Planar optode imaging system
The planar optode imaging instrumentation was developed at the Max Planck Institute in Bremen, Germany. The setup comprised of a 45 cm x 20 cm x 20 cm illumination and imaging unit (Fig 1). It housed a 8.8MP color camera (Flea3 USB from FLIR Systems Inc.), mounted on a linear stage to aid in focusing the field of view on to a sample opening at the top of housing. The camera was equipped with a 25 mm objective lens fitted with a 530 nm long-pass filter (Schott, North America). The top opening allowed samples to be imaged from below with controlled LED illumination (455 nm with a cut-off filter at 470 nm) which were positioned within the housing. Ambient light was eliminated by covering the opening with aluminum foil. The spatial resolution in the focal plane was 25µm/pixel.
The optode foils were prepared according to Larsen et al (2011) . However, a lipophilic coumarin dye Bu3Coum  was used instead of Macrolex yellow in order to avoid leaching of the antenna dye out of the polymer. The sensor material contained Bu3Coum (1.5% wt.), an oxygen indicator Pt(II) porphyrin (0.75 % wt., Frontier Scientific, frontiersci.com) embedded in polystyrene (MW 250000) along with 50% by wt. of TiO2 nanoparticles (P170, Kemira, Finland). This mixture was dissolved in chloroform (4.5g per 300mg polystyrene) (Sigma, USA) and coated in a 2µm thick layer on a clear polyethylene terephthalate sheet (Melinex 505, Pütz, Germany). The optodes were then stored dry and in the dark until used for the sample measurement setup.
Calibration of the optodes
The ratio of the red and green channels in the color images of the optode can be calibrated for dissolved oxygen (DO) using a modified Stern-Volmer equation [25, 27].Stern-Volmer equation to fit intermediate points between the 2 measured points; zero and 100% saturation on the other optodes. The 0% DO solution was made by adding 3 grams of ascorbic acid (C6H8O6, 99.9%), (Sigma, USA) in 50 mL of tap water and adjusting pH to 11 with NaOH. A 100 % DO solution was obtained by bubbling demineralized water with air for 10 min. Since the biofilms were to be grown in a modified brain heart infusion (M-BHI) broth, we used a conversion to account for the lower solubility of DO in BHI . Calibration and measurements were performed at 22 ºC Such optodes have been designed to be insensitive to pH, salinity and CO2 .
Sterilization of optodes
After calibration, the optodes were sanitized in generic mouthwash containing 1.5 % H2O2, 0.07 % cetylpyridinium chloride for 30 min at room temperature. The optodes were then rinsed with dH2O, before being applied to the growth systems. We placed the optode in a petri dish with BHI (Sigma Aldrich, USA) and incubated it for 24hrs at 37ºC to confirm sterility by inspection for turbidity and plating the media on BHI (Sigma Aldrich, USA) agar. If there was no turbidity and no colonies observed, we proceeded with the experiment.
Collection of human saliva / plaque inoculum
This study was approved by the Ohio State University for Human Subject Research (Study Number: 2017H0016). Samples were collected from six consenting healthy adult donors, who did not self-report any known underlying chronic disease and were in good oral health using a protocol adapted from Nance et al. 2013 . The donors had not received antibiotics for at least 3 months prior to collection. Collection of samples was performed in the morning for all volunteers. Volunteers were asked to refrain from eating after dinner the evening prior to and the morning of the collection. No oral hygiene was practiced for a minimum of 8 hours prior to donating. To generate a saliva/plaque inoculum, 10 ml of stimulated saliva was collected in a sterile plastic 50 mL tube (Falcon, Thermo Fisher Scientific, Waltham, MA, USA). Plaque was recovered using a standard toothbrush from the teeth and tongue by brushing but using no toothpaste. The toothbrush was vortexed in 10 mL PBS for 3 min (Vortex-Genie® 2 mixer, Scientific Industries, Inc, Bohemia, N.Y., USA) to transfer the plaque from the brush to the PBS. Vortexing was conducted in an anaerobe chamber (Bactron, USA, with a 5% CO2, 5% H2 and 90% N2 headspace) to avoid overexposure of O2 sensitive anaerobes during the mixing. The bacteria was pelleted by centrifugation (10,000 g for 3 min) and resuspended in the pooled saliva. Saliva and plaque were kept on wet ice through the whole collection procedure. Glycerol was added to a final concentration of 25%. Aliquots of this suspension were stored in 1.5 mL cryogenic tubes (Thermo Fisher Scientific, USA) at −80 °C.
A modified brain heart infusion (M-BHI) broth was used for cultivations. BHI broth (Sigma Aldrich, USA) was supplemented with 5mg/L hemin (Alpha Aesar, USA), 1mg/L menadione (MP Biomedicals, LLC, France), 0.1g/L L-cysteine (Sigma, USA) and 1g/L yeast extract (Sigma, USA). The growth media of current study was absent in. We were particularly interested to see if Gram negative strict anaerobes could survive and proliferate in a biofilm grown under an oxic headspace and we were concerned that adding sucrose would cause S. mutans and other acidogenic bacteruia to dominate since P. gingivalis is not acid tolerant .
Biofilm growth systems
Three growth systems were used. First we used hydroxyapatite (HA) discs to determine the time sequence for the development of the bacterial community under our growth conditions. For the optode work, we used a modified dual chamber slide (Nunc Lab-Tek, USA) and a polystyrene Petri dish (Corning™ Falcon, Thermo Fisher Scientific), both fitted with optode foils.
All biofilms were grown in a 5% CO2 incubator at 37oC (Thermo Fisher Scientific). Conventionally, human plaque biofilms grown in vitro are done under anoxic conditions in an anaerobic chamber with as a low Eh (Redox potential) in the bulk fluid is required for the strict anaerobes . However, we used an oxic headspace under the rationale that 1) the oral cavity is not anoxic and that the natural creation of anoxic microniches by the biofilm was more relevant and 2) to observe differences in DO at the base of the biofilm when disrupted by the microspray required an overlying oxic headspace, recapitulating the in vivo condition.
Hydroxyapatite (HA) discs
One cm diameter hydroxyapatite discs (BioSurface Technologies, Bozeman, MT) were placed into each well of a 12 well plate (Falcon, corning, USA) (Fig 2 A). Then 2 mL of sterile of M-BHI was added to each well followed by 500 µL of the saliva/plaque inoculum. The biofilm was grown for 4 days with daily media exchanges. At each day a triplicate set of dics were sacrificed to yield DNA for 16S RNA gene phylogenetic analysis (see below for details) by transferring them to 50 mL tubes (Falcon, Thermo Fisher Scientific, USA) with 5 mL of sterile phosphate-buffered saline (PBS, Gibco, Thermo Fisher Scientific). Biofilm was removed by sonicating in a sonicator bath (Model # 97043-964, VWR International, West Chester, PA, USA) for 3 min. The supernatant was then centrifuged (Legend micro 21, Thermo Fisher Scientific, USA) at 10,000 g for 10 min. The supernatant was discarded, and the pellet used for DNA extraction.
Dual chamber system
In order to confine the biofilm to a defined area for the shooting experiment, a 2-well chamber slide with removable wells was used (Nunc™ Lab-Tek™ II Chamber Slide™ System, Thermo Fisher Scientific) (Fig 2 B). Each chamber was 2 × 2 cm with a volume of 2 mL. The cover slip that formed the base of the chamber was removed and an optode foil 22 mm × 44 mm was glued in its place using silicon sealer. One chamber was used as an “untreated” control and the other as the experimental HVM treated chamber. The microspray shooting was performed in a laminar flow cupboard to contain possible aerosols. However, the microspray was so forceful in the small chamber that it displaced almost all of the reservoir liquid making the system messy and difficult to work with. Therefore, we also used a larger Petri plate system.
Petri plate system
To overcome the problem of losing the water phase during HVM treatment, we used a 9 cm dia. Petri dish (Corning™ Falcon, Thermo Fisher Scientific) to grow biofilms. A piece of optode film 22 mm × 44 mm was fixed to the bottom of the plate with electrical tape. To prevent the seepage of medium underneath the optode and to prevent the formation of bubbles between the optode and the bottom of the plate, it was important to tightly seal all four sides of the optode (Fig 2C). Electrical tape worked well because of its pliability. Another advantage of using this larger container was that the impinging spray was allowed to spread radically rather than splashing back out of the chamber system. In addition, the larger area of the optode meant that the area of biofilm dislodged by the treated area could be clearly delineated from the surrounding undisturbed areas, thus allowing oxygen mapping over both areas. A further advantage was that the optode was supported by the underlying plate and so the optode remained better positioned during HVM treatment.
Biofilm growth in the dual chamber and petri plate optode systems
The dual chamber and the Petri plate biofilms were grown under similar conditions with daily media exchanges and a daily inoculation from the pooled stock for a duration of 7 days. For the dual chamber system, 2 mL of sterile medium was added to each chamber and 200 µL of the stock inoculum was added to each well. For the Petri plate system, 5 mL of media was added and 500 µL of inoculum. After days 1, 2, 3, 4, 6 and 7, the DO was mapped using the optode system by transferring the growth chamber (either dual chamber or Petri plate) to the optode imager and imaged immediately after exchanging the media, to assure the media to be fully oxic. On days 6 and 7, the biofilm was exposed “shot” with a commercial HVM device using a single pulse (AirFloss, Philips Oral Healthcare). The device was filled with tap water to generate a water microspray which lasted approximately 60 ms and dispensed approximately 130 µl . The HVM device was convenient for these studies since it could generate high velocity fluid flow while minimizing the volume of generated liquid waste [32, 33]. In the dual chamber system, one of the chambers was HVM treated while the other chamber was left as an untreated control. The tip of the HVM device was held approximately 1 cm from the surface of the optode during the microspray. DO images were taken before and after the treatment. After the final optode measurement on day 7, the biofilm growing on the optode in the Petri plate was removed for DNA extraction for identification of species and genera by PCR. A circular area of approximately 1cm diameter in the HVM shot area was sampled by scraping with a sterile loop (Fisherbrand, disposable loop, Thermo Fisher Scientific). A similar area in the adjacent but undisturbed area was sampled in a similar manner to serve as an untreated control.
Biofilm community analysis
To ensure the bacterial community in the saliva/plaque inoculum included representative species (Table 1), we used conventional agarose gel PCR and assessed their relative abundance semi-quantitatively by gel densitometry. For biofilms grown on the HA discs and on the planar optodes in the HVM treated areas and the undisturbed biofilms, we used quantitative PCR (qPCR) to determine the presence and relative abundance of the target species.
DNA from the saliva/plaque inoculum was extracted using a boiling method, which is a simple and cheap method that has been shown to be effective for human dental plaque, although for quantification is less sensitive than qPCR. In this method the scraped biofilm was boiled in double-distilled water (ddH2O) for 10 min. and then chilled for 2 min at 20°C. The sample was then centrifuged (Legend micro 21, Thermo Fischer Scientific, USA) at 16,000 g for 10 min at room temperature. Purity of the extracted DNA was based on the 260/280nm optical density (OD) ratio by spectrophotometry (NanoDrop 1000, Thermo Fisher Scientific). For qPCR, DNA was extracted immediately after the shot. For qPCR, we used the established method used in the “Human Oral Microbiome Identification using Next Generation Sequencing” (HOMINGS) protocols developed at the Forsyth Institute, for which we used an Epicenter, Puremaster Kit (Epicenter, Madison, WI) to extract DNA from biofilm .
Conventional PCR and densitometry for the saliva/plaque inoculum
PCR conditions (annealing temperature and numbers of PCR cycles) were optimized to identify 5 of the 7 target species and genera using the primer sets shown in Table 1. Amplification was performed in a 25 µl mixture containing Mg2+, dNTPs, and recombinant Taq DNA Polymerase for routine PCR of fragments up to 5 kb (Invitrogen, USA), 10µM forward and reverse primers and 2 µl bacterial DNA extract in a thermal cycler (MyCycler™ Thermal Cycler system, BioRad). PCR was carried out using the following conditions: an initial denaturation step for 4 min at 94 °C, with 45 cycles of 30 s at 95 °C, 1 min at 58 °C and 30 s at 72 °C, followed by 5 min at 72 °C. Agarose gel (Sigma, USA) was prepared at a concentration of 1.5% (w/v) in 60 ml Tris-Borate Buffer (TBE). 1 µl of 10 µg/ml ethidium bromide (Sigma, USA) was incorporated into the gel to make a final concentration of 0.5 µg/ml and electrophoresed at 90 V for 60 min. The DNA bands were visualized using a gel documentation system (ChemiDoc XRS, Bio-Rad, USA) under ultraviolet (UV) illumination. Gel densitometry analysis was performed for semi-quantification of the target species and genera using the freely available NIH FIJI image analysis software  where the brightness of the band was measured by positioning a region of interest (of consistent area) within each band and measuring the grayscale value.
qPCR for analysis of HA and optode grown biofilms
qPCR was performed using an iQ5 real time PCR detection system (CFX96, Bio-Rad, USA). Each PCR was performed in a total volume of 20 μL containing 2 μl of 10× iQ SYBR® Green Supermix (Bio‐Rad, Hercules, CA), 0.5 μmol/L each of forward and reverse primers, 7 μl of ddH2O water, and 1 μl (10 ng) of template DNA., depending on the ratio of the individual bacterial species within the biofilm. The qPCR was carried out with an initial incubation of 2 min at 95 °C, followed by 45 cycles of denaturation for 15 s at 95 °C, annealing for 1 min at 60 °C, melting curve 65 °C to 95 °C incremented by 0.5 ºC every 0.05 s. Standard curves were generated for all the target bacteria using DNA from pure cultures and species specific DNA primers. Ten fold serial dilutions of genomic DNA (DNA from 107 to 101 cells), from pure bacterial cultures provided seven data points for standard curve generation. In each standard curve, the concentration of standard sample (known amount of DNA/DNA from known number of bacteria) was plotted against its crossing points. Quantification of the individual target bacteria and total bacteria from the experimental samples were calculated using the standard curves. For each species, a standard curve was generated using defined concentrations of genomic input DNA (10-fold serial dilution). The genomic DNA amount of the target species or genera in the unknown sample was calculated from the standard curve. For the HA grown biofilms triplicate discs were sacrificed for qPCR analysis on each of the 4 days. For the optode experiments biofilms were grown for 6 days with daily nutrient exchanges and then shot with the HVM on days 6 and 7. Triplicate unshot optodes were used for comparative controls. Data were analyzed using Bio-Rad CFX manager software 3.1. qPCR analysis resulted in Ct-values which were converted to copy number using a calibration curve and graphics were made using GraphPad Prism 8.0 (GraphPad Software, Inc., San Diego, CA). Student t-test was used for qPCR data analysis. P < 0.05 was considered significantly different.
Confocal laser scanning microscope (CLSM)
CLSM was used to measure the thickness and visualize the structure of the biofilm on the HA discs and the optodes in the undisturbed areas and the HVM “shot” areas. The dental/plaque biofilm was grown for 7 days on both surfaces, and shot on day 6 and 7. Briefly, on day 7 biofilms on the HA discs and optodes were rinsed two times with sterile saline to remove loose bacteria, and then “shot” with a HMV. Immediately after the shot, the samples rinsed with sterile water and the remaining bacteria were stained with 2.5 μM SYTO 9 green fluorescent nucleic acid stain (480/500 nm; ThermoFisher Scientific) with a 30 min incubation in the dark at room temperature . After incubation imaging was performed using an Olympus FV1000-MPE upright confocal/multiphoton microscope (Olympus Corp., USA) equipped with a ×25, 1.05-numerical-aperture water immersion dipping objective lens. The excitation wavelength was 488 nm, and the emission wavelength filter for SYTO 9 was a 500/530 OlyMPFC1 filter. Each biofilm was scanned at three randomly selected positions in the HVM and undisturbed areas on 3 independent replicates of HA discs and one optode. TheXY-Z series were generated by optical sectioning at each of these positions. Graphics were made using GraphPad Prism 8.0 (GraphPad Software, Inc., San Diego, CA). COMSTAT2  was used to measure the maximum thickness of the biofilm and student t-test run for data analysis. P < 0.05 was considered significantly different.