The Institutional Review Board (IRB) of LV Prasad Eye Institute (LEC/08/110/2009) approved the study, and the data were analyzed anonymously and reported.

Eighteen non-duplicate S. haemolyticus strains isolated from patients with a variety of ocular infections at LV Prasad Eye Institute, Bhubaneswar, India, during 2007–2011 and characterized previously following the standard methods (Panda et al., 2014, 2016) were included in the study. These isolates were from patients with microbial keratitis (n = 8), nasolacrimal duct obstruction (n = 1), blepharitis (n = 1), and keratoconjunctivitis (n = 1). Also, seven strains were isolated from the asymptomatic healthy conjunctiva. The 16S rDNA sequencing and amplification of species-specific nuc gene confirmed the identity of isolates as S. haemolyticus (Panda et al., 2016).

The capacity of biofilm formed by S. haemolyticus was determined by the method modified by Fredheim et al. (2009). Briefly, after diluting bacterial cultures to absorbance 0.05 OD at 595 nm, 200 μl of diluted cultures were added to 96 well polystyrene plates (COSTAR; Sigma) and were allowed to grow for 24 h at 37°C statically. The planktonic cells were removed by inverting the plate on the paper towel and washed the plates once with 1× PBS and then stained with 125 μl of 0.1% crystal violet for 15 min at room temperature. The bound dye was dissolved in the ethanol-acetone mixture (70:30) and recorded the absorbance at 540 nm (Thermo Scientific Varioskan, United States). We performed the biofilm assay in a single run of four wells in biological triplicates. Isolates were considered biofilm positive if the strain had an OD of ≥0.25. Biofilm forming capacities of all isolates were determined in TSB/BHIB, TSB/BHIB with 1% glucose (TSB_glu/BHIB_glu) and also in TSB/BHIB with 3% NaCl (TSB_NaCl/BHIB_NaCl). We used S. epidermidis ATCC 35984 as a positive control for PIA production and biofilm formation in all the experiments.

Diluted bacterial cultures grown in TSB_glu were added to 8 well-chambered slides (Nunc) and allowed to grow statically at 37°C for 24 h. The chambers were inverted, washed with 1× PBS and stained with Acridine Orange (Sigma; 0.02%) for 15 min in the dark at room temperature. The cells were fixed using 10% neutral formalin, kept in the dark for 30 min, then sealed by the coverslip and collected the images by Leica TCS SP5 confocal scanning system (Leica Microsystem, Mannheim, Germany). Excitation at 488 nm detected the fluorescence of Acridine Orange and emission with a 500–530 nm bandpass filter. The z-stacks were collected at 1 μm intervals and exported as.tif files from Leica LAS AF (version 1.8.2) software and imported into the ImageJ (version 1.41) program (USNIH, Bethesda, MD, United States). PHILIP program analyzed the randomly selected five different fields for imaging and quantification of the mean thickness of biofilm.

We performed detachment assays to determine the components of biofilm matrix by the method of Fredheim et al. (2009) using reagent and enzymes (i) 40 mM NaIO₄ (Sigma; S1878), (ii) 0.1 mg/ml proteinase K (Sigma; P2308), and (iii) 0.5 mg/ml DNase I (Sigma; DN25). Diluted bacterial isolates were seeded to 96 well polystyrene plates (COSTAR; Sigma) and were allowed to grow for 24 h at 37°C under static condition. Then the above reagents were added to individual wells at 0 and 24 h. After incubation, the biofilm was disrupted by mixing and remaining adhered cells were quantified by crystal violet staining by the method described above. S. epidermidis ATCC 35984, known to produce PIA dependent polysaccharide matrix, was used as a control.

We performed the detachment assay in duplicate wells in three biological replicates by the method described above (Fredheim et al., 2009). The percent detachment was calculated from the average difference between the treated wells and the control wells and divided the results into three categories: (i) no detachment (<10%), (ii) intermediate detachment (10–50%), and (iii) strong detachment (>50%). S. epidermidis ATCC 35984 and SHN65 grown for 24 h on 8 well-chambered slides were subjected to NaIO₄, proteinase K, and DNase I treatment and allowed to grow for another 24 h at 37°C. Then stained the adherent biofilm comprised of polysaccharides, eDNA, and protein grown in chamber slides with wheat germ agglutinin-Alexa Fluor 488 conjugate (Invitrogen; 1 μg/ml), DAPI (Sigma; 5 μg/ml) and sypro ruby (Invitrogen; 1×) and collected the images by CLSM. WGA-Alexa Fluor 488 and sypro ruby fluorescence detected by excitation at 488 nm, and emission was collected with a 510–570 nm and 600–700 nm bandpass filter, respectively. Excitation at 405 nm detected DAPI fluorescence and emission collected with a 410–475 nm bandpass filter. The z-stacks were collected at 1 μm intervals and exported as.tif files from Leica LAS AF (version 1.8.2) software.

We added Proteinase K and DNase I to the wells containing S. haemolyticus SHN65 and SH1965 and S. epidermidis ATCC 35984 at 0, 1, 2, 3, 4, 6, 8, 10, 12, and 24 h to see the effect on biofilm formation. All wells were incubated for a total of 24 h at 37°C under static condition and the wells treated at 24 h were further allowed for another 12 h after the addition of the reagents.

We determined the inhibitory effect of Sodium polyanethole sulphonate (PAS) on autolysis by the method described earlier (Rice et al., 2007). PAS (0.5 mg/ml) was added to the respective wells at different time points and incubated for a total of 24 h, and the biofilm was allowed to grow for further 12 h after addition of PAS and estimated by the method described above.

Extracellular DNA was extracted by the method described by Rice et al. (2007). Briefly, 100 μl of filter-sterilized supernatant from each isolate was transferred into a chilled tube containing 300 μl of TE buffer and extracted once with an equal volume of phenol/chloroform/isoamyl alcohol (25:24:1) by centrifuging at 18,000 × g for 20 min at 4°C. The supernatant was further extracted once with an equal volume of chloroform/isoamyl alcohol (24:1). The aqueous phase of each sample was then mixed with 3 volume of ice-cold 100% ethanol and 1/10 volume of 3 M sodium acetate (pH 5.2) and stored at −20°C. The next day, the ethanol precipitated DNA was collected by centrifugation for 20 min at 4°C and 18,000 × g, washed with ice-cold 70% ethanol, air dried, and dissolved in 30 μl of nuclease-free water.

The extracted eDNA was diluted 10 times (1:9) using nuclease-free water and used as a template. PCR mixture contained: 4 μl of 5× PCR buffer, 1 μl of 25 mM MgCl₂, 2 μl each of dNTPs (Promega, Madison, WI, United States), 2 μl each of forward and reverse primer (5 pmol/μl) (GCC Biotech, New Delhi), 0.25 μl of 5 U/μl Taq polymerase (Promega, Madison, WI, United States), 1 μl of diluted eDNA, and 7.75 μl of nuclease-free water. The PCR was programmed as follows: initial denaturation at 94°C for 2 min, followed by 25 cycles consisting of denaturation at 94°C for 30 s, annealing at 52°C for 30 s, extension at 72°C for 30 s, and a final extension at 72°C for 5 min. PCR products were electrophoresed on 1.5% agarose gel, stained with EtBr and photographed using a gel documentation system (Bio-Rad, United States). The images were analyzed using ImageJ software and the eDNA quantified by amplifying 16S rDNA, 23S rDNA, and gyrA genes, and taking into account the ratio of released eDNA and biomass of the isolate.

PCR determined the presence of icaA and icaD genes by the method described earlier (Arciola et al., 2001). We used S. epidermidis ATCC 35984 as a positive control for icaA and icaD detection. Also, PCR was used to determine the presence of antibiotic resistance genes among S. haemolyticus (Panda et al., 2016). Supplementary Table S1 showed the cyclic conditions of PCR and primers used for the detection of biofilm-associated gene and amplification of ARG. The amplicons of cidA, cidB, lrgB, sigB, mgrA, lytR, lytS, ebpS, fbp, agrA, agrB, luxS, and srtA genes were sequenced and submitted to GenBank database under accession numbers KU518895–KU518907.

To determine the expression of biofilm-associated genes, the aliquots of 4-h grown culture in TSB_glu diluted to an absorbance of 0.05 OD at 595 nm were seeded into 35 mm Petri plate (Tarson) and incubated at 37°C for 4 h under static condition. The adherent cells were scraped using ice chilled resuspension buffer (50 mM Tris–HCl, 10 mM EDTA, 500 mM NaCl, pH 8.0) and centrifuged at 18,000 g for 5 min at 4°C. The pellet was used to isolate RNA using Trizol reagent (Invitrogen, United States) according to the manufacturer’s instructions, and synthesized cDNA using 500 ng RNA and SuperScript VILO cDNA synthesis kit (Invitrogen, United States). The cDNA was diluted 30 times with nuclease-free water (1:29), and 1 μl of cDNA was used as a template in RT-PCR. The cyclic conditions were as follows: initial denaturation at 94°C for 2 min followed by 25 cycles consists of denaturation at 94°C for 30 s, annealing at 52°C for 30 s, and extension at 72°C for 30 s. The final extension was at 72°C for 5 min. Similarly, RT-PCR determined the expression of srtA in the biofilm of SHN65 grown for 4, 8, 12, 24, 48, and 72 h. Supplementary Table S1 shows the list of primers used in the expression study for biofilm-associated genes.

In this study, we checked the isolates for their ability to form biofilm in different media, instead for examining the differences in the rate of biofilm development. We found all the ocular S. haemolyticus isolates belonging to the infected eye, and healthy conjunctiva formed a biofilm on polystyrene microtiter plates. The capacity of biofilm formation by strains varied with media used (Figure 1A). In general, all the strains efficiently formed biofilms in both media TSB_glu and BHIB_glu. However, one strain from the infected eye (SH1965) and two isolates from healthy conjunctiva (SHN66 and SHN75) formed better biofilm in TSB_NaCl and BHIB, respectively. Isolate SHN65 formed the highest biofilm among the strains tested. The biofilm forming capacity of the isolates was further quantified by estimating the mean thickness of the biofilm formed in TSB_glu by CLSM (Figure 1B). Supplementary Figure S1 showed the representative CLSM images of the biofilm by all the ocular isolates.

All isolates of S. haemolyticus showed decreased capacity to form biofilm in the presence of proteinase K and/or DNase I when added at the initial point of inoculation in both TSB_glu (Supplementary Figures S2A,B) and BHIB_glu (Supplementary Figures S2C,D). These observations thus indicate the role of proteins and eDNA in biofilm development. We then performed detachment assay to see the effect of NaIO₄, proteinase K and DNase I on mature biofilm in both the media TSB_glu (Supplementary Figures S3A–C) and BHIB_glu (Supplementary Figures S3D–F). The extent of detachment of biofilm was media independent. We found that 17 of 18 ocular S. haemolyticus isolates showed detachment after treatment with proteinase K, and 14 isolates each after treatment with DNase I, and NaIO₄, respectively. We noticed that none of the strains showed strong detachment (50–90%) upon NaIO₄ treatment. However, 11 isolates showed moderate (10–50%), and three strains showed low level (0–10%) of detachment, respectively. While studying the effect of proteinase K and DNase I on biofilm detachment, we found 17 isolates showing intermediate and 13 strong detachment, respectively (Figure 2). The positive control S. epidermidis ATCC 35984 displayed strong detachment upon NaIO₄ treatment. Therefore, we selected the best biofilm forming strain SHN65 for the further study.

Confocal laser scanning microscopy imaging of the biofilm formed by ATCC 35984 and SHN65 indicates that the ECM of ATCC 35984 is predominantly composed of polysaccharides which stained green in our assay; whereas, ECM of SHN65 which took both red and blue stains represent protein and DNA as principal components (Figure 3A). CLSM imaging of the detachment assay also inferred that ECM of ATCC 35984 contains polysaccharides and SHN65 contain protein and DNA (Figure 3B). The thickness of the biofilm formed by ATCC 35984 reduced upon NaIO₄ treatment whereas no reduction was observed upon Proteinase K and DNase I treatment. On the contrary, the biofilm thickness of SHN65 decreased after treatment with proteinase K and DNase I and was unaffected by NaIO₄ treatment (Figure 3B).

Time point inhibition studies using proteinase K and DNase I determined the components of the biofilm matrix at different stages of biofilm development in S. haemolyticus strains SHN65 and SH1965. The addition of proteinase K (Supplementary Figures S4A,C) and DNase I Supplementary Figures S4B,D) at all time points diminished the cell adherence without showing any significant effect on bacterial growth.

We used sodium PAS, a known chemical inhibitor of cell lysis which does not affect the growth of the bacteria at different time points of biofilm development. PAS added at 0–3 h of time points drastically reduced the biomass of the adherent cells of S. haemolyticus SHN65 (Figures 4A,B). However, no effect was observed on adherent biomass of positive control S. epidermidis ATCC 35984 (Figures 4A,B).

Although all the ocular isolates produced biofilm, none carried the icaA or icaD genes, irrespective of the sources of isolation. All isolates carried the genes encoding cell surface attachment proteins (ebpS, cbp, fbp, and srtA), quorum sensing (agrA, agrB, agrD, and luxS), eDNA release (cidA, cidB, lrgA, lrgB, lytS, lytR, and atlE), and global regulators (sarA, mgrA, and sigB and araC/xylS) (data not shown). These isolates also showed positive results for mphC, ermC, msrA, aac6′-aph2″, ant4′, aph3′, tetK, dfrG, blaZ, mecA, cat:pC223, sat, lnuA genes.

The expression study of biofilm-associated genes showed that majority isolates expressed genes associated with cell surface attachment, quorum sensing, eDNA release, and global regulation. We then correlated the expression of the genes associated with eDNA release and cell surface attachment proteins with the extent of detachment upon addition of DNase I and proteinase K, respectively. There was a good association between cidA, and cidB expression and detachment upon DNase I treatment. Similarly the expression of fbp, ebpS, cbp, and srtA genes showed an association with higher detachment upon proteinase K treatment among the strains (Supplementary Figures S5A,B). The expression of atl associated with higher detachment upon both DNase I and proteinase K treatment. We did not look for the association of expression of quorum sensing genes, and global regulators with biofilm component as these play a role in both eDNA and protein-rich biofilms. The expression of srtA was elevated at early time points of biofilm formation (Figure 5).

The eDNA present in the mature biofilm of 18 ocular S. haemolyticus isolates and determined by RT-PCR is shown in Supplementary Figure S6A. In comparison to other strains, two of the isolates, SHN65 and SH314 produced more eDNA (Supplementary Figure S6B). Upon using eDNA as a template for PCR, these isolates showed positive results for mphC, ermC, msrA, aac6′-aph2″, ant4′, aph3′, tetK, dfrG, blaZ, mecA, cat:pC223, sat, lnuA genes (Supplementary Figure S7). These findings thus indicate that ARGs released in eDNA can act as a source for the transmission of ARGs.