All experimental procedures were conducted in accordance with relevant ethical regulations and approved by the appropriate institutional review boards or ethics committees.

Human subjects

All human samples (healthy individuals and CF patients) were obtained from adults (22–44 years of age) in the CF program at Yale University under the Yale IRB Protocol 0102012268 and Columbia Protocols AAAR1395. Males and females are represented in an ~50–50% ratio. Sputum samples from healthy adult individuals were collected after nebulization with 3% hypertonic saline for five minutes on three cycles. To reduce squamous cell contamination, subjects were asked to rinse their mouth with water and clear their throat. CF subjects expectorated sputum spontaneously for our studies. Expectorated sputum samples were collected into specimen cups and placed on ice. Of the 7 CF patients that were chronically infected with Staphylococcus aureus (either MSSA or MRSA), 5 patients were co-infected with other agents, such as Pseudomonas (3 patients), Achromobacter (1 patient) or human influenza (1 patient). No CF patients were infected with Aspergillus, a known itaconate producer. None of the CF patients studied were under CFTR modulator therapy (e.g., Kaleydeco, Orkambi). An informed consent was signed by all subjects.

Mouse experiments

All animal experiments were performed following institutional guidelines at Columbia University and Rutgers New Jersey Medical School (NJMS). Animals were housed and maintained at Columbia University Irving Medical Center (CUIMC) and Rutgers Medical Science Building (MSB) under regular rodent light/dark cycles at 18–23 °C and fed with irradiated regular chow diet (Purina Cat#5053, distributed by Fisher). Animal health was routinely checked by an institutional veterinary. The animal work protocols (AC-AABD5602 and PROTO202400003) were approved by the Institutional Animal Care and Use Committee (IACUC). Animal experiments were carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the NIH, the Animal Welfare Act, and US federal law. C57BL/6 mice, purchased from the Jackson Laboratory (stock number 000664), were acclimated in our facilities. Ptenl^−/− mice²⁹ were bred in our facilities. The animals were infected at 8 to 10 weeks of age. In vivo experiments were performed using roughly 50% female and 50% male animals. During studies, animals were randomly assigned to cages.

In silico fumC conservation

We assembled a collection of serially sequenced S. aureus strains from seven published within-host evolution studies of S. aureus during cystic fibrosis (CF)^{36,37,38,39,40,41,42}. The studies were added if: they included at least two S. aureus strains per patient; sequencing reads were publicly available; minimal metadata were reported: patient identifier, specimen source, date of collection. We supplemented the published studies with unpublished data from a CF study at the Children’s Hospital of Philadelphia. Raw reads and metadata were downloaded from NCBI using the Bioproject identifier provided in the publications (Supplementary Data 2). Metadata were additionally retrieved from the publications. Quality control for reads data was performed as described in ref. ⁴³. Reads were excluded if the coverage was below 20, reads classification using Kraken2, v2.1.1⁴⁴ did not yield S. aureus or <50% reads were classified as S. aureus, or if the assembly size exceeded 3.5 million bp. Reads were assembled using Shovill, v.1.1.0, which implements the Spades assembler⁴⁵. Multi-locus sequence type and resistance genes were inferred from the assemblies using Mlst, v2.19.0 and Abricate, v1.0.1, both available at https://github.com/tseemann.

Within-host evolution analysis was carried out as described in ref. ⁴⁶: for each patient, we used Snippy, v4.6.0 (https://github.com/tseemann/snippy) to map isolate reads to the draft assembly of the earliest strain (if multiple strains were available the reference was selected randomly). The output was filtered by removing variants from reference reads and variants at positions for which the reads coverage of the reference was below 10. The amino-acid sequences of the mutated genes across all episodes were clustered using CD-HIT, v4.8.1⁴⁷ and for each cluster representative we searched for a S. aureus FPR3757⁴⁸ homolog (NCBI accession CP000255) using Blastp, v2.10.1. This generated a list of FPR3757 genes with counts of independent episodes with at least one mutation arising within the host. We then excluded synonymous substitutions and limited the list to genes involved in the tricarboxylic acid cycle (TCA) and surrounding pathways (glycolysis, gluconeogenesis, urea cycle) and calculated the relative mutation rate λ for each gene i as (mutations in gene i/length of gene i)/(mutations in all genes/length of all genes). We inferred the significance of the mutations enrichment by fitting two Poisson models: model 0 as the null hypothesis (neutral evolution), where the number of variants per gene was the product of the mean mutation rate across the genome and the gene length; model 1 estimated the number of variants per gene base on a gene-specific mutation rate while variants in the remaining genes are the product of a mean mutation rate. Model 0 and model 1 were compared using a likelihood ratio test. Scripts for the within-host evolution analysis are available at https://github.com/stefanogg.

Bacterial strains

The bacterial strains used in this study are shown in Supplementary Data 6. WT LAC and its derivative strains, ∆fumC and ∆fumC::fumC were constructed as previously described⁴⁹. The transposon insertions sucA::Tn(ermB) and sucD::Tn(ermB) were sourced from the NARSA transposon mutant library⁵⁰. The alleles were transduced into USA300_LAC using bacteriophage 80α⁵¹. The mutant strains were confirmed via PCR using sucAB and sucCD primers. All strains were grown at 37 °C on plates of Luria-Bertani lacking glucose (LB, Becton Dickinson (BD), 244610) supplemented with 1% agar (w/v, Acros Organics, 400400010). Overnight cultures and subcultures (1/100) were grown in LB broth at 37 °C with shaking. Bacterial inocula were estimated based on the optical density at 600 nm, OD_600nm, and verified by retrospective plating on LB agar plates to determine colony forming units (CFUs). S. aureus recovered from the mouse airways were selected on LB agar plates and BBL CHROMagar Staph aureus plates (BD).

Primers

The primers used in this study are listed in Supplementary Data 7. Full-length fumC was amplified from the genome of WT USA300 FPR3757 using KOD One PCR Master Mix -Blue- (TOYOBO, KMM-201) and cloned into the pET-21b(+) vector by Gibson assembly.

Growth curves

A U-bottomed, clear 96-well plate (Greiner Bio-One #650161) was prepared with LB, chemically defined media (CDM)⁵² or artificial sputum media²⁷. For the experiments that used media supplemented with itaconic acid (Sigma, I29204), L-malic acid (Sigma, M1000), sodium fumarate (Sigma, F1506), or hydrogen peroxide, H₂O₂, (Sigma, H1009), the pH of the media was corrected to 7.0 prior to filter sterilization using 0.20 μm filters (Thermo Fisher Scientific,725-2520). Each well was inoculated with 1.5 × 10⁵ bacteria. Absorbance at 600 nm was read every 10 min for 36 h on the Varioskan™ LUX multimode microplate (Thermo Fisher Scientific, VL0000D0), as the plate was maintained at 37 °C with shaking.

Measurement of biofilms

Log phase cultures were diluted to OD_600nm of 0.1 in LB broth supplemented with or without fumarate (Sigma, F1506) prior to incubation in either 96-well plates (Greiner Bio-One, 650161, 100 µl/well) or 4-well chamber slides (Sarstedt, D-51588, 750 µl/well) for 24 h and 12 h respectively. The plates were processed as follows:

1. Crystal violet staining in 96 well-plates: The OD_600nm was measured prior to crystal violet staining. The plates were fixed with 100% methanol, stained with 1% crystal violet (w/v, Sigma, C6158), washed and dried. Subsequently, 33% acetic acid (v/v, Acros Organics, 222140010) was added to solubilize the dye. The OD_540nm was measured and normalized to the bacterial growth at OD_600nm.

2. Wheat germ agglutinin-Alexa Fluor 555 staining and confocal microscopy: Wheat germ agglutinin (WGA) coupled to Alexa Fluor 555 (5 µg/ml, Thermofisher, W32464) was added into each well and the chamber slides were covered with foil. Each chamber slide was placed on a shaker (horizontal orbit, 60 rpm) in a 37 °C room for 5 min, before being kept stationary in a 37 °C room for 12 h. Confocal imaging of viable biofilms was performed using a Zeiss LSM 980 Airyscan with a 37 °C cage incubator. 16-bit images of WGA-positive cells (1.5x crop area, 1024 × 1024 px frame size) were captured using a 40X oil-immersion objective with the 488 nm laser (670 V gain) set at 1% intensity. Z-stack slices were collected at 1 µm intervals. The z-depth of high-density biofilm layers were calculated as the number of slices (1 µm apart) where extracellular spaces occupied no more than 20% of the image. ImageJ 1.54 f. (National Institutes of Health, Bethesda, MD, USA) was used to calculate the % of extracellular space in an image using thresholding.

Isolation of RNA from bacterial cultures

Bacterial cultures grown in LB with and without 10 mM fumarate were standardized to an OD_600nm of 1.0 prior to centrifugation. Bacterial pellets were incubated in lysis buffer (50 μM Tris-EDTA pH 7.5, 8 U ml⁻¹ mutanolysin (Sigma-Aldrich), 0.018 mg ml⁻¹ lysostaphin (Sigma-Aldrich), 0.05 mg ml⁻¹ lysozyme (Sigma-Aldrich)) at 37 °C for 45 min and TRK lysis buffer (Omega Bio-tek) was added. After 10 min at room temperature, 70% ethanol was added, and samples were transferred to E.Z.N.A. RNA isolation columns (Omega Bio-tek). RNA was isolated following the manufacturer’s instructions and treated with DNase using the DNA-free DNA removal kit (Invitrogen).

Complementary DNA synthesis and qRT-PCR

Multiscribe reverse transcriptase (Applied Biosystems) was used to generate cDNA for qRT–PCRs with Power SYBR Green PCR Mastermix (Applied Biosystems). qRT-PCR was performed using primers for hla, lukS, esxA, per, katA, sod, trx, aphF, msrA1, msrB, capA, icaB and 16S (Supplementary Data 7) and a StepOne Plus thermal cycler (Applied Biosystems). The data were analyzed using the ΔΔC_T method.

RNA-sequencing

WT LAC was grown in LB with or without 100 mM fumarate to late exponential phase. Bacterial pellets were incubated in a cell wall lysis mixture (described above) at 37 °C for 45 min. TRK lysis buffer (Omega Bio-tek, R6834-02) and 70% ethanol were added, and samples were transferred to E.Z.N.A. RNA isolation columns (Omega Bio-tek, R6834-02). RNA was isolated following the manufacturer’s instructions and treated with DNase using the DNAfree DNA removal kit (Fisher Scientific, AM1906). The RNA was precipitated with 0.1 volume 3 M sodium acetate (Thermo Fisher, S209) and 3 volumes of 100% ethanol, recovered by centrifugation, and washed with ice cold 70% ethanol. A ribosomal RNA-depleted cDNA library was prepared according to the manufacturer’s instructions using the Universal Prokaryotic RNA-Seq Prokaryotic AnyDeplete kit (NuGEN, 0363-32) and sequenced with Illumina HiSeq. Raw base calls were converted to fatsq files using Bcl2fastqs. Filtered reads were aligned to the LAC_FPR3757 reference genome using STAR-Aligner v2.7.3a. The mapped reads were annotated for read groups and marked for duplicates using the Picard tools v2.22.3. The raw counts were quantified using Subreads:FeatureCounts v1.6.3 and processed for differential gene expression using DEseq2 in R v3.5.3.

In silico protein structure simulation and charge prediction

The predicted structure of S. aureus FumC was constructed using the homology modeling tool from the Schrödinger software package (Schrödinger Release 2022-4: Prime; Schrödinger, Inc: New York, NY, 2022)⁵³. We used the human fumarate hydratase (PDB: 5UPP) with 57% identity as template. The structure of E. coli FumC, which shares a higher identity (60%), could have been utilized, but would have necessitated the de novo construction of longer loops; one loop consisting of 3 amino acids and another loop consisting of 8 amino acids instead of two loops of 3 amino acids and one loop of 1 amino acid. The obtained structure was replicated into four copies, which were superimposed on the four chains of the template tetramer structure and merged. Protonation states of the predicted tetramer were assigned using PROPKA and the structure subjected to a constrained minimization using the OPLS4 force field (Schrödinger Release 2022-4: Protein Preparation Wizard, Schrödinger, LLC, New York, NY, 2022)⁵⁴.

Synthesis of FA-alkyne probe

Monomethyl fumarate (1 g, 1 eq, 7.69 mmol) was dissolved in acetonitrile and cooled to 0 °C. EDCl·HCl (1.6 g, 1.1 eq, 8.46 mmol), HOBt (1.1 g, 1.1 eq, 8.46 mmol) and Et₃N (1.39 mL, 1.3 eq, 2.29 mmol) were added and then stirred for 40 min at 0 °C. Subsequently, propargylamine (0.6 g, 1.3 eq, 10 mmol) was added. The solution was allowed to warm up to room temperature and stirred overnight. The solvent was removed in vacuo, the residue was redissolved in ethylacetate and washed with 1 M HCl, saturated sodium bicarbonate solution and brine. The organic layer was dried over anhydrous sodium sulfate, the resulting crude was dried, and the solvent was removed in vacuo. The resulting crude was purified by flash column chromatography (SiO₂, DCM:MeOH=40:1) to afford the pure compound 1 (820.8 mg, 63.9%) as white solid (Supplementary Fig. 2A). Lithium hydroxide (330.0 mg, 7.86 mmol) was added to a stirring solution of compound 1 (820.8 mg, 4.91 mmol) in a 1:1 mixture of water and tetrahydrofuran (Supplementary Fig. 2B). The reaction mixture was stirred at room temperature and monitored by TLC. After 1 h, the reaction mixture was acidified with 1 M HCl and the solvent was removed in vacuo. The crude product was redissolved in ethylacetate and filtered. The solution was evaporated in vacuo and the crude product was purified by flash column chromatography (SiO₂, DCM:MeOH=20:1 to 10:1) to obtain the pure FAyne (FA-alkyne, 622.1 mg, 82.8%) as white solid. ¹H NMR (400 MHz, DMSO-d₆) δ 8.96 (t, J = 5.5 Hz, 1H), 6.94 (d, J = 15.5 Hz, 1H), 6.58 (d, J = 15.5 Hz, 1H), 4.01 (dd, J = 5.5, 2.6 Hz, 2H), 3.21 (t, J = 2.5 Hz, 1H) (Supplementary Fig. 2C). ¹³C NMR (101 MHz, DMSO-d₆) δ 166.28, 163.35, 135.86, 131.37, 81.77, 72.78, 28.64 (Supplementary Fig. 2D).

Competitive ABPP workflow for S-succinated cysteines in S. aureus

S. aureus USA300 FPR3757 was cultured overnight in tryptic soy broth (TSB), washed with PBS and concentrated to OD_600nm of 40. For the competition group, 800 μl of bacterial suspension was incubated with 80 μl of fumarate (100 mM stock was dissolved in NaOH and pH was adjusted to neutral) on the ThermoMixer (950 rpm, 30 min, 37 °C). For the control group, fumarate was replaced with the same volume of PBS. After competition, both groups were incubated with 8 μl of 1 M FA-alkyne probe on the ThermoMixer (950 rpm, 37 °C, 1 h). The bacterial suspension was then centrifuged (12,000 rpm, 5 min, 4 °C), the supernatant was decanted, and the pellets were washed once with 1 ml of pre-chilled PBS. Washed pellets were resuspended with 200 μl of 0.1% (v/v) Triton X-100 (Amresco, 0694-1 L) in PBS (hereafter, 0.1% PBST). Bacterial suspension was incubated on the ThermoMixer (1200 rpm, 15 min, 37 °C) after the addition of 3 μl of lysostaphin (5 U/μl, Coolaber, CL6941-1mg). Then, the lysate was supplied with 6 μl of 10% SDS (Thermo Fisher Scientific, AM9822) and sonicated (35% intensity). The lysate was centrifuged (20,000 g, 10 min, RT) to remove the debris and the supernatant was transferred into a new 1.5 ml tube (30 μl of supernatant was also aliquoted for the gel-based ABPP assay). Click reaction was carried out with 106 μl of Click reagent mix (60 μl of 50 mM TBTA ligand, 20 μl of 50 mM CuSO₄, 20 μl of freshly prepared 50 mM TCEP and 6 μl of 20 mM acid-cleavable azide-biotin tag (Confluore, BBBD-19-25 mg) and incubated on the ThermoMixer (1200 rpm, 1 h, 29 °C). The samples were then subjected to streptavidin enrichment and dual enzyme digestion (trypsin and Glu-C) steps before mass spectrometry analysis as previously described⁵⁵.

After the dual enzyme digestion, the peptides were further labeled with a dimethylation reagent (“light” for the control group (i.e., FA probe + vehicle) and “heavy” for the fumarate competition group (i.e., FA probe + fumarate))⁵⁶. Streptavidin beads were washed three times with MS-grade water to thoroughly remove dimethylation reagents. The labeled peptides were released from streptavidin beads by incubating with 200 μl of 2% formic acid on the ThermoMixer (1200 rpm, 25 °C, 90 min). Peptides from the control group and fumarate competition group were combined and dried in a SpeedVac (30 °C). They were then analyzed by HPLC-MS/MS and subjected to quantitative proteomic analysis. A competitive light-to-heavy ratio for each target was calculated. A ratio greater than 1 (or greater than 0 after log2 transformation) suggests that the cysteine site (or protein) undergoes succination by fumarate. The corresponding competition ratios for each cysteine site (or protein) were averaged across three biological replicates, and the q-value (false discovery rate/FDR) was calculated.

Protein purification

Plasmid pET21b-FumC-His₆ was transformed into BL21(DE3). Protein expression was performed overnight at 16 °C in the presence of 0.2 mM IPTG (VWR, 0487-100 G). Bacterial pellets were resuspended in suspension buffer (50 mM Tris-HCl, pH 8.0 (Thermo Fisher Scientific, 15568-025), 150 mM NaCl) and disrupted with sonication. After the lysates were clarified via centrifugation (12,000 rpm, 30 min, 4 °C), the supernatant was loaded onto 5 ml HisSep Ni-NTA 6FF column (YEASEN, 20504ES25) and washed with W1 (50 mM Tris, pH 8.0, 150 mM NaCl, 20 mM imidazole, pH 7.4) and W2 (50 mM Tris, pH 8.0, 150 mM NaCl, 40 mM imidazole, pH 7.4). Protein was eluted with elution buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 300 mM imidazole, pH 7.4) and the eluate was centrifuged (4000 g, 4 °C, swing-out) using Amicon® Ultra-15 Centrifugal Filter Unit to remove imidazole and concentrate protein. The concentrated protein was mixed with 20% glycerol and subjected to BCA assay to measure concentration (Thermo Fisher Scientific, 23225). Finally, the protein was supplied with 1 mM DTT, aliquoted, flash frozen in liquid nitrogen, and stored at −80 °C.

FumC activity

Recombinant FumC was diluted with Tris-HCl (20 mM, pH 8.0) to 1 mg/ml, and incubated with either 1 mM itaconate (pH was adjusted with NaOH to 7.4) and 1 mM DTT or PBS and 1 mM DTT as the negative control (2 h, 37 °C). For the enzymatic assay, the reaction was conducted in a 1.5 ml polystyrene cuvette (BIOFIL, CUV010015). The reaction was initiated by mixing 100 μg of post-incubated FumC and the reaction mixture containing 20 mM Tris-HCl (pH 8.0) and 15 mM fumarate (Sigma-Aldrich, 47910) in a total volume of 1 ml at 25 °C for 30 min. The maximal absorbance of fumarate was determined using a spectrophotometer (IMPLEN, P330) in a 1.5 ml cuvette. Absorbance was read at 289 nm. This enzymatic assay was performed on three biological replicates.

AcnA/CitB activity

AcnA assays were conducted as previously described with slight modifications^57,58. Strains were cultured overnight in TSB before diluting them in fresh TSB to an OD_600nm of 0.1. The cultures were diluted in 0.5 ml or 4 ml of TSB in 10 ml culture tubes. The cells were cultured for 8 h, before they were harvested by centrifugation, and the cell pellets were stored at −80 °C. The cells were thawed anaerobically, resuspended with 200 μl of AcnA buffer (50 mM Tris, 150 mM NaCl, pH 7.4), and lysed by the addition of 4 μg lysostaphin and 8 μg DNase. The cells were incubated at 37 °C until full lysis was observed (~1 h). The cell debris was removed by centrifugation, and AcnA activity was assessed by monitoring the conversion of isocitrate to cis-aconitate spectrophotometrically.

Metabolite phenotype microarray (Biolog)

For Carbon Source Phenotype Microarray™ (Biolog) assays, a stock solution of 2 × 10⁷ bacteria/ml was prepared in 1X IF-Oa buffer (Biolog, 72268) supplemented with 1X Redox Dye Mix A (Biolog, 74221). 100 μl of this stock solution (delivering 2 × 10⁶ bacteria) was added to each well of a PM1 Microplate™ (Biolog, 12111) and the plate was incubated at 37 °C overnight with shaking. Absorbance was read at 590 nm on the SpectraMax M2 plate reader (Molecular Devices).

Bacterial extracellular flux analyses

The XFe24 sensor cartridge (Agilent #102340-100) was calibrated as per the manufacturer’s instructions overnight at 37 °C without CO₂. 500 μl of XF base medium (Agilent #102353-100) was added to each well of a Seahorse XF24 well plate (Agilent #102340-100) and inoculated with 3 × 10⁷ bacteria for a 3 h incubation at 37 °C. The oxygen consumption rate (OCR, indicative of oxidative phosphorylation) and extracellular acidification rate (ECAR, indicative of glycolysis) were again measured on a Seahorse XFe24 Analyzer (Agilent Technologies) using Seahorse Wave Desktop v2.6.0. Glucose (Sigma #G7021) was added at a final concentration of 10 mM followed by the addition of fumarate (Sigma, F1506) to a final concentration of 10 mM.

¹³C₄-fumaric acid labeling and stable isotope tracing

WT LAC and the ΔfumC mutant were grown overnight in LB, then inoculated (1/100) into fresh LB supplemented with ¹³C₄-fumaric acid (Sigma, 606014, pH 7.0), grown at 37 °C to late exponential phase and standardized to an OD_600nm of 14. For metabolite extraction, each culture was diluted with 3 volumes of PBS and centrifuged at 2000 × g for 10 min at 1 °C. The supernatant was discarded, and the pellet was washed with PBS. The pellet (30 μl in volume) was resuspended in a 3:1 methanol:water extraction solution and lysed with 10 freeze-thaw cycles by alternating emersion in liquid nitrogen and a dry-ice/ethanol bath. The debris was removed by centrifugation at 14,000 × g for 5 min at 1 °C and the supernatant was stored for analysis. Targeted LC/MS analysis was performed on a Q Exactive Orbitrap mass spectrometer (Thermo Fisher Scientific) coupled to a Vanquish UPLC system (Thermo Fisher Scientific). The Q Exactive operated in polarity-switching mode. A Sequant ZIC-HILIC column (2.1 mm i.d. × 150 mm, Merck) was used for separation of metabolites. Flow rate was set at 150 μl/min. Buffers consisted of 100% acetonitrile for mobile A, and 0.1% NH₄OH/20 mM CH₃COONH₄ in water for mobile B. Gradient ran from 85 to 30% A in 20 min followed by a wash with 30% A and re-equilibration at 85% A. Metabolites were identified based on exact mass within 5 ppm and standard retention times. Relative quantitation was performed based on peak area for each isotopologue. All data analysis was done using MAVEN 2011.6.17. The data in Fig. 4 are presented with correction for natural isotope abundance.

Isolation of murine BMDMs and infection

Murine BMDMs were differentiated as described previously⁹. The femurs and tibias were surgically removed from mice. The bone exterior was sterilized with 70% ethanol and the bone marrow was recovered by flushing with phosphate-buffered saline (PBS, Corning, 20-031). The cell suspension was centrifuged for 6 min at 500 × g and resuspended in ACK lysis buffer (Thermo Fisher Scientific, A1049201) to remove the red blood cells. The lysis solution was quenched with PBS, and the cells were centrifuged again and resuspended in DMEM medium (Corning, 10-013-CV) containing 10% heat-inactivated fetal bovine serum (FBS v/v, Sigma, F4135) and 1% penicillin/streptomycin (P/S, v/v, Corning, 30-002-CI)) supplemented with 20 ng/ml rM-CSF (PeproTech #315-02). The rM-CSF-supplemented media was replenished 3 days after isolation, and the BMDMs were mechanically detached on day 7, centrifuged, resuspended in DMEM medium containing 10% FBS (Sigma), and counted using trypan blue stain (Invitrogen, T10282). The cells were seeded at 1 × 10⁶ cells/ml in a 24 well plate (Agilent, 102340-100) and incubated at 37 °C with 5% CO₂ overnight. The cells were infected at a multiplicity of infection (MOI) of 10 for 30 min and 10 μg ml⁻¹ lysostaphin (Sigma) was then added until the desired time point. The cells were washed, detached using TrypLE Express (Life Technologies), serially diluted, and plated on LB agar. Cell viability was determined by counting live cells using trypan blue (Life Technologies).

Mouse pneumonia model

Mice were infected intranasally with 1 × 10⁷ CFUs of S. aureus in 50 μl PBS. PBS alone was used as a control. The mice were sacrificed at 18 h post infection and the BALF and lung were collected. The lung was homogenized through 40 μm cell strainers (Falcon, 352340). Aliquots of the BALF and lung homogenates were serially diluted and plated on LB agar plates and BBL CHROMagar Staph aureus plates (BD) to determine the bacterial burden. The BALF and lung homogenates were spun down and the BALF supernatant was collected for cytokine and untargeted metabolomic analysis. After hypotonic lysis of the red blood cells, the remaining BALF and lung cells were prepared for fluorescence-activated cell sorting (FACS) analysis as described below.

Flow cytometry of mouse BALF and lung cells

For the identification of immune cell populations, mouse BALF and lung cells were stained in the presence of counting beads (15.45 μm DragonGreen, Bangs Laboratories Inc., FS07F) with LIVE/DEAD stain (Invitrogen, L23105A) and an antibody mixture of anti-CD45-AF700 (BioLegend, 103127, clone 30-F11), anti-CD11b-AF594 (BioLegend, 101254, clone M1/70), anti-CD11c-Bv605 (BioLegend, 117334, clone N418), anti-SiglecF-AF647 (BD, 562680, clone E50-2440 RUO), anti-Ly6C-Bv421 (BioLegend, 128032, clone HK1.4), and anti-Ly6G-PerCP-Cy5.5 (BioLegend, 127616, clone 1A8), each at a concentration of 1:200 in PBS, for 1 h at 4 °C. After washing, the cells were stored in 2% paraformaldehyde (PFA, Electron Microscopy Sciences, 15714-S) until analysis on the BD LSRII (BD Biosciences) using FACSDiva v9. Flow cytometry was analyzed with FlowJo v10. Mouse BAL and lung cells were identified as follows:

alveolar macrophages: CD45⁺CD11b^+/−SiglecF⁺CD11c⁺;

monocytes: CD45⁺CD11b⁺SiglecF⁻MHCII⁻CD11c⁻Ly6G⁻Ly6C^+/−;

neutrophils: CD45⁺CD11b⁺SiglecF⁻ MHCII⁻CD11c⁻Ly6G⁺Ly6C^+/−.

Cytokine analysis

Cytokine concentrations in mouse BALF supernatants were quantified by Eve Technologies (Calgary, Canada) using a bead-based multiplexing technology.

Semi-targeted metabolomics

Metabolites in the BALF were identified and quantified by high-resolution liquid chromatography mass spectrometry (LC-MS) at the Calgary Metabolomics Research Facility (Calgary, Canada). The metabolites were extracted in a 50% methanol (Alpha Aesar #22909):water (v/v) solution. Sample runs were performed on a Q Exactive HF Hybrid Quadrupole-Orbitrap mass spectrometer (Thermo Fisher Scientific) coupled to a Vanquish UHPLC System (Thermo Fisher Scientific). Chromatographic separation was achieved on a Syncronis HILIC UHPLC column (2.1 mm × 100 mm × 1.7 μm, Thermo Fisher Scientific) using a binary solvent system at a flow rate of 600 μl/min. Solvent A consists of 20 mM ammonium formate pH 3 in mass spectrometry grade water and solvent B consists of mass spectrometry grade acetonitrile with 0.1% formic acid (v/v). The following gradient was used: 0–2 min, 100% B; 2–7 min, 100–80% B; 7–10 min, 80–5% B; 10–12 min, 5% B; 12–13 min, 5–100% B; 13–15 min, 100% B. A sample injection volume of 2 μl was used. The mass spectrometer was run in negative full scan mode at a resolution of 240,000 scanning from 50 to 750 m/z. Metabolites were identified by matching observed m/z signals (±10 ppm) and chromatographic retention times to those observed from commercial metabolite standards (Sigma-Aldrich). The data were analyzed using E-Maven and are listed in Supplementary Data 8–9.

Isolation of RNA from mouse lungs

Total RNA was isolated from excised mouse lungs using TRIzol reagent (Thermo Fisher Scientific, 15596026) according to the manufacturer’s instructions. Residual DNA was then degraded using the TURBO DNA-free Kit (Thermo Fisher Scientific, AM1907) prior to reverse transcription, which was performed using the High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific, 4374966). Real-time qRT-PCR was performed on the cDNA generated in the previous step using the primers listed in Supplementary Data 7. The data were analyzed using the ΔΔC_T method.

Quantification and statistical analysis

We modeled the number of independent experiments required to reach significance between groups using the computer program JMP. This simulation was based on experimental design, preliminary data and past experience. Analyses were performed assuming a 20% standard deviation and equivalent variances within groups. Significance <0.05 with power 0.8 was used. Experiments in this study were not performed in a blinded fashion. All analyses and graphs were performed using the GraphPad Prism 7a and 8 software. Values in graphs are shown as average ±SEM and data were assumed to follow a normal distribution. For comparison between average values for more than two groups, we performed one-way ANOVA with a multiple posteriori comparison. When studying two or more groups along time, data was analyzed using two-way ANOVA with a multiple posteriori comparison. Differences between two groups were analyzed using a student’s t test for normally distributed data or Mann–Whitney or Kruskal–Wallis test otherwise. Differences were considered significant when a P value under 0.05 (p < 0.05) was obtained. Statistical details of experiments are indicated in each figure legend. No data points were excluded.