S. aureus laboratory strain Newman and the isogenic derivatives codY (Newman ΔcodY::ermC) (Luong et al., 2011) and ica (Newman ica::tet) (Kropec et al., 2005) mutants were used in this study. In addition, 8 clinical isolates related to USA300 and USA100 clones (Table 1), which were identified previously by spa typing, MLST and SCCmec typing (Lattar et al., 2012) were included in the study. S. aureus SA113 and Reynolds CP5 null (Δcap5) (Watts et al., 2005) strains were also included as reference controls. Bacteria were stored in Trypticase Soy Broth (TSB) (Britania, Buenos Aires, Argentina) with 20 % glycerol at −80°C until use. All cultures were grown in TSB supplemented with 0.25% of glucose (TSBg) medium in the presence or absence of 2 mM salicylic acid (SAL) with or without 50 μM FeSO₄ and incubated for 24 h at 37°C and 200 r.p.m. The iron addition to TSBg did not affect the biofilm formation by S. aureus (Figure S1). Iron-restricted conditions were assessed in iron-depleted TSBg medium (CTSBg) by batch incubation with 3% w/v Chelex 100 (BioRad, Hercules, CA, USA). For selection of the chromosomal marker in the Newman ica and codY mutants 5 μg/ml tetracycline and 10 μg/ml erythromycin were used, respectively. All chemical reagents were purchased from Sigma-Aldrich (St. Louis, MO, USA) unless otherwise indicated.

The chelating capacity of SAL was assessed by the ferrozine assay (Mladenka et al., 2011). Briefly, TSBg medium containing different concentrations of SAL (final volume: 100 μl) was supplemented with 50 μM FeSO₄ and incubated for 5 min. To avoid iron oxidation, 50 μl of 5 mM NH₂OH solution was added to the mixture. Then, 50 μl of 5 mM ferrozine (or distilled water for the blanks) was added. The iron-ferrozine complex formation was determined by determination of the absorbance at 540 nm (Abs₅₄₀).

Aliquots of bacterial lysates generated by incubation of biofilms over 24 h with lysis buffer [100 μl of buffer Tris-EDTA 10:1, (10 mM Tris-HCl pH 8; 1 mM EDTA), 40 μl of 1 mg/ml lysostaphin, 20 μl of 50 mg/ml lysozyme] were treated with HCl/KMnO₄ (a solution of equal volumes of 1.4 M HCl and 4.5 w/v KMnO₄ in distilled H₂O) to promote the release of iron from the bacterial proteins. The Fe²⁺ contents were determined by the ferrozine assay and normalized to the protein concentration on each sample determined by the Bradford method (Bradford, 1976).

Quantitative assessment of biofilm formation was performed as previously described with modifications (Trotonda et al., 2008). Briefly, S. aureus strains were grown for 18 h and diluted 1:100 in TSBg or TSBg with SAL and FeSO₄, when required. Two hundred μl of these cell suspensions were added to sterile 96-well polystyrene microtiter plates. Non-inoculated medium controls were included. After 24 h of static incubation at 37°C, the final culture density (named OD_G) was measured by reading the optical density (OD₅₉₅) using a microplate reader (Multiskan EX, Thermo Electron Corporation, Waltham, MA, USA). The culture medium was then removed from each well and plates were washed twice with phosphate buffered saline (PBS). The biofilms were fixed with 100% methanol for 15 min, stained with 0.5% crystal violet for 20 min, and washed twice again gently under running tap water. The amount of biofilm biomass was measured after addition of 30% glacial acetic acid by reading the OD₅₉₅ (named OD_B). The levels of crystal violet staining were expressed relative to the final culture density measured prior to the biofilm assay (biofilm: OD_B)/OD_G) and named in the text as “biofilm” for the sake of clarity. The induction of biofilm by SAL was defined as the mean value from SAL-treated biofilms relative to the mean value of the biofilm formed in TSBg. The relative amount of biofilm formed by each isolate to that of the SAL-treated Newman biofilm was expressed as a percentage. For detachment assays, biofilms were grown for 24 h in 96-well microplates as described above. The biofilms were washed with PBS and then treated for 2 h at 37°C with 20 μg/ml of Dispersin B (Kane Biotech Inc., Winnipeg, MB, Canada) in PBS (Trotonda et al., 2008). After treatment, the biofilms were washed with PBS, fixed with methanol, stained with crystal violet and quantified as described above.

Overnight cultures of S. aureus grown in TSBg, TSBg with SAL or TSBg with SAL plus iron were adjusted to an OD₆₀₀ of 0.05 and aliquots were utilized to inoculate the wells of a 24-well plate (scanning electron microscopy, SEM) or a 8-well chambered coverglass (ThermoFisher Scientific, Waltham, MA, USA) (confocal laser scanning microscopy, CLSM). After static incubation during 24 h at 37°C the biofilms were processed as follows. For SEM, biofilms were washed with PBS, fixed with 2.5% formaldehyde for 2 h at 4°C and dehydrated in increasing concentrations of ethanol. The glass coverslips were fixed on aluminum stubs, covered with gold-palladium film and examined in a Philips XL30 TMP scanning electron microscope. For CLSM, after washing with PBS, the biofilms were stained using the LIVE/DEAD BacLight Bacterial Viability kit (Molecular Probes, Waltham, MA, USA) to determine bacterial viability, or treated with 90 μg/ml wheat germ agglutinin (WGA) (Oregon Green® 488 conjugate, Molecular Probes) with 5 μg/ml FM 4-64 (Molecular Probes) for PIA visualization, incubated for 15 min in the dark and fixed with 4% paraformaldehyde. After adding PBS, the biofilms were visualized with a Leica confocal laser scanning microscope (model TCS SP5, Germany). With the mixture of the SYTO9® and propidium iodide stains (LIVE/DEAD BactLight Bacterial Viability kit's fluorophores), bacteria with intact cell membranes display green fluorescence, whereas bacteria with damaged membranes exhibit red fluorescence. SYTO9 and propidium iodide were excited at 498–565 nm and their emission was monitored at 600–693 nm. Oregon Green 488® and FM 4-64 were excited at 508–549 nm and their emission was monitored at 650–750 nm. In all cases, 0.7 μm optical sections from the entire biofilm were collected and the stacks of images were analyzed using the Leica LAS AF Lite software. The biofilm parameters were determined using the Comstat 2 software. Images from 2 ramdonly selected positions of 2 independent samples were analyzed.

S. aureus biofilms were grown in 25-cm² polystyrene tissue culture flasks at 37°C statically for 24 h in TSBg or TSBg with SAL and FeSO₄, when required. The biofilms were scraped, suspended in PBS and centrifuged at 8000 × g for 20 min. Planktonic cultures grown up to stationary phase were washed three times in PBS and sedimented by centrifugation. Aliquots of the pellets suspended in deionized water were spotted on a zinc selenite (ZnSe) optical plate and dried during 40 min to yield transparent films. These films were used directly for FTIR spectroscopy, which was conducted with an HTS-XT microplate adapter coupled to a Tensor 27 FTIR spectrometer (Bruker Optics GmbH, Ettlingen, Germany). Infrared spectra were recorded in transmission mode in the spectral range between 4000 and 500 cm⁻¹. Normalized second-derivative spectra of the spectral window from 1200 to 800 cm⁻¹ were selected for principal component analysis (PCA) using the Unscrambler X software (CAMO Software, Oslo, Norway). This spectral region is dominated by C-O-C and C-O-P stretching vibrations of various oligosaccharides and polysaccharides and their specific types of glycosidic linkages and was previously shown to recognize changes in S. aureus surface glycopolymer composition including CP (Grunert et al., 2013; Johler et al., 2016).

S. aureus biofilms were grown in 96-well plates as described above. After measuring the OD₅₉₅ of the cultures (OD_G), the biofilms were washed twice with PBS and fixed with methanol. Then, the quantity of PIA and CP was determined by ELISA or by a fluorometric assay as specified below.

CP from S. aureus Newman (serotype 5; CP5) biofilms was quantified by ELISA. Briefly, 100 μl of blocking solution [1%, w/v, BSA in PBST (PBS+0.05% Tween 20)] was added to fixed biofilms and incubated for 1h at 37°C. After removing the blocking solution 100 μl of CP5 antiserum was added (1:3000 in PBST) and the plate was incubated for 30 min at 37°C. After washing with PBST the plate was dried and 100 μl of protein A-horseradish peroxidase conjugate (HRP-protein A Invitrogen, Carlsbad, CA, USA) (1:3000 in PBST) was added and incubated for 30 min at 37°C. The wells were washed three times for 5 min in PBST. The substrate [1 ml of 10 mg/ml o-phenylenediamine plus 9 ml of citrate (0.1 M Na₂HPO₄ pH = 5) plus 10 μl of H₂O₂ 30V] was then added, the plate was incubated for 5 min at 37°C. The reaction was stopped by addition of 50 μl of H₂SO₄ 12.5% per well and the level of CP expressed was measured by the absorbance at 492 nm (Abs₄₉₂) relative to OD_G.

The PIA produced in biofilms formed by the CBS strain (Table 1) was quantified by ELISA. The procedures were similar to those described above except that, after blocking solution removal, 100 μl per well was added of 75 ng/ml wheat germ agglutinin (WGA)-HRP conjugate, a lectin that binds to PIA sugar residues. The Abs₄₉₂ was measured and related to OD_G. The PIA produced in biofilms formed by the Newman and the BRZ (Table 1) strains was quantified by fluorometry using 100 μl per well of 90 μg/ml WGA-Oregon Green 488® conjugate (Molecular Probes). After 15 min of incubation of the plate in the dark, excess amounts of stain were removed and wells were washed twice with PBS. After adding 100 μl of PBS per well, the fluorescence was measured with a Fluorometer FLx800 (Biotek Instruments Inc., Winooski, VT, USA). Oregon Green was excited at 485 nm and its emission was detected at 528 nm. The level of PIA was expressed relative to the final culture density measured (PIA F/OD_G). For all assays, four independent experiments were performed in sixtuplicate.

S. aureus biofilms were grown in 75-cm² polystyrene tissue (T75) culture flasks at 37°C statically for 6 or 24 h in TSBg or TSBg with SAL and FeSO₄, when required. Flasks were chosen for culture in order to make available a large surface area to support biofilm growth. After static incubation for 6 or 24 h the supernatants were removed from each flask. Biofilm cells were scraped, transferred to an Eppendorf tube and enzymatically disrupted (100 μl of buffer Tris-EDTA 10:1, 40 μl of 1 mg/ml lysostaphin, 20 μl of 50 mg/ml lysozyme). Then, the lysates were exposed to Trizol Reagent® (Invitrogen) and the bacterial RNA extracted according to the manufacturer's protocol.

Bacterial RNA from 6 to 24 h biofilms was obtained as described above. After DNAse treatment using the RQ1 RNAse free DNAse (Promega, Madison, WI, USA), cDNA synthesis was performed with the ImProm-II™ Reverse Transcriptase kit (Promega). qRT-PCR was performed using the HOT FIREPolEvaGreen® qPCR Mix Plus (ROX) (Solis Biodyne, Tartu, Estonia) and Applied Biosystems 7500 instrumental, using the primers detailed in Supplemental material. The gyrB gene was used to normalize data. The number of copies of each sample transcript was determined with the aid of the 7500 system SDS software (Applied Biosystems, Carlsbad, CA, USA). The 2^−ΔΔCt value represents the difference in threshold cycle (Ct) between the target and control (gyrB) genes treated with SAL or SAL with Fe²⁺, minus the difference in Ct between untreated (TSBg) target and control genes (Alvarez et al., 2010).

Protein extracts were prepared from S. aureus biofilm grown in T75 culture flasks containing 40 ml of TSBg in the presence or absence of SAL at 37°C under static conditions for 24 h. The biofilms were scrapped and centrifuged at 10,000 × g for 20 min. Pellets were suspended in 0.6 ml of lysis buffer [30% raffinose in 0.05 M Tris (pH7.5) with 0.145 M NaCl] containing 100 μg/ml lysostaphin and 1 mM PMSF (phenyl-methyl-sulfonil-fluoride) and incubated 1 h at 37°C at 200 r.p.m. The suspension was centrifuged twice at 8000 × g for 15 min at 4°C and supernatants were saved for analysis. The protein concentration of each sample was measured using the Bradford method. An equal volume of 6 × Laemmli sample buffer was added to the protein extracts prior to boiling them for 3 min and separating them by SDS-10% PAGE. SDS-PAGEs were stained with Coomassie blue.

Selected bands were excised from the SDS-polyacrylamide gel. After destaining, the samples were reduced, alkylated and hydrolyzed with 25 ng/μl of trypsin. The samples were analyzed with a MALDI-TOF spectrometer, Ultraflex II (Bruker Daltonics, Billerca, MA, USA) at the Mass Spectrometry Facility (CEQUIBIEM, Buenos Aires, Argentina). The criteria for protein identification were based on the manufacturer's definitions. The protein score is −10 × log p, where p is the probability that the observed match is a random event. Protein scores higher than 81 were considered significant (p < 0.05).

Aconitase activity was quantified according to the method described by Kennedy et al. (1983). Biofilm of Newman strain grown during 24 h with or without 2 mM of SAL or SAL plus 50 μM of FeSO₄, were suspended in PBS and centrifuged at 9300 × g for 5 min and suspended in lysis buffer containing 90 mMTris-HCl (pH 8.0), 100 μM fluorocitrate and 100 μg/ml lysostaphin. The suspensions were incubated for 15 min at 37°C and centrifuged at 21,000 × g for 30 min at 4°C. A 20 μl aliquot of the cell lysate was added to 180 μl assay buffer [100 mM Tris-HCl (pH 8.0), 50 mM trisodium citrate] and incubated for 5 min at 37°C. The amount of aconitate produced was quantified by measuring the Abs₂₄₀. A molar absorption coefficient of 3.6 mM⁻¹ cm⁻¹ was used and 1 U aconitase was defined as the enzyme activity that catalyzes the formation of 1 μm aconitate per min. Protein concentrations were determined by the Bradford assay.

Supernatants from biofilms grown for 24 h were recovered by centrifugation. The concentration of lactate, glucose and acetate were determined using kits purchased from Sigma-Aldrich and following the manufacturer‘s indications. The index of lactate (IL) production by glucose consumption was defined as IL = nm lactate_TSBg/mg glucose_TSBg. The lactate concentration produced by biofilms treated with SAL or SAL plus FeSO₄ was determined as: [lactate]_Treated = IL × [glucose]_Treated.

pH was measured using an Adwa AD12 pH meter (Szeged, Hungary) on supernatants from S. aureus cultured up to exponential phase in TSBg with or without SAL. The iron released from human transferrin was determined as previously described (Friedman et al., 2006). The transferrin-Fe complex concentration was measured at Abs₄₅₀ every 1 min for 15 min upon addition of 40 μM human transferrin to all samples.

Eight-week-old male CF1 outbred mice weighing ~25 g were obtained and maintained at the vivarium of the Instituto de Investigaciones en Microbiología y Parasitología Médica (Universidad de Buenos Aires-CONICET) in accordance with the international guidelines set forth. The experimental procedures were evaluated by the Institutional Animal Care and Use Committee (CICUAL) and approved by resolution N° 901/16 of the School of Medicine, University of Buenos Aires. Thirty minutes before the bacterial challenge, mice randomly split into 2 groups of 5–6 animals received by the intravenous route 200 μl of 2 mM SAL or PBS (control), respectively. Then, a suspension containing an inoculum of 10 μl of 1.6 × 10⁷ CFU of the S. aureus Newman strain was pipetted slowly into the mouse nares. After 4 h, mice were euthanized with CO₂. The area around the nasal region was wiped out with 70% ethanol, and the nose was excised and homogenized in 400 μl TSB using a tissue grinder. Dilutions of the tissue homogenate were plated onto TSA plates for CFU counting (Barbagelata et al., 2011).

Nonparametric data were analyzed with the Mann-Whitney test. Data with normal distribution were compared with the paired t-test, using the Graphpad Software (version 6.0; GraphPad Prism). P-values < 0.05 were considered significant.

Ferrozine assays were conducted to determine the SAL chelating capacity in TSBg medium. As shown the Figure 1, a depletion of 23% of free Fe²⁺ was observed at a concentration of 2 mM of SAL. Increased concentrations of SAL provoked a constant decrease (~80%) of available free Fe²⁺ reaching a plateau at ~12.5 mM. Although 5 mM of SAL depleted the 49% of free Fe²⁺ ions, the concentration of 2 mM of SAL was chosen in this study because this is the serum concentration normally achieved after ingestion of aspirin at doses within the therapeutic range. Furthermore, the concentration of 5 mM produces toxic effects. Similar results were obtained by measuring the total free iron present in TSBg (39.3 ± 7.1 μM) and TSBg with 5 mM of SAL (21.4 ± 5.3 μM) by atomic absorption spectroscopy (Analytical Chemistry Facility, School of Biochemistry and Pharmacy, Univ. Buenos Aires). Interestingly, the intracellular free Fe²⁺ concentration in S. aureus forming biofilms when grown in TSBg with 2 mM of SAL was 27% lower than that obtained in the control without SAL addition, and similar to that observed in TSBg treated with Chelex-100 (Figure 2). Therefore, 2 mM of SAL decreases the free iron content at a similar rate in both intracellular and extracellular environments.

Initially, Newman and CBS (Table 1) strains were chosen to be treated with different SAL concentrations to establish the biofilm formation at various growth times (Figure 3). The biofilms of both strains grown in the presence of 0.36 or 2 mM of SAL showed a significant increase at both time points (24 or 48 h) when compared with those observed in the control groups. Then, S. aureus clinical isolates related to the USA300 and USA100 clones (Table 1) were selected to determine the chelating effect of SAL on biofilm formation since different strains may respond dissimilarly to available iron. Both the MRSA and the MSSA strains showed increased biofilm production under 2 mM SAL exposure whereas iron addition provoked a significant reduction of the biomass (Figure 4). The values of SAL-biofilm induction ranged from 1.25 (AR94) to 3.87 (AR66) for the strains studied. Values of the biomass relative to the Newman strain treated with SAL were as follows: CBS (73%), BRZ (43%), AR94 (54%), AR48 (95%), AR66 (109%), AR56 (60%), AR70 (93%), and AR83 (68%). It should be noted that addition of iron to TSBg did not affect the biofilm formation by S. aureus (Figure S1). Furthermore, the biofilms formed by the Newman strain in the presence of SAL were stained with the LIVE/DEAD BacLight Bacterial Viability kit and visualized by CLSM. The COMSTAT analyses of images from live bacteria with intact cell membranes acquiring green fluorescence established that exposure to 2 mM of SAL provoked the increment in the biomass (1.308 μm³/μm²) and the biofilm maximum thickness (3.356 μm) when these were compared with the controls (biomass: 0.636 μm³/μm² and maximum thickness: 2.666 μm). Addition of iron to SAL containing medium diminished both the biomass (0.688 μm³/μm²) and biofilm maximum thickness (2.112 μm) (Figure 5). The dead-bacteria (red fluorescence) biomass values were similar among the samples studied (control: 0.540 μm³/μm²; SAL: 0.529 μm³/μm²; SAL plus Fe²⁺: 0.485 μm³/μm²). The MTT colorimetric assay was used to quantify the viable bacteria in biofilms following each treatment. As shown in Figure S2, SAL and SAL plus iron treatments did not modify the levels of metabolically active bacterial cells forming biofilm. Taken together all these results suggest that diminution of [Fe²⁺] caused by SAL induced an increment of biofilm formation by S. aureus regardless of the methicillin susceptibility or clonal genomic characteristics.

S. aureus synthesize mainly two polysaccharides, CP and PIA. In our previous work it was demonstrated that CP expression is reduced when planktonic S. aureus is grown in the presence of low concentrations of SAL (Alvarez et al., 2010). Thus, to establish the effect of iron chelation by SAL on S. aureus surface polysaccharides, chemometric-assisted FTIR spectroscopic measurements were performed from S. aureus planktonic cultures and biofilms. The resulting FTIR spectra provided highly specific molecular fingerprints derived from stretching and bending vibrations of all functional groups of the bacterial cell or biofilm matrix (Helm et al., 1991). Based on their specific biochemical constituents, FTIR spectra can be subdivided into several spectral partitions including fatty acid and phosphorus-containing biomolecules of membrane components (e.g., phospholipids), proteins of the bacterial cell and surface-associated polysaccharides. The latter spectral window (1200–800 cm⁻¹) was used for the PCA, an unsupervised multivariate statistical method, to investigate the discriminatory features of the SAL treatment on S. aureus polysaccharide production. The score plot revealed a clear clustering of spectral data according to the treatment conditions (untreated, SAL-treated and SAL plus Fe²⁺). As shown in Figure 6A, SAL treatment of planktonic cultures of the Newman strain caused a perturbation of the surface polysaccharide composition that was not reverted by iron addition. In the light of our previous work and the results described herein it is suggested that CP instead of PIA expression is affected by SAL treatment due to the fact that CP production under planktonic culture conditions is limited (Alvarez et al., 2010). In contrast to planktonic growth, PCA of the Newman biofilms revealed that iron limitation by SAL exposure induced polysaccharide production alterations that were partially reverted by iron addition (Figure 6B). Indeed, ~25% decrease of [Fe²⁺] by SAL chelation is enough to alter the expression of polysaccharides when S. aureus is forming biofilm. In fact, S. aureus did not produce CP5 when bacteria are forming biofilms in any of the conditions investigated (Figure 7). S. aureus CP5-deficient Reynolds strain was utilized as negative control of CP expression. Taken together the results suggest that the polysaccharide perturbations induced by SAL in biofilms -which occurred mainly in PIA rather than in CP5 and in potentially other glycopolymers structures (e.g., LTA, WTA)- might be affected by SAL under both growth conditions.

S. aureus biofilm formation in the presence of SAL was visualized by SEM. Biofilm images revealed that SAL caused more aggregation of Newman cells and that these cells produced more extracellular substances compared with the cells in the control biofilm, which exhibited distinct images of bacteria with sharp contours. The addition of iron to the culture medium with SAL diminished the enclosing material of the bacterial cells (Figure 8A). To gain additional insight into the surface polysaccharides affected by exposure to SAL, the PIA from biofilms was visualized by CLSM after staining with green fluorescent labeled WGA and the red lipophilic membrane dye. As shown in Figure 8B, according to the high contents of green (PIA) and yellow staining (PIA and cell colocalization), SAL exposure of Newman biofilms would induce the PIA production. Interestingly, addition of iron to the culture medium containing SAL resulted in less green and yellow staining sectors suggesting the presence of minor contents of PIA in the extracellular matrices. Moreover, in the presence of iron, the biomasses induced by SAL did not differ from those of the control. To support these evidences S. aureus biofilms were treated with Dispersin B, a β-hexosaminidase that degrades carbohydrates as the PIA. The S. aureus SA113 strain was utilized as reference because PIA is the major extracellular component when this strain adopts the biofilm lifestyle (Di Poto et al., 2009). The enzymatic treatment provoked a significant detachment of the Newman and CBS biofilms grown in the presence of SAL (Figure 8C). Furthermore, the quantity of PIA expressed in the biofilms by Newman, BRZ and CBS strains increased with SAL exposure and significantly diminished with iron addition (Figure 9). These results indicate that exposure to SAL induced PIA-dependent biofilm, which is formed by both MSSA and MRSA strains.

Iron limitation has deep impact on the bacterial metabolism. Since certain enzymes of the TCA cycle, such as aconitase (citB), contain iron-sulfur binding clusters, and because decreased TCA activity in S. epidermidis is associated with high levels of PIA expression (Vuong et al., 2005), it can be speculated that the elevated PIA production observed in S. aureus biofilms treated with SAL may be associated to reduced TCA activity due to the intracellular iron limitation induced by SAL. Initially, the analyses at the transcriptional level showed that SAL decreased the transcription of citB in mature (24 h) biofilms and the relative citB expression increased slightly by addition of iron (Figure 10A). Indeed, the fold change of fur expression was low in the presence of SAL and did not suffer modifications by iron addition in the biofilms grown during 6 or 24 h (Figure 10A). The citB gene expression data were reinforced by functional assays. Aconitase enzymatic activity was measured in biofilms grown on the conditions under study during 24 h. As shown the Figure 10B, SAL exposure significantly decreased the aconitase activity and the addition of iron restored the activity of the enzyme to levels similar to the control. Therefore, the results obtained suggest that iron limitation by SAL negatively affects the activity of the TCA cycle of bacteria forming biofilm.

Since the production of a polysaccharide requires high expenditure of energy and the SAL diminished the TCA cycle activity, we hypothesized that the fermentative pathway would be preferential in bacteria forming biofilm in the presence of SAL. Total RNA extracted from immature and mature biofilms grown in the presence or absence of SAL was utilized to assess the transcriptional levels of enzyme indicators of glycolysis (pyruvate kinase gene pykA), gluconeogenesis (glyceraldehyde-3-phosphate dehydrogenase gene gapB), lactic fermentation (lactate dehydrogenase 1 gene ldh1), and UDP-glucosamine precursor (phosphoglucosamine mutase gene glmM) by qRT-PCR assays. Indeed, the relative level of glmM expression was increased significantly by SAL in immature biofilms (fold change: 1.275 ± 0.09) when compared with that of mature biofilms (fold change: 0.471 ± 0.02) (Figure 10C). The presence of SAL did not increase significantly the transcription of any of the other genes studied. Moreover, SAL reinforced the down-fold change of gapB expression in mature biofilms (Figure 10C). Figure 10D depicts the SDS-PAGE profile of cell wall proteins extracted from mature biofilms treated or not treated with SAL. The results revealed that certain bands were intensified by SAL treatment of biofilm formed by the Newman strain. Each one of these bands were analyzed by MS/MS (MALDI-TOF) and the proteins L- and D-lactate dehydrogenases (Ldh) (score: 134 and 253, respectively; p < 0.05) and acetoin reductase (ButA) (score: 86, p < 0.05) were identified. The acetate, glucose and L-lactate concentrations in supernatants of mature biofilms were also evaluated. There was no acetate accumulation in supernatants of biofilms grown during over 24 h in any of the experimental conditions under study (data not shown). Since 2 mM of SAL causes a slight growth delay of the Newman strain in TSBg (Figure S3) and, consequently, the glucose consumption was diminished, the index of L-lactate (IL) production over glucose consumption (IL = nmol lactate_TSBg/mg glucose_TSBg) in biofilms formed during 24 h was determined. The IL of the Newman strain biofilm was 855 nmol/mg. The amount of L-lactate produced by the Newman strain biofilm treated with SAL was high ([lactate]_SAL = 855 nmol/mg × [glucose]_SAL = 1393.6 nmol) whereas iron addition slightly diminished it ([lactate]_SAL+Fe = 855 nmol/mg × [glucose]_SAL+Fe = 1308 nmol). Finally, SAL exposure of S. aureus during growth diminished significantly the extracellular pH values (p < 0.01, t-test correct for multiple comparisons using Holm-Sidak method) thus promoting the release of iron from human transferrin (Figure 11). The results suggest that SAL only stimulates the lactic fermentation pathway since the acetate levels were undetectable and the Ldh expression was enhanced by SAL.

The main S. aureus polysaccharides coded by the cap and ica operons can be repressed by CodY (Majerczyk et al., 2008, 2010; Thoendel et al., 2011). Hence, the effect of SAL on CodY was investigated. To this purpose, biofilms produced by the Newman codY deficient mutant grown in the presence of SAL or SAL plus iron were subjected to FTIR spectroscopic analysis. As shown in Figure 12A, the PCA revealed remarkable differences of qualitative and quantitative spectral features in the polysaccharide region (1200–800 cm⁻¹) between the wild-type, the codY mutant and the wild-type grown in the presence of SAL. Interestingly, the effect of codY on the surface polysaccharide composition was partially reversed by SAL as well as SAL plus iron. This finding was confirmed by assessment of spectral distance values, a quantitative measure of dissimilarity corresponding to the non-overlapping areas of the spectra derived from the wild-type strain. These values were: (i) wild-type (SAL), 0.76; (ii) wild-type (SAL + Fe²⁺), 0.35; (iii) codY mutant, 0.35; (iv) codY mutant (SAL), 0.18; and (v) codY mutant (SAL+ Fe²⁺), 0.21. Therefore, S. aureus surface polysaccharides were affected by SAL in a codY-dependent manner. The amount of PIA expressed in the codY mutant biofilms significantly increased with SAL exposure as shown in Figure 12B. Furthermore, the SAL-induced increase of the PIA produced by codY mutant biofilms (1.7-fold) was lower than that observed in the biofilms formed by Newman (3-fold). When iron was added to the culture medium containing SAL, the level of PIA expressed by the codY mutant biofilms was not significantly different when compared with that determined in culture medium with SAL and without iron. Conversely, the amount of PIA in Newman biofilms formed in the presence of SAL was diminished by iron addition (Figure 12B).

The role of CodY on the SAL-induced biofilm formation was evaluated. In spite of the low biofilm formation by the codY mutant when compared with the parental strain, SAL treatment increased significantly the amount of biofilm formed by the codY mutant (Figure 13A). In contrast to the effect observed in the Newman strain, addition of iron to SAL-containing medium increased the biofilm production by the codY mutant when compared with that of biofilm grown in the presence of SAL, without iron (Figure 13A).

In order to study the interplay between codY and SAL, the levels of codY, icaA, and capK5 transcripts from immature (6 h) and mature (24 h) biofilms formed by the Newman strain were evaluated by qRT-PCR (Figure 13B). As expected, SAL exposure decreased the transcription of cap5K in biofilms growing at 6 and 24 h and iron addition did not revert these decreases. On the other hand, the presence of SAL in mature biofilms relieved the decrease of icaA expression when biofilms were grown with SAL during 6 h. Consistent with this finding, the relative codY expression was significantly reduced in mature biofilms exposed to SAL. When iron was added, the decrease of icaA and codY expression in mature biofilms remained below the control values. Therefore, SAL exposure seems to relieve CodY repression of ica by diminishing codY transcript levels in mature biofilms.

The initial step in biofilm formation is the adhesion of bacteria to cells or inert surfaces. Because formation of S. aureus biofilm was demonstrated in vivo in both human and murine nasal mucosa (Zernotti et al., 2010; Muthukrishnan et al., 2011; Reddinger et al., 2016) and since PIA contributes significantly to the adherence of S. aureus to nasal epithelial cells (Lin et al., 2015) we investigated the effect of SAL in vivo using a nasal colonization murine model. Groups of mice received 200 μl of 2 mM of SAL or PBS (control) by the intravenous route 30 min before bacterial inoculum. The animals were then inoculated by the intranasal route with 10 μl of a suspension containing ~1.6 × 10⁷ CFU of the S. aureus Newman strain. Mice were sacrificed at 4 h after intranasal inoculation and the number of CFU in the nostrils was determined. The CFU number recovered from SAL-treated mice (median: 1.8 × 10⁴ CFU/nose) was significantly higher than that of the control group (median: 7.5 × 10³ CFU/nose) (p = 0.009, Mann-Whitney test) (Figure 14). These results show that SAL treatment induced a significant increase of S. aureus nasal colonization in mice.

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