Succinic acid and all culture reagents, including Tryptic Soy Broth (TSB), agar, Phosphate-Buffered Saline (PBS), methanol, ethanol, and crystal violet, were procured from Dalian Amioke Biological Technology Co., Ltd. Sterile succinic acid solutions (5, 10, 15, 20, 50 mM) were prepared in triple-distilled H2O, sterilized through 0.22 μm membrane filtration, and stored at 4°C. The P.mirabilis reference strain ATCC^® 29906 was cultured on Tryptic Soy Agar (TSA) plates from -80°C glycerol stock. Single colonies were inoculated into TSB medium and incubated at 37°C with 180 rpm agitation to achieve logarithmic growth phase (OD550 = 0.1, 8×10⁸ CFU/mL) for subsequent experiments.

The minimum inhibitory concentration was determined using broth microdilution method (Manoharan et al., 2020), defined as the lowest concentration inhibiting ≥70% bacterial growth compared to untreated controls. For the biofilm inhibition assay, succinic acid powder was dissolved in triple-distilled H2O to prepare a filter-sterilized 100 mM stock solution. This stock was then serially diluted to create working solutions at 10, 20, 30, 40, and 100 mM. In the assay, 100 μL of each working solution was added to 100 μL of bacterial suspension in a 96-well plate (n=6), resulting in final succinic acid concentrations of 5, 10, 15, 20, and 50 mM, respectively, after equal volume mixing. Control wells received 100 μL of bacterial suspension mixed with an equal volume (100 μL) of triple-distilled H₂O. Following 48-hour biofilm formation at 37°C, planktonic cells were removed by triple PBS washing. Biofilms were methanol-fixed (200 μL, 30 min), stained with 0.1% crystal violet (200 μL, 20 min), and destained with absolute ethanol (200 μL). Absorbance at 550 nm was measured spectrophotometrically, with MBIC defined as the concentration achieving ≥50% inhibition calculated by: Inhibition rate (%) = [1 - (OD550 treatment/OD550 control)] × 100%. Independent samples t-test was used to compare differences between every two groups for continuous data. Data analysis was performed using SPSS 25.0, and bar graphs were generated using GraphPad Prism 9.5.1.

Single colonies of Proteus mirabilis from Tryptic Soy Agar (TSA) plates were inoculated into Tryptic Soy Broth (TSB) and incubated overnight at 37°C. This process was repeated independently to generate three separate biological cultures. Bacterial suspensions were centrifuged at 6,000×g for 15 min at 4°C, then resuspended in fresh TSB to an optical density at 550 nm (OD₅₅₀) of 0.1 for subsequent experiments. For treatment groups, 7.5 mL of bacterial suspension from each independent culture was mixed with an equal volume of a filter-sterilized succinic acid solution to achieve a final treatment concentration equivalent to the predetermined 15 mM. The mixtures were incubated statically at 37°C for 24 h. Control groups received 7.5 mL of bacterial suspension from each independent culture mixed with an equivalent volume of sterile triple-distilled H₂O. This resulted in three biologically independent replicates for both the treatment and control groups. Post-incubation, all samples were centrifuged (6,000×g, 15 min, 4°C). The resulting cell pellets were washed once with cold PBS, immediately flash-frozen in liquid nitrogen, and stored at -80°C until metabolomic analysis. The same sample preparation procedure was followed for all subsequent omics analyses. Metabolomic profiling was performed using Thermo Fisher Scientific’s UHPLC-Q Exactive HF-X system. Liquid chromatography separation was achieved with a C18 column (2.1×100 mm, 1.7 μm) maintained at 40°C. Raw LC-MS data were processed through Progenesis QI software (v2.3, Waters) for baseline correction, peak detection, retention time alignment, and generation of an m/z-retention time-intensity matrix. Metabolite identification was achieved by matching MS/MS spectra against HMDB (v5.0) and METLIN (v2023) databases. Data preprocessing involved removing features with >20% missing values across sample groups, imputing remaining missing values with 10% of the minimum observed intensity, and applying total ion current normalization. Quality control filtering excluded variables exhibiting >30% RSD in QC samples, followed by log10 transformation. Multivariate analysis using the ropls package (v1.6.2) in R included PCA for data structure visualization and OPLS-DA with 7-fold cross-validation to assess model robustness. Significantly altered metabolites were selected based on variable importance in projection (VIP) scores >1.0 and Welch’s t-test p value <0.05 after Benjamini-Hochberg correction. KEGG pathway annotation (Release 107.0) and hypergeometric enrichment analysis via scipy.stats (v1.11.1) identified biologically relevant pathways, with statistical significance determined by Fisher’s exact test followed by false discovery rate (FDR) adjustment (q<0.05).

High-quality RNA samples were obtained by extracting RNA from single bacteria and removing genomic DNA using the CTAB method. The rRNA was removed using the RiboCop rRNA Depletion Kit. The mRNA was randomly cleaved into small fragments of 200bp, and reverse transcription was performed using the mRNA as a template to synthesize double-stranded cDNA. When synthesizing the second strand, dUTP was used instead of dTTP. The repaired cDNA was added to the junction, and finally the second strand containing dUTP was removed with UNG enzyme. The library was established using Illumina Stranded mRNA Prep, and double-ended RNA-seq sequencing was performed using NovaSeq™ X Plus. Bioinformatics analysis processes the data generated by the Illumina platform. The gene expression levels were measured using the FPKM and TPM methods. FPKM was calculated based on the number of reads per thousand bases in every million sequences, while TPM was based on the number of reads from a specific transcription per million reads. FPKM and TPM can eliminate the influence of gene length and sequencing depth, thereby enabling the gene expression level to be used for comparison of differences between samples. Differential expression analysis was conducted by obtaining the Read Counts of genes to identify differentially expressed genes and study their functions. The experiment consisted of three biological replicates, and the difference analysis was conducted using DESeq2 software. This analysis was conducted using the software Goatools for enrichment analysis, and the method employed was Fisher’s exact test. To control the calculated false positive rate, four multiple test methods (Bonferroni, Holm, Sidak and false discovery rate) were used to correct the p value. Generally, when the p value was <0.05, it was considered that there was a significant enrichment of this KEGG function.

Frozen samples were retrieved from -80°C storage and lysed by ultrasonication in ice-cold lysis buffer (8 M urea, 1% protease inhibitor, 3 μM TSA, 50 mM NAM). After centrifugation (12,000 g, 10 min, 4°C), the resulting supernatant was collected for protein quantification using the BCA assay. Equal protein aliquots were precipitated with TCA, washed with acetone, and resuspended in 200 mM TEAB for tryptic digestion (1:50, m/m, overnight). Subsequently, peptides underwent reduction (5 mM DTT, 56°C, 30 min) and alkylation (11 mM IAA, room temperature, 15 min in darkness). Peptide separation was achieved via nanoflow liquid chromatography (NanoElute system) using mobile phase A (0.1% formic acid, 2% acetonitrile in water) and phase B (0.1% formic acid in acetonitrile-water) with a gradient elution (6%-80% B over 20 min, 500 nl/min). Mass spectrometric analysis was conducted on a timsTOF Pro 2 system (ion source voltage: 1.75 kV) in dia-PASEF mode, with MS1 (300–1500 m/z) and MS2 (400–850 m/z) scans at a 7 m/z window width. Protein identification and pathway annotation were performed against the KEGG database using blastp (e-value ≤ 1e-4), retaining the highest-scoring match for each protein.

Frozen specimens were retrieved from -80°C storage and homogenized in lysis buffer (8 M urea, 1% protease inhibitor cocktail, 3 μM TSA, 50 mM NAM) using ultrasonication. After centrifugation, the supernatant was collected for protein quantification. Equal protein aliquots were precipitated with TCA, washed with acetone, and reconstituted in 200 mM TEAB for tryptic digestion (1:50, m/m, overnight at 37°C). Digested peptides underwent sequential treatment with 5 mM DTT (56 °C, 30 min) and 11 mM IAA (room temperature, 15 min in darkness). The peptides were then dissolved in IP buffer (100 mM NaCl, 1 mM EDTA, 50 mM Tris-HCl, 0.5% NP-40, pH 8.0) and incubated with anti-acetyllysine antibody-conjugated beads (PTM0104) at 4°C overnight. After extensive washing, bound peptides were eluted with 0.1% trifluoroacetic acid and desalted using C18 ZipTips prior to LC-MS/MS analysis.

Protein structure prediction was performed using SWISS-MODEL, an automated web-based homology modeling system. The target protein sequence was obtained from the UniProt database, followed by template identification from experimentally determined protein structures. Three-dimensional models were generated through sequence alignment and evaluated based on template identity, amino acid number, and QMEAN score (Patel et al., 2021). The functional domain containing the K6 site was annotated using PyMOL.

The inhibitory effects of succinic acid on P.mirabilis are shown in Figures 3A and B . In the MIC assay ( Figure 3C ), bacterial growth inhibition significantly increased with rising succinic acid concentrations. The most pronounced inhibitory effect (P < 0.001) was observed at 15 mM, which represented the lowest concentration achieving ≥70% growth inhibition compared to the untreated control, thus establishing 15 mM as the MIC against P.mirabilis.

For biofilm formation ( Figure 3D ), succinic acid exhibited concentration-dependent inhibition, with 15 mM demonstrating the most significant suppression (P < 0.001). This concentration achieved >50% biofilm inhibition, thereby being determined as the MBIC against P.mirabilis.

We performed metabolomic analysis on P.mirabilis treated with succinic acid. Figure 4D illustrates the distribution of compounds across different metabolic pathways and biological processes, primarily encompassing metabolism, cellular processes, environmental information processing, genetic information processing, and human diseases. Among these, pathways related to amino acid metabolism, nucleotide metabolism, and membrane transport contained a higher number of compounds, whereas processes such as cellular motility and lipid metabolism exhibited fewer. Figure 4E presents the KEGG pathway enrichment analysis of the metabolomics data. The differentially abundant metabolites were predominantly enriched in nucleotide metabolism, purine metabolism, and arginine biosynthesis, which are primarily associated with bacterial stress adaptation and metabolic regulatory mechanisms.

Among the various metabolic pathways, those related to amino acid metabolism were particularly critical, as their alterations were directly linked to the regulatory effects of D-amino acids (e.g., L-methionine, L-lysine, L-aspartic acid, linatine) and tryptophan on bacterial physiology. D-amino acids significantly inhibit bacterial growth and disrupt biofilm stability through multiple molecular mechanisms. First, at the protein synthesis level, D-amino acids induce tRNA misacylation and feedback inhibition of key metabolic pathways, leading to ribosomal dysfunction and translational errors. Second, D-amino acids directly interfere with peptidoglycan biosynthesis and are aberrantly incorporated into peptidoglycan side chains, resulting in defective cell wall cross-linking and reduced mechanical strength (Kuru et al., 2012). Collectively, these effects cause bacterial growth arrest and eventual lysis. Regarding biofilm regulation, D-amino acids function via a dual mechanism: at subinhibitory concentrations, they disrupt quorum sensing systems and downregulate the expression of genes encoding biofilm matrix components, such as extracellular polysaccharides and pilin proteins (Leiman et al., 2013); at higher concentrations, D-amino acids are specifically incorporated into mature biofilm structures, displacing D-alanine residues in peptidoglycan peptide bridges, thereby disrupting the anchoring connections between amyloid fibers and cell wall lipoproteins and leading to biofilm structural collapse (Kolodkin-Gal et al., 2010).

The tryptophan metabolic pathway plays a pivotal role in bacterial physiological regulation, with its metabolic network modulating genetic stability, virulence expression, and environmental adaptability through multi-layered molecular mechanisms. Metabolomic analysis revealed significant decreases in key tryptophan pathway intermediates, including indole pyruvate, indole acetate, and tryptamine ( Supplementary Table 1 ). Following enzymatic degradation by tryptophanase, tryptophan not only serves as a carbon/nitrogen source for bacteria but also generates indole pyruvate—a critical nodal metabolite that participates in dynamic energy metabolism balance via transamination to produce indole acetate. The depletion of indole pyruvate directly impairs branched-chain amino acid biosynthesis, leading to dysregulated expression of drug resistance-associated proteins (Anyanful et al., 2005), while reduced indole acetate levels compromise its function in stimulating extracellular polysaccharide synthesis, thereby weakening biofilm matrix integrity (Bianco et al., 2006). This metabolic disturbance further extends to quorum sensing systems, where diminished tryptamine—a decarboxylation product of tryptophan—may disrupt neurotransmitter-mimicking mechanisms in bacterial communication (Lee et al., 2007).

The biosynthesis pathways of phenylalanine, tyrosine, and tryptophan occupy central positions in bacterial metabolic networks. By regulating the synthesis of these amino acids, bacteria finely tune intra- and extracellular metabolic homeostasis, influencing essential molecular transport processes, interpopulation signaling, and biofilm formation. Relevant studies demonstrate that these pathways are closely associated with bacterial EPS secretion, which directly governs collective behaviors, antibiotic resistance, and pathogenicity (Guillín et al., 2025). Additionally, these amino acids participate in protein synthesis and energy metabolism. Inhibition or marked reduction of phenylalanine biosynthesis impairs bacterial acquisition and utilization of other essential amino acids, ultimately suppressing growth and proliferative capacity (Zhu et al., 2024).

KEGG enrichment analysis revealed significant alterations in arginine biosynthesis, glutamine family amino acid metabolism, alanine/aspartate metabolism, nucleotide metabolism, ribosome biogenesis, and oxidative phosphorylation pathways ( Figure 5C ). The heatmap displays the expression patterns of the top 30 DEGs between control and treatment groups ( Figure 5D ). Notably, artP and artJ genes comprising the arginine transport system cluster artPIQM-artJ were markedly upregulated in the succinic acid-treated group. artP encodes the ATP-binding protein ArtP of the arginine ABC transporter, which is crucial for transmembrane arginine transport. Given arginine’s importance for bacterial acid resistance, P.mirabilis may employ ABC transporters to accumulate osmoprotectants, thereby mitigating osmotic stress by balancing transmembrane gradients (Grigg et al., 2010).

Multiple genes associated with arginine biosynthesis (argA, argB, gdhA, argH, argE, argC, argF) and stress response (pmi_rs16385, pmi_rs04750, pmi_rs11330, pmi_rs06360) were significantly upregulated ( Supplementary Table 2 ), indicating activation of metabolic adaptation and stress response mechanisms following succinic acid treatment. Furthermore, argR - encoding the arginine repressor ArgR that negatively regulates arginine biosynthetic genes - showed pronounced upregulation. The argR regulon includes its own autoregulatory gene, the carAB operon, and genes involved in arginine transport and catabolism (Cho et al., 2015).

Genes encoding 30S and 50S ribosomal subunit proteins (rplR, rplP, rpsC, rpmI) were significantly enriched among downregulated genes ( Supplementary Table 2 ). Notably, infB—which encodes translation initiation factor IF-2—was also markedly downregulated. IF-2 facilitates GTP hydrolysis to GDP and inorganic phosphate (Pi), releasing initiation factors to enable 50S-30S ribosomal subunit association into the 70S initiation complex. The suppression of infB inhibits ribosomal biogenesis, thereby blocking protein synthesis and exerting bacteriostatic effects.

Major facilitator superfamily (MFS) transporter genes (pmi_rs14645, pmi_rs07650, pmi_rs04170, pmi_rs11915) were significantly downregulated in the succinic acid-treated group. As Gram-negative bacterial efflux pumps, MFS transporters mediate the extrusion of diverse substrates (ions, sugars, amino acids, drugs, and toxins), contributing to multidrug resistance (Du et al., 2018). Additionally, MFS transporters participate in biofilm maturation (May et al., 2009), osmostress protection, and bacterial motility. Prior studies confirm that disrupting MFS transporter genes reduces crystalline biofilm formation in P.mirabilis and increases antibiotic susceptibility (Holling et al., 2014).

The NADH:ubiquinone oxidoreductase (respiratory Complex I)—a core enzyme in bacterial energy metabolism—couples NADH oxidation with ubiquinone reduction to drive proton translocation for ATP synthesis. Succinic acid treatment significantly downregulated nuoI, nuoG, nuoH, and nuoJ ( Supplementary Table 2 ): nuoG: Stabilizes Complex I via its iron-sulfur cluster-binding sites; its absence causes NDH-1 assembly defects, preventing integration of peripheral subunits (except nuoC and nuoD) ( Pohl et al., 2007); nuoH and nuoJ: Form transmembrane domains housing the quinone-binding site and proton channels; their dysfunction impairs electron transport chain (ETC) efficiency (Friedrich et al., 2016); nuoI: A core subunit mediating electron transfer from the NADH dehydrogenase module to the membrane arm; its loss abolishes complex activity (Hedderich and Forzi, 2005). Furthermore, cyoE—encoding heme O synthase—was significantly downregulated. This enzyme catalyzes heme O biosynthesis, a critical component of bacterial terminal oxidases O (Mogi et al., 1994). The coordinated downregulation of these genes disrupts ETC electron transfer, proton pumping, and complex assembly, ultimately blocking ATP synthesis and inhibiting bacterial growth.

Key components of the Type VI Secretion System (T6SS), including TssN, TssC, TssA, IdrA, TssJ, and TssH, were significantly downregulated ( Supplementary Table 3 , Figure 6 ). The T6SS plays a crucial role in the pathogenesis of P.mirabilis by mediating self-recognition and territorial behaviors through interbacterial competition, as well as exerting an indirect influence on swarming motility via strain competition (Saak et al., 2017). Although its direct role in biofilm formation and stress response in P.mirabilis has not been fully elucidated, studies in other bacterial species suggest that T6SS may modulate these processes (Xiong et al., 2024). Specifically, TssA is essential for bacterial fitness during pyelonephritis, with tssA mutants showing impaired colonization in murine bladder infection models (Debnath et al., 2018). TssC and TssB form heterodimers that assemble into the contractile sheath required for T6SS function, while TssJ (ClpV) mediates ATP-dependent sheath disassembly and its deletion affects motility, biofilm formation, and EPS production (Li et al., 2024). TssN connects the membrane complex to the baseplate through interaction with TssK to facilitate effector translocation (Bongiovanni et al., 2024), and TssH is required for effector secretion and virulence, with mutants exhibiting defects in T6SS assembly and regulation (Asolkar and Ramesh, 2020).

Proteomic analysis revealed that P.mirabilis requires three gene clusters (ids, tss, and idr) for self-recognition, with Ids and Idr proteins representing two distinct mechanisms. Ids proteins encode strain-specific identity determinants, while Idr proteins mediate inter-strain competition during population migration (Saak et al., 2017). T6SS is essential for secreting these proteins, enabling self-recognition and territorial behaviors. Through direct cell contact, T6SS delivers effectors into target cell periplasm via its contractile puncturing apparatus, inducing cell death (Hood et al., 2010). This contact-dependent delivery mechanism underlies its bactericidal function. Importantly, T6SS contributes to catheter-associated urinary tract infection progression (Armbruster et al., 2017), suggesting that succinic acid may inhibit biofilm formation by modulating T6SS-related protein secretion or function.

Additionally, the TCA cycle was significantly enriched in KEGG analysis, with key enzymes downregulated ( Supplementary Table 3 , Figure 6 ). In Candida albicans, the TCA cycle is crucial for mature biofilm formation (Sun et al., 2022), influencing matrix synthesis/modification and structural stability. Disrupted TCA cycle enzymes increase dead cells and reduce biofilm biomass, indicating its role in maintaining metabolic homeostasis and biofilm integrity (De Backer et al., 2018). Bacillus subtilis studies show TCA intermediate accumulation during early biofilm formation (12–16 hours) enhances metabolic activity, providing material/energy for growth. These intermediates correlate with extracellular matrix (ECM) synthesis, coinciding with upregulated ECM gene expression and biosynthetic precursor levels (Pisithkul et al., 2019), Proteomic analysis further demonstrated significant downregulation of the siderophore biosynthesis system, including key genes nrpS (nonribosomal peptide synthase) and nrpU (siderophore biosynthesis gene).This specific impairment of siderophore production will critically impair P.mirabilis pathogenicity and environmental adaptation ( Supplementary Table 3 ). The Nrp system maintains bacterial survival under iron-limited host conditions through competitive iron chelation and transport mediated by siderophores (Armbruster et al., 2018). Dysfunction of this biosynthetic system directly compromises iron acquisition capacity, leading to impaired virulence factor expression and disruption of critical pathogenic processes including swarming motility regulation and flagella synthesis. Mouse infection models confirmed the essential role of the Nrp system in host colonization, with mutants showing significantly reduced adherence in bladder and kidney tissues (Schaffer and Pearson, 2015). Furthermore, dysregulation of iron metabolism-related genes disrupts redox homeostasis, inhibits activation of iron acquisition systems under iron-restricted conditions, and weakens bacterial competition against host iron-binding proteins, ultimately impairing metabolic balance and environmental adaptability (Malik and Hedrich, 2021).

Although the direct contribution of T6SS to biofilm formation in P.mirabilis is not yet definitive, our data suggest that succinic acid-induced downregulation of T6SS components may indirectly impair biofilm development by disrupting bacterial competition and microcolony architecture. Furthermore, the proposed deacetylation of H-NS and enhanced repression of T6SS genes represent a novel regulatory mechanism that could potentially influence a network of virulence factors, including those involved in biofilm formation.

The histone-like nucleoid structuring protein (H-NS) functions as a global transcriptional regulator that modulates bacterial environmental adaptation through selective silencing of virulence genes. In Edwardsiella, H-NS has been shown to repress T6SS gene expression by binding AT-rich promoter regions of T6SS gene clusters and forming transcriptional silencing complexes (Ma et al., 2022). The molecular activity of H-NS is determined by its domain architecture: the N-terminal oligomerization domain mediates self-assembly into functional dimers/tetramers, while the C-terminal DNA-binding domain specifically recognizes target gene promoters (Blot et al., 2006). Notably, the transcriptional repression activity of H-NS is highly dependent on its oligomeric state, as non-oligomerized monomers lack the cooperative binding capacity required for effective RNA polymerase inhibition (Liu et al., 2021). However, whether H-NS regulates T6SS genes through a similar mechanism in P.mirabilis remains unreported, presenting a key mechanistic avenue for investigation in this study.

Bacteria dynamically modulate H-NS function via post-translational modifications (PTMs) to adapt to environmental stress (Macek et al., 2019). While existing research on succinic acid’s effects on bacterial phenotypes has primarily focused on succinylation (Wu et al., 2025), studies of H-NS PTMs have largely examined phosphorylation (Liu et al., 2021). Recently, acetylation modifications have emerged as a research focus in metabolite-epigenetic regulation due to their close coupling with central metabolism. Studies reveal that acetylation at K120 within the DNA-binding domain of H-NS abolishes its transcriptional repression activity (Liu et al., 2024). Nevertheless, the functional consequences of acetylation in the N-terminal oligomerization domain and its mechanistic links to metabolic signaling remain unexplored.

This study reveals for the first time that succinic acid specifically modulates the post-translational modification status of P.mirabilis H-NS. Acetylomic analysis demonstrated a significant decrease in ϵ-amino acetylation at lysine 6 (K6) of H-NS in succinic acid-treated groups compared to controls ( Supplementary Table 4 ). Structural modeling (SWISS-MODEL, WP_368934229.1) and PyMOL annotation localized the K6 site within the N-terminal oligomerization domain of H-NS ( Supplementary Figure 1A ). Functional enrichment analysis further indicated that differentially acetylated proteins were predominantly associated with protein complex oligomerization and tetramerization processes, suggesting succinic acid may influence H-NS function by coordinately regulating multiprotein complex assembly ( Supplementary Figure 1B ).

Based on these findings, we propose a cascade regulation hypothesis: Succinic acid specifically reduces K6 acetylation of H-NS, inducing conformational rearrangement in its N-terminal oligomerization domain. This promotes the formation of highly stable H-NS tetramers with enhanced binding affinity for AT-rich promoter regions of T6SS gene clusters, thereby strengthening transcriptional repression of T6SS-related genes ( Figure 7 ).