During the first stages of biofilm assembly in implant medical devices, bacterial cells are exposed to toxic micromolar concentrations of NO produced by cells of innate immunity (Arora et al., 2015; Rinaldo et al., 2018). To elucidate how S. epidermidis copes metabolically with this stress, we determined the end-products excreted from the glucose metabolism ( Figure 1 ) of cells forming biofilms under NO stress. We used strains 1457 and RP62A, which are described as strong biofilm producers ( Table S1 ), grown under static conditions in 24-well plates in high glucose- and fetal bovine serum (FBS)-containing Dulbecco’s modified Eagle medium (DMEM) in the presence of the NO releaser DETANONOate. We analyzed the supernatants of biofilms formed after 24 and 48 h by ¹H-NMR to evaluate substrate consumption and metabolic end-products. The time considered for biofilm formation, i.e., 24 and 48 h, was chosen based on the following rationale: biofilms at 24 h allow observation of the end-products that are accumulated during early-stage biofilm formation, and at 48 h, a mature biofilm is present. DETANONOate is a slow NO donor releaser with a half-life of 20 h, at 37°C and pH 7, which was used at 1 mM concentration so that small micromolar quantities of NO are released over time of biofilm development. We started by evaluating the NO release from DETANONOate through the determination of the amount of nitrite accumulated in the supernatant of biofilms formed after 24 and 48 h ( Figure S1 ). For simplicity, we hereafter designate NO stress as the stress generated by adding DETANONOate. Our data show that DETANONOate (1 mM) leads to a significant decrease in the amount of biofilm mass formed by both strains of S. epidermidis ( Figures 2A , S2 ). We observed that NO also interferes with the viability of biofilm-encased cells ( Figure 2B ). For strain 1457, in mean terms, a 2.5-fold lower biofilm viability was always observed comparatively to the control without NO, i.e., independently of the age of the biofilm. In strain RP62A, the NO-exposed biofilm-encased cells showed 6- and 9.5-fold lower viabilities than those non-exposed, at 24 and 48 h, respectively ( Figure 2B ).

Metabolically, untreated biofilms of strains 1457 and RP62A exhibited after 24 h a homolactic behavior, with lactate (l-lactate+d-lactate) as the major end-product reaching concentrations of 40 ± 2 and 36 ± 0 mM, which account for approximately 80% and 70% of the glucose consumed, respectively ( Figures 1 , 3A–D , left panels; Table 1 ). Acetate is also formed but in lower amounts (3–5 mM) ( Figures 1 , 3E, F , left panels; Table 1 ), together with minor quantities (µM) of formate and succinate (data not shown). It should be noted that S. epidermidis encodes two lactate dehydrogenases, the l- and d-lactate dehydrogenases (l-Ldh and d-Ldh), which produce lactate. In this study, the lactate measured is the sum of l-lactate and d-lactate, as they cannot be distinguished by conventional ¹H-NMR methodology.

NO caused a significant increase in lactate production (~ 20% for both strains), which accounts for approximately 90% and 94% of the glucose consumed in strains 1457 and RP62A, respectively ( Figures 1 , 3A–D , left panels; Table 1 ), denoting a typically fermentative metabolism. The yield and amount of acetate, which accounts for 7% (strain 1457) and 9% (strain RP62A) of the glucose consumed in untreated biofilms, were significantly decreased by NO by approximately twofold in both strains ( Figures 1 , 3E, F , left panels; Table 1 ). The formate and succinate, which are present in the control, were formed in negligible amounts in the NO-exposed cells (data not shown). Interestingly, the strains exposed to NO accumulated a compound, which is absent in non-exposed cells, that resonates in the ¹H-NMR spectrum as a doublet of triplets between 2.8 and 3.2 ppm ( Figure S3 ), most probably due to the formation of a nitrosylated compound, that was not possible to identify by two-dimensional NMR.

The synthesis of ATP by fermentation relies on substrate-level phosphorylation. Although NO caused a significant increase in the lactate amount, the ATP yields obtained by substrate-level phosphorylation were only slightly increased in strain RP62A and were not modified in strain 1457 ( Table 1 ). These results indicate that the decrease in ATP/acetate production in NO-stressed biofilms was counterbalanced by the increase in NAD⁺/lactate production ( Figures 1 , 3A–F , left panels; Table 1 ). Moreover, the carbon recovery and redox balance of NO-treated biofilms were 5% and 10%–15% higher than in 1457- and RP62A-untreated biofilms, respectively ( Table 1 ), showing deviation of carbon from other biochemical processes to the formation of higher amounts of end-products, i.e., lactate/NAD⁺ ( Figure 1 ).

Analysis of the supernatants of biofilms formed after 48 h showed a decrease of the lactate produced when compared to those formed after 24 h, which was independent of the exposure to NO, thus suggesting that lactate is used as a carbon source after glucose is exhausted from the medium ( Figures 3A–D , left panels). The differences in the end-products profile observed at 24 h between NO-treated and untreated biofilms were maintained at 48 h ( Figures 3A–F , left panels).

Interestingly, while the biomass (OD₆₀₀) and the number of viable non-adherent cells in the medium covering the biofilms of strain RP62A decrease in the presence of NO, which is an effect that also occurs under planktonic conditions (data not shown), in strain 1457, the free cell biomass increased at 24 h under NO stress ( Figures 4A, B ), suggesting that in this strain, free cell metabolism may contribute to higher lactate yield. Notably, strains 1457 and RP62A grown planktonically, i.e., as free cells, in the same medium used for the biofilms and in the presence of 1 mM of NO did not show higher lactate amounts, indicating that the higher lactate production caused by NO is biofilm-specific ( Figure 4C ). Moreover, the data in strain 1457 suggest that NO promotes the release of cells from the biofilm and/or impairs their attachment to the biofilm, favoring non-adherent planktonic growth of a population of cells.

To investigate whether the effect of NO on lactate production was associated with strong-biofilm-producing strains, S. epidermidis 1457-M12, which is a mutant of strain 1457 described as a low PIA and biofilm-producing strain (Mack et al., 2000) ( Table S1 ), hereafter designated as M12, was analyzed under NO stress. We observed that strain M12 produced significantly lower biofilm mass than strains 1457 (p-value < 0.0001 at 24 h and < 0.01 at 48 h) and RP62A (p-value < 0.0001 at 24 h and < 0.0001 at 48 h, Welch’s t-tests) ( Figure 2A ). Moreover, S. epidermidis M12 exhibited high lactate yields and a carbon recovery of approximately 100% ( Table 1 ), suggesting that the production of lactate occurred at its maximal. Interestingly, NO did not affect the lactate yield and biofilm mass, contrary to what was observed for strains 1457 and RP62A ( Figures 2A ; S4A–C , Table 1 ). However, in strain M12, the acetate yield also decreased in NO-treated cells ( Figure S4C , Table 1 ). Therefore, in strains 1457 and RP62A under NO stress, the higher lactate concentration measured during biofilm formation most likely reflects the NO inhibition of enzymes that act downstream of pyruvate (e.g., pyruvate-formate lyase and pyruvate dehydrogenase) (Richardson et al., 2008), leading to acetate production, and the higher activity of Ldh that deviates carbon to the production of lactate ( Figure 1 ).

Strains 1457 and RP62A grown in Tryptic Soy Broth (TSB), a low glucose- and high peptide-containing non-physiological medium with unknown chemical defined composition and widely used in in vitro experiments with staphylococci, were exposed to 1 mM of NO. The formation of biofilms with lower mass and viability and with higher lactate production ( Figure S5 ) was also observed, which indicates that the effect of NO on the biofilm characteristics is independent of the medium composition.

Thus, we conclude that NO increases lactate production in S. epidermidis biofilms, and this increase seems to be characteristic of strongly biofilm-producing strains.

As S. epidermidis biofilms exposed to NO showed higher lactate accumulation, we investigated whether this increase could be related to the induced expression of ldh gene encoding the l-lactate dehydrogenase enzyme ( Figure 1 ). l-Ldh has been linked to drug resistance and is considered a possible target of regulatory control in many bacterial species, including in coagulase-negative staphylococci (Richardson et al., 2008; Toyoda and Inui, 2021; Ma et al., 2022; Yuan et al., 2022). For this purpose, biofilms of strains 1457 and RP62A, treated with NO or left untreated and grown for 24 h, were used to extract total RNA and evaluate ldh expression by RT-qPCR. Gene expression analysis of ldh in untreated RP62A and 1457 biofilm cells indicated a modest twofold higher expression in RP62A relative to 1457 cells ( Figure 5A ). However, a substantial approximately 50-fold higher expression of ldh in RP62A relative to 1457 was determined in NO-exposed biofilms ( Figure 5A ). Moreover, while in 1457 biofilm cells NO caused no significant variation in ldh expression, RP62A biofilms exposed to NO expressed ldh approximately 120-fold more than in untreated cells ( Figure 5B ). Importantly, the biofilm amounts and viabilities of 1457 and RP62A strains grown for 24 h with NO did not show significant differences between them ( Figure 2 ).

Altogether, our results show that for biofilms at the same growth stage (after ~ 24 h), NO only induces the expression of ldh in strain RP62A, indicating the occurrence of different regulatory mechanisms regarding the production of lactate between RP62A and 1457.

The role of lactate dehydrogenase in NO resistance of S. epidermidis biofilms was further confirmed in assays performed in the presence of oxamate, which is a pyruvate analog that inhibits the conversion of pyruvate to lactate (Altinoz and Ozpinar, 2022). For this purpose, biofilms were exposed to NO, as described above, in the presence of oxamate, and supernatants were analyzed by ¹H-NMR to determine substrate consumption and metabolic end-products. Additionally, biofilm mass was measured by the crystal violet assay.

The biofilm mass of strain 1457 decreased in the presence of oxamate by approximately 34% ( Figure 2A ). Exposure to NO further lowered the biofilm mass formed at 24 h ( Figure 2A ). Oxamate also caused a reduction in viability of the biofilm-encased cells of strain 1457 by approximately 13-fold, which was further accentuated when the cells were under NO stress, resulting in a viability that was approximately 26-fold lower ( Figure 2B ). However, at 48 h, the pattern observed in strain 1457 changed, as oxamate addition to NO-treated biofilms caused a small increase in biofilm mass and a significant increase (approximately five times) in the viability of the encased cells ( Figure 2 ).

Regarding RP62A cells, oxamate decreased the amount of biofilm mass formed by approximately 40%–46% regardless of the presence of NO and the age of the biofilm ( Figure 2A ). However, at 48 h, the effect of NO became noticeable, as the biofilm mass of cells treated with oxamate was greater when NO was present ( Figure 2A ), suggesting that strain RP62A undergoes modifications in the quantity of the biofilm matrix. In terms of viability, RP62A cells treated with oxamate and NO showed lower colony-forming unit (CFU) counts compared to cells not exposed to NO after 24 h, but this effect was not observed in 48-h biofilms ( Figure 2B ).

Metabolically, biofilms of strain 1457 grown for 24 h in the presence of oxamate denoted a substantial decrease of approximately 44% in consumed glucose and lactate levels (from 40 ± 2 to 8 ± 0 mM) and yields (61% less), relative to untreated biofilms ( Figures 3A, C ; Table 1 ). A significant shift toward acetate formation was also observed ( Figure 3E , right panel; Table 1 ). Lactate and acetate accounted for approximately 31% and 22% of the glucose consumed, respectively ( Figures 3A, C, E right panel; Table 1 ; Figure 1 ). In the RP62A strain, and unlike the 1457 strain, oxamate did not alter glucose consumption, and the decrease in lactate level (from 36 ± 0 to 24 ± 1 mM) and yield (1.5-fold lower) was not so drastic. Concerning acetate, the compound reached a concentration of 9 ± 1 mM, which was twofold higher compared to the control ( Figures 3B, D, F ; Table 1 ). Lactate and acetate accounted for approximately 46% and 17% of the glucose consumed, respectively ( Figures 3B, D, F , right panel; Table 1 ). These results suggest a different regulatory mechanism for the production of lactate and recycling of NADH back to NAD⁺ ( Figure 1 ) that allows strain RP62A to continue glycolysis at a higher extent, which does not occur in strain 1457. The effect of oxamate on lactate production of strains 1457 and RP62A is reflected in 15% lower ATP yields, relative to unexposed biofilms, despite the observed increase in acetate/ATP production ( Table 1 ).

In strain 1457, oxamate/NO-treated biofilms, grown for 24 h, exhibited a significant decrease of approximately 54% in consumed glucose relative to NO-only-treated biofilms that was approximately 10% higher than the decrease caused by oxamate relative to control ( Table 1 ; Figure 3C ). Also, oxamate lowered substantially the lactate yield (by approximately 44%) and increased by approximately fourfold the acetate yields in oxamate + NO versus NO biofilms ( Table 1 ; Figures 3A, C, E ). However, unlike in oxamate-only-treated biofilms, in oxamate/NO-treated biofilms, lactate and acetate accounted for approximately 53% and 12% of the glucose consumed, respectively ( Table 1 ), indicating that NO causes significant deviation of glucose to lactate production.

In biofilms of strain RP62A, no significant effect on glucose consumption was observed upon oxamate addition, but the lactate yield was 22% lower, denoting the negative effect of oxamate on lactate production (oxamate + NO versus NO in Figures 3B, D ; Table 1 ). Nevertheless, this effect was much less pronounced (35% lower) than that observed in 1457 biofilms ( Figures 3A, B ; Table 1 ). For acetate, the trend was the opposite, aligning with the stimulating effect of oxamate on acetate production ( Figure 3F ; Table 1 ).

The addition of oxamate to NO-exposed biofilms lowered the ATP yields of both 1457 and RP62A biofilms by 23% and 13%, respectively, indicating energy production impairment caused by oxamate ( Table 1 ).

In biofilms grown for 48 h, similar results were obtained when oxamate was added to NO-treated cells. The exception was the oxamate/NO-treated biofilms of strain 1457, as glucose consumption was not complete and ~ 40% lower than that observed in biofilms exposed to oxamate or NO alone ( Figure 3C ; Table 1 ). Moreover, the lactate amounts did not increase substantially from 24 to 48 h, contrary to oxamate-only-exposed biofilms, suggesting a reverse effect of NO on lactate production by Ldh in 1457 biofilms in the presence of oxamate ( Figure 3A ; Table 1 ).

Hence, we concluded that inhibition of lactate production of lactate dehydrogenase by the competitive inhibitor oxamate leads to significant impairment in biofilm mass formation of early biofilms (24 h) of S. epidermidis, treated and untreated with NO.

Biofilm mass and end-products amounts are modified by NO ( Figures 2 , 3 ). These results led us to surmise that the levels of metabolite intermediates linked to glycolysis, the amino sugar pathway, and the synthesis of PIA, which is known to be a substantial part of the matrix of several S. epidermidis biofilms ( Figure 1 ), are also affected by NO. Therefore, we used RT-qPCR to determine in biofilms the expression of genes of the amino sugar pathway (glmM and glmU) involved in the production of PIA precursors and those involved directly in PIA production (e.g., icaA). Additionally, we investigated the NO effect on staphylococcal biofilm matrix by combining confocal laser scanning microscopy (CLSM) and staining of the major biofilm matrix components, namely, exopolysaccharides/PIA and proteins, with the WGA and Sypro Ruby fluorophores, respectively. These assays allow the direct visualization of the localization of main matrix components and the determination of their amounts in biofilm aggregates.

In hypoxic (low oxygen) microenvironments, S. epidermidis can grow by fermentation or alternatively by nitrate/nitrite respiration when nitrate/nitrite is available. In the absence of oxygen or under low oxygen tensions, the staphylococcal oxygen-reducing terminal oxidases of the aerobic respiratory chain are replaced by nitrate and nitrite reductases (Nar and Nir), which are involved in the anaerobic dissimilatory nitrate and nitrite reduction ( Figure 1 ). During anaerobic nitrate and nitrite respiration, NO is formed from the activity of nitrite reductase on nitrite ( Figure 1 ). Since we observed that exogenous NO led to an increase in lactate production ( Figure 3 ; Table 1 ), we wondered what would be the effect of NO produced endogenously from nitrite reduction. We observed that during nitrate and nitrite respiration under hypoxic conditions, the biofilm formed by strains 1457 and RP62A decreased by 1.5- and 5-fold relative to the control (i.e., biofilms grown without nitrate or nitrite), respectively ( Figure 8A ). We determined the metabolism of 1457 and RP62A biofilms grown for 48 h under static conditions (control, Ctr) and observed that they consumed all the glucose available in the medium (26 mM) and excreted lactate (approximately 26–36 mM) as the major end-product ( Figure 8B ), which is consistent with a fermentative metabolism for carbon/energy production. Together with lactate, acetate was ubiquitously excreted, but in smaller amounts (approximately 4–7 mM) ( Figure 8B ), and minor quantities of ethanol, 2,3-butanediol, formate, and succinate were detected (data not shown).

When nitrate was used as the terminal electron acceptor, a significantly higher shift to production of acetate occurred (to approximately 20 mM) concomitantly with the decrease in lactate accumulation that reached a concentration of only approximately 20 mM ( Figure 8B ). Despite the synthesis of ATP that accompanies acetate production and the formation of electrochemical gradients generated from the electron transfer systems during anaerobic respiration ( Figure 1 ), which make nitrate respiration energetically more favorable than fermentation, the biofilm masses of the strains were lower than those formed under lactic fermentation ( Figure 8A ).

Strain 1457 had a different response when nitrite was used as an alternative electron acceptor, as the increase in acetate formation was modest relative to the control ( Figure 8B ). However, there was a significant increase in lactate excretion in both strains, compared to the control and nitrate-respiring biofilms ( Figure 8B ). These results indicate that nitrite is not an efficient electron acceptor of the respiratory chain as is nitrate; thus, unlike nitrate reduction, nitrite reduction was not able to efficiently couple to NADH oxidation at the level of the respiratory chain to generate a proton motive force and restore redox balance ( Figure 1 ). Instead, nitrite restores the redox balance at the level of lactic fermentation, an effect that is similar to that observed under fermentative conditions in the presence of NO. In this way, nitrite reduction deviates glucose to the production of lactate, which is detrimental to the production of acetate, a step where ATP, instead of NAD⁺, is produced.

The lactate production by S. epidermidis strains while exposed to murine macrophages J774A.1 and forming biofilms was analyzed. Since macrophages carry out lactic fermentation, lactate was also measured in macrophages cultured in the absence of bacteria to serve as the control. Production of NO by M1 macrophages and the inhibitory effect of iNOS were confirmed by nitrite determination in the macrophages’ supernatants ( Figure 9A ). As expected, macrophages not stimulated or inhibited in NO production, designated as MØ or MØ+l-NMMA, respectively, accumulate negligible amounts of nitrite (approximately 1.5–1.9 mM), while activated macrophages (aka M1/NO) accumulated 35–45 times more nitrite ( Figure 9A ). Inhibition of activated macrophages by l-NMMA (M1/NO+l-NMMA) resulted in non-significant values of nitrite formed, reflecting the non-production of NO ( Figure 9A ).

Analysis of the macrophage/bacteria co-culture showed that, although glucose consumption did not change, significantly more lactate is accumulated when the biofilms of strains 1457 and RP62A are formed (at 24 h) in the presence of activated macrophages (M1/NO) than when in contact with iNOS-inhibited macrophages (M1/NO+l-NMMA) ( Figures 9B, C , S8A, B ). Moreover, the levels of acetate and formate formed were negligible (data not shown).

Importantly, the increase in lactate in co-cultures of bacteria/activated macrophages M1 is directly related to NO production since there is no variation in lactate amount formed in co-cultures of bacteria/non-activated macrophages incubated with the l-NMMA inhibitor ( Figures 9B, C ). We concluded that bacteria are responsible for increasing lactate production, as observed in biofilms from co-cultures of bacteria/NO-producing macrophages ( Figures 9B, C ) in which l-NMMA did not affect the amount of lactate formed or glucose consumed in single cultures of NO-producing macrophages ( Figures 9D , S8C ).

Our data indicate that in S. epidermidis biofilms, the NO released from macrophages promotes a pronounced shift to lactic fermentation. Interestingly, in both staphylococcal strains, the amount of biofilm formed was not affected by the NO-producing macrophages ( Figures 9E, F ), suggesting that the increased lactate production could be one of the factors involved in the NO resistance of bacteria.

Strains of S. epidermidis used in this study, listed in Table S1 , were stored in 25% (v/v) glycerol (Merck, Darmstadt, Germany) at −80°C. Pre-inoculum cultures of S. epidermidis were routinely started at 1% (v/v), from −80°C stocks, in TSB (Difco, Franklin Lakes, NJ, USA) and grown to exponential phase (approximately 4–5 h) at 37°C and 150 rpm in Erlenmeyer flasks.

S. epidermidis biofilms were grown statically, at 37°C, in 2 ml of DMEM containing 4.5 g/L of glucose, 862 mg/L of l-alanylglutamine, 110 mg/L of sodium pyruvate, and 15 mg/L of phenol red (Gibco, Amarillo, TX, USA; ref. 31966021), supplemented with 10% heat-inactivated FBS (Gibco, ref. 10270106) on 24-well polystyrene tissue culture plates (Sarstedt, Nümbrecht, Germany). Where applicable, sodium nitrate and sodium nitrite were added at final concentrations of 40 and 5 mM, respectively. Alternatively, biofilms were grown on 96-well plates (Sarstedt) containing 200 µl of TSB. The media were inoculated with exponential cells of S. epidermidis to an OD₆₀₀ of 0.05 before the filling of the wells (3–4 wells per strain and condition). Following 2 h of incubation at 37°C, the cells (approximately 10⁷ CFUs/ml) were left untreated or exposed to the indicated concentration of DETANONOate (ACROS Organics, Geel, Belgium) and oxamic acid sodium salt (Alfa Aesar, Ward Hill, MA, USA), freshly dissolved in 0.01 M of NaOH and H₂O, respectively, and the plates were incubated longer for 22 or 46 h.

S. epidermidis planktonic cells were grown at 37°C and 150 rpm, in 20 ml of DMEM/FBS medium described above. The media were inoculated with exponential cells of S. epidermidis to an OD₆₀₀ of 0.05, and following 2 h of incubation at 37°C and 150 rpm, the cells were left untreated or exposed to 1 mM of DETANONOate.

S. epidermidis biofilms were grown for 24 and 48 h in 24- and 96-flat well plates in DMEM plus FBS and TSB, respectively, as described above. After the incubation periods, the media covering the biofilms of each well were removed, and whenever relevant, OD₆₀₀ measurements were performed to access free cell biomass. Subsequently, biofilms were heat-fixed at 65°C for at least 2 h and stained with crystal violet (CV), as follows: 2 ml (24-well plates) or 200 µl (96-well plates) of CV (1% (v/v), Merck) was added to wells followed by incubation at room temperature (RT), for 20 min. Biofilms were washed twice with 1 volume of PBS (1×), and images were acquired with an ImageScanner III LabScan 6.0 (600 dpi, GE Healthcare, Chicago, IL, USA). PowerPoint was used to enhance the visualization of the biofilms on the plates, by increasing the brightness by 40% and/or using the sharpen function at 50%. To elute the crystal violet, 1 volume of glacial acetic acid (33% (v/v), Carlo Erba, Milan, Italy) was then added. After 10-min incubation, the CV absorption was measured at 590 nm.

The viability of the biofilm-encased and biofilm-free cells was determined as follows. The medium covering the biofilms in each well was collected into microcentrifuge tubes and homogenized before serial dilutions and plating on Tryptic Soy Agar (TSA). Biofilms extracted from the 24- and 96-well plates were collected in microcentrifuge tubes containing 1 ml and 100 µl of PBS (1×), respectively. Biofilms were vortexed for 30 s, sonicated (Bioruptor^® Plus Sonication System) for 12 cycles (10 s ON, 30 s OFF, on high setting), vortexed again, and centrifuged at 11,337 × g for 1 min at RT. The resulting cell pellets were suspended in 2 ml or 200 µl of PBS (1×), depending if biofilms were grown on 24- or 96-well plates, respectively, and serial dilutions were cultured on TSA medium. The plated biofilm-encased and biofilm-free cells were incubated overnight, at 37°C, and CFUs per ml were counted.

When appropriate, nitrite quantification was performed using the Griess colorimetric method (Tarpey et al., 2004) using freshly prepared solutions. Griess solution (sulfanilamide (1% m/v, Sigma, St. Louis, MO, USA), N-(1-naphthyl)ethylenediamine dihydrochloride (0.1% m/v, Sigma), and H₃PO₄ (2% v/v, Merck) in Milli-Q^® water) was mixed with the media recovered after cell culture in a 1:1 concentration. The absorbance at 450 nm was measured and fitted to a calibration curve obtained with known concentrations of NaNO₂ (Merck).

Silicone catheter pieces (1 or 2 cm) were incubated in 6-well polystyrene tissue culture plates containing 5 ml of DMEM+FBS that had been inoculated with S. epidermidis strains to an OD₆₀₀ of approximately 0.05. After 2 h at 37°C, cells were exposed to 1 mM of DETANONOate and 5 mM of oxamate and incubated for 22 h. The images of the silicone sections were recorded with a digital camera. Then, the media were removed, and the plates were placed at 65°C for 2 h. The silicone pieces in the wells were submerged in 5 ml of 1% crystal violet for 20 min, washed and rinsed twice with PBS (1×), and air-dried. Images of the wells were acquired in an ImageScanner III LabScan 6.0 (600 dpi, GE Healthcare). The silicone tubes were then transferred to falcon tubes and destained by 10-min incubation with 5 ml of glacial acetic acid (33%), after which the CV absorption was measured at 590 nm.

S. epidermidis strains were inoculated in DMEM plus FBS medium (or TSB, when indicated) to form biofilms, as described above. Supernatants (100 µl) were collected immediately after inoculation, after 24 and 48 h of growth, and centrifuged at 1,000 × g for 1 min. Then, 200 µl of cold (−80°C) liquid chromatography (LC)-grade methanol (Merck) was added, and the mixtures were incubated in dry ice for 30 min. To achieve complete FBS precipitation, samples were centrifuged at 16,000 × g for 20 min at 4°C. The resultant supernatants, containing the substrate and excreted metabolites (end-products), were placed under nitrogen flux for approximately 2 h until all the methanol was evaporated. The precipitates were stored at −20°C until further analysis. All precipitates were suspended in 600 µl of D₂O, and samples were transferred to 5-mm NMR tubes for ¹H-NMR analysis. The spectra were acquired in a Bruker Avance II 500-MHz spectrometer (Bruker BioSpin, GmbH, Ettlingen, Germany) operated by TOPSPIN software and using a 5-mm BBIXYZ high-resolution probe head at 16°C to allow the shift of the water peak from the glucose resonances, as described before (Carvalho et al., 2017; Carvalho et al., 2019). To acquire spectra, the standard Bruker pulse program zgprde was used for water presaturation with 12-µs pulse width and a relaxation delay time of 3.5 s (d1 = 2.5 s, d20 = 1 s) for saturated spectra or with an extra relaxation delay time (d20) of 60 s to allow full spin relaxation. Spectra were phase- and baseline-corrected and referenced to the resonance of externally added trimethylsilylpropanoic acid (TSP), which was assigned to 0 ppm. Metabolite concentrations were calculated from the areas of the resonances in ¹H-NMR spectra by comparison to the area of the TSP resonance and after the application of an appropriate factor for correcting saturation of resonances and taking into account the number of protons, giving rise to the NMR signals/resonances.

Biofilms were suspended in RNA Later (Sigma R0901) and PBS (1×) and stored at −20°C prior to RNA extraction. To isolate total RNA, biofilms were first centrifuged at 6,200 × g, 4°C, for 3 min to remove the RNA Later and suspended in 5% cold phenol RNA protective solution and 0.9% NaCl. Following centrifugation at 6,200 × g for 10 min at 4°C, the pellets were resuspended in 10 mM Tris–1 mM EDTA in DEPC water (pH 8). Lysozyme (12.5 mg/ml), mutanolysin (0.25 KU/ml), and proteinase K (2.5 mg/ml), prepared in nuclease-free water, were added to the buffered pellets, and the mixtures were incubated for 2 h at 37°C. Complete lysis was achieved by the addition of 350 µl of lysis buffer from the Aurum™ Total RNA Mini Kit (Bio-Rad, Hercules, CA, USA). For the highest purity, whenever complete lysis was not attained, samples were centrifuged 10 min at 4°C and 16,200 × g to remove non-lysed cells. After the addition of 250 µl of 70% isopropanol, the lysates were transferred to the Aurum RNA Binding Mini Columns, and total RNA was extracted using the respective kit and following the manufacturer’s instructions. RNA samples were concentrated by the NaAcet/EtOH method. In brief, 0.1 vol of 3 M sodium acetate (NaAcet), pH 5.5, and 2.5 volumes of pure cold ethanol were added to samples that were incubated at −80°C for 3 h. After this step, samples were thawed and centrifuged at 16,200 × g for 30 min at 4°C. RNA pellets were washed twice (16,200 × g, 10 min, 4°C) with 2.5 volumes of 75% ethanol and air-dried at RT for approximately 15 min, and RNAs were suspended in nuclease-free water. RNA samples were purified from contaminating DNA by treatment with the Ambion TURBO DNA-free™ DNase Kit (Invitrogen, Carlsbad, CA, USA/Thermo Fisher Scientific, Waltham, MA, USA). The absence of chromosomal DNA was confirmed by PCR using the oligonucleotides 5′-GCTAATGCCTCGTCAATAC-3′ and 5′-TATGGTGCTGGAACAGATAC-3′ for S. epidermidis 16S gene. RNA purity, integrity, and concentration were evaluated by agarose gel electrophoresis and using a NanoDrop 2000c UV-Visible spectrophotometer (Thermo Scientific).

The cDNA was obtained by reverse transcription using approximately 3 μg of total RNA and the Transcriptor High Fidelity cDNA Synthesis kit (Roche, Basel, Switzerland). Samples for real-time RT-qPCRs were prepared with reagents of the Light Cycler 480 SYBR Green I Master Kit, with the primers listed in Table S2 , and reactions were performed in a Light Cycler 96 device (Roche). The expression ratios of the selected genes were normalized relative to the 16S rRNA gene of S. epidermidis, which does not change expression under the tested conditions. Ratios were calculated by the Livak and Schmittgen (Livak and Schmittgen, 2001) method.

Biofilms were grown for 24 h as described above in 24-well µ-plates (82426, IBIDI^®) to avoid background fluorescence in confocal microscopy. After removal of the media of each well, biofilms were stained with SYPRO Ruby (1×, FilmTracer™ SYPRO™ Ruby Biofilm Matrix stain, Invitrogen) and WGA (10 µg/ml, Wheat Germ Agglutinin, Oregon Green 488 Conjugate, Invitrogen) to label the biofilm matrix proteins and PNAG, respectively. The SYPRO Ruby/WGA mixture was incubated with the biofilms for 20  min in the dark, and the stained biofilms were washed one time with NaCl (0.9% m/v, Carl Roth, Karlsruhe, Germany). Confocal Z-series stacks were acquired in a Zeiss LSM 880 point scanning confocal microscope, using a 40× glycine immersion Plan-Apochromat lens (1.2 NA). Emission/excitation wavelengths were chosen according to fluorophores’ manufacturer instructions, and two separate virtual channels were used to avoid signal bleed-through. A minimum of three replicas per condition were performed, and in each well, five different-series images (73 × 73 µm) were taken in 0.5-µm slices. Image analysis was performed using BiofilmQ to determine fluorescence intensities and biofilm dimensions (Hartmann et al., 2021). The roughness of the biofilm matrices of individual confocal images of the biofilms was determined as mean  ( d ) N   . ∑ i N | mean  ( d ) − d i | , with N denoting the total number of pillars and di the thickness of pillar I, as determined by BiofilmQ.

Murine macrophages J774A.1 (LGC Promochem, Teddington, UK) were cultured in DMEM (Gibco, ref. 31966021); supplemented with 10% (v/v) of FBS (Gibco, ref. 10270106), 50 U/ml of penicillin and 50 μg/ml of streptomycin (all reagents from Gibco, Amarillo, Texas, USA), and incubated in 5% CO₂ atmosphere at 37°C. Macrophage cells per ml and viability were determined by resorting to an automatic cell counter (Countess 3, Invitrogen) and staining with 0.4% trypan blue (Sigma). Macrophages were seeded at 5 × 10⁵ cells/ml in wells of 24-well plates, incubated for 48 h, and then left inactivated or activated to the microbicidal M1 state by 5-h incubation with 5 μg/ml of lipopolysaccharide (LPS; Sigma) and 1 μg/ml of interferon-γ (IFN-γ; Sigma), in DMEM without antibiotics. Inhibition of iNOS was achieved by adding 800 µM of l-NMMA (N ^G-monomethyl-l-arginine acetate salt, Sigma). S. epidermidis 1457 and RP62A were grown aerobically in TSB to an OD₆₀₀ of 0.4–0.5, washed three times with PBS (1×), and resuspended in antibiotic-free DMEM medium to an initial bacterial OD₆₀₀ of 0.05 (approximately 10⁷ cells/ml). After the 5-h incubation period for the activation of l-NMMA-exposed and l-NMMA-unexposed macrophages, macrophages were washed three times with PBS (1×), and bacteria were used to infect macrophages at a multiplicity of infection (MOI) of approximately 5, and after 30 min, the extracellular bacteria were eliminated by a 10-min treatment with DMEM supplemented with 50 µg/ml of gentamicin. For the controls without bacteria, macrophages were not only infected but also subjected to treatment with gentamicin. These macrophages were then washed three times with PBS (1×), and fresh DMEM devoid of antibiotics (1 ml), with and without 800 µM of l-NMMA, was added followed by 24-h incubation. The infected macrophages were washed three times with PBS (1×), put in contact with 0.4-μm-pore-sized inserts, and inoculated with 1457 or RP62A to an OD₆₀₀ of 0.05. The volumes of antibiotic-free fresh DMEM, containing or lacking l-NMMA, added to the wells and inserts, were 600 and 400 µl, respectively. In this setup, the inserts allow the passage and diffusion of soluble components (e.g., nitric oxide) between macrophages present in the wells and bacteria in the inserts, but not the phagocytosis of bacterial cells (S. epidermidis cell size: 1–2 μm). After 24 h of growth, the inserts were collected and put in clean 24-well plates, and the amount of biofilm in the inserts was determined with the crystal violet assay as described above. The Griess method (Tarpey et al., 2004) was used to determine the nitrite present in macrophage supernatants, which results from the oxidation of the NO derived from iNOS activity.