Impact of solid surface hydrophobicity and micrococcal nuclease production on Staphylococcus aureus Newman biofilms

Staphylococcus aureus is commonly associated with biofilm-related infections and contributes to the large financial loss that accompany nosocomial infections. The micrococcal nuclease Nuc1 enzyme limits biofilm formation via cleavage of eDNA, a structural component of the biofilm matrix. Solid surface hydrophobicity influences bacterial adhesion forces and may as well influence eDNA production. Therefore, it is hypothesized that the impact of Nuc1 activity is dependent on surface characteristics of solid surfaces. For this reason, this study investigated the influence of solid surface hydrophobicity on S. aureus Newman biofilms where Nuc1 is constitutively produced. To this end, biofilms of both a wild-type and a nuc1 knockout mutant strain, grown on glass, salinized glass and Pluronic F-127-coated silanized glass were analysed. Results indicated that biofilms can grow in the presence of Nuc1 activity. Also, Nuc1 and solid surface hydrophobicity significantly affected the biofilm 3D-architecture. In particular, biofilm densities of the wild-type strain on hydrophilic surfaces appeared higher than of the mutant nuc1 knockout strain. Since virulence is related to bacterial cell densities, this suggests that the virulence of S. aureus Newman biofilms is increased by its nuclease production in particular on a hydrophilic surface.


Results
Effect of solid surface hydrophobicity on bacterial adhesion. Adhesion of S. aureus Newman and its isogenic nuc1 mutant were first investigated on glass (water contact angle 16° ± 21°), silanized glass (water contact angle 96° ± 8°) and Pluronic F-127-coated silanized glass (water contact angle ≤ 25 ± 1) 26 . Planktonic cultures of both staphylococcal strains were allowed to adhere to the solid surfaces for 1 h after which adhered bacteria were imaged with a phase contrast microscope. The obtained results revealed a lack of significant difference between adhesion of the WT and mutant strain on the same solid surface (Fig. 1a, b). Both bacterial strains showed the highest affinity for silanized glass and the lowest number of adhered bacteria was found on Pluronic F-127-coated silanized glass.
Nuclease activity in planktonic cultures and biofilms. Nuclease activity in planktonic cultures of both staphylococcal strains as well as in the biofilms grown on glass, silanized glass and Pluronic F-127-coated silanized glass were measured with a Förster resonance energy transfer (FRET)-based DNAse assay (Fig. 2a). The data obtained revealed that Nuc1 activity persisted during biofilm formation but was decreased by twofold, 1.5-fold and 2.2 fold per CFU respectively on glass, silanized glass and Pluronic F-127-coated silanized glass in comparison to planktonic WT (Fig. 2b). Note that the nuclease activity for the mutant strain was almost zero.
Effect of solid surface hydrophobicity and Nuc1 production on biofilm thickness and structure. Optical coherence tomography (OCT) was employed to acquire cross sectional images of S. aureus Newman WT and S. aureus Newman ∆nuc1 biofilms on all tested solid surfaces after 20 h of growth. As can be seen in Fig. 3a, biofilms appeared as white reflective spots while fluid-filled pores or channels, also referred to as voids, appeared as dark transparent spots on the OCT images (shown by arrows in Fig. 3a). The 2D OCT images showed that the surface of the biofilm from the mutant strain was rougher than that of the WT strain on glass and silanized glass. The outer surface roughness of the biofilm was similar for both staphylococcal strains grown on the Pluronic F-127-coated silanized glass and showed more island like structures (indicated by stars in Fig. 3a) than on the other two surfaces. The biofilm thickness, as determined with a 2D analysis software, revealed a significantly thinner biofilm of the WT strain than the mutant strain on all tested solid surfaces ( Fig. 3b and Fig. S3). The biofilm on Pluronic F-127-coated silanized glass was significantly thinner with respect to silanized glass and glass ( Fig. 3b and Fig. S3). However, there was no significant difference between the thickness of biofilms formed on glass and silanized glass (Fig. 3b).
Effect of Nuc1 production and solid surface hydrophobicities on biofilm density, eDNA and epS polysaccharide. The results revealed that the bacterial density of the WT biofilm is higher than that of the mutant strain on glass (Fig. 4a). The bacterial density of mutant biofilms grown on the Pluronic-coated silanized glass was significantly higher than their counterparts on glass and silanized glass (Fig. 4a). PicoGreen staining showed equivalent eDNA content in biofilms formed by both strains on all tested surfaces (see Supplementary Fig. S2a). A Pearson correlation test showed no significant relation between the eDNA concentration and density in biofilms on glass (r = − 0.72, P = 0.16), silanized glass (r = 0.72, P = 0.11) and Pluronic F-127-coated silanized glass (r = − 0.72, P = 0.17, Fig. 4b) for the WT strain. A Pearson correlation test revealed a positive relation between the concentration of eDNA and bacterial density of biofilms formed on glass (r = 0.75, P = 0.08) and silanized glass (r = 0.91, P ≤ 0.01, Fig. 4c) only for the mutant strain. There were no significant differences between the CFU and EPS production of both staphylococcal strains on all tested solid surfaces (Fig. S2, b and c).

Discussion
As shown in Fig. 1 solid surface hydrophobicity impacted bacterial adhesion. Irreversible adhesion is suspected to have occurred via the hydrophobic effect due to the removal of interfacial water between the bacterium and the solid surface and hydrogen bonding on the hydrophobic and hydrophilic surfaces respectively 27  www.nature.com/scientificreports/ Pluronic F-127 coating was effective in restricting adhesion because an approaching bacterium will press the stretched hydrophilic PEO chains towards the hydrophobic PPO chains and subsequently be repelled due to steric hindrance 28 . This has been found to be valid for other strains as well 21,26,29,30 . Additionally, the presence of a higher number of adhered bacteria on silanized glass ( Fig. 1) may imply a higher number of hydrophobic than hydrophilic patches on the surface of the S. aureus Newman strains, fostering an energetically more favorable condition on hydrophobic surfaces. It is known that DNAse may affect adhesion as was shown by a significant reduction of S. aureus and Pseudomonas aeruginosa, preventing biofilm formation for 14 h 31 . Nucleases in our assays did not show these effects since no differences in numbers of adhered bacteria were found between the wild type and nuclease deficient mutants ( Fig. 1), indicating that cell surface associated eDNA was not affected by the expression of nuclease. In other studies, glucose supplementation to boost biofilm formation resulted in a tenfold decreased production www.nature.com/scientificreports/ and activity of Nuc1 19,32 . In our biofilm assay, it was observed that Nuc1 production was reduced in biofilms with respect to their planktonic counterparts even without glucose supplementation (Fig. 2). S. aureus Newman has been classified as a poor biofilm former due to a point mutation and consequent constitutive activation of the SaeRS two-component system effecting a constant nuc1 expression 24,33 . As expected based on literature 24 we did not find differences in nuclease production per CFU at the respective substrates as a result of the constitutive expression of nuc1. In this study it was found however that biofilm formation by the S. aureus Newman strain was not hindered by the constitutive Nuc1 production (Figs. 2, 3). The minimum biofilm thickness recorded in this study (160 ± 37 μm) was higher than the 30 ± 2 μm thickness recorded for S. aureus MFP03 34 . In this study, the formation of thicker biofilms by the S. aureus Newman strain compared to literature may be partly due to stronger adhesion under the dynamic conditions employed in our biofilm assay 35 . In an earlier report by Kiedrowski et al. 19 , the biomass of biofilms formed by S. aureus Newman was comparable to that of S. aureus USA300 and S. aureus TCH151 and higher than that of S. aureus COL. These results may also suggest that the eDNA in the biofilm is stabilized, possibly via interactions with DNA-binding proteins like Eap 36 and in particular SaeP 36 , which is expressed as an auxiliary protein in the SaeRS two component system 37 .
We further focused on the synergistic effect of solid surface hydrophobicity and Nuc1 activity on biofilms of S. aureus Newman and its isogenic ∆nuc1 mutant. One major finding in this study is that solid surface hydrophobicity affects the internal configuration of biofilms as evidenced by the changes in density with and without www.nature.com/scientificreports/ Nuc1 activity (see Fig. 4). eDNA is essential for biofilm structure 38 , acting as an electronegative string that tether bacteria surrounded by net positively charged matrix proteins and PIA 8 . In essence, eDNA in S. aureus Newman WT biofilms will be cleaved due to Nuc1 activity while eDNA in the mutant strain biofilms remain intact (Fig. 2). This was shown earlier in agarose gels where eDNA from wild type MRSA bacteria appeared as a smear of mainly low molecular weight DNA fragments, whereas the eDNA from mutant strains were of high molecular weight similar to purified genomic DNA 19 . Employing a ball and stick model to represent the 3D net-like architecture www.nature.com/scientificreports/ of a biofilm 8 , eDNA can be represented by sticks. Likewise, bacteria connected to positively charged proteins and PIA are represented by balls. Nuc1 activity will cleave eDNA which will result in relative short "sticks" in the ball-and-stick model. Lack of Nuc1 activity in the mutant strain biofilms will result in maintaining the connecting eDNA (sticks) which will in turn lead to an expansion of the biofilm net-like architecture as related to the WT strain. First, this expanded state is evidenced by the rougher outer surface of the biofilms formed by the mutant strain on glass and silanized glass (Fig. 3a). Second, the expanded state of the biofilm resulted in an increase in biofilm thickness (Fig. 3b) as well as a decrease in density of the mutant biofilms which was significant on glass but not on silanized and Pluronic F-127-coated solid surfaces (Fig. 4a). Also, for the WT strain the density was not affected by higher eDNA content as was the case for the mutant strain on glass and silanized glass, evidenced from significant positive correlations between eDNA content and bacterial cell density on glass and silanized glass (Fig. 4b, c). This correlation was not observed for Pluronic F-127 coated surfaces, signifying again that surface characteristics affected the net-like architecture of biofilms. These differences are also partly in line with increasing amounts of EPS with decreasing hydrophobicity, specifically eDNA and PNAG which are necessary for cell-cell interactions 23 . Thus, nucleases in particular modulate the EPS binding capacity in biofilms.
Bacterial biofilm density has earlier been reported to range between 0.2 and 0.4/μm 3 which is relatively low with respect to a closed packing that would lead to a bacterial density of 1.4/μm 3 39 . This indicates that S. aureus Newman biofilms, the bacterial density of which range up to 0.5/μm 3 , are relatively dense, specifically on hydrophilic surfaces like glass and Pluronic-coated salinized glass. Since bacterial density has earlier been linked to the expression of the agr system which regulates the expression of several S. aureus virulence genes 9,11 , it is suggested that Nuc1 production in combination with the surface hydrophobicity influences the virulence associated with S. aureus biofilms on implants. Mukherjee et al. 40 also predicted that an increase in density may increase the production of an autoinducer of the agr system. In addition to this, the expression of RNA III, the www.nature.com/scientificreports/ main effector of the agr system, was seen to be elevated with increasing cell density in vegetations after S. aureus infection in an experimental endocarditis study 41 . In S. aureus biofilms, agr quorum sensing improves α-toxin production which is harmful to immune cells and promotes cell detachment independent of Nuc1 activity 9,19,42,43 .
Although results from this in vitro study does not depict a perfect representation of events in vivo, the increase in biofilm density due to Nuc1 activity on hydrophilic surfaces, taken together with earlier findings of increased agr expression with increasing bacterial density, strongly suggests a higher virulence of the S. aureus strains on hydrophilic solid surfaces. In summary, this study identified biofilm formation occurring in the presence of Nuc1 activity (Figs. 2,  3). Lack of Nuc1 activity resulted in an expansion of the biofilm net-like architecture due to the presence of uncleaved eDNA. Nuc1 activity caused significant density differences between the WT and mutant strain in particular on the hydrophilic glass (Fig. 4). Augmented agr expression as expected from increasing density suggests that the virulence of biofilm associated infections caused by S. aureus Newman is affected by the production of nuclease in particular on a hydrophilic surface like glass.

Materials and methods
Solid surfaces modifications. Round cover glasses (15 mm diameter, Marienfeld, Lauda-Königshofen, Germany) were cleaned by sonicating for 5 min in 2% RBS in a bath sonicator (Salm en Kipp b.v., Breukelen, The Netherlands). This was followed by washing 3 times in demineralized water and a subsequent 5 min incubation in methanol (CH 3 OH, EMPLURA, Darmstadt, Germany). The cover glass was then washed 3 times in demineralized water and kept submerged until used for glass silanization (i).

Glass silanization.
To activate glass for the silanization step, clean glass cover slips were incubated in a 1:1 mixture of hydrochloric acid (HCl, 12 M, EMSURE, Darmstadt, Germany) and methanol for 40 min and rinsed at least 5 times with ultrapure water (18.2 MΩ). The cover glass was subsequently incubated for 40 min in sulfuric acid (H 2 SO 4 , 95-97%, EMSURE, Darmstadt, Germany) then washed 5 times with ultrapure water. After this, the glass was placed in gently boiling ultrapure water for 1 h and allowed to cool at room temperature. The cover glass was dried with filter sterilized air and carefully placed in a dry petri dish. A small container with 50 µL of propyltrichlorosilane (CH 3 CH 2 CH 2 SiCl 3 , Sigma, Saint Louis, USA,) was placed next to the activated cover glass and kept under vacuum overnight.
Pluronic F-127 coating. Silanized glass was coated with Pluronic F-127 using a modified protocol based on the studies of Nejadnik et al. 25 . In summary, silanized glass was sterilized in 70% ethanol and rinsed in sterile demineralized water. The glass was then placed in a 24-wells plate (Greiner bio-one, USA) and incubated in filter Bacterial strains and culture. S. aureus Newman and its isogenic S. aureus Δnuc1 were kindly donated by Prof. McNamara (Department of Internal Medicine, University of Illinois, USA) and used for all the experiments in this study. S. aureus Newman is an MSSA strain which was isolated from a case of tubucular osteomyelitis in human 44 . The S. aureus Newman ∆nuc1 was constructed by Kiedrowski et al. 19 using the Targetron Gene Knockout system. Single colonies of the strains were obtained by aerobic culturing on blood agar plates for 24 h at 37 °C. To make a preculture, one colony of each strain was inoculated in 10 mL of Tryptone Soya Broth (TSB, OXOID, Basingstoke, UK) and cultured for 24 h at 37 °C. The preculture was transferred to 200 mL TSB and grown for 17 h at 37 °C for the main culture.
Multi-well plate biofilm assay. The bacterial cells were harvested from the main culture by centrifugation (6,250 g 5 min, 10 °C) and washed twice in PBS. Bacteria were sonicated 3 × 10 s (Vibra Cell Model 375, Sonics and Materials Inc., Danbury, CT, USA) in an ice-water bath to disrupt bacterial aggregates and enumerated in a Bürker-Türk counting chamber. The bacterial suspension was diluted to a concentration of 1 × 10 9 mL −1 with sterile PBS and transferred to a 24-wells plate containing the solid surfaces (glass, silanized glass, and Pluronic F-127 coated silanized glass). The bacteria were allowed to adhere for 1 h under rotation in a shaking incubator at 150 RPM at 37 °C after which unattached bacteria were removed by washing with sterile PBS. To evaluate the amount of S. aureus Newman WT and Newman ∆nuc1 that adhered on all tested surfaces, a phase contrast microscope (Olympus) was utilized to image the number of adhering bacteria per unit area. The bacteria were counted manually or with Fiji 45 . To grow biofilms, TSB was added to each well containing a particular solid surface with adhering bacteria and incubated for 20 h (37 °C, 80 RPM). After the incubation period, growth media was removed without exposing the biofilms to air and washed twice with sterile PBS. This was done before further analysis of the biofilms unless otherwise stated.  -6FAM TTT TTT TTTTBHQ1, Sigma, Saint Louis, USA) was added and fluorescence intensity was immediately measured at 1 min intervals during 39 min with a Fluorstar Optima (BMG LABTECH, Offenburg, Germany) plate reader at excitation 485 nm and emission at 520 nm. Wells containing biofilm suspension and Tris-HCl buffer but without the DNA probe were used as a blank as well as a negative control. As a positive control, biofilm suspension was replaced with 25 µL of 0.001 U/mL purified staphylococcal nuclease (Sigma, Saint Louis, USA) reaching a final concentration of 0.000125 U/mL. A calibration curve was performed using purified staphylococcal nuclease at concentrations 0, 0.0001, 0.0002, 0.0004, 0.0008 and 0.001 U/mL (Fig. S1a). Data obtained for S. aureus Newman WT biofilms and purified staphylococcal nuclease were fitted to Eq. (1): where y(t) is the fluorescence intensity, c, a and b are fitting parameters (c = 2,000, a = 1,000 and b = 10) and t is time in minutes. The initial rate at which Nuc1 cleaves the DNA probe was determined from the first derivative at t = 0, being a * b from Eq. (1). The initial rate of activity of the data obtained from Nuc1 mutant strain were determined using a linear regression of the first five measurements due to very low signal. All data fittings were executed in GraphPad Prism version 6 (GraphPad Software, La Jolla California, USA).

Biofilm analysis with optical coherence tomography (OCT). S. aureus Newman WT and S. aureus
Newman ∆nuc1 biofilms were imaged with an OCT Ganymede II (Thorlabs Ganymade, Newton, New Jersey, USA) device using a white light beam of 930 nm. Ten 2D images were taken using the refractive index of water, 1.33, a field of view of 4 mm and 2 µm pixels in the vertical direction. The average thickness of each 2D image was determined with a 2D OCT analysis software (Thorlabs, Newton, New Jersey, USA). In brief, a grey value threshold was determined that separates the biofilm from the surrounding liquid based on the grey value distribution of the particular image. The upper contour line of the biofilm was defined as those pixels that were connected to the bottom of the biofilm by pixels with grey values higher than the grey-value threshold. The software determined the bottom contour line of the biofilm by connecting six points which were manually placed at the solid surface-biofilm interface. The biofilm thickness could then be determined based on the average number of pixels between the bottom and upper contour line 39 . colony forming units. Biofilms formed by S. aureus Newman and its isogenic Nuc1 mutant were washed twice and resuspended in 1 mL sterile PBS. The biofilm suspension was sonicated for 5 min in a bath sonicator (Salm en Kipp b.v., Breukelen, The Netherlands). After which vigorous pipetting was used to further suspend the biofilm. A serial dilution of the biofilm suspension was made in sterile PBS and 100 μL plated on TSB agar plates followed by incubation at 37 °C overnight.
Bacterial density. The bacterial suspensions were diluted 100 times and counted in a Bürker-Türk counting chamber. The values obtained for the total number of bacteria per biofilm were used in calculating the total bacterial density of the biofilms per unit volume by dividing the total cell count by the biofilm volume, calculated through multiplying the biofilm area by its thickness as defined in Eq. (2): picoGreen staining. eDNA was quantified as described by Tang et al. 46 . Briefly, growth media of 20 h old S.
aureus Newman WT and S. aureus Newman ∆nuc1 were removed and the biofilm was washed with sterile PBS. The biofilms were suspended by gently pipetting up and down. 100 µL of the biofilm suspension was mixed with 100 µL freshly prepared PicoGreen solution (1 µL PicoGreen dye in 199 µL TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8) in a black 96-well plate with a clear bottom and incubated for 4 min at room temperature before measuring the fluorescent intensity with a Fluorstar Optima plate reader (excitation/emission 485 nm/520 nm). A calibration curve was performed using a concentration range of 0-1,000 ng/mL of λ DNA (Fig. S1b).
Calcofluor white staining for EPS polysaccharide. 250   www.nature.com/scientificreports/ Statistical analysis. All biofilms were grown in duplicate and repeated with three or four different cultures.
Statistical differences between experimental groups on identical and dissimilar solid surfaces were analysed with a Students t-test and ANOVA test respectively, using GraphPad Prism version 6 (GraphPad software, La Jolla California, USA). Differences were considered significant if P ≤ 0.05.

Data availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.