Pseudomonas aeruginosa senses and responds to epithelial potassium flux via Kdp 2 operon to promote biofilm biogenesis

Abstract

7 162 subtilis toward the biofilm [32]. It was also shown that this phenomenon was species 163 independent by demonstrating planktonic P. aeruginosa attraction to and incorporation into B.
164 subtilis biofilms [32]. The response of P. aeruginosa to potassium from B. subtilis biofilms led 165 us to hypothesize that host potassium efflux from epithelial cells could similarly attract P.
166 aeruginosa and enhance biofilm formation on the epithelial surface. Determining how P.
167 aeruginosa responds to potassium from airway epithelial cells could provide new insights into a 168 novel mechanism by which biofilm biogenesis may be altered by the host through 169 electrochemical signaling.

170
In the current study, we utilized a respiratory epithelial co-culture biofilm live-cell 171 imaging system to investigate if P. aeruginosa biofilm biogenesis is regulated by epithelial 172 potassium efflux. Using potassium channel potentiators and inhibitors, we demonstrate that 173 biofilm biogenesis is mediated by host electrochemical signaling, revealing a novel host-174 pathogen interaction. We further identify a P. aeruginosa operon necessary for the potassium-175 induced biofilm biogenesis on epithelial cells that could serve as a potential target for therapeutic 176 intervention to reduce chronic infections in people with diseases like CF.

P. aeruginosa biofilm biogenesis is increased by potassium efflux
180 from the respiratory epithelium 181 Previous work reported P. aeruginosa motility toward and incorporation into formed 182 biofilms of B. subtilis, an unrelated bacterial species, could be induced by potassium currents . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted June 6, 2023. ; https://doi.org/10.1101/2023.06.05.543669 doi: bioRxiv preprint 206 the vehicle control treated cells (Fig 1B and 1C). After 6 hours of biofilm growth, we observed 207 significantly higher biomass measured when cells were treated with NS19504. To assess if this 208 difference was seen due to the presence of airway cell potassium rather than a direct effect of 209 NS19504 on the bacteria, microtiter dish biofilm assays and planktonic growth curves were 210 evaluated in the presence and absence of NS19504. In the microtiter dish biofilm assay, bacteria 211 form biofilms on wells of a microtiter dish without the presence of airway cells by attaching 212 directly to the plastic wells at the air-liquid interface. Biofilms can then be quantified by staining 213 with crystal violet. In this assay, biofilm growth was not altered by the presence of NS19504 in 214 either LB or synthetic cystic fibrosis sputum (SCFM) media, as seen by the images and 215 measurements of crystal violet staining (Fig 1D and 1E). Moreover, NS19504 did not affect the 216 planktonic growth rate of P. aeruginosa in LB, SCFM, or human cell culture media, minimal 217 essential media (MEM), as seen by the almost identical growth curves in the presence or absence 218 of NS19504 ( Fig 1F). 219 To assess if NS19504 was toxic to CFBE41o-cells, cultures differentiated at air-liquid 220 interface (ALI) were treated with NS19504 for 6 hours and transepithelial resistance (TEER) was 221 measured, as a readout for barrier integrity. NS19504 did not alter TEER, as compared with 222 vehicle control treated cells over the 6-hour time course (Fig S1A). We also assessed if 223 NS19504 altered the cytotoxicity of P. aeruginosa by measuring TEER and using a 3-(4,5-224 dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) based cell viability assay during 225 P. aeruginosa infection of our ALI CFBE41o-cells. There was no significant difference 226 between the DMSO or NS19504 treated cells in the MTT assay or in change in TEER at the end 227 the 6-hour infection (Fig S1B and S1C). Taken together, these data suggest that potassium . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted June 6, 2023. ; https://doi.org/10.1101/2023.06.05.543669 doi: bioRxiv preprint 251 infection with increased potassium efflux from airway cells treated with NS19504, as compared 252 to the DMSO treated (Fig 2A). At the 6-hour time point, there was no significant difference in 253 the number of biofilm aggregates or area of aggregates between when cells were treated with 254 NS19504 versus DMSO treated cells (Fig 2B and 2C). Given that the overall biomass was 255 increased with increased potassium conductance due to NS19504 treatment, and this 256 measurement includes a volumetric measurement, the height of the biofilms must be increased 257 with NS19504 treatment. These data suggest that potassium increases biofilm biogenesis by 258 promoting bacterial attachment, though these current experiments could not assess how 259 potassium may alter microcolony formation. 260 261 P. aeruginosa coalescence into microcolonies is increased by 262 epithelial potassium efflux 263 We hypothesized that potassium gradients generated by the airway epithelium may also was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted June 6, 2023. ; https://doi.org/10.1101/2023.06.05.543669 doi: bioRxiv preprint 274 determine whether biofilms were formed by multiple individual attached bacteria coming 275 together on the epithelial surface (polyclonal) or by replication of a single clone (monoclonal). 276 In these experiments, we observed that P. aeruginosa primarily attached as singleton bacteria at 277 1 hour, but by 6 hours of growth, a mixture of single color and multicolor biofilms were present 278 (Fig 3A and 3B). In these images, more multicolor biofilm aggregates are present in the group 279 treated with NS19504, as compared to those treated with DMSO. We quantified the proportion 280 of multicolor biofilms at 6 hours in Nikon Elements software by defining a binary which 281 contained all three colors of fluorescence and then using the count objects function to count 282 objects that contained one or more than one fluorescent signal above the threshold value.
283 Biofilms of P. aeruginosa in the condition of increased host potassium efflux were more likely to 284 be formed by coalescence of multiple clones (i.e., multicolor biofilms), as compared to cells with 285 basal potassium flux, as seen by the increased proportion of multicolor aggregates noted in the 286 NS19504 treated group (Fig 3C). These findings suggest that epithelial potassium efflux 287 increases biofilm growth by both promoting bacterial association with the respiratory epithelium 288 and increasing microcolony formation by bacterial coalescence on the epithelial surface.

290
Biofilm biogenesis is reduced when potassium efflux is decreased 291 With the finding that potassium efflux increases biofilm formation on airway epithelial 292 cells, we hypothesized that reducing potassium flux would decrease biofilm formation. was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted June 6, 2023. ; https://doi.org/10.1101/2023.06.05.543669 doi: bioRxiv preprint 297 reduced when epithelial potassium efflux was inhibited by paxilline treatment, as compared to 298 DMSO treated controls (Fig 4A and 4B). Paxilline did not affect P. aeruginosa biofilm 299 formation in the absence of airway epithelial cells in LB or SCFM as seen by the similar levels 300 of crystal violet staining in the microtiter dish biofilm assay in both the paxilline and DMSO 301 treated bacteria in both media ( Fig 4C). Additionally, planktonic growth in any of the three 302 growth media (LB, SCFM, and MEM) tested was not altered by paxilline, as seen by the similar 303 growth kinetics for the paxilline and DMSO treated bacteria ( Fig 4D). Paxilline also did not 304 reduce epithelial cell barrier integrity in the absence of bacterial infection as seen by the 305 increasing TEER over the course of 6-hour treatment ( Fig S3A). Additionally, with paxilline 306 treatment during bacterial infection, there was no change in cell viability, as measured by MTT 307 assay, or difference in TEER during infection when paxilline treated cells were compared to the 308 DMSO treated (Fig S3B ad S3C). As expected, based on the finding that potassium efflux 309 increased bacterial attachment, reduced potassium flux from paxilline treatment resulted in 310 decreased bacterial attachment, but did not decrease final biofilm number or area ( Figure S4A-311 C). These data indicate that biofilm biogenesis in association with airway epithelial cells (AECs) 312 can be reduced by reducing the potassium flux from the airway cells.

313
Given the effect of paxilline on reducing biofilms, we examined if paxilline could 314 function as a dispersal agent for established biofilms. In order to assess this, biofilms were 315 grown in the presence of MEM with DMSO for 6 hours and imaged. The media was then 316 changed to MEM containing paxilline and perfused for 30 minutes prior to re-imaging. We have 317 previously published the characterization of this model for studying biofilm dispersal [51].
318 Biomass was unchanged after the 30 minutes of paxilline treatment suggesting that loss of 319 potassium efflux from the respiratory epithelium does not function as a dispersal signal (Fig 4E).
. CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted June 6, 2023. ; https://doi.org/10.1101/2023.06.05.543669 doi: bioRxiv preprint 320 This data suggests that epithelial cell potassium gradients are important in the establishment of 321 biofilms, but do not affect their dispersal. 322 323 Kdp potassium sensing and uptake genes are necessary for 324 chemotaxis toward potassium 325 After demonstrating that epithelial potassium efflux enhances biofilm formation by P. 326 aeruginosa, we next identified the bacterial mechanisms involved in sensing and responding to 327 potassium. Based on the previous work [32] and our findings in this study, we hypothesized that 328 potassium was leading to increased bacterial motility toward the potassium current. To test this 329 hypothesis, we utilized a previously described macroscopic twitching chemotaxis assay with 330 potassium chloride as the chemoattractant [52-54]. We observed twitching chemotaxis toward 331 the potassium gradient with wild-type PAO1, while a pilA strain of PAO1, deficient in 332 twitching motility, did not chemotax toward potassium. This was observed as greater expansion 333 of bacteria toward the potassium gradient on the left side of the growth ring in the wild-type 334 PAO1 and lack of this expansion in any direction in the pilA strain ( Figure 5A). Additionally, 335 this P. aeruginosa showed an increased directional motility in their chemotaxis toward the 336 potassium chloride gradient, as compared to a control water treatment ( Figure 5A, striped bar).

337
Having shown that P. aeruginosa chemotaxes toward potassium and that this chemotaxis 338 is pilus mediated, we interrogated the P. aeruginosa genome for genes annotated to be associated was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted June 6, 2023. ; https://doi.org/10.1101/2023.06.05.543669 doi: bioRxiv preprint 342 strain with a markerless, in-frame deletion of PA5518 were screened for chemotaxis toward 343 potassium using the macroscopic twitching chemotaxis assay [56,57]. We observed that 344 mutants in the kdp gene cluster, particularly the kdpD transposon mutant, had reduced 345 chemotaxis toward potassium, as seen by decreased expansion of these mutants toward the 346 potassium gradient corresponding to a lower directional motility index, while the other mutants 347 screened did not have a reduction in chemotaxis (Fig 5A, 5B, and 5C). kdpFABC encodes a high 348 affinity potassium uptake channel and kdpDE encodes an intracellular potassium sensor kinase,

352
With the evidence that the kdp genes are involved in P. aeruginosa chemotaxis toward 353 potassium, we generated a PAO1 strain with a markerless, in frame deletion of all kdp genes 354 (kdpFABCDE) using a previously described mutagenesis protocol [63] and confirmed this 355 deletion by whole genome sequencing. We assessed this mutant in the chemotaxis assay and 356 observed a significant reduction in chemotaxis toward potassium, as compared to the wild-type 357 PAO1 control (Fig 5D and 5E). With the knowledge that P. aeruginosa chemotaxes toward 358 potassium and that deletion of the kdp operon reduces potassium chemotaxis, we next examined 359 biofilm biogenesis in our respiratory epithelial co-culture biofilm system to determine if the kdp 360 operon may be involved in the potassium-induced biofilm formation in association with the CF 361 airway epithelium.
. CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted June 6, 2023. ; https://doi.org/10.1101/2023.06.05.543669 doi: bioRxiv preprint

364
Before evaluating the role of the kdp genes in our respiratory epithelial co-culture biofilm 365 assay, we investigated if the kdpFABCDE strain was deficient in planktonic or abiotic biofilm 366 growth as compared wild-type PAO1. Planktonic growth curves showed similar growth kinetics 367 between the kdpFABCDE and its wild-type PAO1 parental strain in LB broth, SCFM, and 368 MEM (Fig. S5A). Biofilm growth in the microtiter dish biofilm assay (i.e. abiotic biofilm 369 growth) was similar in LB and SCFM (Fig. S5B). To ensure that the deletion did not alter 370 general twitching motility, a twitching motility assay was performed. The kdpFABCDE was not 371 deficient in twitching motility when compared to wild-type PAO1, as seen by the similar 372 twitching radius in the WT and kdpFABCDE (Fig. S5C).

373
Finally, we tested the kdpFABCDE strain in our respiratory epithelial co-culture biofilm 374 assay. We observed that the kdpFABCDE mutant P. aeruginosa strain was deficient in biofilm 375 biogenesis in association with epithelial cells, as evidence by the reduced GFP signal in imaging 376 and reduced biofilm biomass quantified (Fig 6A and 6B). Additionally, we tested the 377 kdpFABCDE mutant in the presence of NS19504 which did not rescue the phenotype, 378 suggesting that the Kdp operon is necessary to sense airway epithelial potassium gradients and 379 respond by inducing biofilm growth (Fig 6A and 6B). The kdpFABCDE mutant had reduced 380 attachment to the airway epithelium as seen by the reduced bacterial counts at 1-hour post-381 infection, as well as reduced aggregate area and aggregate number at 6-hours of growth (Fig 6C-382 6E). These findings demonstrate that the kdp genes, involved in potassium sensing and uptake, 383 mediate biofilm biogenesis in association with the airway epithelium and the induced biofilm 384 response to enhanced potassium efflux produced by treatment with NS19504.

Discussion
. CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted June 6, 2023. ; https://doi.org/10.1101/2023.06.05.543669 doi: bioRxiv preprint

387
In the current study, we report that potassium efflux from the respiratory epithelium 388 promotes biofilm biogenesis by enhancing P. aeruginosa bacterial attachment and coalescence 389 into microcolonies. Furthermore, we demonstrated that the P. aeruginosa kdp gene cluster, a 390 putative potassium uptake system and its associated potassium sensor and regulator, were 391 necessary for biofilm growth associated with the airway epithelium. Our findings suggest a 392 novel aspect of host-pathogen interactions whereby bacterial biofilm biogenesis is increased by 393 the electrochemical gradient at the mucosal surface. Additionally, our study purports possible 394 new therapeutic targets for impeding chronic bacterial infections in diseases with mucosa-395 associated biofilms by targeting epithelial potassium efflux.

396
The data in the current study shows a link between P. aeruginosa biofilm biogenesis and 397 potassium efflux from host cells. Our study builds on previous work demonstrating that P.   Table S2.
. CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted June 6, 2023. was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made  (Table S1). Chambers were . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted June 6, 2023. ; https://doi.org/10.1101/2023.06.05.543669 doi: bioRxiv preprint 591 treated with MEM without phenol red with 25 M NS19504 or 0.05% DMSO as control. 592 Chambers were imaged at 1, 3, and 6 hours. The number of biofilms containing one or more 593 than one strain of P. aeruginosa was calculated using Nikon Elements software. Nikon Elements 594 uses Boolean logic based on input threshold values for each fluorophore to determine which 595 biofilms contain signal from more than fluorophore. We reported these results as the percent of 596 total aggregates that had more than one strain of P. aeruginosa. The proportion of aggregates 597 with 2 or more colors of P. aeruginosa in the aggregate out of the total number of aggregates 598 was measured in nine visual fields at each time point. These nine measurements were averaged 599 and were considered one biologic replicate. The experiment was repeated four times and the 600 mean proportion of multicolor aggregates is reported. Unpaired t-test was used to determine 601 statistical significance. 602 603 Potassium macroscopic chemotaxis assay 604 A macroscopic twitching chemotaxis assay was adapted from the assay previously was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted June 6, 2023. ; https://doi.org/10.1101/2023.06.05.543669 doi: bioRxiv preprint 614 Overnight cultures of WT PAO1 and kdpFABCDE PAO1 strain were grown in M9 615 broth media. Cultures were then inoculated through M9 media supplemented with 1.5% agar to 616 the plastic surface of the petri plate using a sterile 10 L pipette tip. Cultures were allowed to 617 grow for 48 hours. After growth, agar was carefully removed from the petri dish as to not 618 disrupt the bacteria under the agar. Bacterial growth on the petri plate was stained with 10% 619 crystal violet, then washed three times in deionized water. Experiments were done in duplicate 620 and repeated four separate times for four biologic replicates. Plates were imaged using 621 stereomicroscope and twitching radius was measured using Nikon Elements software. was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted June 6, 2023. ; https://doi.org/10.1101/2023.06.05.543669 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted June 6, 2023. ; https://doi.org/10.1101/2023.06.05.543669 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted June 6, 2023. ; https://doi.org/10.1101/2023.06.05.543669 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted June 6, 2023. ; https://doi.org/10.1101/2023.06.05.543669 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted June 6, 2023. ; https://doi.org/10.1101/2023.06.05.543669 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted June 6, 2023. ; https://doi.org/10.1101/2023.06.05.543669 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted June 6, 2023. ; https://doi.org/10.1101/2023.06.05.543669 doi: bioRxiv preprint