Bacterial strains and growth conditions
Bacterial strains used and generated in this work are detailed in Supplementary Table S1. E. coli and P. syringae strains were grown with aeration in Lysogeny Broth (LB) medium88 at 37ºC for E. coli or 28°C for P. syringae. Antibiotics were used, when necessary, at the following concentration: ampicillin (Amp), 100 mg/ml for E. coli and 500 mg/ml for P. syringae; kanamycin (Km), 50 mg/ml for E. coli and 15 mg/ml for P. syringae derivative strains; gentamycin (Gm), 10 mg/ml; nitrofurantoin 40 mg/ml, and cycloheximide, 2 mg/ml.
To induce the expression of the hrp/hrc genes, bacteria were initially cultured overnight in LB at 28°C, supplemented with the appropriate antibiotic, then, washed twice in 10 mM MgCl2 before being cultured in hrp-inducing minimal medium (HIM), containing 10 mM fructose89. For this study, pH of HIM was adjusted to 7.0 with 10N NaOH. The initial OD was adjusted to 0.13 and cultures were incubated at 28°C with agitation.
Fluorescent labelling of bacterial strains
Bacterial strains carrying a chromosome-located transcriptional fusion of fliC gene to a promoterless tdTomato gene was generated using an adaptation of the method previously described in Zumaquero et al.90. The plasmids used and generated for this purpose are detailed in Supplementary Table S2. The primers used are described in Supplementary Table S3. For the generation of the allelic exchange plasmid, two fragments of approximately 500 pb were amplified from Pph 1448A genomic DNA using Q5 High-Fidelity DNA Polymerase (New England Biolabs, USA); one of these fragments (A) encompasses the 3’ end of the ORF, including the STOP codon, while the other fragment (B) covers the sequence immediately downstream to the STOP codon. All primers used are listed in the key resource table. Each reaction was carried out at 98ºC for 1 minute for the initial denaturation step, followed by 30 cycles at 98ºC for 30 seconds, annealing at 62ºC for 30 seconds, and extension at 72ºC for 30 seconds, followed by 5 minutes at 72ºC for the final extension step. The reaction mixture for each PCR included 0.64 mM deoxynucleotide triphosphate (dNTP) mix, 0.4 ng of each primer, 1 ng of genomic DNA, the appropriate enzyme buffer, and commercial ultrapure water (Nalgene, Rochester, NY, USA). Two µl of each gel-purified PCR product was employed as template for the subsequent fusion PCR, employing primers A1 and B2, in a PCR reaction conducted under the conditions described, with an extended elongation time of 1 min. The resulting bands, comprising the end of each ORF and its downstream sequence separated by an EcoRV restriction site, were A/T cloned into pGEM-T (Promega, USA) and subjected to full sequencing to discard clones carrying mutations. This process rendered the pGT-AB-fliC plasmid needed for generating the allelic exchange plasmid.
The sequence of the promoterless tdTomato (tdT) gene was acquired through PCR amplification from the tdTomato-pBAD plasmid (Michael Davidson & Nathan Shaner & Roger Tsien, addgene plasmid #5485691) using the ProtFluorF and ProtFluorR primers. This PCR generated a tdTomato fragment with a 5´EcoRV restriction site and a 3´EcoRI restriction site. Similarly, the kanamycin cassette was amplified using the P1 EcoRV and P2 EcoRI primers, and plasmid pKD492 served as DNA template to obtain the FRT-nptII-FRT fragment needed for the resistance to kanamycin. This fragment featured a 5´ EcoRV and 3´EcoRI restriction site. Both amplification reactions were performed using Q5 High-Fidelity DNA Polymerase (New England Biolabs, USA) in a 50 µL PCR reaction mixture consisting of 500 ng of plasmid as DNA template, 0.5 µM of each primer, 0.2 mM dNTP´s and 0.5 µL of Q5 High-Fidelity DNA Polymerase. Each reaction followed a thermal cycling protocol that began with an initial step of 98⁰C for 2 minutes, followed by 30 cycles at 98⁰C for 10 seconds, 60⁰C for 30 seconds, and 72⁰C for 45 seconds, concluding with a final extension at 72⁰C for 5 minutes. The gel-purified fragments, both the tdTomato and the FRT-nptII-FRT, underwent digestion with EcoRI and EcoRV enzymes and were cloned into the EcoRV-digested pGT-AB-fliC plasmid, rendering the allelic exchange plasmid named pGT-fliC::tdT.
Strain 1448A fliC::tdT was obtained by introducing the pGT-fliC::tdT plasmid into P. syringae pv. phaseolicola 1448A through electroporation, following the method described in Zumaquero et al.90. Selection was carried out on LB plates with kanamycin. Subsequently, the resulting colonies were replicated onto LB plates with ampicillin (500 µg/ml) to discard the colonies with plasmid integration, which is indicative of a single recombination event. The colonies that exhibited kanamycin resistance but ampicillin sensitivity, were then confirmed through PCR, using the A1 fliC and B2 fliC primers; and through Southern blot analysis, employing the nptII gene as a probe, to confirm proper allelic exchange resulting from a double recombination event occurring at a unique position within the genome. To generate strains carrying two chromosomal transcriptional fusions, the pGT-fliC::tdT plasmid was transformed into the previously generated strains 1448A hrpL::GFP3, 1448A hopAB1::GFP3 and 1448A hrcU::GFP3 and to generate the 1448A hrpL::GFP3 fliC::tdT, 1448A hopAB1::GFP3 fliC::tdT and 1448A hrcU::GFP3 fliC::tdT strains.
Bacterial strains carrying a chromosome-located transcriptional fusion of fliC gene to a promoterless GFP3 gene, 1448A fliC::GFP3 strain, were generated following the method described by Rufián et al., 2018a93 with some modifications. The GFP3-FRT-nptII-FRT fragment was obtained by digesting the plasmid pGT-GFP+ 93 with the EcoRI digestion enzyme. This fragment consists of the promoterless GFP3 gene, complete with its ribosomal binding site (RBS), followed by the kanamycin resistance gene (nptII), flanked by FRT (flipase recognition targets) sites, with the entire construct bordered by two EcoRI restriction sites. The GFP3-FRT-nptII-FRT fragment was then blunt-ended through a PCR procedure and ligated into EcoRV-digested pGT-AB-fliC through blunt-end ligation, leading to the generation of the pGT-fliC::GFP3 plasmid. Subsequently, the resulting plasmid was transformed into P. syringae pv. phaseolicola 1448A to generate the 1448A fliC::GFP3 strain.
Additionally, the constitutively-expressed fluorescent reporter gene eCFP was introduced into the chromosome of the 1448A fliC::GFP3 and 1448A hopAB1::GFP3 fliC::tdT strains using a Tn7 delivery system, as previously described by Lambertsen et al.95 to generate the 1448A fliC::GFP3 eCFP and 1448A hopAB1::GFP3 fliC::tdT eCFP strains.
Generation of mutant bacterial strains
The bacterial strain carrying a deletion of the fleQ gene was generated following the method described in Zumaquero et al.90, which involves the generation of gene knockouts by allelic exchange, replacing the specific ORF by a kanamycin cassette. The allelic exchange plasmid pGT-ΔfleQ was generated as previously described for the generation of the pGT-AB-fliC using primers A1 ΔfleQ, A2 ΔfleQ, B1 ΔfleQ and B2 ΔfleQ and the same experimental settings described above. The FRT-nptII-FRT fragment was obtained by PCR amplification using P1 EcoRI and P2 EcoRI primers and pKD4 as template, and the kanamycin cassette was finally inserted by ligation in EcoRI restriction site, generating the allelic exchange plasmid pGT-ΔfleQ. This plasmid was transformed into 1448A Pseudomonas syringae pv. phaseolicola and mutants were obtained as described in Zumaquero et al.90.
Generation of pFleQ
pFleQ was generated using the backbone of pBBRMCS-496. For its generation, the fleQ ORF was amplified using Q5 High-Fidelity DNA Polymerase (New England Biolabs, USA) in a 50 µL PCR reaction mixture consisting of 500 ng of 1448A Pseudomonas syringae genome as DNA template, 0.5 µM of fleQF and fleQR primers, 0.2 mM dNTP´s and 0.5 µL of Q5 High-Fidelity DNA Polymerase. The reaction followed a thermal cycling protocol that began with an initial step of 98⁰C for 2 minutes, followed by 30 cycles at 98⁰C for 10 seconds, 60⁰C for 30 seconds, and 72⁰C for 45 seconds, concluding with a final extension at 72⁰C for 5 minutes. The gel-purified fragment underwent digestion with KpnI and SacII enzymes and was cloned into the KpnI/SacII-digested pBBRMCS-4 plasmid. The resulting clones were subjected to full sequencing to discard those carrying mutations and transformed into the corresponding bacterial strains using the method described by Zumaquero et al.90 for plasmid transformation of P. syringae.
Plant growth and inoculation
Phaseolus vulgaris bean cultivar Canadian Wonder plants were cultivated under controlled conditions at 23°C, 95% humidity. Artificial light was maintained for periods of 16 hours within the 24 hours of the day. All experiments carried out were performed using 10-day-old plants.
For the preparation of bacterial inoculum, bacterial lawns were cultivated onto LB plates for 48 hours at 28°C. Subsequently, biomass was resuspended in 2 ml of 10 mM MgCl2. The optical density OD600 was adjusted to 0.1 corresponding to the concentration of 5 x 107 colony forming units per millilitre (CFU/ml). Serial dilutions were performed to achieve the desired final inoculum concentration.
The infiltration of bean leaves for visualizing microcolonies using confocal microscopy was performed using a needleless syringe with bacterial suspension at 5·105 CFU/ml.
The inoculation of bean leaves for visualizing bacteria on surface using confocal microscopy was performed by dipping. For that, a bacterial suspension with 5 x 107 CFU/ml was prepared in a 10 mM MgCl2 solution, and the entire leaf was submerged in the inoculum for a few seconds. Visualization was performed 6 hours post-inoculation (hpi).
For infiltrating bean leaves to extract bacteria from the apoplast for subsequent analysis by flow cytometry and microscopy, the method described in Rufián et al.94 was followed. This involved immersing the entire leaf in a bacterial solution with a concentration of 5 x 104 CFU/ml, containing 0.01% Silwett L-77 (Crompton Europe Ltd, Evesham, UK), and using a pressure chamber. Bacteria were recovered from the plant at 4 dpi through apoplastic fluid extraction. This extraction process, as described in Rufián et al.94, entailed pressure infiltrating a whole leaf with 10 ml of a 10 mM MgCl2 solution inside a 20 ml syringe. After applying 5 cycles of pressure, the flow-through was collected and transferred to a fresh 50 ml tube. Three thousand microlitres of the flow-through were directly analysed by flow-cytometry. Simultaneously, the 50 ml tube were centrifuged for 30 minutes at low speed (900 g) at 4ºC. The resulting pellets were resuspended into 1 ml of a 10 mM MgCl2 solution and subsequently analysed by microscopy.
To compare flagellar expression in cells mechanically extracted from or naturally exiting leaves, bean leaves were infiltrated with a 5 x 107 (for 1 dpi) or 5 x 105 CFU/ml (for 7 dpi) bacterial suspension using a pressure chamber, as described above. For natural exit, leaves were detached from the stem by cutting the petiole in the base of the leaf blade at the specified timepoints and incubated for 30 minutes into a 50 ml tube containing 30 ml of 10 mM MgCl2 to analyze natural exit. Mechanical bacterial extraction from the apoplast was carried out as described above.
Flow Cytometry and Cell Sorting
For HIM cultures, five hundred µl of an overnight P. syringae LB culture was washed twice in 10 mM MgCl2, added to 4.5 ml of HIM and incubated at 28ºC for 24h. LB cultures were obtained from an overnight incubation in LB and apoplast-extracted bacterial suspensions were obtained as indicated in Plant growth and inoculation section. Three hundred µl of the cultures in HIM, LB or in plant were analysed using a BD FACS Verse cytometer (BD Biosciences, USA) and graphs were performed with the Kaluza software (Beckman Coulter, USA). FITC-A filter was used to visualise GFP signal and PE-A filter for tdTomato signal. To ensure bleed through was not taking place, strains with transcriptional single fusions to GFP3 and tdTomato were analysed with the PE-A and FITC-A filters respectively, with the observation of fluorescence as the negative control level.
For cell sorting, stationary cultures in LB obtained after an overnight incubation were sorted using a BD FACSAria™ Fusion flow cytometer (BD Biosciences, USA). And exponential cultures in HIM obtained after 24 hours of incubation from 0.13 OD600 were sorted using a MoFloTM XDP cytometer (Beckman Coulter, USA). To initiate the process of sorting, gates were drawn to distinguish cells displaying fluorescence levels overlapping with the 1448A non-GFP bacterial population, which served as negative control, from cells expressing higher GFP levels, as indicated in the corresponding histogram. Based on this analysis, 1 x 105 events were sorted for cells expressing higher GFP levels and lower GFP level. Cells from each gate were collected into separate sterile tubes. After sorting, cells were centrifuged at 12,000 g for 10 minutes, and the resulting pellets were resuspended into 10 mM MgCl2. An aliquot of sorted cells was run again at the cytometer to confirm the differences in expression between the separated populations. Data from cytometry experiments were analysed using the Kaluza Software (Beckman Coulter, USA) for further analysis and visualization.
Confocal microscopy
For single-cell visualization of apoplast-extracted bacteria and cultured bacteria, suspensions of 2 µl were deposited over a 0.17 mm coverslip and an agar-pad square was placed on top of the drop to create a bacterial monolayer, following the method described in Rufián et al.94. To visualize all cells, bacterial suspensions were stained with FM4-64 (Life Technologies) at 20 µM, and bacterial membranes were visualized with fluorescent light, alternatively, in other cases, bright field images were included. Images of single-cell bacteria were acquired using the Zeiss LSM880 confocal microscope (Zeiss, Germany).
For the visualization of P. syringae microcolonies and surface cells, sections of inoculated P. vulgaris leaves (approximately 5 mm2) were carefully excised using a razor blade and mounted on slides in double-distilled H2O positioning the lower epidermis toward objective. A 0.17 mm coverslip was placed over the sample. Images of the leaf mesophyll and apoplast-extracted bacteria were taken using the Leica Stellaris 8 confocal microscope (Leica Microsystems GmbH, Germany) and Zeiss LSM880 confocal microscope (Zeiss, Germany).
Filters for wavelength selection were used for the visualization of the following fluorophores (excitation/ emission): eCFP (405 nm/450 to 500), GFP (488 nm/ 500 to 533 nm), FM4-64 (488 nm/ 604–674 nm), tdTomato (514 nm/570 to 600) and plant autofluorescence (514/ 605 to 670 nm). Image processing was performed using Leica LAS AF (Leica Microsystems, Germany) software. To ensure bleed through was not taking place, strains with transcriptional single fusions to GFP3, tdTomato and strains constitutively labelled with eCFP were observed under the microscope in the conditions of visualization mentioned above, with the observation of no fluorescence. Z series imaging was taken at 1 µm using 40x objectives.
Time-lapse microscopy
Heterogeneous flagellum expression was measured during microcolony formation on HIM + 1.25% agarose pads as follows: 2X HIM medium was mixed with a melted 2.5% agarose solution and immediately placed in the wells of a custom 3D-printed mold (template available here: https://github.com/JLuneau/Pseudomonas_AgarPads_fliC/tree/main/3D_printed_AgarPad_mold) disposed on a 50 mm round coverslip (Epredia, CB005005A140MNZ0). To ensure the flatness of the pads, another coverslip was immediately placed on top of the mold. The pads were solidified for 15 minutes at room temperature. The bottom coverslip was removed and 4 µl of bacterial suspensions adjusted to OD = 0.005 in HIM were dropped on the pads surface. Right after the droplets dried, a new coverslip was placed on the mold and the assembled device was mounted on the microscope. For time-lapse experiments, images were taken every 15 min, starting 4 hours after cells were placed on the pads and for 24 hours at 25°C. Images were acquired using the NIS-Elements software on a Nikon Eclipse Ti2 inverted microscope equipped with a Hamamatsu ORCA-Flash4.0LT Digital camera and a Nikon Plan Apo Lambda 100X/1.45 Oil objective. The 1.5X manual knob was engaged to enhance magnification. Illumination settings: Phase contrast, 100 ms, 50% intensity; GFP (470 nm excitation & 519 nm emission filters), 300 ms, 50% intensity. All imaging data is available upon request.
Time-lapse image analysis
Time-lapse movies were visually inspected using Fiji 2.14.0 to crop the region of interest around microcolonies and to remove later frames when cells overlapped. Cells were segmented and tracked using the DeLTA 2.0 deep learning-based pipeline97 with the default pre-trained models for segmentation and tracking. Time-lapse data analysis was performed using custom Python scripts adapted from Kaczmarczyk et al. (2022)98 (available here: https://github.com/JLuneau/Pseudomonas_AgarPads_fliC). Visual inspection of DeLTA 2.0 output movies showed that while segmentation errors were rare, tracking errors were frequent at late time points. In consequence, similarly to Kaczmarczyk et al.98, we filtered out erroneous cell tracks. Upon division, i) we kept cells for which two sister cells were tracked for at least four frames after division, ii) we excluded sister cells for which the cumulated length at birth differed strongly from the length of the mother cell before division (increase or decrease of more than 20%) and iii) we excluded sister cells which showed unexpectedly large jumps in cell length between two frames (increase or decrease of more than 20%). For all retained cells, the fliC expression level was estimated as the mean fluorescent intensity in the GFP channel for all pixels belonging to a single cell, averaged over the lifetime of each individual cell. The growth rate was obtained by performing a linear regression on the log-transformed cell length over the lifetime of each cell. To estimate the cost of flagellum expression, we grouped cells into two classes: the GFP-high cells showing a mean fluorescence intensity above the median fluorescence intensity of all cells, and the GFP-low cells showing a mean fluorescence intensity below the population's median.
Live-dead stanning
One drop of the propidium iodide solution Ready Probes™ (Thermo Fisher Scientific, USA) was added to 300 µl of the suspension with apoplast-extracted bacteria and live-dead bacteria were identified by flow-cytometry. For live-dead staining, bacteria were syringae-infiltrated with a suspension of 5 x 104 CFU/ml in bean leaves and apoplast-extracted at 4 days post-inoculation.
Competitive index (CI) assay
The competitive index (CI) assay is calculated by determining the ratio between the mutant strain and the wild type in the output sample divided by that on the input (which should be 1.0)99–101. Assays were performed after the mixed strains have been growing in either bean leaves or LB and HIM cultures.
Assays performed in bean plants (Phaseolus vulgaris cv. Canadian wonder) were carried out as detailed in Macho et al., (2007)101. Bean plants were inoculated with 200 µl of a mixed bacterial suspension containing 5 x 104 CFU/ml of each strain, consisting of an equal proportion of wild type and mutant strains. Inoculation was performed using a 1 ml syringae without needle. Samples were extracted for quantification after 4 days of post-inoculation. Bacterial recovery was carried out by taking 5 discs of 1 cm diameter from the infected leaf with a cork borer and homogenising them by mechanical disruption into 1 ml of 10 mM MgCl2. After homogenization, serial dilutions of the bacterial suspensions were prepared and plated onto agar plates supplemented with cycloheximide 2 µg/ml. Bacterial enumeration and CIs were calculated after 2 days of growth at 28ºC. To distinguish wild type from mutant bacteria within the mixed infection, an aliquot from the same dilution was plated onto LB agar and LB agar plates supplemented with kanamycin.
For CIs assays performed in LB cultures, 500 µl of a mixed bacterial suspension with 5x105 CFU/ml was inoculated into 4500 µl of LB liquid in culture tubes. For CIs assays performed in HIM cultures, 500 µl of a mixed bacterial suspension with 5x107 CFU/ml was inoculated into 4500 µl of LB liquid in culture tubes. After 24 hours of incubation with continuous agitation, in both LB or HIM cultures, serial dilutions were prepared and plated onto LB agar and LB agar plates supplemented with kanamycin.
To confirm dosage and relative proportion of the strains, serial dilutions of the inoculum were plated onto LB agar and LB agar plates supplemented with the appropriate antibiotic. After bacterial counting, the ratio of the wild type versus the mutant strain should be close to 1. The competitive indices represent the mean of three independent experiments, each with three replicates. Error bars indicate standard error. Statistical analysis included a two-tailed Student's t-test with a significance threshold of P < 0.05 to assess deviations from a ratio of 1.
In vitro growth curves
Growth curves to analyse growth differences in mutant and overexpressing strains were perform in 96-well plates (Biofil, China), adjusting the bacterial inoculum to an optical density (A600) of 0.13 in HIM in 150 µl of final volume. The inoculum was obtained from an overnight LB culture and cells were washed twice with MgCl2 before adjusting the optical density. Plates were incubated for 50 hours at 28ºC with agitation in a EONC plate reader (Bio Tek Instruments, USA).
Growth curves to compare ΔhrpL growth difference versus the wild type strain were performed in culture tubes in HIM with an initial optical density of 0.13 (Abs600). The inoculum was obtained from an overnight culture in LB, washed twice with MgCl2. Samples were taken at 20, 24, 26, 28, 30, 34, 38, 44, 48 and 50 hours.
To calculate bacterial growth rate, the log10 of absorbance data were calculated and represented versus time. The regression curve was calculated over the zone of exponential growth and the graph slope obtained was used as the growth rate.
Flagellar motility assay
Flagellar motility assays conducted after the sorting of the fliC::GFP3 strain were performed inoculating 2 µl of the aliquots obtained after the cell sorting in HIM plates containing 2.5 g/l agar or in Tryptone plates containing 3%, tryptone 5% MgCl2 and 2.5 g/l agar. Plates were subsequently incubated at 28ºC, and digital photographs were captured to measure the diameter of the swimming halo. Measurements were calculated and the ratio of the high-expressing sorted cells to the low-expressing sorted cells was determined.
Quantification and statistical analysis
All quantification and statistical analysis described in this study was performed using Prism. Details of the analysis used, and level of significance are indicated in the figure legends of each experiment. Software used for data quantification and analysis are further deailed in Supplementary Table S4.