The P. gingivalis ATCC strain 33277 was cultured under anaerobic conditions in Trypticase soy broth (TSB) supplemented with yeast extract (1 μg/ml), menadione (1 μg/ml), and hemin (5 μg/ml), at 37°C and harvested as described previously (Choi et al., 2013). The number of bacteria for infection was determined at mid-log phase using a Klett-Summerson photometer (Klett) and used at a multiplicity of infection (MOI) of 100 in all infection experiments. All the experiments were performed in our Biosafety Level 2 laboratory approved by the institutional biosafety committee, protocol number 530. Primary cultures of GECs were generated and cultured as described previously (Yilmaz et al., 2003, 2008). At least three different collected patient cells were used independently for the experiments and were tested for viability when treated with inhibitors or other chemical reagents in the assays performed for this study. The data represents statistical analyses of the three separate patient cells and their individual replication experiments. No human subject recruitment per-se was done. Adult patients were selected at random and anonymously from those presenting at the University of Florida Dental Clinics for tooth crown lengthening or impacted third molar extraction. No patient information was collected. Gingival tissue that would otherwise be discarded was collected after informed consent was obtained by all patients under the approved guidance of the University of Florida Health Science Center Institutional Review Board (IRB, human subjects assurance number FWA 00005790).

The mRNA expression of each NOX isoform was measured by quantitative real time PCR, using specific TaqMan primers (Applied Biosystems) for each. RNA was isolated using Qiagen RNeasy Plus Kit (Qiagen) and 10 μg per sample was reverse-transcribed from untreated and 3 mM ATP (Sigma) treated primary GECs; the ATP concentration determined and used in our previous studies (Yilmaz et al., 2010; Choi et al., 2013; Hung et al., 2013; Johnson et al., 2015; Atanasova et al., 2016). cDNA were constructed using Superscript III First Strand cDNA Synthesis Kit (Invitrogen) and quantitative PCR was conducted at 95°C for 10 min, followed by 45 cycles at 95°C for 15 s, and at 60°C for 60 s. Gene expression analysis was performed using the CFX Manager Software (BioRad). The expression of reference gene GADPH was used to normalize the measured target gene. The untreated control was used to further normalize the relative quantities for the NOX genes.

TaqMan Primers used from Applied Biosystems.

Cells were seeded in T25 flasks and treated with 3 mM ATP (Sigma) and/or infected with P. gingivalis over a 24 h time period. The cells (4 × 10⁶) were collected using 300 μL of extraction buffer and the NADP/NADPH levels were measured using the NADP/NADPH Assay Kit (Abcam, Cambridge, UK; ab65349). A time course was performed at 30 min intervals to determine optimal incubation time for detection at room temperature and read on a Biotek H1M monochromatic plate reader at 450 nm. The NADP and NADPH measured levels were calculated by comparison with the standard curve. Protein concentration was determined and normalized for each sample prior to analysis.

Cells were incubated with 3 mM ATP in a 24-well plate or 4-well microscopic slide chamber system (Warner instruments). ATP was added to the cells for 30 min and then incubated with the peptide NOX2 inhibitor gp91ds-tat or gp91ds-tat scrambled (non-targeting) control peptide (Anaspec, Fremont, CA) at a concentration of 50 μM determined by dose response assay and the previous literature (Supplementary Figure 1). The effective delivery of this peptide into the GECs was confirmed using FAM-labeled gp91ds-tat (Anaspec, Fremont, CA) peptide (not shown). Cells were pre-labeled with the fluorescent probe 5-(and-6)-chloromethyl-2′,7′-dichlorodihydrofluorescein diacetate, acetyl ester (CM-H₂DCFDA) (Invitrogen, Waltham, MA) which was solubilized in dimethyl sulfoxide (DMSO) (Sigma, St. Louis, MO) and then diluted to 5 μM in Hanks balanced salt solution (HBSS) containing MgCl₂ and CaCl₂ (Invitrogen). The cells were washed twice with HBSS, incubated with 5 μM CM-H₂DCFDA at 37°C for 30 min in a fresh medium. The CM-H₂DCFDA (green) fluorescence intensity was measured using a Biotek H1M monochromatic bottom fluorescence plate reader at excitation 495 nm and emission 525 nm. Results are averaged data from three individual experiments performed in duplicate and normalized to untreated cells as a control. The representative live images of the CM-H₂DCFDA (green) fluorescence intensity were taken using epifluorescence (DM IRE2 HC inverted scope, Leica Microsystems GmbH) microscopy equipped with DIC. The single exposure images were collected sequentially in fluorescence and DIC using a Retiga 4000r CCD camera (Qimaging).

GECs were seeded in 6-well plates (2.5 × 10⁵) and treated with 3 mM ATP for 1 h prior to P. gingivalis infection for 1, 3, and 24 h. The cell lysates were collected and fractionated into cytosolic, membrane, and nuclear fractions according to the manufacturer's protocol using a Subcellular Fractionation Kit (Pierce Technologies, #87790). Samples were loaded on a 10% SDS-page gel and run at 140 V for 1 h. The separated proteins were transferred to a nitrocellulose membrane by wet transfer at 50 V for 1 h and blocked with 5% non-fat milk in 1x Tris-buffered saline and 0.1% Tween-20. The membrane was incubated overnight in 1:1,000 rabbit p67phox (Millipore), 1:1,000 mouse Rac1 (Millipore) and 1:1,000 mouse β-tubulin (Invitrogen) antibodies, washed, and incubated with secondary rabbit or mouse HRP conjugated antibodies (Invitrogen) for 1 h. Membranes were incubated with enhanced chemiluminescence (GE Healthcare Life Sciences) and exposed by autoradiography. This experiment was performed in triplicate.

GECs were seeded at a density of 8 × 10⁴ on 17 mm glass coverslips (Warner Instruments) in four-well plates (Thermo Fisher Scientific). Cells were treated first with 3 mM ATP for 1 h and then infected with P. gingivalis for 3 and 24 h. The cells were fixed with 10% neutral buffered formalin, permeabilized by 0.1% Triton X-100 and blocked with 3% BSA for 45 min. The cells were stained for 1 h at room temperature with a mouse polyclonal antibody against Rac1 (1:500; Millipore) and a rabbit polyclonal antibody against P. gingivalis 33277 (1:1,000). The stained cells were washed and further incubated for 1 h at room temperature with Alexa Fluor 594 conjugated secondary goat anti-mouse polyclonal antibody and Alexa Fluor 488 conjugated secondary goat anti-rabbit polyclonal antibody (1:1,000; Invitrogen). The actin cytoskeleton was labeled using Alexa Fluor 350 phalloidin (1:500; Invitrogen). Coverslips with fixed cells were mounted onto Corning glass microscopy slides (Corning) using VectaShield mounting medium containing DAPI (Vector Laboratories). Images were acquired using Zeiss Axio Imager A1 epifluorescence microscope using QImaging MicroPublisher 3.3 cooled microscope camera and QCapture software.

Cells were seeded in T25 flasks at 70% confluency and treated with 3 mM ATP and P. gingivalis infection for 1, 3, and 24 h. ATP was added 1 h prior to infection. Cell lysates were collected in 1x RIPA buffer with protease inhibitor cocktail (Pierce Technologies) and protein levels determined by Bradford assay. The same protein quantity of lysates were incubated with 1 μg p47phox antibody (Novus Biologicals) for 1 h and further incubated with 20 μL Protein A/G bead plus (Santa Cruz) for 2 h at room temperature. The captured protein was collected in 40 μL 1x Laemmli Sample Buffer and loaded onto a 10% SDS-Page gel following the Western blot protocol described above. The blot was incubated with mouse phosphoserine antibody (1:1,000; Millipore) and mouse β-tubulin loading control (1:1,000; Invitrogen). Clean blot secondary antibody detection system designed to reduce background from immunoprecipitated samples (Pierce Technologies) at a 1:400 final concentration.

MPO was visualized using immunoperoxidase staining in untreated, ATP (3 mM) or H₂O₂ (100 μM) treated cells; the concentrations determined and used in our previous studies (Yilmaz et al., 2010; Choi et al., 2013; Hung et al., 2013; Johnson et al., 2015; Atanasova et al., 2016). The following protocol was used for immunoperoxidase staining: Cells were fixed in 10% NBF for 30 min at RT and permeabilized for 15 min using 0.1% TritonX-100. Endogenous peroxidase activity was quenched by incubating the cells for 30 min in 0.3% hydrogen peroxide in methanol. Cells were washed and then blocked in 3% BSA in PBS-Tween and incubated for 45 min at room temperature. Cells were incubated with the primary mouse MPO antibody (R&D systems; 2.5 μg/mL) overnight at 4°C. After 30 min of 10-min washes, cells were incubated with anti-mouse HRP-labeled secondary antibody (Invitrogen; 1:100) for 1 h. After again washing, DAB-peroxidase solution was added to the cells and incubated for 10 min in the dark. Cells were then imaged using DM IRE2 HC inverted scope, Leica Microsystems at 40x magnification. ImageJ software analysis was used to measure relative intensity of the peroxidase stain and compared to untreated.

Cell lysates of 3 mM ATP treated and P. gingivalis-infected or uninfected GECs were used for the quantitative measurement of MPO by ELISA (R&D systems; DY3174). Cells pre-treated with ATP and/or P. gingivalis were collected in 200 μL of Reagent diluent (1% BSA in PBS), disrupted by sonication three times for 5 s on ice, and clarified at max speed for 5 min. MPO-ELISA was performed following the manufacturers protocol on Corning disposable sterile ELISA plates (Corning). The optical density was read at absorbance of 450 nm using a Biotek H1M monochromatic plate reader with wavelength correction of 540 nm. A standard curve was performed at two-fold dilutions from 2,000 to 31.25 pg/mL.

GECs were seeded at 70% confluence in T25 flasks and treated with 3 mM ATP, 100 μM H₂O₂ or P. gingivalis infection. ATP was added 1 h prior to P. gingivalis infection. The cells were collected in 100 μL assay buffer provided by the MPO activity assay kit and sonicated 3x on ice for 5 s each, then left for 10 min on ice per the manufacturers protocol (Abcam, ab111749). The fluorescence produced was measured on a Biotek H1M monochromatic plate reader kinetically at 2-min intervals for 30 min with an excitation of 484 nm and emission of 525 nm.

GECs (8 × 10⁴) were pre-treated with 3 mM ATP, P. gingivalis for 2 h, 1 μM MPO inhibitor 4-Aminobenzoic hydrazide (Santa Cruz) or 50 μM gp91ds-tat/gp91ds-tat scrambled peptides (Anaspec) and then incubated for 15 min in the dark with 10 μM R19S fluorescent probe (FutureChem, South Korea). The R19S probe fluorescence was measured using the Biotek H1M monochromatic plate reader at an excitation of 515 nm and emission of 545 nm. Live fluorescent images were taken under a (DM IRE2 HC inverted scope, Leica Microsystems) microscopy equipped with DIC.

The intracellular reduced glutathione was quantified using ThiolTracker Violet stain (Invitrogen) in living cells, since glutathione represents the majority of intracellular free thiols in the cell. Briefly, primary GECs were treated with or without 3 mM ATP for 1 h and/or infected with P. gingivalis (MOI100) in 4-well microscopic slide chamber system (LabTek) or 24-well plate. After the designated incubation period, the cells were washed and incubated in Dulbecco's PBS containing 10 μM ThiolTracker Violet for 30 min at 37°C in the dark. The cells were then read on the Biotek H1M monochromatic plate reader at excitation 405 nm and emission 525 nm. GECs seeded in 4-well chamber slides were fixed with 10% neutral buffered formalin, permeabilized by 0.1% Triton X-100 and blocked with 3% BSA for 30 min. P. gingivalis was immunostained using the rabbit polyclonal P. gingivalis antibody (1:1,000) and secondary antibody Alexa Fluor 594 (1:1,000; Invitrogen). The slides were viewed using inverted epifluorescence microscopy (DM IRE2 HC inverted scope, Leica Microsystems GmbH) and images taken using a Retiga 4000r CCD camera (Qimaging). H₂O₂ (100 μM) was used as a positive control for oxidative stress and the antioxidant N-acetyl cysteine (NAC) (50 mM) as an inhibitor of oxidative stress.

Transcription of the Glutathione ligase genes (GCLc and GCLm), Glutathione synthetase and glutathione reductase were measured by quantitative real time PCR using specific primers and the SybrGreen detection system. RNA was isolated using Qiagen RNeasy Plus Kit (Qiagen) and 1 μg per sample was reverse-transcribed from untreated, P. gingivalis infected, and/or 3 mM ATP treated primary GECs. cDNA was constructed using Superscript III First Strand cDNA Synthesis Kit (Invitrogen). Quantitative PCR was conducted at 95°C for 3 min and 40 cycles at 95°C for 30 s and 61°C for 30 s. Gene expression analysis was performed using the CFX Manager Software (BioRad). The expression of the reference gene GADPH was used to normalize the measured target gene and the untreated control was used to further normalize the relative quantities.

SybrGreen Primers.

All assays were performed in at least three separate occasions in duplicate or triplicate each time. Results are expressed as mean ± SEM. Two-tailed Student's t-test was used to calculate the statistical significance of the experimental results between two conditions (significance considered at p < 0.05).

Previously, we have shown that P. gingivalis infection and eATP induce ROS in primary GECs, largely generated by NADPH oxidases (NOX) inside the cell (Choi et al., 2013). It has been previously determined that eATP treatment of primary GECs results in sustained ROS levels as early as 30 min after treatment (Choi et al., 2013), therefore, we initially examined the mRNA expression of the different NOX isoforms with eATP stimulation to evaluate the NOX isoform important for eATP-induced ROS. The mRNA expression of the seven NOX isoforms in GECS with and without 3 mM ATP treatment revealed differential expression levels whereas NOX 1 and 3 isoforms were not expressed. Among the expressed NOXs, eATP significantly increased NOX2 expression only (Figure 2). We then measured the consumption of NADPH in GECs during ATP treatment using specific detection assays to further demonstrate the activity of NOXs in response to danger signal ATP. The consumption of NADPH steadily increased through 24 h of post-ATP treatment in GECs suggesting increased NOX activity (Figure 3A). However, this ATP-induced NADPH consumption is modulated by P. gingivalis infection where the highest increase is at 3 h post-infection that is significantly decreased over 24 h, supporting that P. gingivalis modulates NOX activity (Figure 4). The pattern of NADPH consumption is also consistent with the previously published ROS modulation by P. gingivalis in the presence of ATP (Choi et al., 2013). We further investigated the specific contribution of NOX2 to eATP-induced ROS using a well-established specific peptide inhibitor against NOX2, gp91ds-tat, which prevents the interaction of the required subunit p47phox to bind to NOX2 (Rey et al., 2001; Csanyi et al., 2011; Park et al., 2011). Treatment of ATP stimulated primary GECs with the NOX2 inhibitor significantly decreased ATP-induced ROS as measured by the fluorescent probe, DCFDA (Figures 3B,C). This supports the specific role of NOX2 in ROS produced by the GECs in response to eATP and infection.

Although, the expression of NOX2 was increased upon eATP treatment, this may not necessarily suggest increased NOX2 activity. The activation of NOX2 to produce ROS requires the recruitment of specific subunits (Cross and Segal, 2004; Bedard and Krause, 2007). Therefore, we more closely examined the changes in the active forms and localization of these subunits. The critical organizing subunit for the NOX2 complex, p47phox, requires phosphorylation on multiple serine residues to bind to the other subunits in the complex (Hoyal et al., 2003; El-Benna et al., 2009; Meijles et al., 2014). eATP significantly increases the serine phosphorylation of p47phox. P. gingivalis modulates the phosphorylation state of p47phox with the most significant increase at 3 h of post-infection which is subsequently reduced over 24 h of post-infection (Figure 5A). This result was consistent with marked increases of ROS previously shown at 3 h of infection which is abrogated over time of infection (Choi et al., 2013). In addition, analysis of the localization of interacting subunits Rac1 and p67phox during ATP treatment and P. gingivalis infection reveals similar trends of an increase at 3 h with infection only and decrease at 24 h of P. gingivalis infection (Figure 5B). Micrographs of this localization shift in Rac1, provides visual support of the changes which occur following ATP treatment and the infection (Figure 5C). These changes in subunit activation and localization is consistent with the temporal modulation of NOX activity (Figure 4) and eATP-induced ROS kinetics previously described in human primary GECs (Choi et al., 2013).

Further downstream of NOX-ROS is the generation of hypochlorous acid (HOCl) by myeloperoxidase (MPO) from H₂O₂. Although in neutrophils, MPO is one of the most abundant proteins (Klebanoff et al., 2013), its expression in epithelial cells has not been defined, although is anticipated to be lower. Immunoperoxidase staining in H₂O₂ treated primary GECs shows a significant increase in the relative expression of MPO in stressed conditions (Figure 6A). Quantification of MPO expression using an ELISA assay which showed a marked increase in MPO expression in ATP treated cells, and a significant increase with the addition of H₂O₂ or pre-ATP treatment and P. gingivalis infection (Figure 6B). However, clinical studies reveal that the protein level of MPO may not always provide complete information on the enzymatic activity of MPO which can sometimes vary even if the amount of MPO is similar (Pulli et al., 2013). Therefore, we examined the activity of MPO in primary GECs and show a significant increase in the activity of MPO in ATP treated or P. gingivalis infected cells (Figure 6C). Despite this increase in the MPO activity, our data shows that P. gingivalis is able to significantly inhibit its downstream product antimicrobial HOCl present at 2 h post-infection (Figures 7A,B). Earlier infection times follow a decreasing trend of HOCl production after 30 min of infection with eATP (Supplementary Figure 2). In addition, the lack of HOCl production in P. gingivalis only infected GECs provides evidence for the importance of danger signal ATP signaling in this host innate defense process. Furthermore, the use of the specific inhibitors against MPO (4-aminobenzoic hydrazide) and NOX2 (gp91ds-tat) also significantly decreased HOCl produced in response to ATP. These results suggest that primary GECs induce MPO as a response to infection or stress stimuli such as ATP and that both NOX2 and MPO play major roles in HOCl production against pathogens in primary GECs. However, a well-adapted opportunistic bacterium, P. gingivalis appears to have specific molecular strategies to neutralize host ATP-induced HOCl and possibly evade host pathogen clearance.

To define a potential molecular mechanism for P. gingivalis-mediated reduction in ATP-induced HOCl, we assayed the levels of the critical homeostatic antioxidant glutathione. In its reduced form, glutathione is reactive with HOCl (den Hartog et al., 2002; Venglarik et al., 2003; Gray et al., 2013; Haenen and Bast, 2014). The ThiolTracker Violet dye, which detects the intracellular reduced form of glutathione, revealed a time-dependent oxidative stress response in the presence of ATP in primary GECs, as determined by a decrease in the dye stain (Figures 7C,D). P. gingivalis, in early hours of infection in ATP-stimulated GECs also showed a decreasing trend in the reduced glutathione levels with a significant decline at 2 h of infection which suggests that the antioxidant is scavenging ROS and/or HOCl (Figures 7C,D). Furthermore, depletion of glutathione synthesis using buthionine sulfoximine (BSO; Markovic et al., 2009; Rocha et al., 2014) abrogated P. gingivalis' ability to lessen ATP-induced HOCl which supports a role for glutathione in scavenging the HOCl during early P. gingivalis infection (Supplementary Figure 3). We therefore examined the mRNA expression of key enzymes for glutathione synthesis (Lu, 2013; GCLc and GCLm, glutathione synthetase, and glutathione reductase). The expression levels of GCLc and GCLm, the subunits of the glutamate cysteine ligase (GCL), are significantly increased by P. gingivalis infection at 6 and 24 h of infection even in the presence of ATP. GCL is a critical rate-limiting step in glutathione synthesis (Lu, 2013), therefore the increase of expression suggests that P. gingivalis is inducing more glutathione production. Moreover, the second step in glutathione synthesis requires glutathione synthetase which is significantly increased by P. gingivalis at 3 h post-infection and reduced at 24 h. Glutathione reductase, which reduces oxidized glutathione, follows a similar pattern and is significantly increased at 3 and 6 h post-infection, but returns to normal levels at 24 h (Figure 8). It is important to note that at 3 h of P. gingivalis infection with and without ATP treatment there is a notable increase in glutathione levels at 3 h of infection as compared to the reduced amount at 2 h of infection (Figures 7C,D). The increased expression of glutathione synthetase and reductase at this same time during P. gingivalis infection suggests P. gingivalis can induce host glutathione production as a combative antioxidant response to ATP-induced cellular HOCl. Furthermore, we have previously described (Choi et al., 2013) and also reconfirm using ThiolTracker that 24 h P. gingivalis infection had comparable levels of reduced glutathione detected by ThiolTracker to untreated with and without ATP treatment (not shown). Taken together, these findings suggest that P. gingivalis can control the cellular glutathione synthesis pathway and utilize host glutathione to scavenge cellular HOCl. By further increasing the key synthesizing enzymes necessary, P. gingivalis can restore the reduced glutathione levels in primary GECs and promote an intracellular environment likely conducive for prolonged bacterial survival and persistence.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.