Plasmids and primers
Plasmids and primers used in this work are listed in Supplementary Tables S2 and S3, respectively. qRT-PCR primers are listed in Supplementary Table S5.
Human bladder epithelial cell culture
Human 5637 bladder epithelial cells (HTB-9™, procured from ATCC) were cultured in RPMI 1640 medium (ATCC) supplemented with 10% Fetal Bovine Serum (FBS, Gibco) at 37°C, 5% CO2. Cells were split every other day by gentle detachment with trypsin-EDTA 0.05% (Gibco). Cells were either passaged at a 1:3 ratio in fresh RPMI 10% FBS for cell line maintenance, or diluted in phenol-red-free RPMI medium (Gibco) 5% FBS for microscopy experiments.
Bacterial culture
The UPEC CFT073 strain was originally isolated from a patient with bacteraemia of urinary tract origin (Mobley et al., 1990) and procured from ATCC. CFT073 wild-type and mutant strains were inoculated from frozen stocks in LB Miller medium (Difco) and cultured overnight at 37°C without shaking to induce type 1 pili expression for microscopy experiments (Snyder et al., 2004), or with shaking at 170 rpm for molecular cloning experiments. When required, LB was supplemented with 100 µg/mL ampicillin (Sigma) or 50 µg/mL kanamycin (Sigma).
Construction of the CFT073 ∆ fliC sfGFP parental strain
The fliC locus (between nucleotides 2,151,497 and 2,153,284) was deleted using the λ Red recombineering system (Datsenko and Wanner, 2000). Briefly, CFT073 was first transformed with plasmid pKD46 to express the λ Red recombinase, then with the kanR resistance cassette amplified from plasmid pKD4 using primers fliC_KO_F and fliC_KO_R. Successful chromosomal recombination was confirmed by PCR using primers fliC_F and fliC_R, and the kanR resistance cassette was removed using plasmid pCP20 encoding the FLP recombinase. Loss of swimming motility was assessed by soft-agar assay, as previously described (Lane et al., 2005).
The sfGFP sequence was introduced under control of a strong Pσ70 promoter at the chromosomal attHK022 site (between nucleotides 1,090,927 and 1,090,928) using a recently described method (Eshaghi et al., 2016). Briefly, the kanR-PRhaB-relE toxin cassette was amplified from plasmid pSLC217 using primers P4 and P5 and integrated into the genome of CFT073 ∆fliC using λ Red recombineering. Primers P6 and P7 were then used to amplify the sfGFP-containing fragment from plasmid pSLC293, and these PCR amplicons were used to replace the kanR-PRhaB-relE toxin cassette by λ Red recombineering with negative selection on M9 agar (Sigma) supplemented with 0.2% L-rhamnose (Sigma). Successful recombination at each step was confirmed by PCR using primers attHK_F and attHK_R. The same method was used to introduce the sfGFP sequence in the wild-type (fliC+) CFT073 background.
Construction of the transposon insertion library
Tn5 transposon insertion mutants were generated in the CFT073 ∆fliC sfGFP parental strain using the EZ-Tn5 transposome kit (Epicentre). Briefly, bacteria were electroporated with transposome complexes and allowed to recover in SOC broth for 1h at 37°C / 170 rpm prior to plating on LB agar supplemented with 50 µg/mL kanamycin and incubation overnight at 37°C. A total of 17,600 mutants was obtained, from which 8,184 were picked, grown overnight in LB kanamycin 50 µg/mL, added with 15% glycerol (Fisher Chemical) and archived in 96-well plates.
Spinfection of bladder epithelial cells with UPEC
CellCarrierTM-96 Ultra microplates (Perkin-Elmer) were plasma-activated, coated with an ice-cold solution of 50 µg/mL bovine dermis native collagen (AteloCell) in PBS (Gibco), and incubated for 3h at 37°C for the collagen to polymerize. Wells were washed twice with PBS to remove unbound collagen, seeded with 5637 epithelial cells in phenol-red-free RPMI 5% FBS at a density of 50,000 cells per well, and incubated overnight at 37°C, 5% CO2 to reach confluency. On the day of the experiment, RPMI was removed from the plates and wells were rinsed once with PBS. Overnight UPEC cultures in 96-well plates were diluted 1:10 in phenol-red-free RPMI 5% FBS and added to the epithelial cells at a multiplicity-of-infection (MOI) of 200:1 using a Liquidator 96 (Mettler-Toledo). Plates were spun down at 600 g for 5 minutes and incubated for 20 minutes at 37°C, 5% CO2. Wells were then washed twice with PBS to remove non-adherent bacteria and phenol-red free RPMI 5% FBS was added back.
Immunostaining of bladder epithelial cells
Confluent 5637 bladder epithelial cells in 96-well plates were fixed with 4% paraformaldehyde (PFA) for 30 minutes. Fixed cells were washed three times with PBS, permeabilized with 0.15% Triton X-100 for 15 minutes, washed three times with PBS, and blocked with antibody incubation buffer (PBS supplemented with 1% BSA and 0.01% Triton X-100) for 1 hour. Cells were then incubated with primary antibodies (rabbit monoclonal Alexa Fluor® 647 anti-cytokeratin 8 antibody from Abcam, or mouse monoclonal anti-uroplakin IIIa antibody from Santa Cruz) at a 1:100 dilution in antibody incubation buffer overnight at 4°C, and washed three times with PBS. For uroplakin staining, cells were subsequently incubated with secondary antibodies (donkey anti-mouse Alexa Fluor® 647 highly cross-adsorbed IgG from Invitrogen) at a concentration of 2 µg/mL in antibody incubation buffer for 1 hour at room temperature, and washed three times with PBS. Cell nuclei were stained with a 5 µg/mL DAPI (Invitrogen) solution in PBS for 30 minutes, followed by three additional washes with PBS. Cells were imaged with a 63X oil objective on a Leica SP8 confocal microscope, and images were deconvolved using SVI Huygens (Quality, 0.05; Iterations, 40).
High-content imaging of bladder epithelial cells infected with UPEC
Wells containing infected 5637 bladder cells were imaged in 9 fields of view using a high-throughput Operetta CLS microscope (Perkin-Elmer) on the phase-contrast and green fluorescence channels with a 20X air objective (NA = 0.40). The microplate was maintained at 37°C, 5% CO2 during imaging. Image analysis was performed using the integrated Harmony® high-content analysis software (Perkin-Elmer). Briefly, the mean GFP fluorescence was displayed for each field of view and the average and standard deviation for the 9 fields of view were calculated for each well. For wells where the coefficient of variation (CV) exceeded 10%, images were manually inspected and fields of view with fluorescence artefacts removed from the analysis.
Screening for UPEC mutants with altered adhesion: 1st step
In the first screening step, 96-well plates were arranged with 93 transposon insertion mutants, the ∆fliC parental strain, the ∆fliC dsbA-Tn low-adhesion strain, and one blank well containing LB medium only. After overnight culture, they were used to infect 5637 bladder epithelial cells following the aforementioned procedure. OD600 was also recorded for each well and mutants that did not grow were removed from the analysis. After spinfection and imaging, the average and standard deviation of GFP fluorescence on a particular plate were calculated based on the individual fluorescence values for each well, and mutants with values more than 1.5 standard deviations from the average were selected for retesting. We always made sure that the value for the ∆fliC parental strain was less than 1.5 standard deviations from the average, and that the value for the ∆fliC dsbA-Tn low-adhesion strain was more than 2 standard deviations from the average, otherwise the experiment was repeated for that particular plate. We also visually controlled the integrity of the epithelial cell layer in each well by phase-contrast microscopy.
Screening for UPEC mutants with altered adhesion: 2nd step
In the second screening step, 96-well plates were arranged with 3 wells of each candidate mutant and 30 wells containing the ∆fliC parental strain. After overnight culture, they were used to infect 5637 bladder epithelial cells following the aforementioned procedure. After spinfection and imaging, the average and standard deviation of GFP fluorescence were calculated for the 30 replicates of the parental strain, and mutants with values more than 2 standard deviations from the average were selected for retesting.
Screening for UPEC mutants with altered adhesion: 3rd step
In the third screening step, candidate mutants were cultured overnight in individual flasks along with the ∆fliC parental strain. All cultures were adjusted to the same final OD600 value in pre-warmed phenol-red-free RPMI 5% FBS, so that this would correspond to a 1:10 dilution for the mutant with the lowest overnight OD600. 96-well plates were arranged with 6 wells of each mutant and 6 wells containing the ∆fliC parental strain. The rest of the procedure was conducted as described above. For each mutant, the distribution of GFP fluorescence values after spinfection was compared to the distribution for the ∆fliC parental strain using Welch’s t-test and strains with P < 0.05 were selected for further analysis.
Identification of transposon insertion sites by RATE (rapid amplification of transposon ends) PCR
Tn5 insertion sites were identified using a single-primer PCR protocol adapted from Ducey et al. Briefly, candidate mutants were streaked on LB agar supplemented with 50 µg/mL kanamycin. A single colony was boiled for 10 minutes in 50 µL water and 1 µL was used as template in a standard 50-µL PCR reaction with primer R6K_inv1 using a double amount of Platinum® Taq High Fidelity DNA Polymerase (Invitrogen). The first 30 cycles of the PCR reaction were done at 55°C with a 60-sec extension time; the next 30 cycles were done at 30°C with a similar extension time; and the last 30 cycles were performed at 55°C with a 2-minute extension time. PCR amplicons were then sent for sequencing using the nested R6K_RP1 primer, and the sequence was aligned to the CFT073 chromosome.
Assignment of P pili transposon insertions to the pap1 or pap2 operon
To determine whether a transposon insertion occurred within the pap1 or pap2 operon, we performed a dual set of standard 25-µL Q5® PCR reactions (New England Biolabs) using a first primer specific to the transposon sequence, and a second primer specific to the DNA sequence flanking either of the two operons. Primer pairs were chosen based on the position and orientation of each transposon insertion, as indicated in Supplementary Table S4. PCR amplicons were analysed on 1% agarose gels stained with SYBR® Safe DNA Gel Stain (Invitrogen), and identification of the gene disrupted was based on the brightest amplicon band.
Construction of unmarked UPEC deletion strains
The fimH locus (between nucleotides 5,143,529 and 5,144,440), fim operon (between nucleotides 5,135,689 and 5,144,440), foc operon (between nucleotides 1,188,213 and 1,194,827) and pap operons (pap1 operon between nucleotides 3,429,593 and 3,437,564; pap2 operon between nucleotides 4,940,831 and 4,948,799) were deleted from CFT073 sfGFP and CFT073 ∆fliC sfGFP using the λ Red recombineering system as described above. The kanR resistance cassette was amplified from plasmid pKD4 using primer pairs fimH_KO_F + fimH_KO_R, fimB_KO_F + fimH_KO_R, focA_KO_F + focH_KO_R, and papA_KO_F + papG_KO_R respectively. For deletion of the pap operons, we used the same primer pair for integration of the kanR resistance cassette in place of either operon and then identified the disrupted operon using specific primer pairs. Successful chromosomal recombination was confirmed by PCR using primer pairs fimH_F + fimH_R for ∆fimH, fimB_F + fimH_R for ∆fim, focA_F + focH_R for ∆foc, papI_F + papG_R for ∆pap1, and papI2_F + papG2_R for ∆pap2. The double pap deletion mutants were obtained by successively deleting the pap1 and pap2 operons from CFT073 sfGFP and CFT073 ∆fliC sfGFP. Adhesion phenotypes of all mutants were assessed using the same method as in the third step of the screen. To be able to pool data coming from different experiments, results were normalized to the ∆fliC parental strain in each experiment.
Quantitative real-time PCR (qRT-PCR)
CFT073 mutant strains cultured overnight in static conditions were pelleted and RNA was extracted using the High Pure RNA Isolation kit (Roche) according to the manufacturer’s instructions. RNA samples were DNAse-treated using the TURBO DNA-free™ kit (Invitrogen), and cDNA was obtained using the SuperScript™ IV reverse transcriptase with random hexamers (Invitrogen). qRT-PCR reactions were set up using the SYBR® Green PCR Master Mix (Applied Biosystems) with 2.5 µM primers and 30 ng cDNA. Reactions were run on an ABI PRISM7900HT Sequence Detection System (Applied Biosystems). For each gene and each experiment, relative expression levels were calculated by normalizing to the expression level of the housekeeping gene gapA and to the expression level in the ∆fliC parental strain.
Biotinylation of surface-exposed proteins and protein affinity purification
The bacterial surface proteome was isolated as described by Monteiro et al. Briefly, CFT073 mutant strains cultured overnight in static conditions were pelleted, washed twice with PBS, and incubated for 30 minutes in PBS supplemented with 1% (m/m) sulfo-NHS-SS-biotin (Thermo Scientific Pierce™) under gentle agitation at room temperature. Excess of the biotinylation reagent was quenched by three washes with a 500 mM glycine (AppliChem) solution in PBS, and bacteria were resuspended in PBS supplemented with 1% (v/v) Triton X-100 and Protease Inhibitor Cocktail (Roche). Cell lysis was performed using FastPrep (MP Biomedicals) with two steps of 20 sec at 6 m/s, and cell debris were pelleted by centrifugation at 20,000 g for 30 minutes. Biotinylated proteins in the supernatant were affinity-purified using High Capacity NeutrAvidin™ Agarose (Thermo Scientific Pierce™) according to the manufacturer’s instructions. Briefly, agarose slurry was packed into 2 mL Centrifuge Columns (Thermo Scientific Pierce™), washed with 3 column volumes of PBS, and incubated with protein samples for 60 minutes at 4°C under gentle rotation. Unlabelled proteins were washed away with 5 column volumes of PBS, and biotinylated proteins were eluted with DTT-modified Laemmli buffer (2% SDS, 20% glycerol, 62.5 mM Tris-HCl, 50 mM DTT, 5% ß-mercaptoethanol). The protein concentration in eluted samples was assessed using the Pierce™ 660nm Protein Assay (Thermo Scientific Pierce™) according to the manufacturer’s recommendations for samples eluted in Laemmli buffer. To assess the efficiency of the surface biotinylation and affinity-purification steps (Supplementary Figure S8A), CFT073 ∆fliC sfGFP protein samples were prepared as described above, except that lysis was performed with 2% SDS (Sigma). Samples before affinity-purification (total cell extracts) were then directly analysed by mass spectrometry along with affinity-purified samples (surfaceomes).
Protein sample preparation
Mass spectrometry-based proteomics-related experiments were performed by the Proteomics Core Facility at EPFL. Protein samples (10 µg per dataset) were loaded on an SDS-PAGE gel and allowed for a short migration. The gel pieces containing the concentrated proteins were excised, washed twice with 50% ethanol in 50 mM ammonium bicarbonate (AB, Sigma) for 20 minutes, and dried by vacuum centrifugation. Proteins were reduced with 10 mM dithioerythritol (Merck-Millipore) for 1 hour at 56°C followed by washing/drying. Reduced proteins were alkylated with 55 mM iodoacetamide (Sigma) for 45 minutes at 37°C in the dark followed by washing/drying. Alkylated proteins were digested overnight at 37°C using mass spectrometry grade Trypsin Gold (Promega) at a concentration of 12.5 ng/µL in 50 mM AB supplemented with 10 mM CaCl2. Resulting peptides were extracted in 70% ethanol, 5% formic acid (FA, Merck-Millipore) twice for 20 minutes, dried by vacuum centrifugation and stored at -20°C until further analysis.
Mass spectrometry analysis
Peptides were desalted on C18 StageTips (Rappsilber et al., 2007) and dried by vacuum centrifugation prior to LC-MS/MS injections. Samples were resuspended in 2% acetonitrile (Biosolve), 0.1% FA and nano-flow separations were performed on a Dionex UltiMate™ 3000 RSLCnano UPLC system (Thermo Scientific™) on-line connected with an Orbitrap Exploris™ 480 mass spectrometer (Thermo Scientific™). A capillary precolumn (Acclaim™ PepMap™ C18, 3 µm-100Å, 2 cm x 75 µm i.d.) was used for sample trapping and cleaning. A 50cm long capillary column (75 µm i.d., in-house packed using ReproSil-Pur C18-AQ 1.9 µm silica beads, Dr. Maisch) was then used for analytical separations at 250 nL/min over 150-minute biphasic gradients. Acquisitions were performed through Top Speed Data-Dependent acquisition mode using a cycle time of 2 seconds. First MS scans were acquired with a resolution of 60’000 (at 200 m/z) and the most intense parent ions were selected and fragmented by High energy Collision Dissociation (HCD) with a Normalized Collision Energy (NCE) of 30% using an isolation window of 2 m/z. Fragmented ions were acquired with a resolution of 15’000 (at 200 m/z) and selected ions were then excluded for the following 20 seconds.
Bioinformatic analysis
Raw data were processed using MaxQuant 1.6.10.43 (Cox and Mann, 2008) against the Escherichia coli CFT073 Uniprot database (5336 entries, last modification 210307) Carbamidomethylation was set as fixed modification, whereas oxidation (M), phosphorylation (S, T, Y), acetylation (Protein N-term), CAMthiopropanoyl (K and Protein N-term) and glutamine to pyroglutamate were considered as variable modifications. A maximum of two missed cleavages were allowed and “Match between runs” option was enabled. A minimum of two peptides was required for protein identification and the false discovery rate (FDR) cutoff was set to 0.01 for both peptides and proteins. Label-free quantification and normalisation was performed by MaxQuant using the MaxLFQ algorithm, with the standard settings (Cox et al., 2014).
The statistical analysis was performed using Perseus 1.6.12.0 (Tyanova et al., 2016) from the MaxQuant tool suite. Reverse proteins, potential contaminants and proteins only identified by sites were filtered out. Protein groups containing at least three valid values in at least one group were conserved for further analysis. Empty values were imputed with random numbers from a normal distribution (width: 0.3, down shift: 1.8 sd). A two-sample t-test with permutation-based FDR statistics (250 permutations, FDR = 0.01, S0 = 0.5) was performed to determine significant differentially abundant candidates. Further graphical displays were performed using homemade programs written in R (https://www.R-project.org/).
Scanning electron microscopy (SEM) sample preparation
5637 bladder epithelial cells spinfected with UPEC in µ-dishes (Ibidi) or CFT073 mutant strains left to adhere for 20 minutes on poly-L-lysine-coated coverslips were fixed for 30 minutes with 1.25% glutaraldehyde in PBS 0.1M pH 7.4. Samples were post-fixed for 30 minutes with 0.2% osmium tetroxide in 0.1 M cacodylate buffer followed by washing with distilled water. Next, samples were dehydrated in graded ethanol series and dried in an automated critical point dryer (Leica Microsystems). Finally, samples were attached to an adhesive conductive surface followed by coating with 3–4 nm of gold/palladium (Quorum Technologies). Images of the samples were acquired using a field emission scanning electron microscope (Zeiss NTS).
Automated quantification of single-cell piliation using Ilastik
The Ilastik software (https://www.ilastik.org) was trained by manually annotating background, pili and bacterial body on four SEM images of the ∆fliC parental strain. After applying the trained software on an SEM image, the channel corresponding to the bacterial body was further processed using Fiji: a threshold was applied to keep only pixels with intensity values from 0.30 to 1, holes were filled, and particles were analysed to keep only circular particles with an area above 250,000 pixels. The channel corresponding to the pili was then subtracted with the processed bacterial body, a threshold was applied to keep only pixels with intensity values between 0.55 and 1, and particles were analysed to keep non-circular particles (circularity = 0.00-0.20) with an area above 100 pixels, resulting in the processed pili image (see also Supplementary Figure S9). Single-cell piliation was calculated for each image by simply dividing the area of the pili by the area of the bacterial body. To be able to pool data coming from different experiments, results were normalized to the ∆fliC parental strain in each experiment.
Gentamicin protection assay to assess bacterial invasion
Invasion assays were performed essentially as described by Martinez et al. Briefly, confluent 5637 bladder epithelial cells in two sets of triplicate wells were spinfected with UPEC in RPMI 5% FBS at a MOI of 50:1. Plates were incubated for 20 minutes at 37°C, 5% CO2 and washed twice with PBS to remove non-adherent bacteria. One set of triplicate wells was lysed by the addition of trypsin-EDTA and Triton X-100, both at final concentration of 0.05%, and bacteria in these lysates were titered on LB agar plates, representing the adherent UPEC fraction. The second set of triplicate wells was added back with RPMI 5% FBS, incubated for 2 hours, washed twice with PBS, and incubated for another 2 hours in RPMI 5% FBS containing 100 µg/mL gentamicin (Gibco) to kill extracellular bacteria. Wells were then washed twice with PBS, lysed and titered, representing the invasive UPEC fraction. Invasion frequencies were calculated by dividing the number of bacteria in the invasive fraction by the number of bacteria in the adhesive fraction.