Susceptibility patterns and the role of extracellular DNA in Staphylococcus epidermidis biofilm resistance to physico-chemical stress exposure

Background Over 65% of human infections are ascribed to bacterial biofilms that are often highly resistant to antibiotics and host immunity. Staphylococcus epidermidis is the predominant cause of recurrent nosocomial and biofilm-related infections. However, the susceptibility patterns of S. epidermidis biofilms to physico-chemical stress induced by commonly recommended disinfectants [(heat, sodium chloride (NaCl), sodium hypochlorite (NaOCl) and hydrogen peroxide (H2O2)] in domestic and human healthcare settings remains largely unknown. Further, the molecular mechanisms of bacterial biofilms resistance to the physico-chemical stresses remain unclear. Growing evidence demonstrates that extracellular DNA (eDNA) protects bacterial biofilms against antibiotics. However, the role of eDNA as a potential mechanism underlying S. epidermidis biofilms resistance to physico-chemical stress exposure is yet to be understood. Therefore, this study aimed to evaluate the susceptibility patterns of and eDNA release by S. epidermidis biofilm and planktonic cells to physico-chemical stress exposure. Results S. epidermidis biofilms exposed to physico-chemical stress conditions commonly recommended for disinfection [heat (60 °C), 1.72 M NaCl, solution containing 150 μL of waterguard (0.178 M NaOCl) in 1 L of water or 1.77 M H2O2] for 30 and 60 min exhibited lower log reductions of CFU/mL than the corresponding planktonic cells (p < 0.0001). The eDNA released by sub-lethal heat (50 °C)-treated S. epidermidis biofilm and planktonic cells was not statistically different (p = 0.8501). However, 50 °C-treated S. epidermidis biofilm cells released significantly increased eDNA than the untreated controls (p = 0.0098). The eDNA released by 0.8 M NaCl-treated S. epidermidis biofilm and planktonic cells was not significantly different (p = 0.9697). Conversely, 5 mM NaOCl-treated S. epidermidis biofilms exhibited significantly increased eDNA release than the corresponding planktonic cells (p = 0.0015). Further, the 50 μM H2O2-treated S. epidermidis biofilms released significantly more eDNA than the corresponding planktonic cells (p = 0.021). Conclusions S. epidermidis biofilms were less susceptible to physico-chemical stress induced by the four commonly recommended disinfectants than the analogous planktonic cells. Further, S. epidermidis biofilms enhanced eDNA release in response to the sub-lethal heat and oxidative stress exposure than the corresponding planktonic cells suggesting a role of eDNA in biofilms resistance to the physico-chemical stresses.


Background
A bacterial population exists either as planktonic (freefloating cells) or as biofilm [1]. Bacterial biofilm refers to a sessile cluster of bacterial cells encased in a selforiginating extracellular matrix (ECM) [2] composed of polysaccharides, proteins, water, lipids and nucleic acids [1]. Bacterial biofilms are ubiquitous and cause over 65% of human infections [3]. Moreover, bacterial biofilms are less susceptible to antibiotics and host immunity [4,5]. Staphylococcus epidermidis is a Gram-positive coagulase negative bacteria most commonly linked with nosocomial and biofilm-related infections [6]. S. epidermidis adopts a biofilm lifestyle that enables resistance to antibiotics and host immunity, which potentially could lead to severe conditions such as bacteremia and sepsis, and if left untreated could result in death [7]. Up to 80% of infections of medical implant devices such as central venous catheters, cardiac pacemakers, tissue fillers, intrauterine devices and joint prostheses are caused by S. epidermidis biofilms [8].
Several studies have compared the susceptibilities of biofilm and planktonic forms of bacterial species to various physico-chemical stresses. For instance, studies have reported that biofilm forms of Mycoplasma bovis [9], Vibrio cholerae O1 [10], Salmonella enterica [11], Burkholderia cenocepacia [12], Lactobacillus plantarum subsp. planturum [13] Pseudomonas aeruginosa [14], Mycobacterium avium, Mycobacterium intracellulare [15] and Klebsiella pneumoniae [16] are more resistant to various conditions of heat, osmotic and oxidative stress exposure than the respective planktonic forms. In contrast to the bacterial species above, S. epidermidis biofilm is the most clinically relevant and a model of bacterial biofilm infections [17]. However, the susceptibility patterns of S. epidermidis biofilms to physicochemical stresses commonly recommended for disinfection of food, drinking water, surfaces and medical equipment in domestic and human healthcare settings remains largely unknown.
The molecular mechanisms underlying bacterial biofilms resistance to stress agents remains largely unclear [18]. It is thought that mechanisms such as limited diffusion across the ECM barrier, the slow growth rate of biofilm cells, overproduction of antibiotics' neutralizing enzymes, physiological heterogeneity of biofilms, the presence of persister cells and adaptive stress responses contribute to the high bacterial biofilm resistance [1,3] against antimicrobials. However, the above mechanisms not only inconclusively explain the resistance of bacterial biofilms to antibiotics but also apply to a limited bacterial species [1,6]. Growing evidence is beginning to link extracellular DNA (eDNA) of genomic origin released via active secretion or controlled cell lysis [19] with microbial biofilm resistance to various stress agents. For instance, eDNA has been shown to protect S. epidermidis, Staphylococcus aureus and Actinobacillus pleuropneumoniae biofilms against antibiotics such as vancomycin, βlactams and penicillin G respectively [20][21][22]. Furthermore, a study reported that eDNA protects Candida albicans biofilm against 5 mM hydrogen peroxide (H 2 O 2 ) stress [23]. The release of eDNA via lysis of a subpopulation of cells is a common phenomenon throughout the life cycle of biofilm-forming bacteria such as S. epidermidis [19]. However, the role that the released eDNA plays in S. epidermidis biofilms resistance to physico-chemical stress exposure is yet to be understood. Therefore, the present study aimed to evaluate the susceptibility of and the eDNA release by S. epidermidis biofilm and planktonic cells to physico-chemical stress exposure.

Sample collection, bacterial isolate and growth conditions
The study was conducted between November 2015 and January 2017. S. epidermidis isolates were obtained by swabbing arm joints of seventy-one outpatients at Kisumu County Hospital, a referral health facility in western Kenya using a protocol described in [24]. A swab from each outpatient was plated on mannitol salt agar (MSA) (HiMedia Laboratories Pvt. Limited, Nashik, India) and then incubated for 24 h (hours) at 37°C. One S. epidermidis isolate (largest colony) from the MSA plate was inoculated into 2 mL Tryptic Soy Broth (TSB) (Sigma Aldrich Chemie GmbH, Steinheim, Germany) and incubated at 37°C with shaking at 120 revolutions per min (rpm) for 18 h to form S. epidermidis suspension. Biofilm-forming ability of each of the S. epidermidis suspension was assessed using the tube method biofilm assay as described in [25]. Only sixty-two S. epidermidis suspensions that were capable of forming biofilms were used in the present study.
A pair of S. epidermidis biofilm and planktonic cultures was generated from the biofilm-forming suspension as described in [26] with few modifications. Briefly, to generate S. epidermidis planktonic culture, 100 μL of the S. epidermidis suspension was inoculated into 10 mL fresh TSB (Sigma Aldrich Chemie GmbH, Steinheim, Germany) in conical polystyrene tube and then incubated at 37°C with shaking at 120 rpm for 18 h. To generate S. epidermidis biofilm culture, 100 μL of the S. epidermidis suspension was transferred to a conical polystyrene tube containing 10 mL fresh TSB (Sigma Aldrich Chemie GmbH, Steinheim, Germany) supplemented with 1% glucose (Unilab Limited, Nairobi, Kenya) to induce biofilm formation and then incubated at 37°C with shaking at 120 rpm for 24 h. In this study, S. aureus American Type Culture Collection (ATCC) 29,213 was used as a reference strain due to its good biofilm-forming ability within 24 h [27].

Susceptibilities of S. epidermidis biofilm and planktonic cells to physico-chemical stress induced by the commonly recommended disinfectants
Based on the recommended guidelines for routine disinfection in domestic and human healthcare settings, the following physico-chemical stress conditions were used: heat (60°C) [28], 1.72 M sodium chloride (NaCl) [29], 150 μL of waterguard in 1 L of water [30] and 1.77 M H 2 O 2 [31]. Waterguard contains 0.178 M sodium hypochlorite (NaOCl) as the disinfectant [30].
Procedure of exposure of S. epidermidis biofilm and planktonic cells to the physico-chemical stresses commonly recommended for disinfection The effectiveness of the physico-chemical stress induced by the commonly recommended disinfectants against S. epidermidis biofilm and planktonic cells was determined as described in [16] with few modifications. Briefly, 1 mL of S. epidermidis biofilm or planktonic culture diluted to an OD 600 of 1.0 was transferred to 9 mL of 1. 72 M NaCl (Unilab Limited, Nairobi, Kenya), solution containing 150 μL waterguard in 1 L of water (Supersleek, Nairobi, Kenya) or 1.77 M H 2 O 2 (RFCL Limited, New Delhi, India) vortexed for 2 min then incubated at 37°C with shaking at 120 rpm for 60 min. For heat stress, 1 mL of S. epidermidis biofilm or planktonic culture diluted to an OD 600 of 1.0 was added to 9 mL of sterile distilled water and then placed in a water bath model JSWB-11(T) (JS Research Inc., Gongju-city, Korea) at 60°C. At 0, 30 and 60 min of exposure to each disinfectant, 1 mL sample was drawn for colony-forming units (CFUs) enumeration. To neutralize the waterguard and H 2 O 2 -treated cultures, 2 g/L sodium thiosulphate (Unilab Limited, Nairobi, Kenya) was placed in the first dilution tube. For NaCl stress-treated cultures, sterile distilled water was used instead of sodium thiosulphate. For heat stress, sterile water at 4°C was placed in the first dilution tube to lower the temperature. Three replicate experiments were conducted.
Enumeration and normalization of CFUs of S. epidermidis biofilm and planktonic cells exposed to the physicochemical stresses commonly recommended for disinfection The S. epidermidis biofilm and planktonic cells sampled at 0, 30 and 60 min were enumerated as described in [32] with few modifications. Briefly, 1 mL sample obtained at each time point (0, 30 and 60 min) was serially diluted 5fold. Then, 100 μL of the 10 − 5 dilution was plated in duplicate on Nutrient agar (HiMedia Laboratories Pvt. Limited, Mumbai, India) and then incubated for 20 h at 37°C. The CFUs were counted using Colony Counter SC6 plus (Bibby Scientific Limited, Staffordshire, United Kingdom) and then converted into CFU/mL. The CFU/mL values were normalized into log reduction of CFU/mL as described in [16]. Briefly, log reduction is defined as the negative log 10 of the quotient of CFU after treatment and before treatment [−log 10 (CFUafter treatment / CFU before treatment )]. The log reduction of CFU/mL for three replicate experiments were averaged and standard error of the mean (SEM) calculated. A log reduction value is directly proportional to the difference between the number of CFUs after and before treatment.
Quantification of the effects of sub-lethal physicochemical stress exposure on eDNA release by S. epidermidis biofilm and planktonic cells It is recommended that bacterial biofilm samples exhibiting high resistance to disinfectants should be selected for further molecular analyses of the resistance mechanism(s) [33]. Therefore, a subset of S. epidermidis biofilm cultures (n = 12) that showed high resistance (smaller log reduction of CFU/mL units) to physicochemical stress induced by the four commonly recommended disinfectants and the corresponding planktonic cultures were selected for eDNA experiments. Sub-lethal stress exposure induces bacterial eDNA release by partial repairable cell lysis or lysis-independent mechanisms and not via mass cell die off or explosive lysis [34] hence suitable for evaluating the effects of a stress agent on bacterial biofilm mechanisms such as eDNA release.

Determination of the sub-lethal physico-chemical stress conditions
Sub-lethal concentrations of NaCl, NaOCl, and H 2 O 2 were determined as described in [35] with some modifications. Briefly, 200 μL aliquot of S. epidermidis planktonic culture (pooled together from five different planktonic cultures) diluted to an OD 600 of 1.0 was inoculated into 2 mL of increasing concentrations of NaCl (Unilab Limited, Nairobi, Kenya), NaOCl (Supersleek, Nairobi, Kenya) or H 2 O 2 (RFCL Limited, New Delhi, India) for 60 min. At the 60th min, 1 mL of NaOCl and H 2 O 2 stress-treated S. epidermidis cultures were neutralized with 200 μL of 2 g/L sodium thiosulphate (Unilab Limited, Nairobi, Kenya) and then serially diluted 8-fold. A 100 μL of the 10 − 8 dilution was plated in duplicate on Tryptic Soy Agar (Sigma Aldrich Chemie GmbH, Steinheim, Germany) at 37°C for 18 h. Then, CFUs were enumerated using Colony Counter SC6 plus (Bibby Scientific Limited, Staffordshire, United Kingdom). The following physico-chemical stress concentrations that induced stress without severe growth inhibition were used: 0.8 M NaCl, 5 mM NaOCl and 50 μM H 2 O 2 . For sublethal heat stress, 50°C that has been shown to induce stress without degrading eDNA was used [36].

Procedure of exposure of S. epidermidis biofilm and planktonic cells to the sub-lethal physico-chemical stress conditions
The S. epidermidis biofilm and planktonic cells were challenged with the four sub-lethal physico-chemical stresses as described in [16] with slight modifications. Briefly, 200 μL of S. epidermidis biofilm or planktonic culture diluted to an OD 600 of 1.0 was inoculated into 400 μL of TSB (Sigma Aldrich Chemie GmbH, Steinheim, Germany) supplemented with 700 μL of 0.8 M NaCl (Unilab Limited, Nairobi, Kenya), 5 mM NaOCl (Supersleek, Nairobi, Kenya) or 50 μM H 2 O 2 (RFCL Limited, New Delhi, India), vortexed for 2 min and then incubated at 37°C with shaking at 120 rpm for 60 min. At the 60th min, NaOCl and H 2 O 2 stress-treated S. epidermidis biofilm and planktonic cultures were neutralized by adding 200 μL of 2 g/L sodium thiosulphate (Unilab Limited, Nairobi, Kenya) to the tubes. For S. epidermidis biofilm or planktonic cultures exposed to 0. 8 M NaCl, sterile distilled water was added instead of sodium thiosulphate. The untreated controls for NaCl, NaOCl and H 2 O 2 comprised of 200 μL of S. epidermidis biofilm or planktonic culture inoculated into 400 μL of TSB (Sigma Aldrich Chemie GmbH, Steinheim, Germany) supplemented with 700 μL of sterile distilled water. For heat stress exposure, 200 μL of S. epidermidis biofilm or planktonic culture was inoculated into 400 μL of TSB (Sigma Aldrich Chemie GmbH) supplemented with 700 μL of sterile distilled water and then transferred to a water bath model JSWB-11(T) (JS Research Inc., Gongju-city, Korea) at 50°C for 60 min. At the 60th min, 200 μL of sterile distilled water at 4°C was added to lower the temperature. The untreated control was 200 μL of S. epidermidis biofilm or planktonic culture inoculated into 400 μL of TSB (Sigma Aldrich Chemie GmbH) supplemented with 700 μL of sterile distilled water then transferred to water bath model JSWB-11(T) (JS Research Inc., Gongju-city, Korea) at 25°C for 60 min.

Isolation of eDNA
To minimize variations associated with DNA precipitation, eDNA was obtained directly from the supernatant [37]. The exopolymeric substances were separated from bacterial cells by high-speed centrifugation which does not cause cell lysis [38]. Free and bound eDNA were then obtained from the exopolymeric substances using Tris-Ethylenediaminetetraacetic acid (EDTA) (TE) buffer [39,40].
The eDNA released by the sub-lethal stress-treated S. epidermidis biofilm and planktonic cultures and their respective untreated controls was obtained as described in [20] with few modifications. Briefly, sub-lethal physicochemical stress-treated culture or untreated control was centrifuged at 20,000 rpm at 4°C for 20 min. Then, 1 mL of the supernatant was pipetted into 1 mL of TE buffer (10 mM Tris, 1 mM EDTA, pH 8.0) and then centrifuged at 13,000 rpm for 3 min. Finally, 30 μL of the supernatant was suspended in 100 μL of TE buffer and then stored at − 20°C until further use.

Quantification of eDNA
The eDNA in the supernatant was quantified using Qubit™ dsDNA high sensitivity (HS) assay kit (Invitrogen, Paisley, United Kingdom) and Qubit® 2.

Data analyses
Data obtained were stored in Microsoft Office Excel and analyzed using Prism 5 for windows version 5.03 (GraphPad software, Inc., California, USA). Data normality was verified using Shapiro-Wilk test. Normally and non-normally distributed data were presented as mean (± SEM) and median (25th and 75th percentiles) respectively. Differences in log reduction of CFU/mL between S. epidermidis biofilm and planktonic cells subjected to each of the four commonly recommended physico-chemical disinfectants were compared by paired t-test. Comparisons of log reductions of CFU/mL of S. epidermidis biofilm or planktonic cells among the four commonly recommended physico-chemical disinfectants were performed using repeated measures analysis of variance (ANOVA) with Tukey's post hoc. Differences in eDNA release between S. epidermidis biofilm and planktonic cells treated with each of the four sub-lethal physicochemical stresses were conducted by Wilcoxon matched-pairs signed rank test. Similarly, differences in eDNA yield between sub-lethal physico-chemical stress-treated S. epidermidis biofilm or planktonic cells and their respective untreated controls were analyzed by Wilcoxon matched-pairs signed rank test. Statistical significance was considered at p < 0.05.

Results
Susceptibility patterns of S. epidermidis biofilm and planktonic cells to physico-chemical stress induced by the commonly recommended disinfectants One of the specific aims of the present study was to determine the susceptibility patterns of S. epidermidis biofilm and planktonic cells to physico-chemical stress induced by the four commonly recommended disinfectants (heat, NaCl, NaOCl and H 2 O 2 ) in domestic and human healthcare settings.
S. epidermidis planktonic cells are more susceptible to heat stress than the biofilm cells The S. epidermidis biofilms exposed to heat (60°C) stress for 30 min exhibited a significantly lower log reduction of CFU/mL than the corresponding planktonic cells (p < 0.0001) (Fig. 1a). Similarly, the log reduction of CFU/mL of S. epidermidis biofilms subjected to the same stress for 60 min was statistically lower than the planktonic cells (p < 0.0001) (Fig. 1a). These results indicated that 60°C was less effective against S. epidermidis biofilms compared to the corresponding planktonic cells. Further analyses showed that S. epidermidis biofilm or planktonic cells subjected to 60°C stress for 30 min had a significantly lower log reduction of CFU/mL compared to their respective cells exposed to 60°C stress for 60 min (p < 0.0001) ( Table 1). This indicated that the S. epidermidis biofilm or planktonic cells killing by the heat stress were directly proportional to the exposure duration.
S. epidermidis biofilms are less susceptible to NaCl stress than the planktonic cells When challenged with 1.72 M NaCl for 30 min, S. epidermidis biofilms exhibited a significantly lower log reduction of CFU/mL than the analogous planktonic cells (p < 0.0001) (Fig. 1b). Treatment of S. epidermidis biofilm and planktonic cells with 1.72 M NaCl for 60 min yielded a similar pattern (p < 0. 0001) (Fig. 1b). These results showed that 1.72 M NaCl was less effective against S. epidermidis biofilms than the corresponding planktonic cells. Further analyses revealed that S. epidermidis biofilm or planktonic cells subjected to 1.72 M NaCl stress for 30 min had a significantly lower log reduction of CFU/mL than their respective cells exposed to 1. 72 M NaCl stress for 60 min (p < 0.0001) ( Table 1). These implied that the effectiveness of 1.72 M NaCl on S. epidermidis biofilm or planktonic cells is dependent on the exposure duration.
NaOCl stress is more effective against S. epidermidis planktonic than biofilm cells S. epidermidis biofilms exhibited significantly lower log reduction of CFU/mL when exposed to a solution containing 150 μL of waterguard in 1 L of water for 30 min a b c d  1c). A similar pattern was observed upon exposure of S. epidermidis biofilm and planktonic cells to a solution containing 150 μL waterguard in 1 L of water for 60 min (p < 0.0001) (Fig. 1c). These results indicated that S. epidermidis biofilms were more protected against a solution containing 150 μL of waterguard in 1 L of water than the corresponding planktonic cells. Further analyses showed that S. epidermidis biofilm or planktonic cells subjected to a solution containing 150 μL of waterguard in 1 L of water for 30 min had a significantly lower log reduction of CFU/mL than their respective cells exposed to the same stress for 60 min (p < 0.0001) ( Table 1). These results implied that the effectiveness of NaOCl stress against S. epidermidis biofilm and planktonic cells was proportional to the exposure duration.
H 2 O 2 stress is less effective against S. epidermidis biofilms than planktonic cells S. epidermidis biofilms treated with 1.77 M H 2 O 2 for 30 min had a statistically lower log reduction of CFU/ mL than the analogous planktonic cells (p < 0.0001) (Fig.  1d). Similarly, S. epidermidis biofilms challenged with 1. 77 M H 2 O 2 for 60 min showed a significantly lower log reduction than the corresponding planktonic cells (p < 0.0001) (Fig. 1d) Comparison of the effectiveness of physico-chemical stress induced by the commonly recommended disinfectants against S. epidermidis biofilm and planktonic cells The present study also compared the effectiveness of physico-chemical stress induced by the four commonly recommended disinfectants against S. epidermidis biofilm or planktonic cells. Repeated measures ANOVA showed that the log reductions of CFU/mL of S. epidermidis biofilm cells differed significantly among the four commonly recommended disinfectants upon exposure for 30 or 60 min (p < 0.0001). Tukey's post hoc revealed that when S. epidermidis biofilm cells are exposed to the four commonly recommended disinfectants for 30 min, the log reduction of CFU/mL was significantly highest for 1.72 M NaCl followed by 1.77 M H 2 O 2 , 60°C and a solution containing 150 μL of waterguard in 1 L of water in that order (Fig. 2a). Conversely, when S. epidermidis biofilms were subjected to the four commonly recommended disinfectants for 60 min, Tukey's post hoc showed that the log reduction of CFU/mL was significantly highest for 1.77 M H 2 O 2 , followed by 1.72 M NaCl, 60°C and a solution containing 150 μL of waterguard in 1 L of water in that order (Fig. 2a). These results indicated that susceptibilities of S. epidermidis biofilm cells to physico-chemical stress induced by the four commonly recommended disinfectants for 30 min was not dependent on the diffusion rate (molecular weight) (NaCl > H 2 O 2 > heat > NaOCl). However, susceptibility of S. epidermidis biofilm cells to physicochemical stress induced by the four commonly recommended disinfectants for 60 min was dependent on the diffusion rate of the disinfectants (H 2 O 2 > NaCl > heat > NaOCl). Furthermore, repeated measures ANOVA revealed that the log reductions of CFU/mL of S. epidermidis planktonic cells differed significantly among the four  (Fig. 2b). These results implied that the susceptibilities of S. epidermidis planktonic cells to the four commonly recommended disinfectants for 30 or 60 min were dependent on the diffusion rate of the disinfectants.
Effects of sub-lethal physico-chemical stress exposure on eDNA release by S. epidermidis biofilm and planktonic cells To effectively control, manage and eradicate bacterial biofilms in domestic and healthcare settings it is necessary to understand the mechanisms that mediate their survival against commonly reccommded physico-chemical disinfectants. Consequently, a second specific aim of the present study was to evaluate the eDNA release as a potential mechanism that underlies S. epidermidis biofilms resistance to physico-chemical stress exposure.

Effect of heat stress on eDNA release by S. epidermidis biofilm and planktonic cells
The percentage change in eDNA yield by 50°C-treated S. epidermidis biofilms was slightly more than the corresponding planktonic cells albeit not statistically different (p = 0.4697) (Fig. 3a). Further analysis showed that 50°C -treated S. epidermidis biofilm cells released significantly increased eDNA than the respective untreated controls (p = 0.0098) ( Table 2). On the contrary, 50°Ctreated planktonic cells yielded more eDNA than the untreated controls but the difference was not statistically significant (p = 0.7910) ( Table 2).
Impact of NaCl stress on eDNA release by S. epidermidis biofilm and planktonic cells S. epidermidis planktonic cells exposed to 0.8 M NaCl stimulated higher percentage change in eDNA yield than the analogous biofilm cells although the difference was not statistically significant (p = 0.8501) (Fig. 3b). Further analysis showed that 0.8 M NaCl-treated S. epidermidis biofilm cells released less eDNA than the untreated controls although not statistically different (p = 0.3804) ( Table 2). However, 0.8 M NaCl-treated S. epidermidis planktonic cells yielded more eDNA than the untreated controls but the difference was not statistically significant (p = 0.6772) ( Table 2). These results showed that osmotic stress does not significantly affect eDNA release by S. epidermidis biofilm and planktonic cells.

Effects of oxidative stresses on eDNA release by S. epidermidis biofilm and planktonic cells
Significantly increased percentage change in eDNA was released by S. epidermidis biofilms exposed to 5 mM NaOCl compared to the corresponding planktonic cells (p = 0.0015) (Fig. 3c). Further analysis showed that 5 mM NaOCl-treated S. epidermidis biofilm cells released significantly more eDNA than the untreated controls (p = 0. 0005) ( Table 2). On the contrary, 5 mM NaOCl-treated S. epidermidis planktonic cells produced slightly more eDNA than the untreated controls but the difference was not significant (p = 0.9097) ( Table 2) When challenged with 50 μM H 2 O 2 stress for 60 min, the S. epidermidis biofilm cells exhibited significantly increased percentage change in eDNA than the analogous planktonic cells (p = 0.0210) (Fig. 3d). Further analyses revealed that 50 μM H 2 O 2 -treated S. epidermidis biofilm cells produced significantly more eDNA than the untreated controls (p = 0.0005) ( Table 2). In contrast, the eDNA yield between 50 μM H 2 O 2 -treated S. epidermidis planktonic cells and the respective untreated controls was not significantly different (p = 0.7910) ( Table 2). Taken together, these results indicated that S. epidermidis biofilm cells enhance eDNA release in response to 5 mM NaOCl and 50 μM H 2 O 2 stress exposure.

Discussion
Susceptibility patterns of S. epidermidis biofilm and planktonic cells to physico-chemical stress induced by the commonly recommended disinfectants Bacterial biofilms are most frequently encountered in domestic and healthcare environments [41,42]  prevention of acquisition, spread and establishment of biofilm-forming bacteria such as S. epidermidis using effective disinfection guidelines [12] in these environments is necessary. Therefore, the present study evaluated the effectiveness of four commonly recommended physico-chemical disinfectants against S. epidermidis biofilm and planktonic cells.
The present study showed that the commonly recommended heat stress was less effective against S. epidermidis biofilms compared to the corresponding planktonic forms. The present findings are in agreement with previous reports on different bacterial species such as M. bovis [9], S. enterica [11] and a fungus, Cryptococcus neoformans [43]. However, the previous studies reported on different bacterial species such as M. bovis which are cell wall-less, S. enterica that overproduce protective cellulosic polymer and a fungal biofilm hence may not inform bacterial biofilm response to heat stress. The observed increased heat stress resistance of S. epidermidis biofilms compared to the corresponding planktonic cells could probably be explained in two ways. First, since bacterial biofilms overexpress heat stress-related genes [44] resulting in heat shock proteins that consume excess energy in form of adenosine triphosphate [45], it is likely that the S. epidermidis biofilm cells overexpressed heat stress-related genes to protect them against the deleterious effects of heat stress. Second, an increase in temperature has been shown to switch the staphylococcal biofilm cells fatty acid profile such that the anteiso-C19 fatty acids known to have high melting point rises, leading to decreased membrane fluidity resulting in increased resistance to heat stress [32].
Furthermore, the present study showed that biofilm forms of S. epidermidis are more resistant to the commonly recommended NaCl stress than the analogous planktonic cells which is consistent with a previous report on V. cholerae O1 by Wai et al. [10]. However, the previous study used V. cholerae a bacterium highly adapted to high salinity [46] and utilized high salt concentration (2.5 M NaCl) that may not inform routine bacterial biofilm disinfection. The observed increased resistance of S. epidermidis biofilm cells to osmotic stress could possibly be due to the osmotic stress-enhanced production of exopolysaccharides and proteins that formed a water-retaining layer around S. epidermidis biofilm cells thus protecting the cells against dehydration [23,47]. An alternative explanation could be linked to the previous observation that osmotic stress enhances quorum sensing in bacterial biofilms [48] which correspondingly increases resistance against osmotic stress [49] by the S. epidermidis biofilm cells.
The present study also showed that the commonly recommended NaOCl stress is less effective against S. epidermidis biofilms compared to the analogous planktonic cells. The present finding is in consonance with several previous reports on different bacterial species that were either overproducing protective pellicles [11], were subjected to high [13] or low NaOCl concentrations [12], were overgrown for six days [16] or had protective mycolic acid rich membranes [15]. Thus, the previous reports may not inform general bacterial biofilm response to NaOCl stress. The observed increased S. epidermidis biofilm resistance could partly be due to the reaction of NaOCl with the ECM components such as proteins and polysaccharides and/or the slow diffusion across ECM barrier thus reducing the effect of NaOCl on most S. epidermidis biofilm cells [16].
The present study also demonstrated that S. epidermidis biofilms are more protected against the commonly recommended H 2 O 2 stress than the analogous planktonic cells. The present finding concurs with previous reports on S. epidermidis [50], V. cholerae [10] and B. cenocepacia [12] exposed to low concentrations of H 2 O 2 that are not routinely used for disinfection. In addition, Glynn et al. [50] reported on the effect of H 2 O 2 on S. epidermidis biofilm formation using a semiquantitative approach which may not present a clear picture of S. epidermidis biofilms response to H 2 O 2 stress. The observed increased resistance of S. epidermidis biofilms could possibly be attributed to neutralization of H 2 O 2 by the ECM components such as polysaccharides and proteins and/or overexpression of catalase-producing genes resulting in overproduction of catalase enzymes that decompose the H 2 O 2 [12] thereby reducing its effects on most of the inner S. epidermidis biofilm cells.
Taken together, the results presented in the present study showed that the susceptibility of S. epidermidis biofilm and planktonic cells was proportional to the duration of exposure to each of the four commonly recommended disinfectants. Generally, a disinfectant producing a log reduction unit above three (99.9% bacterial reduction) is considered effective against a bacterial biofilm [33]. However, in the present study, S. epidermidis biofilm cells exhibited log reduction units below three (< 3) when subjected to the four commonly recommended disinfectants. This implied that the four physico-chemical stresses commonly recommended for disinfection in domestic and human healthcare settings were ineffective against the S. epidermidis biofilm cells hence creating a healthcare concern.
Reaction-diffusion inhibition mechanism inadequately explains the increased S. epidermidis biofilms resistance to physico-chemical stresses commonly recommended for disinfection Bacterial biofilm resistance to osmotic and oxidative stress exposure is mostly attributed to the stress agent's reaction with and/or slow diffusion across the ECM (reaction-diffusion inhibition mechanism) [16,51]. The movement of the stress agents probably occur via waterfilled channels on the bacterial biofilm's ECM [2]. The observation that the susceptibility of S. epidermidis biofilm cells to the commonly recommended disinfectants for 60 min depended on the diffusion rate (molecular weight) (H 2 O 2 > NaCl > heat > NaOCl) appear to support the reaction-diffusion inhibition mechanism. It has been shown that NaOCl (with slowest diffusion rate) diffuses across the ECM in about 50 min [16]. Accordingly, all the four disinfectants should have crossed the ECM within the 60 min of exposure and killed an equal number of S. epidermidis biofilm and planktonic cells. Taking into account the observation that more S. epidermidis planktonic cells were killed compared to the corresponding biofilm cells at 60 min of exposure and that susceptibilities of the biofilm cells to the commonly recommended disinfectants for 30 min did not correspond to the diffusion rate, it is reasonable to surmise that the reaction-diffusion inhibition mechanism inadequately explains the increased resistance of S. epidermidis biofilm cells to the physicochemical stress induced by the commonly recommended disinfectants. This lends credence to the existence of complementary mechanism(s) of resistance against physicochemical stress exposure in S. epidermidis biofilms such as eDNA release [19] and upregulation of biofilm-specific stress resistance genes [3].
Role of eDNA in S. epidermidis biofilms resistance to physico-chemical stress exposure The reaction-diffusion inhibition mechanism did not fully account for the increased S. epidermidis biofilms resistance to physico-chemical stress induced by the commonly recommended disinfectants relative to the planktonic forms. In the recent years, some studies have linked eDNA release with survival of bacterial [20][21][22] and fungal biofilms [23] against antibiotics and H 2 O 2 respectively. Thus, the present study evaluated eDNA release as a potential complementary mechanism underlying S. epidermidis biofilms resistance to physico-chemical exposure.
The present study did not reveal any significant difference in eDNA release between the S. epidermidis biofilm and planktonic forms exposed to sub-lethal heat stress. However, the present study showed that the sub-lethal heat stress-treated S. epidermidis biofilms significantly enhanced eDNA release than the untreated controls. For S. epidermidis planktonic cells, no difference was observed in eDNA release between the treated cells and untreated controls. This set of results can be interpreted in two ways. On the one hand, the finding that eDNA released by S. epidermidis biofilm and planktonic cells were not statistically different could mean that eDNA release was slightly stimulated by a rise in temperature via active secretion or controlled cell lysis [19] in both the biofilm and planktonic cells and not necessarily as a resistance mechanism against the heat stress exposure. On the other hand, the observation that unlike the planktonic forms, the biofilm forms of S. epidermidis subjected to heat stress released significantly increased eDNA than the untreated controls strongly suggested a central role of eDNA in S. epidermidis biofilm cells resistance against the heat stress.
Further, there was no difference between eDNA released by S. epidermidis biofilm and planktonic cells subjected to sub-lethal NaCl stress. Unexpectedly, the NaCl stress treated S. epidermidis biofilms released less eDNA than the untreated controls. A possible explanation for this unexpected observation could be inferred from previous studies which have shown that NaCl stress stimulates increased exopolysaccharide production in the ECM [23,47] which might have resulted in strong bond formation between the eDNA and polyssacharides [52] rendering it largely inaccessible for quantification. Considered together, the present findings show that NaCl stress does not affect eDNA release by both S. epidermidis biofilm and planktonic cells. The present finding concurs with a previous report on C. albicans biofilm subjected to 2 M NaCl [23]. However, C. albicans is a fungus hence may not inform bacterial biofilms response to osmotic stress. Further, the present findings are consistent with a previous report which showed that high salt concentration affects exopolysaccharide release by Halomonas variabilis and Planococcus rifietoensis [47]. However, the previous report only evaluated one component of the ECM that is the exopolysaccharide. The observation that neither S. epidermidis biofilm and planktonic cells nor the respective treated and untreated controls showed a significant difference in eDNA release suggested that eDNA does not play a role in osmotic stress resistance. In support of this interpretation, a study showed that osmotic stress does not produce eDNA richer ECM instead, exopolysaccharide and protein yield is enhanced to form a water-retaining layer around the biofilm cells [23] thus protecting the cells against dehydration. Moreover, it has been shown that autolysin (atl) gene, which is often associated with eDNA release is not affected by osmotic stresses [53] further implying that osmotic stress does not induce eDNA release.
Further, the present study showed that S. epidermidis biofilms enhance eDNA release in response to sub-lethal oxidative (NaOCl and H 2 O 2 ) stress exposure. This is consistent with previous reports, which showed that H 2 O 2 stress exposure induced eDNA release by Streptococcus gordonii [34,37] and C. albicans biofilm [23]. However, the previous reports only focused on planktonic forms of S. gordonii and a fungus, C. albicans hence may not inform eDNA release by bacterial biofilms in response to H 2 O 2 .
Considered together, oxidative stresses damage genomic DNA triggering eDNA release by a subpopulation of bacterial cells [37]. Why oxidative stress-treated S. epidermidis biofilms released more eDNA than the corresponding planktonic cells is still not clear. One possible explanation could be related to the extracellular DNases released alongside eDNA in the following ways. First, unlike the planktonic cells, bacterial biofilm cells form small-protected pockets [21] that could be protecting most eDNA from DNases degradation. Second, bacterial biofilms eDNA is mostly bound to ECM [52] hence may not be easily accessible to the DNases. Third, bacterial biofilms produce relatively fewer DNases than the planktonic cells [54] thus minimizing the eDNA degradation. Fourth, bacterial biofilms induce release of proteolytic exoenzymes that inactivate the DNases locally [55]. Taken together, the explanations above suggest that eDNA and DNases release by bacterial biofilm cells are highly regulated processes. This implies that bacterial planktonic cells majorly release eDNA to be degraded for nutrients whereas bacterial biofilm cells induce eDNA release both as a nutrient source and for protection against the lethal effects of oxidative stresses.
An alternative explanation for the relatively increased eDNA release by S. epidermidis biofilm cells could be related to the high production of catalase enzyme [56] and ECM [12,57] which neutralizes and reacts with the H 2 O 2 respectively. It has also been shown that NaOCl reacts with the organic components of the ECM thereby reducing its concentration [16]. Thus, the relatively increased eDNA release by S. epidermidis biofilm cells despite the potential exposure to lower concentrations of NaOCl and H 2 O 2 strongly suggested an integral role of eDNA in the biofilm cells resistance to the oxidative stresses.

Limitations
A potential limitation of the present study is that although DNase appears to provide a more plausible explanation for the increased eDNA release by S. epidermidis biofilms challenged with the sub-lethal oxidative stresses, the presence of DNase was not quantitatively measured. Further studies with DNase (+) controls and treatment groups may be necessary to confirm the direct link between eDNA and bacterial biofilm resistance to physico-chemical stresses.

Conclusions
In summary, S. epidermidis biofilms were less susceptible to physico-chemical stress induced by the four commonly recommended disinfectants than the corresponding planktonic forms. Thus, there is need to review the current disinfection guidelines to improve S. epidermidis biofilm disinfection efficiency. Further, S. epidermidis biofilms enhanced eDNA release in response to sub-lethal heat and oxidative stress exposure than the corresponding planktonic cells, suggesting a role of eDNA in the bacterial biofilms resistance to physicochemical stress exposure. Therefore, eDNA may be a potential target for novel anti-bacterial biofilm control and eradication strategies.

Funding
This study was supported by a grant from National Commission for Science and Technology (NACOSTI), Kenya. The funding body had no role in study design, data collection, analysis and interpretation of data and in writing of the manuscript.

Availability of data and materials
All data generated or analyzed during this study are presented within this manuscript. All materials used in this study including raw data shall be availed upon reasonable request.
Authors' contributions COO, PAO, DMO, and DOO conceived and designed the experiments. COO collected samples, conducted the experiments and data analysis. COO, PAO, DMO, and DOO interpreted the data and developed the manuscript arguments, reviewed and revised it. All authors read and approved the final manuscript.
Ethics approval and consent to participate The use of samples from human participants and all the experimental protocols were reviewed and approved by the Maseno University Ethics Review Committee (Reference number: MSU/DRPI/MUERC/000187/15). Further, permission to recruit outpatients at the Kisumu County Hospital in this study was granted by the Kisumu County Hospital, Kenya (Research file number: 74-2015). Written informed consents were obtained from all the participants for sample collection and further analyses before they participated in this study.