Mice aged 8–12 weeks were housed in the SPF animal facility of the Experimental Animal Department at the Shanghai Public Health Clinical Center. Bitter taste receptor-deficient mice (mTas2r104^-/-/105^-/-, mTas2r105^-/-/114^-/-, mTas2r104^-/-/105^-/-/114^-/-, Gnat3^-/-, Gnat3-Tas2r104^-/-/105^-/-) on the C57BL/6 background were generated as previously described (10). Pathogen-free C57BL/6 mice were purchased from Shanghai Laboratory Animal Center (SLAC, Shanghai, China). All animal experiments were approved by the Animal Care and Use Committee of Shanghai Public Health Clinical Center (Shanghai, China).

The frozen S. aureus Newman strain was thawed and inoculated onto a mannitol agar plate by streaking and allowed to grow overnight at 37 °C. After picked, S. aureus was grown Trypto-casein soja broth (TSB BD Biosciences) at 37 °C with shaking (200 rpm) until the log phase. The bacteria were subsequently centrifuged twice and adjusted to a suspension of 5 ×10¹⁰ bacteria/ml. After anesthetized with isoflurane, these mice were infected intranasally with 50 μl (10⁹ CFU) bacteria, and maintained in the Animal Biosafety Level II laboratory.

The infected lung tissue was excised, weighed and homogenized in sterile PBS at a ratio of 1 ml of PBS per gram of tissue. After 10-fold dilutions. the samples were then plated and incubated at 37 °C for 48 hours. Bacterial load was obtained by counting colony (7).

An RNA extraction kit (19211ES60, Yisheng, China) was used to extract and purify the total RNA of the mouse lung and tracheal tissues. The above mouse tissue RNA was reverse transcribed at 42 °C for 30 min via a reverse transcription kit (1123ES60, Yisheng, China). The results of the qPCR experiments were analyzed via a SYBR Green Master Mix kit (11201ES80, Yisheng, China) and a quantitative PCR instrument (FQD-96A, Bioer, China), and the sequences of the primers used and other relevant information are listed in Supplementary Table S1 .

The total protein was extracted and quantified from the lung samples as previous reports (24). In brief, equal amounts of protein samples were separated by SDS-PAGE and transferred to PVDF membranes (ISEQ00010, Merck Millipore, UK). The PVDF membranes were then blocked and incubated with primary antibodies overnight at 4 °C ( Supplementary Table S2 ). After washed and incubated with the corresponding HRP-conjugated secondary antibodies (HS101-01, TransGen, China) for 2 h, immunoreactive bands were detected with an enhanced chemiluminescence detection reagent. Band intensities were quantified via ImageJ software, and the results were normalized to β-actin levels and reported as intensities relative to those of the controls.

Lung samples were fixed overnight in 4% paraformaldehyde, cryosectioned and subjected to immunofluorescence staining, as previous reports (24). After being permeabilized in 0.3% TX-100 and blocked in 5% bovine serum albumin (BSA), the sections were incubated with primary antibodies ( Supplementary Table S2 ) at 4°C overnight. The sections were then incubated with the secondary antibody Alexa-488-conjugated mouse anti-rabbit (Invitrogen, A-21202, 1:800) for 1 h at room temperature and subsequently incubated with 4’,6-diamidino-2-phenylindole (DAPI) for 0.5 h. All images were collected with a confocal microscope (Leica TCS SP5).

Lung tissues were fixed in 4% paraformaldehyde, embedded in paraffin, and sectioned at a thickness of 5 μm. Lung histopathological injury was assessed by hematoxylin and eosin (H&E) staining. Entire lung sections stained with H&E were automatically were captured under an artificial pathology section scanner (KF-PRO-120, KFBIO). At least 5 points each slides were assessed by a histopathologic inflammatory scoring system as described previously in a hamster M. pneumoniae infection model (25). This scoring system has been previously used in other mouse models of respiratory infections (7, 26, 27). Briefly, a final score per mouse on a scale of 0 to 26 was obtained by assessing the quantity and quality of peribronchiolar and peribronchial inflammatory infiltrates, luminal exudates, perivascular infiltrates, and parenchymal pneumonia. A. Peribronchial/peribronchiolar infiltration: 0 = no infiltration; 1 = minimal infiltration (<25%); 2 = moderate infiltration (25%~75%); 3 = complete infiltration (>75%). B. Qualitative assessment of peribronchial/peribronchiolar infiltration: 0 = no infiltration; 1 = mild infiltration; 2 = moderate infiltration; 3 = severe infiltration. C. Bronchial/bronchiolar lumen dissection: 0 = no exudation; 1 = mild exudation (<25%); 2 = severe exudation (>25%). D. Perivascular infiltration: 0 = no infiltration; 1 = minimal infiltration (<10%); 2 = moderate infiltration (10%); 3 = extensive infiltration (>50%). E. Parenchymal pneumonia: 0 = no pneumonia; 3 = mild pneumonia; 5 = severe pneumonia. Total score = A + 3(B + C) + D + E. The scores for all images of each lung section were then averaged, and the composite score for that animal’s lung was considered as 1 data point for statistical analysis and individual data point in graphs.

A single-cell suspension was obtained from lungs and thymus as previous describe (7). The cells were then stained with the following antibodies. Zombie Aqua™ Fixable Viability Kit (Biolegend), APC/Cyanine7-conjugated anti-CD45 (clone 30-F11), FITC-conjugated anti-CD3 (clone 145-2C11), Pacific blue-conjugated anti-CD4 (clone GK1.5), Brilliant violet605-conjugated anti-CD8 (clone GK1.5), PE/Cyanine5-conjugated anti-CD19 (clone 6D5), PE-conjugated anti-CD11b (clone M1/70), and FITC-conjugated anti-LY6G (clone 1A8) antibodies were used. The data were analyzed via FlowJo v10.6.2.

SPSS 21.0 software was used to analyze significant differences. A homogeneity of variance test and normality test were performed for the experimental data. If multiple sets of variables were consistent with homogeneity of variance, analysis of variance (ANOVA) was used to compare multigroup variables, and the least significant difference (LSD) test was used to compare intergroup variables. If homogeneity of variance was not assumed, Dunnett’s T3 test was used to compare the intergroup variables. The values are presented as the means ± SDs. Values of p < 0.05 were considered significant.

The airway has been used as a phenotypical tissue to explore the pathophysiology of extraoral TAS2Rs. In a previous study, we generated mTas2r105-Cre/GFP transgenic mice (28). This mouse line was crossed with R26:loxP-StoploxPLacZbpA transgenic mice and can be used to track the expression of the mTas2r105 gene cluster in extraoral tissues (28, 29). The frozen sections from the lung tissues of the double transgenic mice were stained for LacZ. LacZ-positive cells were observed in the epithelium of the trachea ( Figure 1A ) and primary bronchi ( Figure 1B ) but rarely in the epithelium of the bronchioles ( Figure 1C ). To further study the function of mTas2r105 in extraoral tissues, we generated several mTas2rs mutant mice by targeting mTas2105 gene clusters via the CRISPR/Cas9 technique (10). We profiled the expression of several mTas2rs highly expressed in large airways and the lungs (30, 31). In the trachea, the expression level of Tas2r117, mTas2r130 and was downregulated in mutant mice ( Figure 1D ). In the lungs, we observed upregulated expression of mTas2r106, 107, 130, 136, and 143 in mTas2r105^-/-/114^-/- mice and downregulated expression of mTas2r108 in all mutant mice ( Figure 1E ). These changes indicated a compensatory or regulatory network among Tas2r genes, though the exact pathway remains unclear.

Previous studies have shown the expression of bitter taste receptors including mTas2r105, 108, 131 and 143 in the thymus (31, 32). We employed flow cytometry to investigate whether genetic mutations in bitter receptors affect the composition of lymphocytes in the lung and thymus ( Figure 2A ). The results revealed that the lymphocyte composition did not change in the thymus ( Figures 2B, C ) or lung ( Figures 2D–G ). Moreover, the number of neutrophils in the lung was also unchanged ( Figure 2G ), indicating that genetic mutation of the bitter receptor did not alter the composition of immune cells in the lung or thymus, and could not cause limited damage to T-cell development.

To obtain insight into the role of bitter receptors in controlling infectious diseases caused by bacteria, we investigated the role of bitter receptors in a previously established murine model of S aureus-induced pneumonia (7). We compared the survival of wild-type (WT) and mutant mice (mTas2r104^-/-/105^-/-, mTas2r105^-/-/114^-/-, and mTas2r104^-/-/105^-/-/114^-/-) after the intranasal instillation of 10⁹ bacteria per mouse ( Supplementary Figure S1A ). After 7 days, 90% of both groups of mice survived and recovered completely from the infection ( Supplementary Figure S1B ). Weight loss was more severe in mTas2r104/105 mice ( Supplementary Figure S1A ). Pulmonary infection significantly increased the mRNA level of IFNγ ( Supplementary Figure S2A ), TNFα ( Supplementary Figure S2B ) and IL-6 ( Supplementary Figure S2C ) at D1 in WT and mutant mice but not in mTas2r105^-/-/114^-/- mice, indicating an impaired innate immune response in mTas2r105^-/-/114^-/- mice.

QPCR analysis was further employed to investigate whether genetic mutation of bitter receptors affects the expression of virulence factors ( Supplementary Figures S2D-F ). Higher expression of the superantigens staphylococcal enterotoxin A (SEA) ( Supplementary Figure S2D ) and RNAIII ( Supplementary Figure S2F ) was observed only in mutant mice. We also observed a high expression level of 16 s at D1 in the lung homogenates from the WT and mutant mice ( Supplementary Figure S2E ). At D3, the SEA level returned to baseline ( Supplementary Figure S2D ), and a relatively high RNAIII level was still observed ( Supplementary Figure S2F ). Moreover, the bacterial load was further investigated in lung homogenates. Most CFUs were detected in mTas2r104^-/-/105^-/- mice at D1 post-infection ( Supplementary Figure S2G ). These results indicate that bitter receptor signals may be involved in the regulation of virulence factors expression.

To investigate the impact of S. aureus on the pulmonary parenchyma after intranasal inoculation, the mice were euthanized, and the lung histological analysis at D1 (early stage), D3 (middle stage) and D7 (late stage) was also performed. The different histopathological lesions at the three stages of infection are shown in Figure 3A . A histopathological scoring system was further used to assess the pathology of pulmonary infection ( Figure 3B ). A higher score was obtained at D3 in mTas2r105^-/-/114^-/- mice, indicating that different pulmonary lesions may be observed among the three mutant strains after S. aureus-induced pneumonia.

The above results suggested that genetic mutation of bitter taste receptor aggravated pathological injury of lung. To further investigate the role of bitter taste signaling in S. aureus-induced pneumonia, we established a murine model of S. aureus-induced pneumonia in Gnat3^-/- and Gnat3^-/–mTas2r104^-/-/105^-/- mice, which can genetically mutate G protein α subunit (Gnat3) and is thought to eliminate most bitter taste signals (10, 11, 33). Similar to the previous situation, lower mortality was observed in these mutant mice. After 7 days, 80% of the WT and mutant mice survived and recovered completely from the infection ( Supplementary Figure S1B ). However, as shown in Supplementary Figure 1C , weight loss was significantly lower for mTas2r105^-/-/114^-/-, mTas2r104^-/-/105^-/-/114^-/-, and Gnat3^-/- animals. In contrast, Gnat3^-/–mTas2r104^-/-/105^-/- mice and WT mice regained their initial weight on day 7 ( Supplementary Figure S1B ).

The levels of immunoreactive cytokines and chemokines were measured in lung homogenates harvested 24 h and 72 h after inhalation of S. aureus ( Figures 4A-P ). Infection-related induction of most chemokines was evident in both WT and mutant mice, but responses were significantly blunted in Gnat3-mTas2r104^-/-/105^-/- animals. The expression of IL-6 ( Figure 4B ), TNFα ( Figure 4I ) and IFNγ ( Figure 4J ) did not differ significantly between mutant and WT mice. A significant increase in MCP-1 expression was observed at D1 and D3 in the lungs of mutant mice ( Figure 4K ). More interestingly, the expression levels of IL-10 ( Figure 4C ) and MIP-2 ( Figure 4L ) were significantly increased in the lungs of mTas2r104^-/-/105^-/-/114^-/- and Gnat3^-/- mice, but no significant change, especially for MIP-2, was detected in the lungs and trachea of Gnat3^-/–mTas2r104^-/-/105^-/- mice, indicating that genetic mutation of the bitter receptor signaling pathway has a vital impact on the expression of MCP-1, IL-10 and MIP-2.

We also investigated the expression of antimicrobial peptides (AMPs) at D1 in the lungs and trachea after S. aureus infection ( Figures 5A–J ). Indeed, S. aureus challenge induced increased expression of Defβ14 ( Figures 5C, D ), RegIIIg ( Figures 5G, H ), and LCN2 ( Figures 5I, J ) in both the lungs and trachea, which have been shown to play important roles in S. aureus-induced pneumonia (34–37). Compared with that in the WT infection group, Defβ14 expression in the lungs of mutant mice was significantly lower ( Figure 5C ). In contrast, RegIIIg expression was significantly greater in the lungs of mutant mice than in those of WT-infected mice ( Figure 5G ). Increased expression of cathelicidin (Camp) was observed only in the trachea of WT and mTas2r105^-/-/104^-/-/114^-/- mice ( Figure 5B ). These results suggest that bitter taste receptor signaling deficiency disrupts the expression dynamics of key AMPs during early S. aureus pneumonia, with tissue- and peptide-specific effects.

Moreover, we investigated the expression profile of mTas2rs in the lungs and trachea after S. aureus infection ( Supplementary Figures S3A, B ). In the lungs, pulmonary infection significantly increased the levels of most mTas2rs in mutant and WT mice, similar with previous reports (41). Increased mTas2r135 expression was detected only in WT mice ( Supplementary Figure S3A ). In the trachea, S. aureus infection increased the expression of mTas2r106, 126 130, 135, 136 and Gnat3 in WT mice. No significant changes were detected in the mutant mice ( Supplementary Figure S3B ).

To evaluate the role of the bitter taste signaling pathway in the severity of lung pathology, we analyzed pulmonary inflammation and injury in lung tissue slides at D3 post-infection ( Figures 6A, B , Supplementary Figure S4 ). Compared with WT mice, Gnat3^-/–Tas2r104^-/-/105^-/- mice had higher pathology scores ( Figure 6C ). Detailed immunohistochemical analysis with anti-CD68 indicated that CD68+ macrophages had an increasing trend in D3 lung tissues from mutant mice than in those from WT mice ( Figures 6B, D ).

Immunostaining analysis was further employed to investigate the immune response and histological changes in S. aureus-induced pneumonia. A large number of S. aureus were distributed in lung tissues and very rarely colocalized with CD68+ macrophages ( Figure 7A ) or Ly6G+ neutrophils ( Figure 7B ). Ki-67 is commonly used as a marker to assess cell proliferation (38). Weaker expression of Ki-67 was detected in the lungs of the WT mice ( Figure 8A ). However, Ki-67 expression was significantly upregulated in the surrounding bronchioles ( Figure 8A the first column and 8B) and alveolar region ( Figure 8A the second column and 8C) of infected mutant mice, indicating that bitter taste signaling deficiency may affect cell proliferation in the lungs via unknown molecular mechanisms. Ki-67 was colocalized with surfactant protein C (SFTPC), indicating the proliferation of type II alveolar cells ( Supplementary Figure S5 ).

Zonalula occludens 1 (ZO-1) expression is clinically associated with inflammation in human lung tissue (39, 40). We also investigated the expression of ZO-1 in infected lung tissues by immunostaining. Representative images are shown in Figure 8A . ZO-1 expression was detectable in normal tissues, including bronchioles and alveoli ( Figure 8A the third column). An intense signal was detected in bronchioles from mTas2r105^-/-/114^-/- and Gnat3-mTas2r104^-/-/105^-/- mice. We also observed a strong signal in hypertrophic connective tissue around bronchioles ( Figure 8A the third column), indicating that upregulated expression of ZO-1 is associated with increased levels of inflammation in mutant mice.

Considering the increased expression of Ki-67 and ZO-1, we speculated that bitter taste signaling deficiency could influence the expression of several critical pathways in lung tissues from infected mice at D14 post-infection ( Figures 9A–G ). No significant changes in E-cadherin ( Figure 9B ), fibronectin ( Figure 9C ) or ZO-1 ( Figure 9E ) were detected via western blot analysis. However, the expression level of the mTOR protein was significantly decreased, and the phosphorylation level of the mTOR protein was significantly increased in the mutant mice compared with the WT-infected mice ( Figure 9D ). Interestingly, we also observed a significant decrease in the level of the endothelial nitric oxide synthase (eNOS) protein in Gnat3^-/–mTas2r104^-/-/105^-/- and mTas2r104^-/-/105^-/-/114^-/- mice compared with WT mice ( Figure 9F ). The ratio of p-AMPK/AMPK was significantly increased in mTas2r104^-/-/105^-/-/114^-/- mice compared with WT-infected mice ( Figure 9G ). Representative histological images of H&E-stained mouse lungs revealed thickened alveolar walls and hemorrhagic lesions in mutant mice ( Figures 9H–M ).

In addition, immunofluorescence staining for Ki67, fibronectin, collagen I and E-cadherin was further performed in lung tissues at D14 post-infection ( Figure 10 ).

Ki67 was less expressed in alveolar region in all groups ( Figure 10 the first column). We can observe the increased expression of Ki67 in proliferation tissues ( Figure 10 the first column). The histological distributions of fibronectin in alveolar region ( Figure 10 the second column) or bronchial region ( Figure 10 the third column) were not significantly different between the mutant and WT mice. The expression of collagen I increased in connective tissues of bronchiole from these mutant mice ( Figure 10 the fourth column). We observed intense signals for E-cadherin in bronchioles from Tas2r104^-/-/105^-/-/114^-/- mice ( Figure 10 the fifth column). Collectively, these findings demonstrate that bitter taste signaling deficiency leads to the activation of mTOR pathway, reduces eNOS expression, and retards the recovery process at D14 post-infection.