MDM2-P53-V-ATPases axis driven by dehydroevodiamine to fight intracellular bacterial infection

Abstract

Host-directed antibacterial compounds remain underdeveloped for intracellular pathogens. Here, we identify Dehydroevodiamine (DEHD) as a broad-spectrum host-directed antibiotics that inhibits intracellular bacterial replication (Salmonella, E. coli, S. aureus, etc.) and synergizes with antibiotics in vitro and in vivo. Structural analyses reveal DEHD directly binds MDM2 (KD=68.34 μM), activating the MDM2-P53-V-ATPases axis to maintain lysosomal acidity through V-ATPase activity and induce mTOR-dependent autophagy. This mechanism enhances antibiotic efficacy against resistant pathogens, reducing mortality from 90% to 10% in lethal murine infections. Our work establishes lysosomal activation via the MDM2-P53-V-ATPases axis as a potent host-directed strategy, with DEHD providing a promising lead compound against intracellular infections.

Lysosomal activation via the MDM2-P53-V-ATPase axis unveils a potent host-directed strategy for eradicating intracellular pathogens and potentiating conventional antimicrobial therapy.

Introduction

Bacterial infection has long been a leading source of clinical illness with persistently high mortality and high morbidity worldwide¹. Such infection was temporarily overcome following the discovery of penicillin, as well as other antibiotics. However, the overuse of antibiotics has led to a dire issue of antibiotic resistance. Multidrug-resistant (MDR) bacterial infection now poses a global health threat, with an estimated 10 million people projected to die annually by 2050^2,3. Even more alarmingly, intracellular bacteria, such as Salmonella, have shown significant resistance to conventional antibiotic treatment and could thrive in an intracellular environment protected from antibiotics and immune response^4,5. In addition, the interaction of pathogens with bacteria-directed compounds within the host cell further altered the ecological environment of the host cell as well as the metabolic reprogramming of cytoplasmic pathogens⁶. Such interaction usually resulted in antibiotic treatment failure and recurrent or stubborn infections. Bacterial resistance to cellular processes highlighted the relevance of the other side of the coin: natural immune pathways that enhance the efficacy of existing medicines^7,8. Thus, robust therapeutic intervention strategies are urgently to eliminate the cytosolic pathogen.

The host's natural immune response in macrophages was used as one of the most powerful weapons to constantly eliminate intracellular bacteria^9,10. However, this strategy was impaired as intracellular bacteria have developed countermeasures that overcome the natural immune response of macrophages to promote their proliferation^11,12. Host-directed antibacterial compounds intervening with these countermeasures could effectively limit the propagation of intracellular bacteria and further prevent bacterial infection. The remodeled of host cell membrane by host-directed antibacterial compounds effectively protected host cells from intracellular bacteria-mediated early infection¹³. In their recent study, Abrams et al. reported that oxysterols, a cholesterol product from cholesterol 25-hydroxylase (CH25H) oxidation, provide innate immunity to bacterial infection by mobilizing cell surface accessible cholesterol¹⁴. Moreover, arsenite, anisomycin and hydrogen peroxide mediated intracellular protein reaction safeguarded intestine cells against intracellular pathogens¹⁵. Thus, host-directed antibacterial compounds displayed the potential to compensate for traditional medicine deficiencies in the treatment of intracellular pathogens.

The present development of host-directed antibacterial compounds mainly directly fights bacterial invasion/survival in host cells to mitigate infection. Furthermore, the identification of combined therapy of host-directed antibacterial compounds with classic antibiotics may robustly improve clinical outcomes¹⁶. Recent reports have demonstrated that acceleration of ROS accumulation in host cells by andrographolide, metformin and berberine promoted the removal of intracellular bacteria^17–19. However, overproduction of ROS may lead to cellular and tissue damage and chronic inflammation in a multitude of neurodegenerative, cardiovascular and metabolic diseases. In contrast, enhancement of lysosomal activity and autophagy biogenesis was recognized as an alternatively promising strategy with efficient activity against bacterial infection, especially with antibiotics [8].

In this study, we are demonstrating that the MDM2 inhibitor Dehydroevodiamine (DEHD) and its analogues are emerging as potent agents against intracellular bacterial infections—including Salmonella, Escherichia coli, Shigella Castellani, Staphylococcus aureus, and Listeria monocytogenes. These compounds are exerting their antibiotics effects by activating the MDM2-P53-V-ATPases axis, thereby initiating host cell autophagy in both antibiotics-treated and untreated conditions. Our findings are uncovering the critical role of the MDM2-P53-V-ATPases axis in orchestrating host defense mechanisms, while simultaneously identifying a series of promising therapeutic candidates for combating intracellular pathogens.

Results

Inhibition of intracellular bacteria proliferation by DEHD in combination with or without antibiotics

By establishing the Salmonella typhimurium (S. Typhimurium) infected host cells model (Fig. 1A), we first examined the inhibitory effect of S. Typhimurium proliferation in the presence of Dehydroevodiamine (DEHD) (Fig. 1B), an isoquinoline alkaloid, or DEHD plus with cefixime. As expected, DEHD significantly reduced S. Typhimurium SL1344 colonization in both Caco-2 cells (Fig. 1C) and RAW264.7 cells (Fig. 1D) in a concentration-dependent manner. Importantly, such reduction was also observed for the treatment of DEHD in combination with 20 μg/mL cefixime, which almost resulted in no visible inhibition of SL1344 colonization, in both Caco-2 cells (Fig. 1E) and RAW264.7 cells (Fig. 1F). However, in the samples received 16 μg/mL DEHD combined with 20 μg/mL cefixime, the SL1344 colonization decreased by 91.06% vs 56.90% for Caco-2 cells (Fig. 1E) and 96.73% vs 55.82% for RAW264.7 cells (Fig. 1F) compared with those treated with 16 μg/mL DEHD. To further confirm the synergistic bactericidal effect of DEHD with cefixime, SL1344 in the host cells were observed under microscopy. Consistently, few bacteria were found in the samples treated with DEHD in combination with cefixime, suggesting that this combination was effective in killing intracellular Salmonella (Fig. 1G−I). To validate such a synergistic effect, the intracellular bactericidal effect of DEHD with various antibiotics was determined. As shown in Supplementary Fig. 1, although no bactericidal effect was observed for the treatment of various antibiotics, DEHD showed excellent potential performance in combination with or without cephalothin (Supplementary Fig. 1A), cefpirome (Supplementary Fig. 1B), cefaclor (Supplementary Fig. 1C), cefuroxime (Supplementary Fig. 1D), amoxicillin (Supplementary Fig. 1E), cefotaxime (Supplementary Fig. 1F), polymyxin (Supplementary Fig. 1G), minocycline (Supplementary Fig. 1H) or ciprofloxacin (Supplementary Fig. 1I) against intracellular Salmonella replication. More importantly, As shown in Supplementary Fig. 1J−M, the intracellular antibacterial efficacy of both the cefixime-treated (Supplementary Fig. 1J) and penicillin G sodium-treated (Supplementary Fig. 1K) groups significantly decreased with increasing passage number. However, the DEHD combination group markedly slowed the development of bacterial resistance. Furthermore, the minimum inhibitory concentrations (MICs) against Salmonella after 15 serial passages (Supplementary Fig. 1L, M) demonstrated that combining DEHD with antibiotics effectively reduced the emergence of resistance compared to antibiotic treatment alone. In addition, we have assayed a diverse range of common intracellular pathogens to verify such synergistic effect. The results showed that DEHD increased antibiotics efficacy against intracellular pathogens growth by 5-50-fold for Shigella Castellani (Supplementary Fig. 2A–C), Listeria monocytogenes (Supplementary Fig. 2D–F), Escherichia coli (Supplementary Fig. 2G–I) and Staphylococcus aureus (Supplementary Fig. 2J–L). Taken together, our results revealed that DEHD showed a trend towards lower bacteria proliferation in host cells, as well as potentiated antibiotics against bacteria colonization and reduced the emergence of bacterial resistance, which providing effective strategies and agent to Fight intracellular pathogen infection by targeting host.

Fig. 1. DEHD effectively limited S. Typhimurium replication in macrophage cell with or without cefixime.

A Schematic diagram of gentamicin protection experiment. S. Typhimurium SL1344 was co-cultured with macrophage cells for 1 h and eliminated with 100 μg/mL gentamicin treatment for 45 min. After replacing the fresh cell culture medium supplemented with 10 μg/mL gentamicin, the cells were further cultured for another 4 h with the indicated treatment. The intracellular S. Typhimurium was counted by agar plating assay or stained with Alexa Fluor® 488 (Abcam). B The chemical structure of Dehydroevodiamine (DEHD). The total number of intracellular S. Typhimurium colony-forming unit (CFU) in Caco-2 cells (C) and RAW 264.7 macrophages (D) was quantified using the Gentamicin protection experiment during treatment with various concentrations of DEHD (0, 16, 32 and 64 μg/mL). The S. Typhimurium CFU in Caco-2 Cells (E) and RAW264.7 (F) treated with various concentrations of DEHD (4, 8 and 16 μg/mL) in combination with cefixime (20 μg/mL). G Fluorescence values in (H) were quantified using Image-J with three replicate images in each group. H Fluorescence images of S. Typhimurium in RAW264.7 macrophages treated with control (DMSO), DEHD (16 μg/mL), cefixime (20 μg/mL) or combination (DEHD and cefixime). Blue fluorescence, RAW264.7 nucleus; Red fluorescence, S. Typhimurium. Scale bars, 50 μm. I The intracellular S. Typhimurium at 4 h post induction with control (DMSO), DEHD (16 μg/mL), cefixime (20 μg/mL) or combination (DEHD and cefixime) was observed under transmission electron micrographs. Scale bars, 5 μm. **, P ≤ 0.01; ns no significance.

Potentiation of antibiotics efficacy by quinoline alkaloids

DEHD is an isoquinoline alkaloid extracted from Euodia rutaecarpa (Juss.) Benth. Then, the above DEHD mediated the potentiation of antibiotics efficacy was also investigated for others quinoline alkaloids. Most unexpected, several quinolines were found to synergize with cefixime against intracellular Salmonella. For instance, the isoquinoline-type natural compound, cephaeline hydrochloride, showed similar potentials profiles with DEHD (Supplementary Fig. 3A). In addition, quinoline derivatives, such as quinine (Supplementary Fig. 3B), berberine (Supplementary Fig. 3C), tetrandrine (Supplementary Fig. 3D), camptothecin (Supplementary Fig. 3E), sinomenine (Supplementary Fig. 3F) or halofuginone (Supplementary Fig. 3G), exhibited similar ability to enhance the efficacy of cefixime, which may be determined by their common parent nuclei (Supplementary Fig. 3B–G). Then, a checkerboard experiment was employed to determine whether these isoquinoline alkaloids could synergize with cefixime against Salmonella SL1344 in vitro. However, none of the tested compounds synergized with cefixime against Salmonella SL1344, as all showed no synergistic, effects (FIC index =2) (Supplementary Fig. 4A–H). Together, our results established that DEHD and other isoquinoline alkaloids boosted the efficacy of antibiotics by targeting host cells rather than bacteria.

DEHD synergized with cefixime mitigated Salmonella infection in vivo

The in vitro improvement of antibiotics activity by DEHD prompted us to further examine such an effect if observed in vivo. The safety and efficacy of DEHD were first evaluated. As expected, incubation with DEHD for both 6 h and 12 h was not cytotoxic against Caco-2 cells (Supplementary Fig. 5A) and RAW264.7 macrophages (Supplementary Fig. 5B) within the effective action concentration (0–64 μg/mL). In line with these observations, the addition of DEHD (0–128 μg/mL) almost showed no hemolysis for the rabbit erythrocyte (Supplementary Fig. 5C). Additionally, the growth of SL1344 was not visibly affected by the treatment with 8–32 μg/mL DEHD, suggesting that DEHD is likely to Fight Salmonella infection through host targeting (Supplementary Fig. 5D). Furthermore, DEHD treatment remarkably enhanced the performance of cefixime against Salmonella-mediated host cell injury, as revealed by much lower lactate dehydrogenase released (Supplementary Fig. 5E) and dead cells observed (Supplementary Fig. 5F) in the co-infection system with Salmonella and RAW264.7 cells. Thus, DEHD, as an agent, is an un-toxic and effective antibacterial synergist in vitro.

We first conducted a comprehensive assessment of the pharmacological toxicity of DEHD in mice to select the optimal dosage. Comprehensive toxicity assessment in mice demonstrated that DEHD (0–50 mg/kg) caused no hematological abnormalities (Supplementary Fig. 6A), specifically, the types and numbers of white blood cells in the blood remain stable. Further studies on intestinal inflammatory factors revealed that high concentrations (50 mg/kg) of DEHD promote the secretion of IFN-γ, IL-1β, and TNF-α in the intestines of mice (Supplementary Fig. 6B). Although no significant differences were observed in HE-stained sections (Supplementary Fig. 6C), considering the subsequent combination treatment with antibiotics, we selected a moderate dose of DEHD as the effective concentration for animal experiments. To further assess the efficacy of DEHD in conjunction with cefixime in the clearance of bacterial infection, a mouse model of Salmonella-infected colitis was developed, whereby the bacteria were administered orally at a concentration of 5 × 10⁷ CFUs/mouse or 1 × 10⁸ CFUs/mouse to establish sublethal infection or lethal infection, respectively (Fig. 2A). Based on the results of preliminary experiments (Supplementary Fig. 6D), DEHD, at the dose of no less than 50 mg/kg, significantly reduced the bacteria load in the liver and colon of the infected mice. Then, to assay the potential syngenetic effect of DEHD in vivo, DEHD (25 mg/kg), cefixime (25 mg/kg) or DEHD (25 mg/kg) plus with cefixime (25 mg/kg) was orally administered at 4 h after Salmonella infection and then at 24-h intervals thereafter for a total of five doses (Fig. 2A). As revealed, the treatment with DEHD or cefixime showed on significant influence on the weight (Fig. 2B) and diet (Fig. 2C) of infected mice. However, combined therapy significantly reversed the weight loss and appetite loss associated with Salmonella infection (Fig. 2B, C). In agreement with the above results, DEHD treatment or cefixime treatment displayed no effective therapeutic effect, whereas, combined treatment remarkably reduced the mortality of infected mice with lethal infection by 80% (Fig. 2D). Additionally, DEHD effectively attenuated Salmonella-mediated pathological damage (Fig. 2E and Supplementary Fig. 6E) and bacteria load in liver (Fig. 2F), spleen (Fig. 2G), caecum (Fig. 2H) and colon (Fig. 2I) in combination with cefixime. Tissue immunofluorescence demonstrated that autophagy was markedly activated in the liver and caecum of the DEHD-treated mice, as evidence by increased LC3B was observed (Supplementary Fig. 7). Thus, the above results indicated that a certain degree of autophagy activation by DEHD in the animal organism may facilitate the defense against intracellular pathogenic bacteria. Although cefixime treatment effectively inhibited inflammatory response in the Salmonella-infected mice, such response was significantly lower in the infected mice treated with DEHD in combination with cefixime by comparing those mice that received monotherapy with DEHD or cefixime, as evidenced by increased levels of IL-10 (Fig. 2J) and decreased level of IL-1β (Fig. 2K), IFN-γ (Fig. 2L), IL-6 (Fig. 2M) and TNF-α (Fig. 2N) in colon. In line with these observations, DEHD (25 mg/kg) in conjunction with tetracycline markedly diminished the number of bacteria colonies in the liver (Supplementary Fig. 8A) and spleen (Supplementary Fig. 8B) of mice infected with Listeria monocytogenes. Furthermore, the administration of DEHD with tetracycline effectively alleviated the pathological damage in the liver and spleen of infected mice (Supplementary Fig. 8C), exhibiting a mild inflammatory reaction and a notable reduction in necrotic tissue within the microscopic field of view. Taken together, our results indicated that DEHD effectively improved the therapeutic performance of antibiotics for infection by both Gram-negative bacteria and Gram-positive bacteria.

Fig. 2. DEHD protected mice from S. Typhimurium infection.

A Schematic diagram of the establishment of S. Typhimurium infection model. BALB/c mice received 5 g/L of streptomycin in drinking water for three days were orally inoculated with 5 × 10⁷ CFUs/mouse for sublethal infection or 1 × 10⁸ CFUs/mouse for lethal infection, respectively. DEHD (25 mg/kg), cefixime (25 mg/kg), or DEHD (25 mg/kg) in combination with cefixime (25 mg/kg) were orally administered to infected mice at 4 h after infection and then every 24 h for five doses. All the mice with sublethal infection were euthanized at 5 days after infection. Then, the targeted organs were collected for pathology evaluation. Changes in body weight (B) and dietary intake (C) were monitored in each group of mice (n = 9) during the sublethal infection. D A log-rank test was used to monitored the survival of mice with lethal infection over 7 days (n = 10). E After 5 days of infection, the liver, spleen, caecum and colon from sacrificed mice in each group were collected for microscopic pathology evaluation with HE staining analysis. Scale bars, 100 μm. The viable bacteria in the homogenates of liver (F), spleen (G), cecum (H) and colon (I) of each group mice (n = 9) were determined by plating. In addition, ELISA analysis was used for the measurement of inflammatory factors IL-10 (J), IL-β (K), IFN-γ (L), IL-6 (M) and TNF-α (N) in colon tissue of each group mice (n = 3). **, P ≤ 0.01; ns no significance.

Activation of the lysosome by DEHD is essential for the eradication of Salmonella

Lysosome, as a dense vesicle, is an important executor in clearing bacteria in macrophage²⁰, which was observed under transmission electron microscopy in the Salmonella-infected host cells received DEHD plus with cefixime (Fig. 1I), suggesting that combined therapy prompted the autophagy in the host. Then, mono-dansylcadaverine (MDC) and lyso-tracker-red were used to label lysosomes. As expected, like autophagy agonist Earle's Balanced Salt Solution (EBSS), DEHD induced robust activation of cytosolic lysosome with increased lyso-tracker-red fluorescence compared to untreated cells (Fig. 3A, B). Furthermore, inhibition of lysosomal acidification was observed with the treatment of late-stage autophagy antagonist bafilomycin A1 (BafA1) or lysosomal acidification antagonist chloride (NH₄Cl) in both samples with or without DEHD treatment (Fig. 3A, B), suggesting that DEHD showed a potential in activating autophagy in host cells. Consistently, the treatment with 16 μg/mL DEHD caused a dramatic reversal of early Salmonella infection-mediated inhibition of macrophage lysosomal activity (Fig. 3C, D), which is important for macrophage clearance of intracellular Salmonella. Acidification of lysosomes usually requires ATP²¹. To elucidate the lysosomal kinetics of DEHD-treated cells, the ATP, NADH and NADPH contents were determined at different time points. As shown in Fig. 3E, the level of cellular ATP downward fluctuated greatly after on hour of DEHD treatment and normalized after three h. Differing from ATP, the levels of NADH (Fig. 3F) and NADPH (Fig. 3G) have not been influenced by DEHD treatment, which may be due to the occurrence of intrinsic cellular compensatory mechanisms. To further verify whether lysosomal activation is a key component of DEHD action, the clearance of intracellular Salmonella by DEHD synergized with cefixime was evaluated. The results, as shown in Fig. 3H, showed that late-stage autophagy antagonist BafA1 and lysosomal acidification antagonist NH₄Cl, but not the early-stage autophagy inhibitor chloroquine (CHQ) and the ATPase inhibitor oligomycin (OLI), successfully reversed the synergistic effect of DEHD, suggesting that lysosomal activation is a key factor in enhancing antibiotic efficacy intracellularly by DEHD. In addition, an increase of intracellular electron-dense structures and appearance of autophagic lysosomes after DEHD treatment was observed under electron microscopy (Fig. 3I). Thus, our results suggested that DEHD treatment increased the activity of lysosome for effective eradication of Salmonella in host cells.

Fig. 3. DEHD activated lysosomes and reversed lysosomal inhibition by S. Typhimurium early infection.

A Lysosomal activation by DEHD. RAW264.7 macrophages were treated with DEHD (16 μg/mL, 4 h), EBSS (starvation control), bafilomycin A1 (BafA1, 100 nM), BafA1 + DEHD, NH4Cl (200 nM), or NH4Cl + DEHD. Lysosomal activity was assessed by co-staining with monodansylcadaverine (MDC-green, autophagosomes) and LysoTracker Red (lysosomes). Scale bars, 10 μm. B Quantification of LysoTracker Red fluorescence intensity for treatments in (A). C DEHD reverses S. Typhimurium-induced lysosomal dysfunction. Cells were infected with S. Typhimurium (MOI = 25, 1 h), treated with or without DEHD (16 μg/mL), and stained with LysoSensor Red to measure lysosomal activity. Nuclei were counterstained with DAPI. Scale bars, 5 μm. D Fluorescence quantification for (C). Metabolic changes in macrophages. ATP levels (E), NADH (F), and NADPH (G) were measured in macrophages treated with control (DMSO), DEHD (16 μg/mL), cefixime (20 μg/mL), or DEHD + cefixime for 4 h. H DEHD-mediated suppression of intracellular S. Typhimurium replication requires lysosomal function. Infected macrophages (MOI = 25, 1 h) were treated with DEHD ± BafA1 (100 nM), NH4Cl (200 nM), chloroquine (CHQ, 1 μM), or oritavancin (OLI, 1 μM). Intracellular bacteria were quantified by CFU assay. I TEM visualization of autophagic lysosomes. DEHD (16 μg/mL, 4 h) increased autolysosome formation (red arrows), indicating enhanced autophagic flux. Scale bars, 5 μm. J–L DEHD modulates autophagy markers. LC3B-II accumulation and SQSTM1 degradation were assessed by Western blot after DEHD treatment at varying concentrations (J) or time points (K). Quantification by ImageJ is shown in (L). M, N LC3B puncta formation in infected cells. Immunofluorescence staining revealed increased LC3B aggregation (red) post-S. Typhimurium infection with DEHD treatment (16 μg/mL, 4 h). Nuclei were stained with DAPI (blue). Scale bars, 5 μm. O, P Autophagic flux analysis. GFP-mCherry-LC3B-transfected macrophages treated with DEHD showed increased red puncta (autolysosomes) and decreased yellow puncta (autophagosomes), quantified in (P). Scale bars, 5 μm. Q, R LAMP1-LC3B colocalization. Confocal microscopy and ImageJ analysis confirmed enhanced lysosome-autophagosome fusion (colocalization ratios in R). S Lysosomal inhibitors block DEHD effects. LC3B/SQSTM1 expression in macrophages treated with DEHD ± NH4Cl (200 nM) or BafA1 (100 nM). *, P ≤ 0.05; **, P ≤ 0.01; ns no significance.

To further validate the above observations, the extent of autophagy was assessed by determination of the autophagy related proteins LC3B and SQSTM1 expression. As expected, DEHD treatment promoted macrophage autophagy in a dose-dependent (Fig. 3J, L) and time-dependent (Fig. 3K) manner by upregulating LC3B expression and downregulating SQSTM1 expression. Immunofluorescence staining analysis also confirmed such promotion effect, as evidenced by remarkable activation of autophagy for the uninfected cells and reversal of Salmonella infection mediated autophagy inhibition following the treatment with DEHD (Fig. 3M, N). Adenovirus expressing mCherry-GFP-LC3B fusion protein was transfected with macrophages to monitor the autophagic flux of cells treated with or without DEHD. Increased autophagic flux and fluorescent co-localization of LC3B with LAMP1 (Fig. 3P), a hallmark of autophagic lysosome formation, was found in the macrophages transfected with mCherry-GFP-LC3B under DEHD treatment, suggesting that DEHD exhibited a robust activation of autophagy and autolysosome formation (Fig. 3O). As shown in Fig. 3R, the co-localization of LC3B with LAMP1 overlap coefficient of 0.87 in the DEHD-treated group was much higher than that of 0.26 in the untreated group under similar exposures (Fig. 3R). The fluorescence of co-localization results further verified the enhancement of autophagic flux in macrophages by DEHD treatment.

Although the formation of autophagic lysosomes and the expression of marker proteins are the most commonly used indicators to detect autophagy, the blockage of autophagic pathways should be further examined. As revealed, the expression of autophagy related proteins in the DEHD-treated samples was similar to that of EBSS with decreased SQSTM1 and increased LC3B (Fig. 3S). In line with previous results, the treatment with BafA1 or NH4Cl effectively inhibited the autophagic flux of macrophages, a reversal effect was observed in the samples treated with DEHD with BafA1 rather than NH4Cl, further indicating that DEHD could boost the autophagic flux of SQSTM1 transferring into LC3B (Fig. 3S). Taken together, our results established that DEHD is a potent activator of macrophage autophagy.

Identification of the mechanism of DEHD mediated the potentiation of antibiotics efficacy

The mechanism of action of DEHD to potentiate autophagy was further validated based on combined transcriptomics and metabolomics analysis. The hierarchical clustering method was used to cluster genes with similar expression patterns, indicating that DEHD treatment could induce significant changes in transcription levels of host cells (Fig. 4A). The differential gene volcano map directly showed the overall distribution of genes with significant differences, as evidenced by 415 genes were down-regulated and 589 genes were up-regulated (Fig. 4B). Bar graphs and pie charts for the 20 KEGG pathways with the most significant gene clustering were further plotted. The results, as shown in Fig. 4C–E, showed that DEHD treatment resulted in the most significant increase in the levels of cellular transcription factors, as well as p53 signaling pathway and AMPK signaling pathway. By further analyzing the results based on the STRING protein interaction database, we constructed interaction networks for the interactions of the target genomes and imported them into Cytoscape software for visualization and editing. The results, as shown in Fig. 4F, indicated that DEHD may exert its antibiotics function by activating autophagy through the P53-mediated mTOR-dependent pathway.

Fig. 4. Identification of the mechanism of DEHD on macrophages by transcriptomics and metabolomics analysis.

RAW264.7 macrophages treated with or without 16 μg/mL DEHD for 4 h were used for transcriptomics and metabolomics analysis. A Differential genes expression in macrophages with heatmap. B Distribution of differentially expressed genes with volcano plot. C–E The top 20 most significant pathways from the KEGG database were selected and plotted as KEGG enrichment results. F The STRING protein interaction database was applied to obtain the protein interaction networks based on transcriptomics analysis. The importance of secondary metabolites was analyzed using volcano plot (G), correlation heatmap (H) and clustering heatmap (I) with metabolomics analysis. Red, a positive correlation; blue, a negative correlation. J KEGG pathway analysis showed the correlation coefficients between differentially expressed genes and metabolites. K The effect of DEHD on the expression of indicated genes was determined by RT-PCR analysis. Red, an up-regulation effect; blue, a down-regulation effect.

Evaluation of the intracellular distribution of differential metabolites and the changing trends could facilitate us to further understand the mechanism of drugs action. Thus, the secondary differential metabolites in host cells treated with or without DEHD were analyzed, which suggesting that dozens of metabolites were found to be significantly different in these two groups of samples (Fig. 4G). Based on the synergistic or mutually exclusive relationship between different metabolites, the Pearson correlation coefficients of differential metabolites were calculated to determine the correlation of individual metabolites, as well as the samples and differential metabolites were simultaneously clustered in both directions. Fumaric acid and L-histidine was upregulated by DEHD treatment with remarkable correlation (Fig. 4G–I). Notably, the mTOR signaling pathway was significantly induced in the secondary differential metabolite identification analysis (Fig. 4I) as well as in the differential metabolite-differential genome co-analysis (Fig. 4J), revealed that DEHD most likely perform the function in an mTOR signaling pathway-dependent manner. To verify the results of the transcriptomic and interaction networks analysis, the key genes associated with the P53-AMPK-mTOR-autophagy pathway by RT-PCR assay. As expected, DEHD treatment significantly increased the expression of mTOR, AMPK, Trp53 and MDM2 (Fig. 4K). Together, our results indicated that DEHD treatment activated P53-AMPK-mTOR-autophagy pathway for the Fight against bacteria with antibiotics.

DEHD promoted P53 entry into the nucleus to initiate the transcription process

The P53 protein, as a major intracellular transcription factor, can regulate multiple cellular pathways such as pyroptosis, apoptosis, autophagy and DNA repair. Previous work has reported that P53 initiates subsequent cellular processes mainly through nucleation²². Thus, in this study, immunoblotting assay was used to detect nucleoprotein, cytoplasmic protein and total protein of P53 in DEHD treated macrophages for 4 h. The results, as shown in Fig. 5A, B, revealed DEHD concentration-dependently increased the intracellular expression of P53 with reduced cytoplasmic P53 and, however, robustly elevated intranuclear P53. Then, the intracellular localization of P53 protein was examined by fluorescence labeling. As shown in Fig. 5C, P53 gradually aggregated in the nucleus with increasing time from 1 to 4 h, suggesting that DEHD treatment prompted the entry of P53 from cytoplasm into the nucleus. Notably, the signaling pathway associated with P53 was activated by DEHD for promotion of autophagy, evident by increased expression of phosphorylated AMPK with an effective reversal of phosphorylated mTOR mediated inhibition for phosphorylated ULK1 (Fig. 5D). In addition, DEHD treatment inhibited the level of phosphorylated mTOR/mTOR and increased the level of phosphorylated AMPK/AMPK (Fig. 5E), suggesting that DEHD could activate the phosphorylation at the Ser555 site of ULK1 protein. To prove suppose that above, we treated samples with the AMPK inhibitors dorsomorphin and BAY3827 (Fig. 5F–H), as well as the mTOR agonists MHY1485 and 3BDO (Fig. 5I–K). Both AMPK inhibitors reversed DEHD’s synergistic antibacterial activity and attenuated DEHD-induced lysosomal acidification. Similar effects were observed with mTOR agonists. These results indicate that DEHD activates autophagy against intracellular Salmonella infection by inducing lysosomal acidification via the AMPK-mTOR pathway.

Fig. 5. DEHD facilitated the entry of P53 into the nucleus to initiate macro-autophagy.

A The expression of P53 protein in the nucleus or cytoplasm of RAW264.7 macrophages with the addition of DEHD (0, 4, 8 or 16 μg/mL) for 4 h incubation. B Quantification of the P53 protein expression in (A). C Intracellular localization of P53 protein in RAW264.7 macrophages treated with or without DEHD (16 μg/mL) for the indicated times. Green fluorescence, P53 protein; blue, nucleus. Scale bars, 20 μm. D The expression of protein associated with MDM2-P53 pathway in RAW264.7 macrophages treated with DEHD (0–16 μg/mL) for 4 h. E Quantification of phosphorylated MTOR/MTOR and phosphorylated AMPK/AMPK using Image J based on (D). Infected macrophages (MOI = 25, 1 h) were pretreated with AMPK inhibitors Dorsomorphin, BAY3827 (F), or mTOR agonists MHY1485 and 3BDO (I) before receiving combined treatment with DEHD and cefixime. Intracellular bacterial counts were quantified using the colony-forming unit (CFU) assay. Representative images of lysosomal fluorescence in cells pretreated with AMPK inhibitors (G) or mTOR agonists (J). Total cellular fluorescence intensity in G (H) and J (K). Scale bars, 10 μm. L DEHD synergizes with cefixime to combat intracellular Salmonella infection in ATG5^flox/flox and ATG5^−/− macrophages. M Representative images of lysosomal fluorescence in DEHD treated in ATG5^flox/flox and ATG5^-/- macrophages. Scale bars, 10 μm. N Cell integrated density in (M). *P ≤ 0.05, **P ≤ 0.01; ns no significance.

DEHD increased the dissociation of MDM2-P53 by targeting MDM2

The dissociation of MDM2-P53 is required for the activity of P53. Thus, an indirect ELISA was used to determine the MDM2-P53 dissociation coefficient. As shown in Fig. 6A, B, DEHD dramatically promoted MDM2-P53 dissociation in a concentration-dependent and time-dependent manner. The surface plasmon resonance technique was employed to confirm whether DEHD could bind with MDM2. And such binding was observed with the KD of 68.34 μM Fig. 6C, D).

Fig. 6. DEHD binds directly to MDM2 and dissociates the MDM2-p53 complex.

Indirect ELISA assay was utilized to validate the concentration-dependent (0, 4, 8 and 16 μg/mL) (A) and time-dependent (0, 1, 2, 3, 4 and 5 hours) (B) dissociation of the MDM2-P53 complex by DEHD treatment. C The protein affinity analysis of MDM2 with different concentrations of DEHD (0–800 μM) by SPR assay. D The kinetic profile of DEHD binding to MDM2 was used to calculate the dissociation constant. E RMSD of the protein backbones of free MDM2 from the initial coordinates as a function of time. F RMSD of the protein backbones of MDM2-P53 and MDM2-DEHD complexes from the initial coordinates as a function of time. The stable 3D structure of MDM2-P53 (G) and the binding free energy decomposition of the binding sites in MDM2- P53 complex (H). The stable 3D structure of MDM2-DEHD (I) and the binding free energy decomposition of the binding sites in MDM2-DEHD complex (J). K The Potential of Mean Force (PMF) of DEHD with wild-type MDM2 and its mutants (V41A, G42A, N45A, M50A and I99A). L The number of hydrogen bonds over the 600 ns simulation in the MDM2-DEHD complex. M The percentage of the hydrogen bonds formed by DEHD with MDM2 over the 600 ns simulation. N The radial distribution function of MDM2 residues in the binding sites around the DEHD. *P ≤ 0.05, **P ≤ 0.01.

A molecular docking analysis was performed to evaluate the affinity of DEHD for MDM2. Then, to further elucidate the interaction mechanism between the small molecule inhibitor DEHD and MDM2, we conducted standard molecular dynamics simulations on three systems: the isolated MDM2 binding domain protein, the MDM2-P53 complex, and the MDM2-DEHD complex. The MDM2 binding domain protein, constructed using a deep learning online server (I-TASSER-MTD), reached equilibrium after a 2-microsecond simulation (Fig. 6E). Utilizing the equilibrated receptor protein 3D structure, both the MDM2-P53 and MDM2-DEHD complexes achieved a converged equilibrium state after 600 ns of simulation (Fig. 6F). The stable structure of the MDM2-P53 complex simulation revealed that the substrate P53 binds to the MDM2 binding pocket, with the main binding sites being residues 38–57, 86–103, and 145–150 (Fig. 6G). The binding free energy decomposition results indicated that the binding energies of LEU-38, VAL-41, MET-50, ILE-53, ILE-54, VAL-88, LYS-94, and LEU-149 within the binding pocket were all less than −1.0 kcal/mol, identifying them as the principal binding site residues (Fig. 6H).

For the MDM2-DEHD complex, the stable structure showed that the inhibitor DEHD also binds to the MDM2 binding pocket and can form potential hydrogen bond interactions with ASN-45 (Fig. 6I). As shown in Fig. 6J, the binding free energy decomposition further indicated that VAL-41, GLY-42, and ASN-45 within the binding pocket contributed significantly to the binding free energy (< −1.0 kcal/mol), with ASN-45 exhibiting a stronger binding free energy value of −5.0 kcal/mol. To validate these findings, we calculated the Potential Mean of Force (PMF) for both the wild-type MDM2- DEHD complex and the mutant MDM2- DEHD complexes using Adaptive Steered Molecular Dynamics (ASMD). As shown in Fig. 6K the binding free energy of the ligand DEHD with wild-type MDM2 was significantly higher than with mutants (V41A, G42A, N45A, M50A, I99A), confirming that the main binding sites of the MDM2- DEHD complex are VAL-41, GLY-42, and ASN-45, consistent with the binding free energy decomposition results.

Notably, in the MDM2- DEHD complex, the inhibitor DEHD can form potential hydrogen bond interactions with ASN-45. Hydrogen bond analysis of the simulation trajectory demonstrated that 1–2 hydrogen bonds can form between the receptor protein and the ligand molecule when the system structure reaches equilibrium (Fig. 6L). Among the potential amino acid residues that can form hydrogen bonds, only ASN-45 formed stable interactions throughout the simulation (Fig. 6M). The radial distribution function (RDF) results also indicated that only ASN-45 had a high density at a radial distance of 3.71 Å from the ligand (Fig. 6N), suggesting that DEHD forms stable hydrogen bond interactions with ASN-45 in the mdm2-ddd complex. In summary, the main binding sites in the MDM2- DEHD complex are VAL-41, GLY-42, and ASN-45, overlapping with the binding pocket region of the substrate P53. The inhibitor DEHD acts as a typical competitive inhibitor of MDM2.

MDM2-P53-Vatpase axis activated by DEHD inhibited Salmonella infection

To verify the function of P53 against intracellular bacterial infection, the potential inhibitory effect of the P53 activators Tenovin-1 and Nutlin-3 on intracellular Salmonella SL1344 replication in RAW264.7 macrophages as well as Balb/c mouse peritoneal macrophages was further examined. As expected, both Tenovin-1 (10 μM) and Nutlin-3 (5 μM) had a similar synergistic effect as DEHD with cefixime to inhibit Salmonella colonization in RAW264.7 cells (Fig. 7A) and mouse peritoneal macrophages (Supplementary Fig. 94A). Then, siRNA gene interference against Mdm2 or Trp53 was employed to verify such inhibitory effect. As revealed, interference of MDM2 enhanced the efficacy of cefixime against Salmonella infection in vitro with or without DEHD treatment (Fig. 7B). Whereas interference of Trp53 abolished the synergistic effect of DEHD with cefixime against Salmonella infection (Figs. 7B and S9B), suggesting that P53 is required for the inhibition of Salmonella infection by DEHD in combination with cefixime. To further substantiate this conclusion, we constructed Mdm2 or Trp53 overexpression plasmids (Supplementary Fig. 9C, D) using pcDNA3.1 in RAW264.7 macrophages. As expected, overexpression of P53 protein increased the intracellular antibiotics activity of cefixime with or without DEHD, but such activity was not observed in Mdm2-overexpressing macrophages, even in the presence of DEHD (Fig. 7C). Notably, although activation of Trp53 by interference of Mdm2 or overexpression Trp53 improved cefixime performance, the inhibition of Salmonella replication combination with DEHD was more effective in these cells without DEHD (Fig. 7B, C), further indicating DEHD enhanced the activity of cefixime with a P53 dependent manner. To validate these results, the CRISPR-Cas9 system was used to generate both Trp53 and Mdm2 knockdown cell lines. In agreement with the results of siRNA interference analysis, knockdown of P53 protein significantly reversed the synergistic effect of DEHD with cefixime against Salmonella replication (Fig. 7D). However, knockdown of Mdm2 enhanced the ability of cefixime against intracellular Salmonella (Fig. 7E).

Fig. 7. Driving of MDM2-P53-Vatpase axis by DEHD effectively inhibited intracellular Salmonella infection in macrophages.

Macrophages or gene-edited macrophages were infected with S. Typhimurium SL1344 at a MOI of 25 for 1 h, incubated with gentamicin for 45 mins and treated with the indicated treatment for 4 h. Then, the intracellular S. Typhimurium was counted by agar plating assay. A The influence of DEHD (16 μg/mL), 100 nM BafA1 and the MDM2 inhibitors Tenovin-1 (10 μM) or Nutlin-3 (5 μM) on intracellular S. Typhimurium replication in primary mouse peritoneal macrophages with or without cefixime. B The S. Typhimurium replication in RAW264.7 macrophages with or without siRNA interference for MDM2 or Trp53 genes in the presence of control (DMSO), DEHD (16 μg/mL DEHD), cefixime (20 μg/mL cefixime) or combination (16 μg/mL DEHD plus with 20 μg/mL cefixime) treatment. C The S. Typhimurium replication in RAW264.7 macrophages with or without Mdm2 or Trp53 overexpression plasmids in the presence of the indicated treatment as (B). The S. Typhimurium replication in Trp53 knockdown (D) or Mdm2 knockdown (E) RAW264.7 macrophages in the presence of the indicated treatment as (B). F The S. Typhimurium replication in RAW264.7 macrophages with Trp53 overexpression plasmids, ATP6V0D1 knockdown plasmids or Trp53 overexpression plasmids and ATP6V0D1 knockdown plasmids in the presence of the indicated treatment as (B). G The effect of gene-edited macrophages on the expression of autophagy related proteins LC3B and SQSTM1. H The S. Typhimurium in macrophages as in panel F was further fluorescently labeled. Scale bars, 10 μm. I The degree of lysosomal acidification in macrophages as in F without any infection or treatment. Scale bars, 10 μm. J Schematic diagram of the establishment of S. Typhimurium infection Trp53-KO or ATP6V0D2-KO mice model. K Detection of autophagy-related proteins LC3B and SQSTM1 expression in KO mice liver. L The survival of mice with lethal infection was monitored over 7 days (n = 10). Changes in dietary intake (M) and body weight (N) were monitored in each group of mice (n = 6) during the sublethal infection. The viable bacteria in the homogenates of liver (O), spleen (P), cecum (Q) and colon (R) of each group mice (n = 6) were determined by plating. *P ≤ 0.05; **P ≤ 0.01; ns no significance.

Noteworthily, all the above observations have not provided sufficient evidences uncovering the mechanism of bafilomycin A1 rescuing the synergistic effect of DEHD. In this context, we constructed a knockdown cell line for ATP6V0D, a key gene required for lysosomal acidification, based on both RAW264.7 macrophages and RAW264.7 macrophages with pCDNA3.1-Trp53. Detection of lysosomal acidity revealed that p53 overexpression significantly activated the acidification of lysosome and the process of autophagy with an effective eradication of Salmonella in the presence of cefixime by comparing the WT cells without pCDNA3.1-Trp53 (Fig. 7G–I). Knockdown of Atp6v0d2 strikingly inhibited lysosomal acidification with increased Salmonella in both RAW264.7 macrophages with or without pCDNA3.1-Trp53 (Fig. 7G–I), suggesting that inhibition of lysosomal acidification reversed the effect of DEHD improving cefixime activity. Next, to verify the key role of Trp53 and Atp6v0d2 in activating autophagy against pathogenic bacterial infection (Fig. 7J), we constructed Trp53 and Atp6v0d2 KO mice (Supplementary Fig. 9E). Unsurprisingly, the knockdown of Trp53 significantly decreased the expression of LC3B-II, but the knockdown of Atp6v0d2 further increased the expression of LC3B-II (Fig. 7K). However, in combination with the results of Fig. 7I, we concluded that the knockdown of Atp6v0d2 likely exerted the role of BafA1 to hinder autophagy. Next, vivo experiments showed that the knockdown of Trp53 or Atp6v0d2 reversed the increased survival (Fig. 7L), appetite recovery (Fig. 7M) and weight (Fig. 7N) gain associated with DEHD combination therapy in infected mice. Additionally, Trp53 or Atp6v0d2 knockdown effectively increased bacteria load in the liver (Fig. 7O), spleen (Fig. 7P), caecum (Fig. 7Q) and colon (Fig. 7R) compared to DEHD combination with cefixime. Taken together, our results established that activation of the MDM2-P53-Vatpase axis by DEHD upregulated lysosomal acidification and thus enhanced the intracellular antibiotics activity of antimicrobials.

Discussion

Throughout evolution, host and pathogenic bacteria have developed a unique “tacit understanding” [6]. Host cells Fight pathogens via their natural immune mechanisms such as iron death, apoptosis, pyroptosis and autophagy. However, as the other side of the coin, pathogens have also developed sophisticated “spear” to penetrate the host cell “shields” to escape or confuse the immune system through a variety of immune evasion mechanisms including ion channel disorders, activated efflux pump and developed containing vacuole²³. Thus, bacteria-directed antibiotics remain optimal solution for intracellular bacterial infection. However, the efficiency of antibiotics has been remarkably limited with the emergence of bacteria resistance²⁴. Moreover, our results indicated that the intracellular MICs of various antibiotics for Salmonella, Shigella, Listeria monocytogenes, Escherichia coli and Staphylococcus aureus increased by 4-25-fold compared with the extracellular MICs (Table S2). This phenomenon may be due to reduced antibiotic penetration of the host membrane, activation of bacterial efflux pumps, or vacuolar retention—factors that collectively limit drug entry into cellular pathogens and further weakening the effectiveness of antibiotics against intracellular bacteria. Thus, novel strategies and agents are urgently needed in the war against bacterial infection, especially for intracellular bacteria.

Recently, Zhu et al. have reported that endocytosis-mediated redistribution of nano-antibiotics is a promising alternative to eradicate intracellular bacterial infection. However, the unexpected toxicological and off-target of nanocarriers challenged the clinical therapy efficacy²⁵. Another perspective focused on the antibiotic-cell-penetrating peptide strategy in Fighting intracellular pathogen. Modification of existing antibiotics by combining them with cell-penetrating peptides holds promise for restoring cellular defense mechanisms and reducing bacteria evasion in host cells. Again, the toxicity and uncontrollability associated with exogenous materials still were inescapable²⁶. Then, recommission immunomodulatory agents represented a greatly attractive strategy to combat intracellular bacterial infection, especially in combination with traditional antibiotics. Metformin, a maturely drug used for type-2 diabetes, enhanced the efficacy of standard antibiotics with remarkably reduced disease pathology²⁷. In additionally, eight commercially available statins have been identified as potent adjunctive, host-directed therapy, in combination with first-line tuberculosis drugs to effectively eliminate Mycobacterium tuberculosis in host²⁸.

Here, we identified a series of isoquinoline derivatives as broad-spectrum antibiotics against intracellular pathogens. Interestingly, these isoquinoline alkaloids also significantly enhanced the activity of several classes of antibiotics against intracellular pathogens at low concentrations (4–16 μg/mL) without synergistic bactericidal effect in vitro with the tested antibiotics. In the in vivo experiments, DEHD significantly improved the therapeutic effect of cefixime against S. typhimurium and reduced the intestinal colonization of bacteria. Importantly, DEHD synergized with cefixime reduced the mortality of Balb/c mice with lethal infection from 90 to 10%. Additionally, our results further demonstrated that DEHD treatment promoted lysosomal acidification and thus synergizes with antibiotics against intracellular pathogens by activating innate immunity. The loss of DEHD's efficacy in ATG5-deficient macrophages conclusively establishes autophagy as the essential effector mechanism. Thus, these data provided compelling evidence for enhancing host cell intrinsic immunity against intracellular pathogens as a potentially promising therapeutic strategy.

In this study, we have tapped into a novel autophagy activator, DEHD, exerting significant efficacy against infection at the cellular level as well as in mice. Lysosomal acidification against recalcitrant infection mediated by intracellular pathogenic mycobacteria further supports the important role of mTOR-dependent antimicrobial autophagy in host innate immunity. Unlike rapamycin—which induces autophagy via direct mTORC1 inhibition but impairs lysosomal acidification—or metformin—which activates AMPK yet requires high doses risking lactic acidosis—DEHD uniquely synchronizes autophagy initiation (via P53-AMPK-mTOR) with lysosomal acidification (through V-ATPase activation). This dual functionality enables synergistic antibiotic efficacy at low doses (16 μg/mL) without metabolic toxicity (Fig. S5), offering a mechanistically distinct and therapeutically advantageous host-directed strategy. Here, our observation validated that the mTOR-mediated autophagy process is driven by P53-activated AMPK phosphorylation due to DEHD binding to the C-terminal (VAL-41, GLY-42 and ASN-45) of MDM2, thereby inhibiting MDM2 activity and preventing the degradation of P53 by MDM2. It must be acknowledged, however, that although the DEHD demonstrates high affinity for MDM2's P53-binding cleft, we cannot rule out interactions with structurally homologous domains (e.g., MDMX). Given that quinoline alkaloids like berberine exhibit pleiotropic effects, future chemical proteomics studies (e.g., affinity pull-down + MS) will revealed that map DEHD's interactome [18].

Furthermore, our results suggested that P53-driven activation of V-ATPase enzymes synergizes with broad-spectrum antibiotics against a wide range of intracellular pathogens, which may be applicable in the perioperative period after prophylactic treatment of clinical diseases and further expanding the importance of P53 in translational self-healing of diseases²⁹. Nevertheless, the pharmacokinetic profile of DEHD and its broad-spectrum activity against diverse bacterial species require further elucidation. Therefore, we future work will validate DEHD in primary human macrophages and clinical isolates to confirm translational potential, leveraging the conserved pathway architecture revealed here.

From our work DEHD was identified as MDM2 inhibitors to effectively activate the MDM2-P53-Vatpase axis and, then, multi-factorially Fight intracellular bacteria with or without antibiotics. In conclusion, the present study suggested a possible regulatory network that exhibited potent antipathogenic functions through increased lysosomal acidity by DEHD, which may be a promising host-directed therapy for intracellular bacteria.

Materials and methods/experiment

Mammalian cell culture

Caco-2 epithelial cells and RAW264.7 macrophage cells were maintained at 37 °C with 5%CO₂ in Dulbecco’s Modified Eagle Medium (DMEM, high glucose, Cytiva) containing 10% (v/v) heat-inactivated fetal bovine serum (FBS, CellBox). Cells were seeded at 2 × 10⁴ cells/well in 96-well plates or seeded at 2 × 10⁵ cells/well in 24-well plates 16 h before bacterial infection (MOI = 25).

Peritoneal macrophages were derived from wild-type BALB/c mice pretreated with 5% sodium mercaptoacetate for 3 days³⁰. The peritoneal of mice immersed in alcohol (75%) was cut and flushed to collect peritoneal lavage with a 5-mL syringe. And the collected peritoneal lavage suspension was lysed with eBioscience buffer (Thermofisher) to remove red cells, plated and cultured in RPMI 1640 containing 10% FBS at 37 °C with 5%CO₂ for 4 h.

Bacterial strains

S. aureus USA300, S. Typhimurium SL1344, E. coil ATCC25922, L. monocytogenes ATCC19115, Sh. Castellani 2a 2457T were maintained by our laboratory. Bacteria were cultured in Lysogeny broth (LB) or Tryptic Soy Broth (TSB) supplemented with appropriate antibiotics, where necessary with 100 μg/mL streptomycin or 50 μg/mL ampicillin.

Plasmid construction

Plasmids and primers used in this work were listed in the Supplementary information and Supplementary data 2. Primer and siRNA/sgRNA sequences were purchased from Sangon Biotech (China, Shang Hai). Standard cloning techniques were used to generate the plasmids. The complete coding sequences of Mdm2 or Trp53 were obtained through PCR amplification using the C57BL/6 mouse liver genome and cloned into the pcDNA3.1-Myc-N plasmid to construct overexpression cell lines. The gene encoding Mdm2 was cloned into PET-28a for the expression of MDM2 protein. sgRNA of Trp53/ATP6V0D1/Mdm2 was subcloned into the EcoRI and XbaI sites of pLV3-U6-Trim47(mouse)-sgRNA1-Cas9-EGFP-Puro and then transferred into 293T cells with pMD2.G and psPAX2.

Preparation of mouse peritoneal macrophages

Peritoneal macrophages were derived from wild-type C57, Atp6v0d2-KO, or Atg5^−/− mice injected with 5% sodium mercaptoacetate after 3 days³¹. The peritoneal of mice immersed in 75° alcohol was cut and flushed to collect peritoneal lavage with 5-mL syringe, and the collected peritoneal lavage suspension was lysed with eBioscience buffer (Thermofisher) to remove red cells and plated in RPMI 1640 containing with 10% FBS. After 4 hours of culture, replacement of culture medium to remove anchorage-independent cells to purify macrophages.

Gentamicin protection assay

Cells or gene-edited cell lines were infected with the indicated bacteria at a MOI of 25 for 1 h and further treated with 100 μg/mL gentamicin for 45 min. Then gentamicin-containing medium was replaced with a complete medium supplemented with the indicated quinoline alkaloids, antibiotics or quinoline alkaloids plus antibiotics. The indicated agonists or inhibitors were used as the control. Following an additional incubation at 37 °C in 5% CO₂ for 4 h, the cells were lysed in 2 ‰Triton X-100 for plating³².

To assess resistance development, Salmonella was serially infected RAW264.7 macrophages passaged 15 times under cefixime, penicillin G sodium, or their combination with DEHD, with each passage subculture onto respective group-specific XLD agar plates. Antibiotic pressure was maintained by supplementing cultures with 1/4 MIC cefixime or penicillin G throughout passages. Minimum inhibitory concentrations (MICs) against Salmonella SL1344 were determined following Clinical and Laboratory Standards Institute (CLSI) guidelines via microdilution assays using 5 × 10⁵ CFU/ml inoculate incubated with antibiotic gradients for 16 h at 37 °C.

Transmission electron microscopy analysis

RAW264.7 cells were infected with or without S. typhimurium as described above treatment [19]. Briefly, pretreated cells were cultured with 16 μg/mL DEHD, cefixime or DEHD plus with cefixime for 4 h. The cells were collected by centrifugation at 3000 × g for 5 min at room temperature, inactivated and fixed with 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer, dehydrated with a gradient of 50%, 70%, 90% and 100% ethanol solutions and transferred to a mixture of ethanol and acetone (1:1, v/v) for observation under transmission electron microscopy (Tecnai T12).

Fluorescence microscopy analysis

RAW264.7 cells were infected S. typhimurium as described above. Then, cells and bacteria were observed under the laser confocal microscope (Olympus FV3000), which were stained with DAPI and the indicated primary antibodies, as well as reciprocal secondary antibodies. In addition, lyso-senor/tracker, MDC and Ad-mCherry-GFP-LC3B were used for the determination of lysosomes in the cells.

RAW264.7 cells were cultured in DMEM containing 16 μg/mL DEHD, then rinsed three times with PBS, fixed with 4% paraformaldehyde for 3 min, saturated with 0.4% Triton X-100 for 10 min and blocked with 5% albumin bovine V (Thermo Fisher). Cells were then incubated overnight with LAMP1 and LC3B primary antibodies, followed by an incubation with Goat Anti-Rabbit (Alexa Fluor® 647) and Goat Anti-mouse (Alexa Fluor® 488) at room temperature for 2 h and staining with DAPI for 15 min. Then, the co-localization of LAMP1 with LC3B was observed under the laser confocal microscope (Olympus FV3000).

For the localization of the P53 protein, DEHD-treated RAW264.7 cells were fixed, saturated and blocked as above. Then, the cells were incubated with an anti-P53 primary antibody and Goat Anti-Rabbit (Alexa Fluor® 488) and observed with the laser confocal microscope³³.

Cytotoxicity assay

The RAW264.7 macrophages or Caco-2 epithelial cells were incubated with increasing concentrations of DEHD in DMEM for 6 h or 12 h. The viability of cells was determined using a CCK-8 detection kit (Solarbio, Beijing). Fresh rabbit erythrocytes were treated with the indicated concentrations of DEHD, followed by centrifugation at 8000 rpm for 5 min at room temperature, and the absorbance of the supernatant was measured to evaluate the hemolysis of DEHD.

In vitro infection assay

RAW264.7 cells were infected with S. typhimurium as described above and then treated with DEHD, cefixime or DEHD plus with cefixime for 6 h. The supernatant of this co-infection system was collected for the determination of released lactate dehydrogenase using the Lactate Dehydrogenase cytotoxicity detection kit (Solarbio, Beijing) on a microplate reader (Bio Tek). Additionally, the dead/live cells were observed under fluorescence microscopy (IX83, Olympus) with ethD-1 and Calcein.

Antibacterial activity analysis

The S. typhimurium growth in the presence of DEHD was monitored by measuring the absorption value (OD600nm) of each sample at different time points. The synergistic activity of the indicated quinoline alkaloids with various antibiotics against S. typhimurium was determined by Fractional Inhibitory Concentration assay³⁴. Briefly, SL1344 was inoculated in each well of the 96-well plate with the indicated quinoline alkaloids or cefixime ranging from 0 to 128 μg/mL. The cultures were incubated at 37 °C for 16 h, after which the results were recorded.

In vivo infection assays

All procedures carried out in this study were approved by the Institutional Animal Care and Use Committee of Jilin University (number of permits: SY202309036). We have complied with all relevant ethical regulations for animal use. BALB/c mice aged 6–8 weeks were obtained from Changsheng Biotechnology Co., Ltd (Benxi, China) and the knockdown of Trp53 or Atp6v0d2 mice were constructed from Shanghai Model Organisms Center, Inc. For the pre-experiment, each mouse received 5 g/L of streptomycin in drinking water for three days and was orally inoculated with 5 × 10⁷ CFUs SL1344 and administered with the indicated doses of DEHD (0, 10, 25 or 50 mg/kg) at 4 h after infection and then every 24 h for five doses. Animal inclusion followed a priori criteria: (1) post-infection weight loss <10% (humane endpoints); (2) confirmation of bacterial colonization (baseline CFU > 10⁴/organ); Data analysis included all surviving animals meeting inclusion criteria. The bacteria load in the liver and colon of sacrificed mice was measured with plating on LB agar plates supplemented with streptomycin.

Each mouse received 5 g/L of streptomycin in drinking water for three days and was orally inoculated with 5 × 10⁷ CFUs SL1344 for sublethal infection or 1 × 10⁸ CFUs SL1344 for lethal infection, respectively. Infected mouse was randomly assigned to groups which control (CMC), DEHD (25 mg/kg), cefixime (25 mg/kg), or DEHD + cefixime (both 25 mg/kg) suspended in 0.5% carboxymethylcellulose to oral administration at 4 h post-infection, then every 24 h for 5 doses (100 µL/mouse via 8-gauge gavage needle). Following the determination of the weight and diet of infected mice for five days, the liver, spleen, caecum and colon in sacrificed mice were collected and homogenized in sterile PBS. Viable counts were obtained by serial dilution and plating on LB agar plates supplemented with streptomycin. In addition, the level of inflammatory factors IL-10, IL-β, IFN-γ, IL-6 and TNF-α in the homogenates of colon tissue was examined by ELISA analysis as previously [27]. Alternatively, parts of the indicated tissue were used for gross pathology evaluation and microscopic pathology evaluation with hematoxylin and eosin (HE) staining assay as previously³⁵.

For the establishment of a mouse infection model with Listeria monocytogenes, bacteria were cultured in TSB medium overnight until mid-logarithmic growth (OD₆₀₀ = 0.8), centrifuged (1000 r/m, 10 min), washed and resuspended in PBS. Then, each mouse was orally inoculated with 2 × 10⁶ CFUs Listeria monocytogenes and orally administered with DEHD (25 mg/kg), tetracycline (20 mg/kg), or DEHD (25 mg/kg) in combination with tetracycline (20 mg/kg) at 4 h after infection and then every 24 h for five doses. The bacteria load in the liver and spleen of sacrificed mice was measured with plating. In addition, microscopic pathology of the liver and spleen was determined using an HE staining assay.

ATP/NADH/NADPH determination

Overnight cultured RAW264.7 cells were treated with DMSO, DEHD, cefixime or DEHD plus with cefixime and then grown for 4 h. The level of ATP/NADH/NADPH was assayed in accordance with the instructions provided by the ATP Assay kit, NAD+/NADH Assay Kit with WST-8 and NADP+/NADPH Assay Kit with WST-8 (Beyotime, Shanghai).

Immunoblotting assay

Overnight cultured RAW264.7 cells with the indicated treatment were lysed by RIPA buffer (Sigma) containing protease and phosphatase inhibitors (Sigma). The protein in the nuclear/cytosolic fraction was isolated with the Nuclear and Cytoplasmic Protein Extraction Kit (Beyotime, Shanghai). Then, the targeted protein in each sample was detected with an immunoblotting assay³⁶. The antibodies used in this work are shown in the Supplementary Data 2.

Fluorescence quantitative PCR analysis

Total RNA from RAW264.7 cells co-cultured with 16 μg/mL DEHD at 37 °C for 4 h was extracted using RNA Trizol reagent (Sangon Biotech) and reverse transcribed with StarScript III All-in-One RT Mix (Gene star) according to the manufacturer's instructions. qRT-PCR was then performed on a QuantStudio 1 qPCR instrument (Thermofisher) using 2 × RealStar Universal SYBR qPCR Mix (Gene star). The relative fold expression of each tested gene based on was calculated with the comparative threshold cycling (2^-ΔΔCT) method using Actb as the housekeeping gene. The primers used in this study be found in Supplementary data 1.

Transcriptomics and metabolomics analysis

The total RNA from RAW264.7 cells treated with or without 16 μg/mL DEHD at 37 °C for 4 h was extracted as described above for transcriptomics analysis^37,38. RNA-seq data processing and analysis were performed as follows. Raw sequencing reads were first subjected to quality control using FastQC (v0.11.9). Adapter sequences and low-quality bases were then trimmed using Trimmomatic (v0.39) with the following parameters: SLIDINGWINDOW:4:15, LEADING:3, TRAILING:3, and MINLEN:36. The cleaned high-quality reads were aligned to the [Insert reference genome, e.g., GRCh38] genome using STAR (v2.7.10a) in two-pass mode. Read counts for each gene were generated by feature Counts (v2.0.3) based on the annotation file. Differential expression analysis was carried out using the DESeq2 package (v1.34.0) in R, with a design formula that accounted for [e.g., batch effects if applicable]. Genes with an adjusted p-value (FDR) < 0.05 and absolute log2 fold change > 1 were considered significantly differentially expressed. Functional enrichment analysis of Gene Ontology (GO) terms and KEGG pathways was performed on the differentially expressed genes using the cluster Profiler package (v4.2.2), with an FDR cutoff of 0.05 for determining significant enrichment. For metabolomics analysis, lysates from DEHD-treated cells were stored at 4 °C in a methanol-acetonitrile mixture (1:1). Then, the follow-up samples were processed by Hangzhou Kaitai Biotech (Hangzhou Kaitai Biotech Co., Ltd).

Construction of genes disruption/knockdown/overexpression cells

Mouse Trp53, Mdm2 or ATP6V0D genes were attenuated in RAW264.7 macrophages using CRISPR-Cas9 system³⁹. Briefly, Cas9 and sgRNA expressing plasmids were generated by digesting the pSpCas9 (BB)-2A-Puro (PX459) vector with BbsI (Thermo Fisher Scientific). The resulting plasmids were sequenced using the primer U6-CRISPR to verify the correct ligation of the vector and targeting oligos. RAW264.7 macrophages were plated overnight and transfected with 5 μg of each plasmid using Lipo3000 (Thermo Fisher Scientific). Puromycin selection (5 μg/mL) was started at 36 h after transfection³³. For the siRNA interference assay, RAW264.7 cells were reverse-transfected with 20 pmol of siRNA against mouse Trp53 or Mdm2 (Oligo were shown in Supplementary information) for the following assays.

Cellular thermal shift (CESTA) assay

The CESTA experiment was performed according to the general CETSA protocol⁴⁰. Briefly, overnight cultured RAW264.7cells were treated with 16 μg/mL DEHD for 4 h, harvested and resuspended in phosphate-buffered saline and heated at different temperatures for 3 min in a PCR plate, followed by the addition of protein inhibitor cocktail. Then, each sample was lysed by freeze-thaw cycles and further used for the determination of MDM2 using an immunoblotting assay.

Surface plasmon resonance (SPR) assay

MDM2 protein was expressed with PET-28a-MDM2 vector in E. coli BL21, purified by affinity chromatography with His tag, encapsulated in the CM5 chip (pH 4.0 acetate buffer, 5000 RU target density) and incubated with different concentrations of DEHD (0, 12.5, 25, 50, 100, 200, 400 or 800 μM) by injecting DEHD in PBS buffer at a 30 µL/min flow rate for 60 s association time and 120 s dissociation time at pH = 4.0. Then, the individual binding (ka) and dissociation (kd) rate constants of DEHD with MDM2 were determined using a Biacore X100 (Cytiva) and calculated with Biacore X100 evaluation software⁴¹.

MDM2-P53 complex dissociation analysis

The dissociation degree of the MDM2-P53 complex was verified through an indirect enzyme-linked immunosorbent assay⁴². Briefly, RAW264.7 cells were treated with or without DEHD for 4 h and lysed by ultrasonication. Then, the supernatant of lysates was added to the 96-well plates pre-coated with anti-MDM2 antibody for 2 h incubation at room temperature. After three washs with TBST, the plate was incubated with anti-P53 antibody for 2 h at room temperature. Finally, the absorbance value of each well was detected on a microplate reader (BIO-TEK ELX800) at 450 nm after the addition of Tetramethylbenzidine (TMB) substrate.

Molecular dynamics simulations

The three-dimensional structure of the protein receptor MDM2 was constructed using the I-TASSER-MTD deep learning online server and preprocessed using the leap module in Amber Tools^43,44 by adding missing hydrogen atoms and ligand molecules and applying appropriate force field parameters. The ff19SB force field was used for the receptor, while the Generalized Amber Force Field (GAFF) was applied for the ligand. The system was solvated in a cubic box of TIP3P water molecules with a minimum distance of 10 Å between the box boundary and the nearest protein atom. Additionally, an appropriate number of Na⁺ and Cl^- ions were added to neutralize the system and simulate physiological ionic strength. Before performing molecular dynamics (MD) simulations, the system underwent energy minimization to eliminate any unreasonable initial conformations. The minimization was carried out in two stages: first, the protein and ligand molecules were fixed and only the solvent and ions were minimized; then, the entire system was minimized. The system was then gradually heated from 0 to 300 K throughout 50 ps, employing weak positional restraints to prevent drastic conformational changes. Following the heating, a 500 ps equilibration simulation was conducted to ensure system stability. Finally, a 600 ns production MD simulation was performed under the same conditions.

Statistics and reproducibility

Sample sizes were determined based on established standards for similar microbiological and animal studies, with n = 3 for in vitro replicates and n = 6–10 for in vivo groups. Data analysis was conducted with GraphPad Prism (version 8.3.0) and the data expressed as the mean ± SD. The log-rank test was used for the statistical significance analysis of survival rates and two-tailed Student’s-test was used for other assays. *P < 0.05; **, P < 0.01.

General

The authors are grateful to home for researchers for the preparation of parts of Figures in this paper.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

Designed the research: H.W. and J.F.W.; Performed the experiments: J.F.W., X.Y.H., N.W., L.X., and L.C.K.; Date analyzed and Date curation: H.W., Y.W., and X.M.D.; Visualization: H.W., L.Q.W., H.X.M., and X.D.N.; J.F.W. and H.W. wrote the manuscript with input from all the authors.

Peer review

Peer review information

Communications Biology thanks S. Dinesh Kumar and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary Handling Editors: Suan Sin (Jolin) Foo and Tobias Goris. A peer review file is available.

Data availability

The raw data for all graphs appearing in this paper are available in the OSF database: 10.17605/OSF.IO/4KV9S⁴⁵. Uncropped and unedited blot/gel images are available in the Supplementary Fig. 10. The RNA-seq raw data have been deposited to National Center for Biotechnology Information (NCBI) under the BioProject number PRJNA1363118. All other data are available from the corresponding author on reasonable request.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s42003-025-09251-w.