 hours post infection, and the recovered bacteria were enumerated on selective media for each strain. The competitive index (CI) was calculated as described in the materials and methods. Each dot represents one mouse, and the solid horizontal lines indicate the median. The limit of detection (LOD) of 40 CFU is indicated by the dotted line. Two biological replicates were carried out for each infection model, with 3-5 mice per experiment for both competitive infections. Mice that lost the catheter or died prior to the time of sacrifice were omitted. Statistical analysis was performed using the Mann-Whitney test.

Static infection of bladder cells.
The human bladder cell line PD07i was cultured to confluence in 24well cell culture plates at 37 ˚C in 5% CO2 in EpiLife + HKGS. Bacterial strains were grown statically overnight at 37 °C in 5ml LB medium to promote fimbrial expression. Bacterial cultures were diluted to 10 9 cells/ml (OD600=1). To infect PD07i cells, 0.1 µl/mm 2 of bacterial culture was added to each well, to give an MOI of 100. Cell culture plates were centrifuged for 5 min, 600 g at 24°C to force the bacteria onto the cell surface then incubated at 37°C with 5% CO2 for 2 hours. For the adhesion assay, cells were washed with PBS 6 times. To lyse the mammalian cells, 1ml of lysis solution (0.5 % Trypsin-EDTA, 0.1% Triton X-100) was added and incubated for 15 minutes at 37 ˚C in 5% CO2, and bacterial cells were resuspended by pipetting. Dilutions of the bacteria were plated onto LB agar plates to count the CFU/ml. For bacterial internalisation and intracellular growth assays, the EpiLife media was removed and replace with 4 µl/mm 2 of EpiLife + HKGS + 100 µg/ml gentamycin. This was incubated statically for 1 h (internalisation) or 24 h (intracellular growth). The media was then removed and the wells washed with PBS four times. The bladder cells were lysed and plated as for the adhesion assay. Averages of technical replicates were calculated and the data was normalised to the wild -type control before means and standard errors of the means were calculated for the biological replicates. Unpaired student T-tests were then performed to test for significant differences between strains.
Murine CAUTI and UTI model. The urinary tract infection model was carried out as previously described 5 . Briefly, bacterial strains were grown in 15 mL LB media for 16-18 hrs with shaking at 37°C. Cells were pelleted, resuspended in 5 mL sterile 1× PBS and the OD was normalized to 2 × 10 8 CFU/mL. For competitive CAUTI, equal volume of each strain was mixed prior to infection. Firstly, groups of 5 isofluorane-anesthetized female wild-type C57BL/6 mice (7-8 weeks old, 22 to 25g; InVivos, Singapore) were infected via intraurethral catheterization (polyethylene catheter, 5mm long, 0.61mm) with 50 µL of bacteria inoculum. Mice were euthanized by carbon dioxide inhalation and cervical dislocation 24 hpi after transurethral challenge, and bladder, kidney pairs were aseptically excised, weighed and homogenized in 1 mL and 0.8 mL 1× PBS, respectively using a homogenizer (Pro200, SPD scientific, Singapore) for approximately 10 sec at high speed. The catheter implant was also retrieved from each bladder and placed in 1mL 1× PBS prior to water bath sonication at r.t.p for 5 mins. Serial homogenate dilutions were plated onto MacConkey agar selection plate with and without kanamycin for CFU enumeration. The limit of detection is 40 CFU. The murine CAUTI model was performed twice independently and murine that lose the catheter at the time of sacrifice was omitted and a P value of less than 0.05 was considered significant.
For competitive UTI, the mice were infected with 50 µL of 2× 10 8 CFU/mL of mixed bacterial inoculum and the steps for euthanization, excision of bladder and kidneys as well as the preparation of the homogenate is the same as above.