Micron Scale Spatial Measurement of the O2 Gradient Surrounding a Bacterial Biofilm in Real Time

O2 is a fundamental environmental metabolite that affects all life on earth. While toxic to many microbes and obligately required by others, those that have appropriate physiological responses survive and can even benefit from various levels of O2, particularly in biofilm communities. Although most studies have focused on measuring O2 within biofilms, little is known about O2 gradients surrounding biofilms. Here, we developed electrochemical methodology based on scanning electrochemical microscopy to measure the O2 gradients surrounding biofilms in real time on the micron scale. Our results reveal that P. aeruginosa biofilms produce a hypoxic zone that can extend hundreds of microns from the biofilm surface and that this gradient remains even after the addition of antibiotic concentrations that eradicated 99% of viable cells. Our results provide a high resolution of the O2 gradients produced by P. aeruginosa biofilms and reveal sustained O2 consumption in the presence of antibiotics.

IMPORTANCE O 2 is a fundamental environmental metabolite that affects all life on earth. While toxic to many microbes and obligately required by others, those that have appropriate physiological responses survive and can even benefit from various levels of O 2 , particularly in biofilm communities. Although most studies have focused on measuring O 2 within biofilms, little is known about O 2 gradients surrounding biofilms. Here, we developed electrochemical methodology based on scanning electrochemical microscopy to measure the O 2 gradients surrounding biofilms in real time on the micron scale. Our results reveal that P. aeruginosa biofilms produce a hypoxic zone that can extend hundreds of microns from the biofilm surface and that this gradient remains even after the addition of antibiotic concentrations that eradicated 99% of viable cells. Our results provide a high resolution of the O 2 gradients produced by P. aeruginosa biofilms and reveal sustained O 2 consumption in the presence of antibiotics. KEYWORDS biofilm, oxygen, antibiotics, electrochemistry, Pseudomonas aeruginosa, antibiotic resistance M olecular oxygen (O 2 ) is one of the most important molecules dictating bacterial lifestyle and behavior. For organisms capable of tolerating O 2 , it can provide a means to remove excess electrons formed during metabolism. While general fundamentals of O 2 consumption are well established, the role of O 2 is complex in bacterial communities, including those associated with human infection, since O 2 levels vary tremendously based on the infection site and the host response (1-3). In addition, bacteria in many infections grow as sessile communities called biofilms (4), and the three-dimensional structure of these communities can affect O 2 levels throughout the biofilm.
Previous work has shown that O 2 gradients within biofilms affect their biology (5,6). This has prompted an examination of O 2 levels within and surrounding biofilms. In particular, stagnant biofilms rapidly deplete O 2 and waste material buildup occurs as a result of mass transport limitation at the surface of biofilms (7). Although it is clear that O 2 levels are decreased within the biofilm, the levels immediately adjacent to the biofilm surface have not been thoroughly investigated in static biofilms, in part due to the difficulties in robustly measuring O 2 with high spatial precision (5,(8)(9)(10)(11)(12)(13). To address this gap in knowledge, we developed a system to spatially measure O 2 levels above a microbial biofilm in real time at the micron scale. We chose the facultative anaerobe Pseudomonas aeruginosa strain PA14 for these studies since this opportunistic pathogen preferentially utilizes aerobic respiration (14), and its physiology and behavior are highly influenced by O 2 availability (14,15). A significant challenge that was overcame is the inherent difficulty with continuously measuring O 2 over extended time periods. To address this challenge, we developed a system using electrochemical methods to measure O 2 in real-time with micronscale spatial resolution (Fig. 1A). O 2 can be detected electrochemically through a four-electron reduction on a platinum ultramicroelectrode (UME) (Fig. 1A) (16). How- ever, platinum UMEs readily deactivate which leads to long wait times for O 2 current stabilization and sub-nA current (see Fig. S1 in the supplemental material). To address this challenge, we optimized a platinization protocol that coats the UME surface with platinum particles that actively reduce O 2 while avoiding severe changes in the geometry of the UME surface (see Fig. S2). Importantly, our platinum UMEs had a higher electroactive area and could continuously monitor O 2 levels over several hours without loss of sensitivity (Fig. 1B). This is especially important because the current measured in bulk was approximated to be 205 M; since current measured is directly proportional to O 2 concentration, stability ensures accurate O 2 measurement despite each platinized UME used only once per experiment and having slightly variable degrees of platinization or size after polishing.
We next sought to measure O 2 levels surrounding a P. aeruginosa biofilm using scanning electrochemical microscopy (SECM). The P. aeruginosa strain chosen for this work (fliC9::MrT7) (17) has an inactivated flagellar motor protein, rendering the strain unable to leave the biofilm via swimming motility. Biofilms of the P. aeruginosa fliC mutant were formed on polycarbonate membranes as previously described for electrochemical studies (18). Membrane biofilms were grown for 8 h on Todd Hewitt broth (THB) agar, yielding an ϳ3-mm-diameter nascent biofilm containing ϳ4 ϫ 10 7 bacteria (see Fig. S3). These biofilms contain fewer cells than those used in previous studies (5,8) focused on O 2 consumption to better mimic biofilms observed in human infections. After formation, the membrane containing the biofilm was removed from the agar plate and attached to the bottom of a glass vial using double-sided tape and covered with ϳ5 ml of morpholinepropanesulfonic acid (MOPS)-glucose minimal medium.
To measure O 2 levels above the biofilm, a 10-m-diameter platinized UME was approached to 40 m above the biofilm surface using ferrocenyl methyl trimethylammonium (FcMTMA ϩ ; the toxicity and stability are assessed in Text S1 in the supplemental material) as the redox mediator ( Fig. S4 and S5) using SECM (19). The UME tip was then poised at Ϫ0.5 V (O 2 reduction potential), with a wait time of 5 min; afterward, the UME was retracted at 6 m/s while continually measuring O 2 until bulk O 2 levels were detected, ϳ1,400 m above the biofilm surface (the O 2 gradient calculations are detailed in Text S1). The O 2 levels above the biofilm resembled a sigmoidal curve with no O 2 detectable until ϳ200 m above the biofilm ( Fig. 2A; see also Fig. S6). Assuming a 10-pA minimal background current, the detection limit of the UME is ϳ1 M.
For comparison, we created an O 2 gradient without a biofilm using a platinum electrode the same size as the biofilm as the SECM substrate. The 3-mm platinum electrode was held at three potentials (0.1, 0, and Ϫ0.5 V versus Ag/AgCl) for 5 min, and then the O 2 gradient was measured as described for the biofilm (detailed in Text S1). The biofilm O 2 gradient was similar to the Ϫ0.5 V poised electrode gradient, which is the potential at which O 2 reduction is mass transport limited at the surface of the electrode. The biofilm O 2 gradient was distinct from the other potentials at which O 2 was being consumed at a submaximal rate (i.e., limited in part by kinetics and not predominantly by mass transport). Using Comsol Multiphysics to digitally simulate O 2 consumption ( Fig. 2B; see also Fig. S7), we approximated the flux of O 2 at the surface of the biofilm to be 8.2 ϫ 10 Ϫ7 mol/cm 2 /s (detailed in Text S1). Assuming each cell has a dimension of 1.5 m ϫ 0.8 m, 9.8 ϫ 10 Ϫ15 mol/s O 2 or 5.9 ϫ 10 9 molecules of O 2 per second were consumed by each bacterium. Collectively, these results reveal that P. aeruginosa biofilms produce a hypoxic zone that can extend hundreds of microns from the biofilm surface within minutes, and the biofilm consumes O 2 at a maximum rate.
To assess the effect of antibiotic treatment on biofilm O 2 consumption, we treated our biofilms with 400 times the MIC of the antibiotic ciprofloxacin (40 g/ml) and then measured the O 2 gradient above the biofilm. We first confirmed that ciprofloxacin does not interfere with the electrochemical signal for O 2 quantification ( Fig. S8; see also Text S1). Ciprofloxacin treatment of the biofilm was performed by initially adding 20 g/ml ciprofloxacin and measuring the O 2 response 1.5 h after submersion in MOPS-glucose (Fig. 2C). After we observed no immediate change in signal, we treated the biofilm with another 20 g/ml. After addition of the second dose of ciprofloxacin, the O 2 gradient was measured for 50 min with no observable change in the O 2 gradient, a total of 1 h and 35 min after ciprofloxacin was first added. Despite the fact that addition of ciprofloxacin reduced the number of viable bacteria in the biofilm by 100-fold to ϳ2.4 ϫ 10 5 bacteria, there was no change in the O 2 gradient or O 2 consumption rates ( Fig. 2A and B). While prior work has primarily measured bulk O 2 at the biofilm/air interface, we show that, in contrast, at a stagnant biofilm/liquid interface a hypoxic region forms several hundred microns above the biofilm surface. Containing only ϳ4 ϫ 10 7 bacteria, our biofilms consumed O 2 at maximum rates and continued to do so despite 99% killing by ciprofloxacin. These data corroborate similar findings that bulk respiratory activity and carbon consumption persists despite antibiotic exposure (20,21). Given this high O 2 consumption rate and the observation that biofilms of this size exist in human implant/catheter infections (22), we propose that biofilms are capable of rapidly depleting local O 2 in chronic infections even during antibiotic challenge. Ultimately, the experimental system developed in this work provides a valuable framework for studying biofilm O 2 consumption.

MATERIALS AND METHODS
Instrumentation. Initial electrochemistry experiments were performed using a BioLogic SECM (model M470). Biofilm experiments measuring O 2 gradients by scanning distances were done using a CHI model 920D scanning electrochemical microscope (CH Instruments). For all experiments, a threeelectrode setup was used. This consisted of a 10-m-diameter platinum UME (working electrode), Ag|AgCl|Saturated KCl (reference electrode to which all potentials are referred to in all experiments), and platinum wire (counter electrode). An in-depth protocol for UME fabrication may be found elsewhere (19).
Ultramicroelectrode fabrication and SECM cell setup. An in-depth protocol for UME fabrication may be found elsewhere (19). Briefly, platinum (99.9% purity) wire, 10-m diameter, temper: hard (Goodfellow Metals, Cambridge, United Kingdom; product PT005107) was used for the preparation of the SECM UME tip. The metal wire was heat sealed with a heating coil under vacuum in a glass capillary. The tip was sharpened to an RG of ϳ10, where RG is the ratio of the glass diameter to wire diameter. Prior to electrochemical experiments, UMEs were sonicated in a water bath for 30 s. Platinization significantly alters the surface and UMEs were seldom repolished and replatinized for reuse.
Bacterial strain culture and preparation. P. aeruginosa (PA14) fliC9::MrT7 mutant was obtained from a PA14 nonredundant transposon insertion mutant set (http://ausubellab.mgh.harvard.edu/cgi-bin/ pa14/home.cgi) (17). Biofilms were grown in THB agar for 8 h at 37°C, at which point an ϳ3-mm-diameter biofilm formed before transfer to the SECM cell. All SECM experiments were performed using MOPS minimal media (23) containing 20 mM glucose. CFU were enumerated at the end of experimentation by removing media above the biofilm, substantially vortexing the biofilm off the polycarbonate membrane, and plating on THB agar plates overnight at 37°C.
Platinizing UMEs. Handmade 10 m platinum UMEs (as described above) were sonicated in water, acetone, and water. A modified protocol for platinizing UMEs was used that may be found elsewhere (24) with an adjusted recipe for the platinization solution containing 0.250 ml of H 2 PtCl 6 and 0.4 mg of Pb(NO 3 ) 2 up to a final volume of 7.36 ml in 1ϫ phosphate-buffered saline (pH 7.4). Geometric and electroactive effects on the UME surface resulting from platinization were measured to confirm the stability and reproducibility of platinization. An in-depth review of platinizing electrodes can be found elsewhere (25).
Measuring O 2 . PA14 tn::fliC biofilms were grown as described above. After 8 h growth, the polycarbonate membrane was removed and attached to the bottom of a custom glass vial using double sided tape. UMEs were cycled in platinizing solution (same as above) from 0.2 V to Ϫ0.3 V versus Ag/AgCl at 100 mV/s until the maximum limiting current increased ϳ1.2ϫ for FcMTMA ϩ oxidation (the synthesis is detailed in Text S1). After platinization and ensuring proper geometric area of the UMEs, 1 mM FcMTMA ϩ was added to MOPS-glucose media, and approximately 5 ml was added to the vial containing the biofilm. A three-electrode setup using a platinum wire counter, and Ag/AgCl reference electrodes were connected. Platinized UMEs were precisely positioned with micron-scale accuracy using SECM. SECM

FIG 2 Legend (Continued)
analysis module in two-dimensional axial symmetry using stationary conditions with a parametric sweep of the "d" or distance between UME tip and substrate (Fig. S5, and detailed in Text S1). (C) Changes in O 2 concentration 600 m above a biofilm measured as a response to ciprofloxacin treatment. At 120 s, the first dose of 20 g/ml ciprofloxacin was added (designated by red arrow). Each line represents a biological replicate. The y axis (ordinate) is the ratio of the tip current at each time point divided by current measured before ciprofloxacin addition (i.e., a value of 1 indicates no change in current after ciprofloxacin addition). There were changes immediately after ciprofloxacin addition (peaks at red arrow), likely a result of the mixing caused by addition of ciprofloxacin to the growth media above the biofilm. Importantly, the current quickly stabilized.
Measuring O 2 Surrounding a Biofilm ® September/October 2020 Volume 11 Issue 5 e02536-20 mbio.asm.org 5 positions UMEs at defined distances from the biofilm surface using an electroactive mediator while observing tip current changes as a function of distance. For this work, we chose the electroactive mediator FcMTMA ϩ since it is neither consumed by nor is toxic to P. aeruginosa. UMEs were poised at 0.5 V to oxidize FcMTMA ϩ , approached within ϳ40 m above the surface (within the hindered diffusion region corresponding to a decrease in signal to ϳ95% limiting current), and then poised at Ϫ0.5 V and retracted at 6 m/s to measure O 2 gradients. For antibiotic chronoamperometry curves, UMEs were positioned first approximately 600 m for the first addition of ciprofloxacin or control, approximately 1 h and 30 min elapsed after MOPS-glucose was added over the biofilm. Ciprofloxacin was added slowly during this time at ϳ7.5 mm above the biofilm, and the stage was attached to the vial containing the biofilm was rotated 10 times in a circular motion immediately after addition. The UME was then approached approximately 300 m above the biofilm for the second addition of ciprofloxacin, approximately 2 h elapsed after MOPS-glucose was added over the biofilm. Ciprofloxacin or control was then added quickly at ϳ7.5 mm above the biofilm. For both additions, a 2-min window was given before antibiotics were added to the media and current was measured for a minimum of 1,000 s in total. O 2 concentration gradients were immediately measured in triplicate during each biological replicate after the second addition of antibiotics to determine the O 2 consumption rates.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. TEXT S1, PDF file, 0.2 MB.