In conclusion, our discovery of a significant impact of the nasal microbiome on a widespread neurological disease highlights the importance of the nose-brain axis and nasal bacterial colonization for such diseases.
The research complied with all applicable ethical regulations. The human clinical study was approved by the ethics committee of Renji Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China (approval number KY2022-139-B). All animal procedures followed the ethical guidelines outlined in the Guide for the Care and Use of Laboratory Animals proposed by the Institute for Laboratory Animal Research of the National Academy of Sciences and all protocols were approved by the Animal Welfare Committee of Renji Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China (approval number RJ2023-025B).
Healthy volunteers were recruited from the Physical Examination Center of Renji Hospital. Patients who had been newly diagnosed with depression were recruited from the Department of Psychological Medicine of Renji Hospital. All participants were ethnically Han Chinese. Each participant was provided with a detailed written questionnaire by the research physicians. Informed consent was obtained from all human research participants. Patients with depression and healthy volunteers were enrolled between November 2022 and November 2024. All participants were between 18 and 44 years of age. The patients with depression were diagnosed according to the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) by clinical psychologists and had no other psychiatric disorders or family history of any psychiatric disorders. The severity of depression and anxiety was evaluated using PHQ-9 and GAD-7 scores, respectively. The healthy controls were physically healthy and lacked any neurological illness or related family history, with PHQ-9 and GAD-7 scores both below four. Other exclusion criteria for both control and depression groups included nasal or oral diseases, previous use of any type of antidepressant medication within the past three months, use of antimicrobial drugs within the past four weeks, thyroid dysfunction, diabetes, hypertension, autoimmune diseases, tumours and pregnancy or lactation.
Due to the minimal risks involved in the sampling process of this study, only non-economic compensation measures were adopted. Professional psychologists were provided for participants when they experienced significant discomfort or needed immediate psychological support during the research process.
Two nasal swab samples were collected from the nasal cavity of each participant using Copan swabs. With the participant's head tilted slightly backward, a trained physician gently inserted the swab along the nasal cavity to a depth of 2-3 cm without endoscopic guidance or local anaesthesia. The swab was rotated firmly against the mucosal surface for 15-20 s to ensure adequate specimen collection before slowly withdrawing. For microbiota analysis, the collected swabs were immediately immersed in 1 ml sterile saline. All collected swabs were maintained at 4 °C and processed within 4 h of collection. For metabolome analysis, the collected swabs were immediately stored at -80 °C without sterile saline. We collected 119 serum samples from participants who agreed to provide blood samples. After collection, the serum samples were immediately aliquoted and stored at -80 °C.
Nasal swabs were vortexed in 1 ml sterile saline for 2 min. Aliquots of 100 μl from each swab were diluted, plated on 5% sheep blood agar and incubated at 37 °C for 24 h. Twenty-four random colonies were isolated and identified by matrix-assisted laser desorption ionization-time-of-flight mass spectrometry (MALDI-TOF-MS, Bruker Daltonics) in each sample.
Nasal swabs were immersed in 1 ml sterile saline and vortexed for 2 min. Each sample (500 μl volume) was centrifuged at 13,000g and 4 °C for 10 min. The pellets were resuspended in 180 μl Buffer ATL (QIAamp DNA mini kit, QIAGEN, 51306) with 5 μl lysozyme (50 mg ml; Sigma, L6876) and 5 μl lysostaphin (1 mg ml; Sigma, L4402), and incubated at 37 °C for 30 min. Next, DNA was extracted according to the manufacturer's instructions. The full-length 16S rRNA gene was amplified using the primers 27F (5'-AGRGTTYGATYMTGGCTCAG-3') and 1492R (5'-RGYTACCTTGTTACGACTT-3'). The amplicons were then sequenced on the PacBio platform. For negative controls, a full sequencing protocol was applied to a sterile swab.
Sequence analysis was performed using the DADA2 workflow in the QIIME2 software pipeline. Initially, primers and adaptors were removed and sequences with a quality score of less than three or expected errors greater than two were filtered out, retaining sequences with lengths between 800 and 1,800 base pairs (bp). Further processing with DADA2 included the removal of duplicate sequences, learning error models, inferral of amplicon sequence variants (ASVs) and removal of chimaeras to obtain an ASV feature table. To mitigate potential contamination effects, all ASVs detected in the negative controls were removed from subsequent analyses (Supplementary Data 1). Batch correction was performed with MMUPHin. Only ASV sequences present in at least two samples were retained to eliminate spurious features. The taxonomy of ASV sequences was analysed using RDP Classifier version 2.13 against the NT_16S (v20221012) database. Taxonomic α-diversity was estimated using Shannon and Simpson indices. The β-diversity between groups was measured using PCoA based on Bray-Curtis dissimilarity and compared using the PERMANOVA method with the vegan R (v2.6-4) package. Microbiome taxonomic abundance data were analysed using general linear models by MaAsLin2 (ref. ), adjusted for demographic variables (age, sex and body mass index), socioeconomic factors (education and income), family relationships, adverse childhood experiences, social support (availability of supportive friends) and technical factors (batch). The P values were adjusted for multiple comparisons using FDR based on the Benjamini-Hochberg method. Differentially abundant genera and species were determined using FDR < 0.05. Absolute quantification of S. aureus genomic copies was achieved through quantitative PCR analysis utilizing a standard curve developed via successive 1:10 dilutions of linearized plasmid DNA templates containing cloned nuc sequences with pre-determined copy number.
A single S. aureus isolate per specimen was randomly selected for whole-genome sequencing. Genomic DNA was isolated from the bacterial cell pellets using a Bacterial DNA kit (OMEGA) according to the manufacturer's instructions. Paired-end libraries with insert sizes of 150 bp were prepared following Illumina's standard genomic DNA library preparation procedure. The qualified Illumina paired-end library was used for Illumina NovaSeq 6000 sequencing (150 bp × 2; Shanghai BIOZERON Co., Ltd). The raw paired-end reads were filtered using fastp v0.12.5 and de novo assembly was performed using SPAdes v3.15.4. The resulting scaffolds were annotated using Prokka v1.14.6-. Multilocus sequence typing was performed using mlst v2.23.0 (https://github.com/tseemann/mlst) based on whole-genome sequencing data.
Extract solution (1,000 μl; 3:1 methanol:water containing isotope-labelled internal standard mixture) was added to the nasal swab sample and the mixture was sonicated for 10 min in an ice-water bath. The samples were incubated for 1 h at -40 °C and centrifuged at 13,800g and 4 °C for 15 min. The resulting supernatant was transferred to a fresh glass vial for analysis. Liquid chromatography with MS/MS (LC-MS/MS) analyses were performed using an ultra-high-performance liquid chromatography (UHPLC) system (Vanquish, Thermo Fisher Scientific) with a UPLC HSS T3 column (2.1 mm × 100 mm; 1.8 μm) coupled to an Orbitrap Exploris 120 mass spectrometer (Orbitrap MS, Thermo Fisher Scientific). The mobile phase consisted of 5 mmol l ammonium acetate and 5 mmol l acetic acid in water (A) and acetonitrile (B). The auto-sampler temperature was 4 °C and the injection volume was 2 μl. The mass spectrometer was used to acquire MS/MS spectra based on information-dependent acquisition mode in the control of the acquisition software (Xcalibur, Thermo Fisher Scientific). Raw data were converted to the mzXML format using ProteoWizard and processed using the XCMS R package (v3.22.0) for peak detection, extraction, alignment and integration. An in-house MS2 database (BiotreeDB) was used for metabolite annotation. The cutoff for peak annotation was set at 0.3. The identification level of metabolites was annotated based on the method by Alseekh et al., and metabolites identified at the MS2 level with definitive KEGG compound assignments were used for further analysis. The raw peak areas of metabolites were normalized using internal standards. Differentially abundant metabolites were determined based on the variable importance for the projection value calculated by OPLS-DA analysis and FDR using the ropls R package (v1.30.0), with a cutoff of variable importance for the projection > 1 and FDR < 0.05. KEGG pathway enrichment analysis was performed using MetaboAnalyst 6.0 (ref. ). HAllA-based multivariate analysis (v0.8.20) was conducted to explore associations between species showing differential abundance in MaAsLin2 analysis and metabolites meeting statistical significance criteria.
Specific-pathogen-free C57BL/6J mice were purchased from GemPharmatech and bred in-house under specific-pathogen-free conditions. The vendor specifically guaranteed that the mice were free of S. aureus. The mice were provided with food and water ad libitum and housed at consistent ambient temperature (22 ± 1 °C) and humidity (50 ± 5%) with a 12-h light-dark cycle. The mice were maintained on a ɣ-gamma-irradiated standard rodent diet (Suzhou Shuangshi Laboratory Animal Feed Technology Co., Ltd). The diet contained corn, wheat bran, soybean meal, fish meal, alfalfa meal, calcium hydrogen phosphate, sodium chloride, vitamin premix and mineral premix, with the following nutritional composition: crude protein, ≥200 g kg; crude fat, ≥40 g kg; crude ash, ≤80 g kg; crude fibre, ≤50 g kg and moisture content, ≤100 g kg.
Nasal microbiota transplantation was performed in male and female C57BL/6J mice that were about 6 weeks old. Briefly, 10 μl antibiotic cocktail (0.5 g l ampicillin, 0.5 g l metronidazole and 0.25 g l vancomycin) was instilled daily into each mouse's nostril (total of 20 μl per mouse) for seven days. Nasal microbiota transplantation was performed three days after the last antibiotic instillation. Male mice received nasal transplants from male donors and female mice from female donors (one donor per mouse). Donor nasal microbiota samples were randomly selected from the healthy and depression cohorts, and were not expanded through cultivation before inoculation. Each mouse was inoculated with the microbiome of a different donor. The nasal microbiota sample was dropped into the nostrils (10 μl per nostril; 20 μl total per mouse) of antibiotic-treated mice every day for seven days to improve microbial community stability. Four mice were housed in one cage. Behavioural assessments were performed afterwards. To evaluate the bacterial load, noses were collected at different time points, homogenized in 0.5 ml cold PBS, and homogenates were serially diluted and plated on 5% sheep blood agar plates for enumeration.
DNA was extracted according to the manufacturer's instructions (EZNA soil DNA kit, Omega). The V3-V4 region of the 16S rRNA gene was amplified using the primers 338F (5'-ACTCCTACGGGAGGCAGCAG-3') and 806R (5'-GGACTACHVGGGTWTCTAAT-3'). The amplicons were then sequenced on the Illumina platform. Sequence analysis was performed using the DADA2 workflow in the QIIME2 software pipeline. The taxonomy of ASV sequences was analysed using the SILVA v138.2 database. Taxonomic α-diversity was estimated using Shannon and Simpson indices. The β-diversity between groups was measured using PCoA based on Bray-Curtis dissimilarity and compared using the PERMANOVA method with the vegan R (v2.6-4) package. Microbiome taxonomic abundance data were analysed using general linear models by MaAsLin2.
S. aureus strain P24-2 (ST398) or S. epidermidis strain P24-1, which were isolated from the nasal microbiome of patient 24 in the depression cohort, were cultured in TSB at 37 °C for 8 h, harvested by centrifugation at 13,000g and 4 °C for 2 min, and then resuspended in sterile PBS. Each nostril of male and female C57BL/6J mice (approximately 6 weeks old) received 10 μl of an antibiotic cocktail (0.5 g l ampicillin, 0.5 g l metronidazole and 0.25 g l vancomycin; 20 μl total per mouse). After three days of daily antibiotic pretreatment, the mice were subjected to treatment with 10 μl bacterial solution in each nostril (total of 20 μl per mouse containing approximately 1 × 10 CFU) once every two days for seven total treatments. For S. aureus strain JSNZ (ST88), mice without antibiotic pretreatment directly received two treatments of 2.5 μl bacterial solution seven days apart (total of 5 μl per mouse containing approximately 1 × 10 CFU). The mice were euthanized at pre-determined intervals throughout the experimental period. Nose and lung tissues were collected and homogenized in 0.5 ml cold PBS. Caeca and faeces were collected and homogenized in 5 ml cold PBS. For CFU counts, homogenates were serially diluted and plated on 5% sheep blood agar plates, ChromAgar Staph aureus selective plates, or selective plates containing 0.5 mg ml streptomycin (for ST88), which were then incubated at 37 °C for 24 h. S. aureus and S. epidermidis colonies were identified using MALDI-TOF-MS (Bruker Daltonics).
Male C57BL/6J mice were exposed to one of the following various low-intensity social/environmental stressors sequentially for a total duration of seven days based on the method by Yu and colleagues. The stressors were: (1) food deprivation for 24 h, (2) overnight illumination for one night, (3) absence of sawdust in the cage for 24 h, (4) moistened sawdust for 24 h, (5) water deprivation for 24 h, (6) physical restraint for 6 h and (7) 45° cage‐tilt along the vertical axis for 3 h.
To eliminate the effects of physiological hormonal fluctuations, a bilateral ovariectomy was performed on female C57BL/6J mice and a bilateral orchidectomy on male C57BL/6J mice (all approximately 6 weeks old). After 14 days of postoperative care and no antibiotic pretreatment, 5 μl JSNZ bacterial solution (containing about 1 × 10 CFU) was instilled into the right nostrils of the mice. The next day, the mice received hormone replacement therapy using a nasal hydrogel for sustained release of estradiol or testosterone. The female (ovariectomy) mice received 1 μg estradiol in 5 μl in situ nasal hydrogel. The male (orchidectomy) mice received 0.1 mg testosterone in 5 μl in situ nasal hydrogel.
To study the effects of intranasal administration of pure sex hormones, female mice colonized with JSNZ (ST88) received 0.1 μg estradiol or vehicle in 5 μl in situ nasal hydrogel every two days, and CUMS-treated male mice colonized with JSNZ (ST88) received 10 μg testosterone or vehicle in 5 μl in situ nasal hydrogel every two days.
The thermosensitive in situ nasal hydrogel was prepared as follows. Solubilization of estradiol, testosterone or FITC was achieved through hydroxypropyl-β-cyclodextin (HP-β-CD) complexation, sonication (40 kHz, ice bath) in 2% HP-β-CD in PBS (pH 6.5), followed by magnetic stirring (500g, 25 °C) and purification via centrifugation. PLGA-PEG-PLGA and CS-PEG polymers were dissolved in chilled ultrapure water and acetate buffer (pH 5.0), respectively, and allowed to electrostatically self-assemble (2:1 vol/vol) with drug-loaded HP-β-CD under continuous stirring (4 °C for 6 h). The gel was reconstituted by blending complexes with 20% chilled poloxamer 407 (1:9 vol/vol), which is a thermosensitive molecule widely used in mucosal delivery. The gelation time was about 30 s at 37 °C. In vitro release studies employed dynamic dialysis (molecular weight cutoff, 3.5 kDa) in the synthetic nasal medium SNM3 (pH 6.5) containing 0.5% HP-β-CD, with sustained drug release monitored over 48 h using LC-MS/MS. Anaesthetized mice received unilateral intranasal administration of 5 μl thermosensitive hydrogel (containing about 10 μg FITC) or control liquid formulation (FITC dissolved in 2% HP-β-CD/PBS, pH 6.5). The animals were euthanized at pre-determined intervals post administration, with major organs excised for fluorescent imaging using an IVIS in vivo imaging system.
Behavioural tests were performed based on the previously reported methods.
The TST test was performed by wrapping the tail of each mouse with tape, with the tip sticking out by about 1 cm, and then suspending the mouse head-downwards at a height of 15 cm above the floor. The animals were videotaped from the front for 6 min. The total duration of immobility within the last 4 min was analysed using the VisuTrack software (Shanghai XinRuan Information Technology Co., Ltd.). Duration of immobility was defined as the time when animals did not seem to struggle.
The FST test was performed by placing individual mice in a water-filled cylinder (diameter of 12 cm and a height of 25 cm; water temperature of 25 ± 1 °C). The animals were videotaped from the front for 6 min. The time of immobility during the last 4 min was counted using the VisuTrack software. It was defined as the time when the animals remained floating or motionless with the only observed movements being those necessary for keeping their balance in the water.
The OFT test was performed by placing the mice in the centre of an arena (50 cm × 50 cm × 40 cm). The animals were videotaped from above for 6 min. The time spent in the central area during the last 4 min was counted using the VisuTrack software. Additional data related to all OFT experiments are in Supplementary Data 2.
Mice nose and lung tissues were collected and homogenized in 0.5 ml RIPA solution (containing 0.1 mM phenylmethylsulfonyl fluoride). After centrifugation at 4 °C and 10,000g for 15 min, the supernatants were collected and assessed using a bicinchoninic acid assay for protein quantification (Sangon Biotech). Concentrations of interleukin-6 and 1β, tumour necrosis factor-α and CXCL1 were determined using an enzyme-linked immunosorbent assay kit (CUSABIO).
Mouse noses and lungs were fixed in 4% paraformaldehyde at 4 °C for three days. The noses were then decalcified in 0.12 mol l EDTA solution (pH 7.4) for 7-14 days at room temperature. Hematoxylin and eosin staining was then performed following standard procedures.
Nasal tissues were collected four days after the final colonization (before behavioural testing) and fixed in 4% paraformaldehyde at 4 °C for 72 h. Following EDTA decalcification (0.12 M, pH 7.4; 7-14 days with solution replacement every 48 h), the tissues were dehydrated through graded ethanol, cleared in xylene and paraffin-embedded for 5-μm sectioning. Immunofluorescence protocols included sequential dewaxing with eco-friendly agents, EDTA-based heat-induced epitope retrieval (98 °C for 20 min), blocking with 10% donkey serum and incubation with primary antibodies -- anti-IBA1 (guinea pig polyclonal antibody; Oasis Biofarm, OB-PGP049) or anti-OMP (rabbit polyclonal antibody; Bioss, bs-19568R) at a dilution of 1:100 -- and the fluorescent secondary antibodies Alexa Fluor 594 goat anti-guinea pig IgG (Oasis Biofarm, G-GP594; 1:200) and Alexa Fluor 488 donkey anti‐rabbit IgG (Thermo Fisher Scientific, A21206; 1:400). Nuclei were counterstained with 4',6-diamidino-2-phenylindole (1:500), followed by mounting with antifade medium and fluorescence microscopy analysis.
Lung tissues were harvested four days after the final colonization (before behavioural testing) and fixed in 4% paraformaldehyde at 4 °C for 72 h. Paraffin sections underwent sequential dewaxing with eco-friendly agents and graded ethanol, followed by antigen retrieval using either EDTA-based pressure cooking (1,200 W, 1.5 min boiling) or microwave-mediated retrieval (medium power, 12 min). Endogenous peroxidase activity was quenched with 3% HO before blocking with 10% goat serum. Anti-LY6G primary antibody (rabbit monoclonal antibody; Abcam, ab238132, EPR22909-135; 1:500) was applied overnight at 4 °C, followed by corresponding horseradish peroxidase-conjugated secondary antibody (goat anti-rabbit IgG; Abcam, ab205718; 1:2,000) at 37 °C for 45 min. Immunoreactivity was visualized using 3,3'-diaminobenzidine chromogen with reaction monitoring, counterstained with haematoxylin for 1 min, and differentiated with acid-alcohol. The sections were dehydrated, cleared and mounted with eco-friendly medium for bright-field microscopy analysis.
Sickness scores were determined according to Huet et al., where eight parameters (fur aspect, activity, posture, behaviour, respiration, chest sounds, eyes and body weight) were given a score between one, representing the minimum (healthy), and four, representing the maximum. Total scores thus ranged from eight for entirely healthy mice to 32.
Midbrain tissues were collected, weighed and homogenized in 80 μl extract solvent (0.1% formic acid in acetonitrile; precooled to -20 °C)/20 μl water. The samples were kept at -20 °C overnight and then centrifuged at 13,800g and 4 °C for 15 min. The 80-μl supernatant samples were incubated with 40 μl of 100 mmol l sodium carbonate solution and 40 μl 2% benzoyl chloride in acetonitrile solution for 30 min. The samples were then centrifuged at 13,800g and 4 °C for 15 min after the addition of 10 μl internal standard. Next, 40-μl aliquots of the supernatants were added to 20 μl water and transferred to an auto-sampler vial for UHPLC-MS/MS analysis (Waters ACQUITY Premier or SCIEX Triple Quad 6500+ MS). The mobile phase A was 0.1% formic acid and 1 mmol l ammonium acetate in water, and the mobile phase B was acetonitrile. SCIEX Analyst Work Station Software (version 1.6.3) was employed for multiple-reaction-monitoring data processing.
Mice received nasal antibiotics, microbiome transplants or bacteria and were euthanized by cervical dislocation under anaesthesia using isoflurane inhalation.
Serum cytokine concentrations were measured using a cytokine detection kit (BD Biosciences). Serum concentrations of thyroid-stimulating hormone (TSH), free triiodothyronine (FT3) and free thyroxine (FT4) were determined using a Roche Cobas fully automated electrochemiluminescence immunoassay system.
Steroid hormones were detected using LC-MS/MS. For serum or supernatant samples of bacterial late-exponential-phase cultures (TSB), 200 μl of the samples was mixed with 20 μl internal standard solution and 400 μl methanol, followed by vortexing for 30 s and centrifugation for 10 min at 13,800g and 4 °C. Next, 600 μl water was added to 500-μl supernatant aliquots, followed by vortexing for 30 s and centrifugation under the same conditions. A 950-μl aliquot of the obtained supernatant was further purified with solid phase extraction cartridges (Agela). The cartridges were washed with 200 μl methanol and then equilibrated with 200 μl water. After sample application loading, the cartridges were washed with 200 μl of 10% acetonitrile in water (vol/vol) and 200 μl hexane. The cartridge was then rinsed with 40 μl of 90% acetonitrile in water (vol/vol) and 60 μl water was added to the eluent. All samples were shaken for 3 min, after which the mixed solution was subjected to UHPLC-MS/MS analysis (Waters ACQUITY UPLC/Xevo TQ-S MS). The mobile phase A was 0.5 mmol l ammonium fluoride in water, and the mobile phase B was methanol. Waters MassLynx V4.1 was employed for multiple-reaction-monitoring data acquisition and processing.
For tissue samples, nose and midbrain tissues were collected, weighed and homogenized in 200 μl water. After the addition of 20 μl internal standard solution and 400 μl methanol, the samples were processed as described for the serum samples.
Nasal bacteria were cultured in TSB medium containing 100 ng ml estradiol or testosterone at 37 °C. After 24 h of cultivation, the supernatant was collected, and testosterone and estradiol concentrations were determined using LC-MS/MS. For time-dependent degradation of testosterone and estradiol by S. aureus supernatant, the same conditions were used and the sampling time points were set to 0, 24, 48 and 60 h.
Nasal mucosal tissues and midbrain samples were carefully isolated using microscopy forceps. Total RNA was extracted using TRIzol reagent according the manufacturer's instructions (Invitrogen) and genomic DNA was removed using DNase I (TaKara). RNA-seq transcriptome libraries were prepared following RNA preparation with a TruSeqTM RNA sample preparation kit (Illumina). Paired-end libraries were sequenced by Illumina NovaSeq 6000 sequencing (150 bp × 2; Shanghai BIOZERON Co., Ltd). The raw paired-end reads were trimmed and quality-controlled using Trimmomatic v0.39. Next, clean reads were separately aligned to the reference genome with orientation mode using the hisat2 v2.2.1 software. FeatureCounts v2.0.3 was used to count each gene read. Differentially expressed genes were determined using edgeR v3.42.4 based on FDR and fold change, with a cutoff of FDR < 0.05 and |log(fold change)| >1. KEGG pathway enrichment analysis was performed using clusterProfiler v4.8.2. For bacteria, RNA was prepared from bacteria cultured to logarithmic growth phase and RNA-seq was performed as described above.
Total RNA was extracted from the midbrain using an RNeasy kit (Qiagen) according to the manufacturer's instructions. The extracted RNA was reverse-transcribed into cDNA using a PrimeScript RT reagent kit with gDNA Eraser (TaKaRa). Real-time PCR was performed using the Hieff UNICON Universal Blue qPCR SYBR Green Master Mix (Yeasen) on a 7500 real-time PCR system (Applied Biosystems). Gene expression was calculated using the 2 method.
The SDR superfamily encompasses multiple Pfam entries: PF00106, PF01073, PF01370, PF05368, PF08659 and PF1356 (ref. ). Hidden Markov models of these Pfam profiles were extracted from the Pfam database and hmmer v3.0 was used to scan protein sequences of the P24-2 (ST398) S. aureus genome for proteins belonging to these Pfam protein families. Proteins with an E-value of less than 1 × 10 were recorded as positive hits. The sequences of putative SDR superfamily proteins were compared using blastp with the Uniprot database to exclude candidates with known other functions.
The plasmid pET28a was used as an overexpression vector to express SDR proteins from S. aureus in E. coli. Genomic DNA of S. aureus P24-2 was used as a template to amplify putative SDR genes via PCR (primers in Supplementary Table 5). Amplified genes were cloned into the pET28a plasmids and the resulting plasmids were transformed into E. coli BL21 (DE3). The recombinant E. coli strains were cultured in 3 ml Luria-Bertani medium containing 100 ng ml testosterone or estradiol at 37 °C. When the optical density at 600 nm reached a value of 0.6-0.8, 0.5 mM isopropyl β-d-1-thiogalactopyranoside was added to the medium to induce protein expression. After 24 h of cultivation, the supernatant was collected, and the concentrations of testosterone and estradiol and their degradation products androstenedione and estrone, respectively, were determined using LC-MS/MS.
The hsd12-deletion mutants were generated in the strains S. aureus P24-2 (ST398) and JSNZ (ST88) using the pKOR1 allelic replacement strategy. Briefly, sequences of approximately 1 kb flanking the hsd12 gene were amplified by PCR. The recombinant plasmids pKOR1-HSD12 (for deletions in ST88 and ST398) were constructed using a ligation-independent cloning method and transformed into E. coli Top10 cells as the cloning host. The plasmids purified from E. coli Top10 cells were electroporated into S. aureus P24-2 and JSNZ using S. aureus RN4220 as an intermediary host. Allelic replacement was induced by temperature shift. The pKOR1 transformants were selected by plating on tryptic soy agar containing 10 μg ml chloramphenicol and incubation at 30 °C, followed by incubation in TSB containing 10 μg ml chloramphenicol at 43 °C to allow plasmid integration into the chromosome. Non-plasmid-carrying mutants were selected by plating on tryptic soy agar containing 1 μg ml anhydrotetracycline. Successful deletion of hsd12 was verified by PCR. To verify the growth behaviour of the hsd12-deletion strain (SAΔhsd), the same amount (1 × 10 CFU ml) of mid-logarithmic phase S. aureus P24-2 and its hsd12-deletion mutant were cultured in 200 μl of fresh TSB medium in 96-well plates. The plates were incubated at 37 °C in a BioTek Synergy H1 multi-mode microplate reader with continuous shaking for 14 h. The optical density at 600 nm was measured every 15 min.
GraphPad Prism version 10.2.0 for Macintosh was used for simple statistical analyses including comparisons of two or more groups and correlation analyses. Two-group comparisons were performed with unpaired two-sided Student's t-tests or Mann-Whitney tests. Comparisons of three or more groups were performed using a one-way or two-sided ANOVA, or a Kruskal-Wallis test, as appropriate, depending on the assessments of normal distribution using the Shapiro-Wilk test, with Tukey's and Dunn's post-tests, respectively. Spearman's correlation was used to assess the association between two quantitative variables, interpreted according to Akoglu et al. and Chan et al., and Fisher's exact tests or χ tests were used to assess contingency, as indicated. Principal coordinate analyses were performed and evaluated using the vegan R (v2.6-4) package. Further tests are described in their specific method sections. Each independent experiment was performed with at least three biological replicates. Values were expressed as the mean ± s.d., unless otherwise indicated in the figure legend.
Where indicated, randomization was performed using the randomization functions of Microsoft Excel 2019, which randomly assigns a number between zero and one to every sample, after which the numbers are sorted in ascending order and the top samples of the desired n selected. Animals were randomly assigned to experimental groups using computer-generated random numbers following the completion of an acclimation period. Experimental conditions (including bacterial administration sequence and behavioural testing order) were randomized across subjects.
No statistical method was used to pre-determine sample size. The sample sizes were similar to those reported in previous publications. No data were excluded from the analyses except when using the described data filtering processes. The investigators were not blinded to allocation during experiments and outcome assessment, except for histological analysis, where slides were examined independently by a histopathologist who was blinded to the treatment.
All unique biological materials are available from M.L. ([email protected]) on request.
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.