The bacterial strains and plasmids used in this study are listed in Table 1. Experiments were performed using S. xylosus strain C2a or its isogenic mutants. S. xylosus was routinely cultured in Tryptic Soy Broth (TSB, Difco) under aerobic conditions (1:10 volume to flask ratio, 150 rpm) or on Tryptic Soy Agar (TSA, Difco) at 30°C. When required, strains were grown in a minimal medium (MX) as already described (Fiegler and Brückner, 1997). The MX medium, when needed, was supplemented with amino acids (100 mg/L). For genetic manipulation, Escherichia coli was cultured in Luria–Bertani (LB, Difco) broth with shaking aeration (150 rpm) or on LB agar (Difco) at 37°C. Antibiotics, when required to maintain plasmids or select recombinants, were added at the following concentrations: for E. coli, 100 μg/mL ampicillin; for S. xylosus, 20 μg/mL chloramphenicol, or 2.5 to 10 μg/mL erythromycin. All chemicals were from Sigma–Aldrich.

To compare growth kinetics under microaerobic conditions, overnight cultures of S. xylosus were diluted to an optical density (OD) of 0.05 to 0.10 at 600 nm and incubated in 100-well microtiter plates with shaking at 30°C in a Bioscreen C plate reader (Labsystems France) while the turbidity was monitored every 30 min for 24 h. Three independent experiments were done for each set of conditions.

Genomic DNA from S. xylosus was prepared from overnight cultures. Briefly, cells were resuspended in Tris-EDTA-sucrose buffer containing 0.1 mg/mL lysostaphin (Sigma-Aldrich) and incubated for 30 min at 37°C. Cells were lysed with sodium dodecyl sulfate and treated with RNase A. Following extraction with phenol-chloroform-isoamyl alcohol (25/24/1) and chloroform, DNA was precipitated with ethanol and resuspended in Tris-EDTA buffer (pH 8.0). Plasmid DNA from E. coli was isolated using the NucleoSpin Plasmid Quick-Pure kit (Macherey-Nagel). Restriction digests were performed using high-fidelity restriction enzymes (New England Biolabs). Ligations were performed using T4 DNA ligase (Roche). Amplifications were performed using Phusion High-Fidelity DNA Polymerase (New England Biolabs) for cloning and GO Taq DNA polymerase (Promega) for standard purposes. PCR products were visualized by 0.8% agarose gel electrophoresis and ethidium bromide staining, and imaged on a Gel Doc 2000 (Bio-Rad). As required, they were purified using the QIAquick Gel Extraction kit (Qiagen). DNA sequencing was performed by GATC Biotech (Mulhouse, France).

The oligonucleotides used in this study were designed based on the genome sequence of S. xylosus C2a (GenBank accession no. LN554884) and are listed in Table 2. A chromosomal nos disruption mutant was constructed by deletion of the nos coding sequence corresponding to amino acids 11 to 345 (out of 354) and insertion of an erythromycin resistance cassette (ermB) using the temperature-sensitive vector pBT2. Briefly, a 756-bp upstream fragment and a 783-bp downstream fragment were separately amplified by using the genomic DNA of S. xylosus C2a as template. The ermB gene was amplified from the pEC4 vector generating an amplicon of 1,366-bp. The three purified PCR products were annealed by overlapping PCR using the outside primers and inserted into pBT2 using appropriate restriction enzymes. The resulting plasmid was constructed in competent E. coli TOP10 and introduced into S. xylosus C2a by electroporation as described (Götz et al., 1983; Brückner, 1997). Gene replacement was allowed to take place by incubating the temperature-sensitive plasmid at 42°C as described (Brückner, 1997). Successful C2aΔnos mutant was screened for resistance to erythromycin and sensitivity to chloramphenicol, indicating loss of the pBT2 backbone, and this was confirmed by PCR and sequencing.

Complementation of the nos mutation was performed by amplifying the nos gene with its putative promoter. The corresponding amplicon was cloned into the shuttle vector pRB473 using appropriate enzymes, generating pRBnos. Similarly, an nos-pdt complementation vector (pRBnospdt) was constructed. A complementation vector pRBpdt was constructed, placing the expression of pdt under the constitutive promoter of pRB474. All PCR-amplified fragments were verified by sequencing. The resulting plasmids were constructed in competent E. coli TOP10 and introduced into S. xylosus C2aΔnos by electroporation according to an established protocol (Brückner, 1997).

A 20 mg/mL metmyoglobin solution was freshly prepared from equine heart myoglobin (Sigma-Aldrich) as described (Gündoğdu et al., 2006). Overnight S. xylosus cultures were inoculated in TSB supplemented with 2 mg/mL metmyoglobin to a final OD_{600 nm} of 0.5 (8 log CFU/mL). Cultures were incubated either under limited oxygenation (covered with mineral oil and without stirring) or under aerobic conditions as described above. After 24 h of incubation at 30°C, OD_{600 nm} and pH of the cultures were measured and cells were serial-diluted and plated on TSA for enumeration. Cultures were centrifuged and supernatants were used to measure the absorbance spectrum between 500 and 600 nm (BioMate 3, Thermo Fisher Scientific) using as control a solution of sterile TSB supplemented with metmyoglobin and incubated in the same conditions. Finally, supernatants were treated with acetone 1:4 (v/v) to extract the nitrosoheme from nitrosomyoglobin, and the absorbance spectrum was acquired between 450 and 640 nm (V-770 UV/visible, Jasco). Sterile TSB supplemented with metmyoglobin and extracted with acetone was used as control. Experiments were carried out in three independent biological replicates.

After overnight growth in TSB, S. xylosus strains were plated on TSA. After 2 days of growth (when pigmentation is strongest), 100 mg of each strain was collected from plates and pigments were extracted and quantified according the method of Morikawa et al. (2001). Briefly, cells were washed once in water, centrifuged for 5 min at 15,000 g, re-suspended in 1 mL of methanol and heated for 5 min at 55°C. Cells were centrifuged and the supernatant containing pigment was retrieved. The absorbance spectrum of methanol-extracted pigment was measured between 400 and 600 nm to evaluate carotenoid content. A peak at λ = 460 nm was detected and results are expressed as relative optical density at 460 nm with normalization to C2a. Experiments were performed with three independent biological replicates.

The S. xylosus strains were grown in TSB under aerobic conditions. To assess sensitivity to oxidative stress, overnight cultures were diluted to an OD_{600 nm} of 0.05 with fresh medium and incubated until OD_{600 nm} was approximately 1. One milliliter of each culture was collected for serial dilutions and determination of colony-forming units (CFU)/mL on TSA plates. The remaining culture was treated with 150 mM hydrogen peroxide (Sigma-Aldrich) and incubated for 1 h at 30°C. 4,000 U/mL of catalase (Sigma–Aldrich) was added to quench residual H₂O₂ as described (Barrière et al., 2002). Cells were serial-diluted and plated in duplicate on TSA for enumeration of CFU from surviving cells.

To evaluate the impact of saline and acid stresses, overnight cultures of S. xylosus strains were freshly diluted (1/100) in TSB and grown under aerobic conditions at 30°C until OD_{600 nm} was approximately 0.3. Strains were diluted in NaCl-supplemented TSB (10 and 20% NaCl) or diluted in TSB with adjusted pH (5.0 and 6.0) to an OD_{600 nm} of 0.1 and incubated in 100-well microtiter plates with shaking at 30°C in a Bioscreen C plate reader, where turbidity was monitored every 30 min for 20 hours. Assays were performed with two independent cultures for each strain.

Staphylococcus xylosus strains were grown in TSB under aerobic conditions for 6 h and 24 h. Cell pellets were immediately frozen in liquid nitrogen to stabilize the bacterial RNA and stored at –80°C. For RNA extraction, cell pellets were thawed on ice and resuspended in 500 μL of cold Tris-EDTA buffer. Samples were transferred into tubes containing 600 mg zirconia-silica beads (0.1 mm diameter), 50 μL of sodium dodecyl sulfate (10%), 500 μL of acid phenol and 3.5 μL of β-mercaptoethanol. Cells were disrupted in a FastPrep^TM machine (MP Biomedicals). After addition of 200 μL of chloroform and centrifugation, the aqueous phase containing RNA was purified with the Nucleospin RNAII kit (Macherey Nagel) according to the manufacturer’s instructions. A supplementary treatment was performed with Turbo DNAse (Ambion) to remove residual DNA. Absence of contaminating DNA was checked by PCR targeting the sod gene (Table 2). Total RNA isolated was quantified using a Nanodrop 1000 (Thermo Fisher Scientific). RNA samples were stored at –80°C. RNA isolated from three independent biological replicates for each strain or condition (wild-type and its mutants at 6 and 24 h) was reverse-transcribed to cDNA with a SuperScript Reverse Transcriptase kit following the manufacturer’s instructions (Invitrogen).

To determine if the two adjacent genes, nos and pdt, were co-transcribed, cDNA from the mRNA species transcribed from C2a 24-h cultures was subject to PCR using a pair of primers targeting the nos-pdt junction region (Table 2). The conditions for the amplification were 5 min at 95°C, followed by 25 cycles of 30 s at 95°C, 30 s at 60°C and 120 s at 72°C and finally 5 min at 72°C. As controls, this amplification was also performed on C2a genomic template DNA and no template sample.

Expression of genes of interest was carried out by quantitative real-time PCR (qRT-PCR) using iQ^TM SYBR^® Green Supermix (Bio-Rad) and the MasterCycler RealPlex (Eppendorf). Thermal cycling consisted of 30 s at 95°C, followed by 40 cycles of 15 s at 95°C and 60 s at 60°C. The primers used are listed in Table 2. The relative fold change of gene expression, using measured sod housekeeping gene expression, was determined by the Livak (2^-ΔΔCt) method (Livak and Schmittgen, 2001).

The data were analyzed by using GraphPad Prism software (version 5.01). The significance of experimental differences in bacterial growth, formation of myoglobin derivatives, and colony pigment production were analyzed by one-way analysis of variance (ANOVA) with Tukey’s multiple comparison.

To define the role of NOS activity in S. xylosus C2a, a mutant strain (Δnos) was created by insertion of an erythromycin resistance gene cassette into the nos coding sequence. Deletion of the nos gene was confirmed by PCR analysis and sequencing (data not shown). When grown in TSB under aerobic conditions, the Δnos mutant displayed a slight growth defect compared with the wild type (Figure 1A). Complementation of the Δnos mutant with a plasmid expressing the nos gene (ΔnospRBnos) showed restored growth (Figure 1B). The wild type and the Δnos and nos complemented mutants did not differ in growth in microtiter plates under microaerobic conditions (Supplementary Figure S1).

Growth was measured in minimal medium under microaerobic conditions for the wild-type C2a, the Δnos mutant and the nos complemented Δnos mutant (Figure 2A). The Δnos mutant was not able to grow in this minimal medium and growth was not restored by plasmid complementation with the nos gene (Figure 2A). Since the nos gene is in a cluster with the pdt gene, we tested the hypothesis that both genes were co-transcribed (Supplementary Figure S2). RT-PCR experiments were carried out using a primer annealing specifically with nos and a primer annealing specifically with pdt (Table 2). Amplification of a 1,421-bp fragment indicated a co-transcription of both nos and pdt (data not shown). We therefore speculated that the deletion of nos and the insertion of ermB had a polar effect on the downstream gene pdt, leading to a frameshift in the pdt open reading frame (ORF). The pdt gene encodes a prephenate dehydratase (EC4.2.1.51), which catalyzes the penultimate reaction in the phenylalanine biosynthesis. Supplementation with phenylalanine restored the growth of the Δnos and the nos complemented (ΔnospRBnos) mutants (Figure 2B), while no restoration of growth was observed with other amino acids such as tryptophan or tyrosine, which, like phenylalanine, are derived from chorismate (data not shown). Moreover, complementation of the Δnos mutant with pdt (ΔnospRBpdt) or with the nos-pdt operon (ΔnospRBnospdt) restored growth to the wild-type C2a level (Figure 2C).

The formation of the red myoglobin derivatives occurred for S. xylosus C2a under limited oxygenation (Figure 3A) and aerobic conditions (Figure 3B), where different forms of myoglobin came from the conversion of metmyoglobin (brown) to some red derivatives, such as oxymyoglobin and nitrosomyoglobin. The red pigment content was sharply decreased, but not completely abolished, in the Δnos mutant compared with the wild type, and was restored to the wild-type level in the nos complemented Δnos mutant under both conditions (Figure 3). Under limited oxygenation after 24 h of incubation, the cellular population was close to initial level, 8 log CFU/mL, for C2a (7.8 ± 0.04), Δnos mutant (7.6 ± 0.40) and nos complemented Δnos mutant (7.2 ± 0.01). Under aerobic conditions, the red pigment level was higher (Figure 3B) and the cellular population reached 9 log CFU/mL for C2a (9.1 ± 0.04), Δnos mutant (9.2 ± 0.20) and Δnos complemented mutant (9.2 ± 0.10).

To estimate the production of nitrosomyoglobin, the nitrosoheme of culture supernatants from S. xylosus C2a and its mutants was extracted and measured (Figure 4). Under limited oxygenation (Figure 4A), where the pH of cultures was about 6.0, the nitrosoheme formation of the Δnos mutant was reduced (A_540nm = 0.060 ± 1.1E-03) compared with the wild type (A_{540 nm} = 0.090 ± 7.1E-04) and the nos complemented Δnos mutant (A_{540 nm} = 0.090 ± 1.1E-03). Under aerobic conditions (Figure 4B), where the pH of the cultures was about 8.0, maximum absorption shifted, as described (Yu et al., 2016), to λ_{580 nm} instead of λ_{540 nm}. In these conditions, no nitrosoheme was formed for the Δnos mutant (A_{580 nm} < 0.005) compared with the wild type (A_{580 nm} = 0.060 ± 2.8E-03) and the nos complemented Δnos mutant (A_580nm = 0.080 ± 1.4E-03).

The Δnos mutant and the pdt complemented Δnos mutant had the same activity, while for the nospdt complemented Δnos mutant the formation of red myoglobin derivatives was restored to the wild-type level (data not shown). These results clearly demonstrated the contribution of NOS and not PDT to the production of NO in S. xylosus C2a.

When grown on TSA plates, the Δnos mutant displayed enhanced pigmentation relative to the wild type. This pigmentation returned to wild-type level upon complementation with the nos and nos-pdt genes, but not with the pdt gene alone (Figure 5A). Pigments of the wild type and the Δnos and ΔnospRBnos mutants were extracted and absorbance at 460 nm was measured. Production of carotenoid pigment in the Δnos mutant was significantly higher than in the wild type, and returned to the wild-type level in the nos complemented mutant (Figure 5B). These results clearly demonstrated that solely NOS was involved in the pigmentation phenotype. This difference in carotenoid pigment production between the wild type and the Δnos mutant was, however, not observed with the bacterial pellets after liquid cultures in TSB, even after prolonged growth for several days (data not shown).

Stress challenge assays with 150 mM H₂O₂ were performed on the wild-type C2a and the Δnos and nos complemented Δnos mutants to investigate a possible oxidative stress-sensitive phenotype in the Δnos mutant. After 1-h treatment, the Δnos mutant presented a significant reduction in cell viability (5 log) compared with the wild type (Figure 6). The phenotype was restored to the wild-type level in the nos complemented mutant (Figure 6). To evaluate the specificity of NOS in protecting against oxidative stress, stress challenge assays in saline and acid conditions were performed on the wild type and the Δnos and nos complemented Δnos mutants (Supplementary Figure S3). As the salt concentration increased (Supplementary Figure S3A) or the pH decreased (Supplementary Figure S3B), bacterial growth was affected and decreased for all strains with no significant differences between strains. These results demonstrated that NOS was not required to overcome those stresses.

To determine whether increased H₂O₂-sensitive phenotype and defects in growth under aerobic conditions of the Δnos mutant were due to differential expression of catalase genes, we evaluated expression of the katA, katB, and katC genes by qRT-PCR. The katA gene was overexpressed 6- and 10-fold at 6 h and 24 h, respectively in the Δnos mutant related to the wild type (Figure 7). Expression of the two genes katB and katC was not modified in the Δnos mutant at 6 h and was sharply downregulated at 24 h by comparison with the wild type (Figure 7). In the nos complemented mutant, only the nos gene was highly overexpressed at the two times of incubation, while the expression of the three catalase genes was not modified by comparison with the wild type. However, there was no difference in survival between the wild type and the nos complemented Δnos mutant when treated with 200 mM H₂O₂ (4 log reduction) and 250 mM (6 log reduction).