OprF Impacts Pseudomonas aeruginosa Biofilm Matrix eDNA Levels in a Nutrient-Dependent Manner

ABSTRACT The biofilm matrix is composed of exopolysaccharides, eDNA, membrane vesicles, and proteins. While proteomic analyses have identified numerous matrix proteins, their functions in the biofilm remain understudied compared to the other biofilm components. In the Pseudomonas aeruginosa biofilm, several studies have identified OprF as an abundant matrix protein and, more specifically, as a component of biofilm membrane vesicles. OprF is a major outer membrane porin of P. aeruginosa cells. However, current data describing the effects of OprF in the P. aeruginosa biofilm are limited. Here, we identify a nutrient-dependent effect of OprF in static biofilms, whereby ΔoprF cells form significantly less biofilm than wild type when grown in media containing glucose or low sodium chloride concentrations. Interestingly, this biofilm defect occurs during late static biofilm formation and is not dependent on the production of PQS, which is responsible for outer membrane vesicle production. Furthermore, while biofilms lacking OprF contain approximately 60% less total biomass than those of wild type, the number of cells in these two biofilms is equivalent. We demonstrate that P. aeruginosa ΔoprF biofilms with reduced biofilm biomass contain less eDNA than wild-type biofilms. These results suggest that the nutrient-dependent effect of OprF is involved in the maintenance of P. aeruginosa biofilms by retaining eDNA in the matrix. IMPORTANCE Many pathogens form biofilms, which are bacterial communities encased in an extracellular matrix that protects them against antibacterial treatments. The roles of several matrix components of the opportunistic pathogen Pseudomonas aeruginosa have been characterized. However, the effects of P. aeruginosa matrix proteins remain understudied and are untapped potential targets for antibiofilm treatments. Here, we describe a conditional effect of the abundant matrix protein OprF on late-stage P. aeruginosa biofilms. A ΔoprF strain formed significantly less biofilm in low sodium chloride or with glucose. Interestingly, the defective ΔoprF biofilms did not exhibit fewer resident cells but contained significantly less extracellular DNA (eDNA) than wild type. These results suggest that OprF is involved in matrix eDNA retention in biofilms.


Growth rate assays
Cells from stationary phase cultures were seeded into flat-bottom polystyrene 96-well plates at an OD600 of 0.001. Absorbance at OD600 was read in a Synergy Hybrid HTX Microplate Reader (Bio-Tek Instruments) every 5 min for 16 h at 37˚C with shaking. The average of three biological replicates, which is the average of 4 technical replicates per strain, was reported. Growth rates were calculated by determining the slope of the growth curve in exponential phase.
Transmission electron microscopy of P. aeruginosa Cell pellets were harvested by centrifugation at 8000 g at 4°C for 15 min. Cells were incubated overnight at 4°C in a premixed fixative solution of formaldehyde-glutaraldehyde (2.5% each) in 0.1 M sodium cacodylate buffer at pH 7.4 (Electron Microscopy Sciences). Cells were subsequently washed with a solution containing 1.2% sucrose and 4 mM CaCl2 in 0.1 M sodium cacodylate buffer, then subjected to a secondary fixation for 1 h at room temperature with 2% OsO4 (prepared in the above solution). After another wash, cells were stained for 1 h at room temperature with UranyLess, a solution containing lanthanides, and an alternative to uranyl acetate (Electron Microscopy Sciences). Cells were serially dehydrated with the following ethanol solutions (10 min each): 50%, 70%, 95% (two times), 100% (three times). Cells were gradually infiltrated with epoxy-type resin (EMbed 812 kit with BDMA, Electron Microscopy Sciences) through the following steps: ethanol was exchanged with 100% acetone (three times 10 min), then cells were exposed to a 1:1 acetone-resin mix for 1 hr at room temperature, to a 1:3 acetone-resin mix overnight, and to 100% resin (three times for 1hr). After a final exchange with fresh resin, samples were cured for 48 hr at 60°C. Ultrathin sections (70 nm) were cut using a Leica EM UC7 ultramicrotome, recovered on 200-mesh Cu grids, and double-stained with uranyl acetate and lead citrate for 10 and 4 min, respectively. Transmission electron microscopy was performed at 120 KeV on a JEOL TEM 1400Plus at the Electron Microscopy Core Laboratory of the University of Utah. Images were acquired with a Gatan Orius TM SC-1000 CCD camera (1 sec acquisition time).
For negative stain grid preparation and imaging, 5 μl of WT or ∆oprF culture grown in TSB or LB and grown to static or exponential phase was applied onto glow-discharged, carbon-coated copper grids (Agar Scientific). After incubating the sample for 2 min at room temperature, grids were rapidly washed in one drop of DI water and subsequently exposed to two successive short drops and one long incubation (30 sec) drop of UranyLess stain solution (Electron Microscopy Sciences). Images were recorded on a JEM1400 TEM (JEOL) equipped with a Matataki flash (JEOL) at 5000x and 12000x magnifications for all conditions.

Medium osmolarity
Medium osmolarity for 8 base media and variants (TSB, TSB 10g/L NaCl, TSB no glucose, TSB no K2HPO4, LB, LB 5 g/L NaCl, LB w/ glucose, LB w/ K2HPO4) was measured using a Vapro 5520 vapor pressure osmometer (Wescor). 10 µL of each medium was loaded onto sample discs (Wescor SS-033) and subjected to a 73 sec vaporization. The average of 3 readings per medium was reported.

Planktonic cell staining with crystal violet
Overnight culture was diluted 1:100 in fresh TSB and grown to mid-log growth phase (0.5 OD600). 10-fold serial dilutions of mid-log cultures were spread on LB agar and incubated overnight for verification of cell counts (CFU/mL). Mid-log culture volumes corresponding to 5.0x10 7 , 7.5x10 7 , 1x10 8 , 1.25x10 8 , 1.5x10 8 , 2.0x10 8 CFU/mL were centrifuged at 6000g for 8 min. Supernatant was removed by pipetting, and cell pellets were stained with 0.1% crystal violet, vortexed, and incubated for 15 min. Stained cells were centrifuged at 6000g for 8 min and were subsequently washed with 1mL deionized water three times (with centrifugation between washes). Stained biomass was eluted using 30% acetic acid, transferred to a flat-bottom 96-well plate (Greiner Bio-One, #655090), and the absorbance at OD550 was read in a Synergy Hybrid HTX Microplate Reader (BioTek Instruments). Absorbance from stained blank media tubes was subtracted from raw OD550 readings. Three biological replicates were averaged.

SUPPLEMENTAL FIGURES
Supplemental Figure S1. OprF is expressed upon induction in the ∆oprF + oprF restoration strain.
Supplemental Figure S3. No major cell morphology differences between wild type and ∆oprF grown in TSB or LB.  (A) 24-hour static microtiter biofilm assays were performed with a tomato plant isolate (E2), a water isolate (MSH10), a UTI isolate (X24509), and their respective isogenic ∆oprF mutants in TSB. (B and C) 24-hour static microtiter biofilm assays were performed with an oprF interruption mutant (H636) and its parental strain (H103) in TSB and LB media. Biofilm formation was normalized to that of each respective parental strain grown in the same medium. Error bars, SEM (N = 3); dot, each biological replicate, which is the average of 3-6 technical replicates; asterisk, statistically different from the respective parental strain (p < 0.01; Student t-test).   Figure S5. ∆oprF forms less biofilm in low sodium chloride and with glucose. 24-hour static microtiter biofilm assays were performed with P. aeruginosa PAO1 (WT, black), ∆pslD (blue), ∆oprF (red), and a ∆oprF attTn7::PBAD-oprF restoration strain (∆oprF + oprF) without (white) and with (gray) 0.5% arabinose (Ara) in the indicated media. Left column, TSB base medium and variants; right column, LB base medium and variants. (A) Biofilm formation in altered sodium chloride (NaCl); (B), glucose; and (C), dipotassium phosphate (K2HPO4) concentrations. Concentrations of altered medium component denoted in top row of each graph in g/L. Biofilm formation is normalized to WT in each respective medium. Error bars, SEM (N = 3); dot, each biological replicate, which is the average of 6 technical replicates; asterisk over bar, statistically different from WT in the same medium (p < 0.05; two-way ANOVA with post hoc Bonferroni). Statistical difference between ∆oprF strains in different media are indicated by a bar and asterisk. See Table S3 for full statistical comparisons.
Supplemental Figure S6. Weak correlation that is not statistically significant between osmolarity and ∆oprF biofilm.
Osmolarity of each base medium and its variants compared to ∆oprF static microtiter biofilm biomass formed in respective medium. Osmolarity of original TSB base medium (black, circle), TSB 10 g/L NaCl variant (black, square), TSB with no glucose variant (black, triangle), TSB with no K2HPO4 variant (black, diamond), LB base medium (white, circle), LB 5 g/L NaCl variant (white, square), LB with 2.5 g/L glucose variant (white, triangle), and LB with 2.5 g/L K2HPO4 variant (white, diamond) were correlated to ∆oprF biofilm biomass in each respective medium (gray dotted line, R 2 = 0.40541; p > 0.05, Pearson's correlation).