Extracellular DNA facilitates bacterial adhesion during Burkholderia pseudomallei biofilm formation

The biofilm-forming ability of Burkholderia pseudomallei is crucial for its survival in unsuitable environments and is correlated with antibiotic resistance and relapsing cases of melioidosis. Extracellular DNA (eDNA) is an essential component for biofilm development and maturation in many bacteria. The aim of this study was to investigate the eDNA released by B. pseudomallei during biofilm formation using DNase treatment. The extent of biofilm formation and quantity of eDNA were assessed by crystal-violet staining and fluorescent dye-based quantification, respectively, and visualized by confocal laser scanning microscopy (CLSM). Variation in B. pseudomallei biofilm formation and eDNA quantity was demonstrated among isolates. CLSM images of biofilms stained with FITC-ConA (biofilm) and TOTO-3 (eDNA) revealed the localization of eDNA in the biofilm matrix. A positive correlation of biofilm biomass with quantity of eDNA during the 2-day biofilm-formation observation period was found. The increasing eDNA quantity over time, despite constant living/dead ratios of bacterial cells during the experiment suggests that eDNA is delivered from living bacterial cells. CLSM images demonstrated that depletion of eDNA by DNase I significantly lessened bacterial attachment (if DNase added at 0 h) and biofilm developing stages (if added at 24 h) but had no effect on mature biofilm (if added at 45 h). Collectively, our results reveal that eDNA is released from living B. pseudomallei and is correlated with biofilm formation. It was also apparent that eDNA is essential during bacterial cell attachment and biofilm-forming steps. The depletion of eDNA by DNase may provide an option for the prevention or dispersal of B. pseudomallei biofilm.


Introduction
Biofilm provides shelter for various pathogens and its formation is clearly essential for microbial survival in diverse environments, potentially leading to increased virulence [1,2]. The a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 isolated for this study but had been collected as part of a previous study of the epidemiology of B. pseudomallei in Khon Kaen Province. Patients cannot be identified: the isolates were anonymous and de-identified when we received them. Approval for the study was given by the Khon Kaen University Ethics Committee for Human Research (HE490324).

Bacterial strains and growth conditions
B. pseudomallei isolates used in this study are listed in Table 1. Each bacterial isolate from glycerol stock at -80˚C was grown on Ashdown's agar and incubated at 37˚C for 48 h. A single colony of B. pseudomallei was inoculated into 3 mL of Luria-Bertani (LB) broth containing appropriate antibiotics and incubated at 37˚C with shaking (200 rpm) for 18 h. A 2% inoculum was added into 25 mL fresh LB broth and further incubated until OD 540 = 0.8-0.9 to provide bacterial starter culture [27,31,32].

Quantification of biofilm and eDNA in biofilms in 96-well plates
Two-day biofilm formation of B. pseudomallei was determined by crystal violet staining in 96-well microtiter plates as previously described by Taweechaisupapong et al, 2005 and Kunyanee et al, 2016 [27,31]. Two hundred microliters of starter culture was dispensed, in replicates of eight, into each well of a 96-well flat-bottomed polystyrene plate (Nunclon #167008, Thermo Scientific, Denmark) and incubated at 37˚C for 3 h to allow bacterial adhesion. Negative controls containing no B. pseudomallei were included. Following incubation, non-adhering bacteria were removed, then fresh LB medium was added and plates further incubated for another 21 h. Non-adhering bacteria were again removed before the biofilms were carefully washed with sterile distilled water and the wells refilled with fresh LB medium. After incubation for an additional 24 h, biofilms were carefully washed three times with sterile distilled water. The 2-day biofilm in each well was fixed with 99% methanol for 15 min and air dried. The biofilms were stained with 2% w/v crystal violet for 5 min. The excess stain was removed with running tap water. After air-drying, adherent crystal violet stain was dissolved C17 Biofilm complemented strain of M10 C17 was constructed by restoring the bpsl0618 gene in M10. [27] in 200 μL 33% (v/v) glacial acetic acid and the optical density at 620 nm of each sample was measured by a microplate reader (TECAN Safire, Port Melbourne, Australia). The quantity of eDNA associated with B. pseudomallei biofilm was examined in 96-well black plates (SPL Life Sciences, Korea) in triplicates concurrent with the biofilm quantification. The 2-day biofilm culture was rinsed three times with sterile distilled water. eDNA in each well was mixed with 200 μL of freshly prepared QuantiFluor dsDNA dye in TE buffer for 5 min (QuantiFluor dsDNA System, Promega, Madison, WI, USA) before measuring the fluorescence intensity (excitation 504 nm/emission 531 nm) using a fluorometer (Varioskan Flash Multimode Reader, Singapore) with SkanIt Software 2.4.3 RE for Varioskan Flash. Lambda DNA (QuantiFluor dsDNA System) was used to generate a standard curve for each run.

Confocal laser scanning microscope observation
Burkholderia pseudomallei biofilm architecture and quantity of eDNA associated with the biofilm were evaluated on glass coverslips immersed in bacterial culture in 24-well plates (Costar #3524, Corning, NY, USA) using an Amsterdam Active Attachment (AAA) model slightly modified from previous descriptions [27,[33][34][35]. In brief, the set of 12 mm-diameter round glass coverslips attached to the sterile AAA model's lid were autoclaved. The glass coverslips held on the lid were immersed into 1 mL of bacterial starter culture in each well of a 24-well plate and incubated at 37˚C for 3 h. The lid was then transferred to a new 24-well plate containing fresh LB medium and incubated for another 21 h. The adhered bacteria on the coverslips were subsequently washed with sterile distilled water before being submerged into fresh LB medium and incubated for another 24 h to produce a 2-day biofilm. Three-hour, 24-h and 2-day biofilms on the glass coverslips were rinsed three times with sterile distilled water prior to staining with 50 μg/mL fluorescein isothiocyanate-concanavalin A (FITC-Con A) (Sigma, Saint Louis, Missouri, USA), which binds extracellular polysaccharide (representing biofilms) and 2 μM TOTO-3 (Thermo fisher Scientific, Oregon, USA), which stains eDNA present within biofilm, for 20 min according to the manufacturers' instructions. Bacterial viability within 2-day biofilms was examined after staining with 3.34 μM SYTO 9 (live cells) and 5 μg/ mL propidium iodide (PI) (dead cells) (Invitrogen, Thermo fisher Scientific, Oregon, USA) for 15 min. Stained biofilms were subsequently fixed with 2.5% glutaraldehyde for 3 h before being washed three times with sterile water and air-dried. The structure of the stained biofilm, eDNA present and bacterial viability on coverslips were visualized by confocal laser scanning microscope (CLSM, LSM 800, Carl Zeiss, Jena, Germany). The biofilm intensity was analyzed from z-stack processing using Zen blue software [13,27]. Biomass of adherent cells and eDNA quantity were calculated with the COMSTAT computer program [36]. The bacterial viability is presented as live/dead ratio.

DNase I treatment of B. pseudomallei biofilms and addition of exogenous DNA to biofilms
The role of eDNA in B. pseudomallei biofilm formation was investigated by depletion of eDNA using DNase I (Roche, Mannheim, Germany). DNase at a final concentration of 0.01, 0.1 or 1 U/mL was used, following Kim et al [37] with slight modification. DNase I was added into bacterial cultures at different time points representing various steps of biofilm development: 0 h (initial attachment), 24 h (biofilm formation) and 45 h (biofilm maturation). The selected DNase I concentrations were constantly maintained in the medium for up to 48 h at 37˚C before the 2-day biofilms and eDNA were quantified by crystal violet staining and the QuantiFluor dsDNA System kit, respectively, and images obtained using CLSM. Untreated controls using LB media were also used for direct comparison.
Chromosomal DNA of B. pseudomallei H777 was extracted using a popular method described by Sambrook et al [38]. The DNA pellet was re-suspended with Tris-HCl buffer. Exogenous DNA, namely chromosomal DNA of B. pseudomallei or salmon-sperm DNA (Sigma, St Louis, MO, USA) (0.1 μg/mL), was added either simultaneously to the starting culture or the DNase-treated biofilm after removal of DNase by washing twice with sterile distilled water. DNase I enzyme, DNase I buffer and Tris-HCl buffer (diluent of the exogenous DNA) were also used as controls.

Statistical analysis
Statistical analyses were performed using SPSS software, version 23 (SPSS Inc., Chicago, IL, USA). Biofilm formation and eDNA quantities produced by clinical and environmental isolates were analyzed using the nonparametric Kruskal-Wallis test followed by Dunn's post-hoc test for comparison between pairs. Pearson correlation analysis of equal variance data was used to determine the relationship between biofilm biomass and eDNA quantity. Biofilm, eDNA and live/dead ratios were analyzed for statistical significance using the one-way ANOVA followed by Tukey's post-hoc test, or the Games-Howell post-hoc test to correct for variance heterogeneity. Variance heterogeneity was assessed by Levene's tests. The levels required for statistical significance were � p < 0.05 and �� p < 0.001.

Variations of biofilm formation and eDNA quantities of B. pseudomallei
The static 2-day biofilm and eDNA associated with that biofilm of 10 different B. pseudomallei isolates (Table 1), stained with crystal violet and QuantiFluor dsDNA reagent, demonstrated the variation of biofilm formation ( Fig 1A) and eDNA quantity ( Fig 1B). Included were 4 clinical isolates (B1, L1, P1 and H777), 4 environmental isolates (3E, 8E, 23E and ST39), the biofilm mutant of H777 (M10) and the biofilm complemented strain of M10 (C17) in LB medium. Notably, P1 (isolated from pus) exhibited the greatest ability to form biofilm followed by L1 (from lung) and H777 (from blood) whilst L1 produced the most eDNA. Strains isolated from lung (L1) and pus (P1) produced significantly more biofilm than did blood isolates (B1 and H777) (p < 0.05) and all environmental isolates (p < 0.001).

eDNA localized in B. pseudomallei biofilm and correlated with biofilm formation
The clinical B. pseudomallei isolates (L1, P1, H777); the biofilm mutant (M10) and the biofilm complement strains (C17) were chosen for further investigation. We examined the biofilm architecture and eDNA quantity on glass coverslips after staining with FITC-ConA and TOTO-3 using CLSM. FITC-ConA stains biofilm green, whereas TOTO-3 stains eDNA red (Fig 2A). We observed an increase in adhered cells as well as eDNA signal intensities on the glass coverslips over time as biofilm developed. Moreover, the CLSM images revealed morphological differences of biofilm and eDNA signal intensities among isolates (Fig 2A) which are in line with our initial findings of variation in biofilm biomass and eDNA production (Fig 1). The biofilm complement strain, C17, produced a flattened biofilm architecture which covered the coverslips but the wildtype biofilm phenotype was not fully restored. The biofilm mutant failed to form the tower structures of mature biofilm. CLSM images clearly demonstrated the localization of eDNA within the biofilm matrix of B. pseudomallei. The highest red fluorescence of eDNA was seen in the wild-type isolates (L1, P1 and H777), especially in the 48 h biofilm.  To further investigate the hypothesis that eDNA is important for development of biofilm, we also analyzed the Z-stack confocal images with the COMSTAT image-analysis software. This revealed significant increase of biofilm biomass and of eDNA concentration in biofilm over time in several strains: the biofilm wild-type B. pseudomallei L1, P1 and H777 (Fig 2B and  2C). Pearson correlation analysis demonstrated a significant positive correlation in these strains between eDNA production in biofilm and biofilm biomass (p < 0.001) ( Table 2). Higher magnification CLSM 2D images (100×) clearly demonstrated diffuse eDNA surrounding bacterial biofilms, emphasizing the extracellular location of this DNA (Fig 2D).
Biofilms of strains L1, P1, H777, M10 and C17 varied greatly in thickness according to the serial-section gallery of the 2-day FITC-ConA/TOTO-3-stained biofilms (Fig 3A and 3B). The thickest biofilms and the greatest quantity of eDNA were produced by strain P1, followed by L1, H777 and C17. In addition, the relative slice gallery revealed eDNA particularly at the base of the biofilm, as indicated by red fluorescence on the bottom of the images, while the upper layer displayed a predominance of green cells.

eDNA released from living B. pseudomallei cells
We determined whether eDNA was released during biofilm formation primarily from living or dead bacterial cells. To do this, live/dead staining of the 3, 24 and 48 h B. pseudomallei H777 biofilm was employed. CLSM imaging and COMSTAT analysis revealed a constant live/dead ratio during the observation period (Fig 4A and 4B). Given that eDNA accumulates through time in biofilm, this implies the liberation of eDNA from living bacterial cells.

DNase I reduces B. pseudomallei biofilm
The impact of DNase I on biofilm formation by B. pseudomallei at different time points was investigated in microtiter plates followed by crystal-violet staining for biofilm and fluorescence intensity for the eDNA associated with the biofilm. DNase I (0.01, 0.1, and 1 U/mL) was added to the B. pseudomallei biofilm culture in LB at 0, 24 and 45 h, and maintained in the culture for up to 48 h. Biofilms of B. pseudomallei strains L1, P1 and H777 were considerably reduced when DNase I had been present in the bacterial culture since initial adhesion (0 h) and biofilm formation (24 h) stages compared with the untreated controls ( � p < 0.05 and �� p < 0.001) (Fig 5). These results emphasize the role of eDNA in initial adhesion and biofilm formation stages. When DNase I was added at the 45 h preformed-biofilm stage, biofilm was reduced in the L1 and P1 strains, but not in the H777 strain (Fig 5). DNase I noticeably lowered eDNA concentrations in biofilm if the enzyme was added into the starting inoculum (0 h). The remaining eDNA at the 24 h and 45 h preformed biofilm stages may have been released from dispersed bacterial cells in the wells.
The effects of 0.01 U/mL DNase I on initial attachment and biofilm-formation stages were further witnessed using CLSM after staining with FITC-ConA and TOTO-3. The CLSM images showed that continuous presence of the enzyme from the starting inoculum (0 h) or 24 h (Fig 6) until the end of the experiment at 48 h clearly diminished biofilm formation of all tested B. pseudomallei strains relative to controls (Fig 6A). Biofilms treated with DNase I eDNA facilitate early stages of Burkholderia pseudomallei biofilm formation appeared thinner than in untreated controls. Statistical analysis of images showing DNase Itreated biofilm and eDNA at 0 and 24 h using the COMSTAT software confirmed the suppressive effect of the enzyme on initial adhesion (0 h) and biofilm formation (24 h) compared to untreated controls ( � p < 0.05) (Fig 6B). This suggests that DNase I downgrades B. pseudomallei biofilm development, in particular during initial biofilm formation.

Exogenous chromosomal DNA did not alter B. pseudomallei biofilm formation pattern
Given the association of eDNA with the biofilm matrix and the sensitivity of biofilm development to DNase, we further questioned whether exogenous DNA may raise B. pseudomallei biomass. Therefore exogenous chromosomal DNA of B. pseudomallei and salmon sperm DNA were added into biofilm cultures of B. pseudomallei H777, either at 0 h or after eDNA depletion using DNase I followed by washing steps to remove DNase I. The results demonstrated that exogenous DNA could not alter biofilm-formation ability of B. pseudomallei either in normal conditions or after DNase treatment (Fig 7A and 7B). Notably, structure of DNase-treated biofilm was spontaneously rebuilt without any addition of exogenous DNA after removal of  Fig  3A).
https://doi.org/10.1371/journal.pone.0213288.g003 the enzyme. These data may indicate that eDNA released from B. pseudomallei cells present at the time is adequate for biofilm reconstruction.

Discussion
Growing as a biofilm contributes not only to the survival of bacterial cells in unfavorable environments but also shields them from antimicrobial agents and host immune defenses. Numerous attempts have therefore been made to find effective means of biofilm dispersal or to prevent biofilm formation, thus increasing susceptibility of bacterial pathogens to antimicrobial agents and host defenses. For that reason, biofilm composition, mechanisms of formation and structure need to be understood. One of the key biofilm components is eDNA, which essential for biofilm development during bacterial adhesion and provides structural support for biofilm formation in both Gram-negative and Gram-positive bacteria [5,7,10,37,[39][40][41]. Biofilm of B. pseudomallei is known to act as a barrier for antimicrobial agents and is associated with relapsing melioidosis [21,24]. It has not yet been established whether eDNA plays a role during B. pseudomallei biofilm formation. The knowledge from this study indicates that eDNA can be a primary target for eradication strategies against B. pseudomallei biofilm.
In this study, we demonstrated variations in biofilm formation and eDNA quantities among clinical and environmental B. pseudomallei isolates by staining with crystal violet and the QuantiFluor dsDNA System. Similar variation in biofilm-forming capacity among 50 strains of B. pseudomallei was previously reported by Taweechaisupapong [31]. Clinical isolates of B. pseudomallei from lung and pus showed greater ability to form biofilm, in line with the higher eDNA amounts produced. However, eDNA levels were unpredictable in strains less capable of biofilm formation, including the biofilm mutant, M10. CLSM was performed for further quantitative and qualitative investigation of biofilm biomass and eDNA. CLSM micrographs of B. pseudomallei biofilm on coverslips stained with FITC-ConA (biofilm) and TOTO-3 (eDNA) demonstrated the different B. pseudomallei biofilm architectures among isolates. The eDNA was shown to be present primarily at the base of the matrix of B. pseudomallei biofilm. These findings hint at the participation of eDNA in initial steps of biofilm formation. Visualization and quantification of biofilms and eDNA using CLSM when isolates were grown on coverslips for visualization and quantitative analysis using CLSM gave more consistent data than did crystal violet staining and fluorescence detection in wells. CLSM images also provided direct evidence of eDNA in biofilm structure, indicating that eDNA is essential for biofilm formation. In addition, we have provided the first report of significant positive correlation of B. pseudomallei biofilm biomass with eDNA quantity, emphasizing the contribution of eDNA to biofilm formation.
The drastic increase of eDNA quantity through time (Fig 2C) despite the constant and high live/dead ratios of bacterial cells in 3, 24 and 48 h biofilm (Fig 4) points toward the liberation of eDNA from living B. pseudomallei. This is consistent with a previous report that demonstrated accumulation of B. pseudomallei eDNA on murine gastric tissues without bacterial cell lysis [28]. Mechanisms of liberation of eDNA from B. pseudomallei cells remain to be elucidated: knowledge of these might help in devising strategies to counteract biofilm formation by this pathogen.
CSLM images and the COMSTAT analysis demonstrated that depletion of eDNA by DNase I treatment considerably reduced biofilm formation, depending on the biofilm stage at which DNase treatment was started. The presence of DNase in biofilm culture from 0 h or 24 h, with maintenance of DNase I in the cultures up to the end of the experiment at 48 h, significantly reduced B. pseudomallei biofilm. This emphasizes the importance of eDNA as an intercellular connector during initial attachment and early biofilm development. However, DNase intervention after biofilm has reached the mature stage of development led to reduction of eDNA in only 2 of 3 clinical B. pseudomallei isolates. It is possible that the DNase could not gain access to eDNA in the mature biofilm matrix [10,42]. Our findings are consistent with previous reports that demonstrated the essential role of eDNA during the initial step of biofilm formation by other bacteria. Harmsen and colleagues demonstrated, by use of DNase I, that eDNA was involved in the initial attachment of Listeria monocytogenes biofilm [15]. Kim and colleagues demonstrated DNase I can inhibit the initial step of biofilm formation of Campylobacter strains isolated from raw chicken [37]. Several reports have demonstrated that DNase I can inhibit the biofilm formation of Pseudomonas aeruginosa [10,43], Escherichia coli, Staphylococcus aureus [44], Campylobacter sp. [37], Xanthomonas citri subsp. citri [8], Neisseria meningitidis [15] and S. epidermidis [45]. Svensson and colleagues demonstrated DNase I treatment decreased the biofilm formation and stress tolerance of C. jejuni [46]. Mann and colleagues demonstrated that DNase I inhibits biofilm formation and biofilm maturation in Staphylococcus aureus [13] In addition, Kim and colleagues demonstrated that DNase I significantly inhibits the mature biofilm of Campylobacter strains when treated with DNase I after 72 h of biofilm formation [37]. However, Whitchurch and colleagues demonstrated that DNase I could not abolish 84 h P. aeruginosa biofilm [10]. Similar findings for Helicobacter pylori suggest that eDNA may not be the main component of H. pylori biofilm [12].
To our knowledge, there are two previous reports about eDNA of Burkholderia species. Austin and colleagues demonstrated that B. pseudomallei strain K96243 produced eDNA without bacterial cell lysis and that eDNA enhanced bacterial colonization of stomach tissues of BALB/c mice [28]. Meanwhile, Garcia and colleagues revealed that eDNA is not required for the initial attachment step but essential for cellular interaction of 16 h biofilm formation in B. thailandensis [29].
The addition of exogenous B. pseudomallei genomic DNA and salmon-sperm DNA during biofilm inoculation, prior to or after DNase I treatment, did not produce a detectable effect on B. pseudomallei biofilms. This is in contrast with observations by Harmsen and colleagues, who demonstrated that the addition of genomic DNA and salmon-sperm DNA restored the biofilm of L. monocytogenes at the initial attachment stage, indicating that eDNA is required for the adhesion step of biofilm formation in that species [15]. Similarly, Carrolo and colleagues demonstrated that the addition of salmon-sperm DNA could restore the ability of mutant strains of Streptococcus pneumoniae to form biofilms [47].
In conclusion, our findings may have important biological implications by pointing out that eDNA is a key component of B. pseudomallei biofilm, in particular during the early stages of biofilm growth. eDNA could be an attractive target for prevention of B. pseudomallei biofilm formation by using DNase digestion. This points the way to novel strategies to destabilize B. pseudomallei biofilm formation: perhaps a combination of DNase I and antibiotics will lead to an effective treatment for relapsing melioidosis.