Metabolomics Analysis Provides Insights into the Antibacterial Activity of Curcumin-Based Photodynamic Treatment Against Staphylococcus aureus

TL;DR

This study uses metabolomics to explore how curcumin-based photodynamic treatment affects the metabolism of Staphylococcus aureus, revealing key metabolic changes that could help improve food safety.

Contribution

The study provides new insights into the metabolic mechanisms of S. aureus under sub-lethal photodynamic treatment using untargeted metabolomics.

Findings

Sub-lethal photodynamic treatment induced significant metabolic perturbations in S. aureus.

97 differential metabolites were identified, linked to amino acid, lipid, and nucleotide metabolism.

Metabolic changes suggest both stress adaptation and potential bactericidal vulnerabilities.

Abstract

Staphylococcus aureus is a major foodborne pathogen that poses persistent challenges to food safety. Antimicrobial photodynamic treatment (PDT) has emerged as a promising non-thermal antimicrobial strategy capable of effectively inactivating S. aureus, though accumulating evidence suggests that the bacteria may initiate adaptive responses to the PDT or even develop tolerance. However, the metabolic mechanisms underlying bacterial responses to PDT exposure, particularly under sub-lethal conditions, remain poorly understood. Thus, in the current study, untargeted metabolomics based on liquid chromatography–tandem mass spectrometry (LC–MS/MS) and gas chromatography–mass spectrometry (GC–MS) were employed to characterize intracellular metabolic alterations in S. aureus following curcumin-mediated PDT (40 µM curcumin, 425 nm blue light at intensity of 0.198 J cm−2). The obtained results revealed a clear separation between the control and PDT-treated groups, indicating sub-lethal PDT induced pronounced metabolic perturbations while preserving partial cellular viability. A total of 97 significantly differential metabolites were identified, encompassing a range of key metabolites associated with amino acid biosynthesis, lipid metabolism, and nucleotide metabolism, indicating the PDT-induced metabolic alterations in pathways could be associated with stress adaptation, membrane integrity, and energy metabolism. Collectively, these findings demonstrate that PDT, even at sub-lethal doses, induces extensive metabolic dysregulation in S. aureus, which potentially represent critical bactericidal vulnerability points target by PDT, yet may paradoxically participate in adaptive metabolic responses that support bacterial survival under PDT exposure. Further investigations are therefore warranted to elucidate the relationships among PDT-induced metabolic perturbations, bacterial inactivation, and long-term phenotypic adaptation. Overall, this work may provide mechanistic insight into PDT-induced antimicrobial action and support further optimization of PDT as an effective non-thermal intervention for food safety and preservation.

Figures

Full text

1. Introduction

Staphylococcus aureus is one of the most important foodborne pathogens and is frequently implicated in food contamination and staphylococcal food poisoning due to its ability to survive under diverse environmental conditions and produce heat-stable enterotoxins [1]. A recent systematic analysis of the global burden of 34 key pathogens in the elderly population demonstrated that S. aureus was the leading pathogen, responsible for an estimated 325,165 deaths and 4766,188 disability-adjusted life years in 2021 [2]. Notably, studies further indicate that S. aureus is widely prevalent across diverse food matrices, including ready-to-eat products, meat, and dairy items. This widespread distribution underscores its persistent presence in the food chain and highlights the ongoing challenges associated with its effective control [3]. In particular, although thermal treatment is effective at S. aureus inactivation, it is often associated with significant quality deterioration of foods, such as losses in nutrients, texture, flavor, and bioactive compounds [4]. Consequently, there has been increasing interest in alternative, non-thermal antimicrobial technologies that ensure microbial safety while preserving food quality.

Antimicrobial photodynamic treatment (PDT) has attracted growing attention as a promising non-thermal intervention for food preservation [5]. PDT relies on the excitation of a photosensitizer by visible light in the presence of oxygen, leading to the generation of reactive oxygen species (ROS) capable of inactivating microorganisms [6]. Compared with conventional antimicrobial approaches, PDT offers several advantages, including broad-spectrum antimicrobial activity, low likelihood of resistance development, and mild processing conditions. In particular, previous studies have demonstrated the effectiveness of PDT against S. aureus in both planktonic cultures [7] and real food-related matrices [8].

Although the antimicrobial efficacy of PDT against S. aureus has been well documented, most existing studies have primarily focused on microbial inactivation kinetics [9] and oxidative damage to cellular components, including the cell membrane [10], key enzymes [11], and genomic DNA [12]. In contrast, the intracellular metabolic responses of S. aureus to PDT-induced stress remain largely unexplored. In particular, emerging evidence suggests that S. aureus possesses a notable capacity to develop tolerance to photodynamic inactivation through adaptive mechanisms. For example, Snell et al. [13] reported that repeated PDT challenges can select for S. aureus subpopulations that tolerate subsequent treatments, a phenotype linked to elevated superoxide dismutase activity and qsrR mutation; similarly, Shi et al. [14] demonstrated that staphyloxanthin production in S. aureus can attenuate PDT efficacy by enhancing resistance to oxidative stress.

Metabolomics, as a high-throughput analytical technology identifying and quantifying metabolites, offers a comprehensive and sensitive approach for characterizing global biochemical changes in microorganisms subjected to processing-induced stresses. By profiling low-molecular-weight metabolites, metabolomics has been successfully applied to elucidate microbial responses to thermal treatments [15], oxidative stress [16] and antimicrobial interventions [17]. However, metabolomics-based investigations of PDT-induced stress responses in foodborne pathogens, particularly S. aureus, remain scarce. Therefore, elucidating the metabolic perturbations associated with sublethal PDT exposure may provide preliminary evidence identifying critical bactericidal vulnerability points targeted by PDT, as well as for exploring potential bacterial stress adaptation mechanisms, thereby guiding the rational optimization of photodynamic antimicrobial strategies.

In the present study, both untargeted LC-MS/MS and GC-MS-based metabolomics approaches were employed to investigate the metabolic responses of S. aureus to curcumin-based PDT. Notably, LC-MS/MS preferentially captures lipids, nucleotides, and redox-related metabolites, whereas GC-MS provides robust coverage of central carbon metabolism and amino acid pathways; thus, this combined application of LC-MS/MS and GC-MS-based metabolomics provides complementary metabolite coverage and enables comprehensive profiling of bacterial metabolic responses to PDT. By comparing untreated cells with PDT-treated bacteria, this work aimed to identify key metabolic pathways affected by photodynamic stress and elucidate potential biochemical signatures associated with PDT as a non-thermal antimicrobial approach.

2. Materials and Methods

2.1. Bacterial Strain and Culture Conditions

The strain of Staphylococcus aureus (ATCC 29213) was purchased from MingZhou Biotechnology Co., Ltd. (Ningbo, China). The strain was stored at −80 °C in tryptic soy broth (TSB) containing 20% (v/v) glycerol. Prior to the experiments, bacteria were streaked onto TSB plates and incubated at 37 °C for 18 h. A single colony was transferred into liquid TSB and cultured overnight at 37 °C with orbital shaking at 220 rpm. Then the overnight culture was diluted 1:100 into fresh TSB and grown to mid-logarithmic phase (OD_600_ ≈ 0.5). Bacterial cells were collected by centrifugation at 8000× g for 5 min, followed by two washes with sterile phosphate-buffered saline (PBS, pH 7.4), and resuspended to approximately 10^8^ CFU/mL.

2.2. Photosensitizer Preparation

Curcumin was dissolved in absolute ethanol (Kunshan Jincheng Reagent Co., Ltd., Kunshan, China) to obtain a 10 mmol/L stock solution. Prior to use, the stock was diluted with phosphate-buffered saline (PBS) to the desired concentrations. Working solutions were freshly prepared and protected from light.

2.3. Antimicrobial Photodynamic Treatment

Bacterial suspensions were incubated with the photosensitizer in the dark for 30 min before irradiation was conducted using a light-emitting diode (LED) system with a peak wavelength at 425 ± 20 nm. The light intensity was set as 0.11 mW cm^−2^ per well, which equals to a total energy of 0.198 J cm^−2^ per well.

For colony-forming unit (CFU) determination, the bacterial samples after treatment were diluted serially, spread on Plate Count Agar plates, and incubated at 37 °C for 24 h. The number of colonies on Plate Count Agar plates was analyzed by a colony counter (SCAN1200, Paris, French).

For metabolomics analysis, the bacteria were harvested by centrifugation (8000× g for 5 min at 4 °C) and rapidly quenched by snap-freezing in liquid nitrogen. Metabolites were extracted using a methanol-water-chloroform-based protocol. Briefly, 1 mL of pre-cooled methanol–water solution (4:1, v/v) containing a mixed internal standard (4 μg/mL) was added to each sample and transferred into a 1.5 mL Eppendorf tube in two aliquots. Subsequently, 200 μL of chloroform was added, and the mixture was thoroughly dispersed by repeated pipetting. Cell disruption was further achieved by sonication in an ice bath (1620 W for 10 min with pulse mode: 6 s on, 4 s off), after which the samples were maintained at −40 °C for 2 h to enhance metabolite extraction. Following incubation, samples were centrifuged at 13,000× g for 10 min at 4 °C. An aliquot of 400 μL of the supernatant was carefully transferred into glass derivatization vials and evaporated to dryness.

2.4. LC-MS/MS-Based Metabolomics Analysis

For LC-MS analysis, dried extracts were re-dissolved in 200 μL of methanol–water (1:4, v/v). The solution was vortexed for 30 s and sonicated in an ice–water bath for 3 min. Samples were then maintained at −40 °C for 2 h, followed by centrifugation at 13,000 rpm for 10 min at 4 °C.

Subsequently, 150 μL of the supernatant was collected using a syringe, passed through a 0.22 μm organic-phase syringe filter, and transferred into LC autosampler vials. LC-MS/MS based metabolomics profiling was performed using a mass spectrometry platform consisting of a Waters ACQUITY UPLC I-Class Plus system coupled with a Thermo Scientific Q Exactive high-resolution mass spectrometer. Chromatographic separation was achieved on an ACQUITY UPLC HSS T3 column (100 mm × 2.1 mm, 1.8 μm) maintained at 45 °C. The mobile phase consisted of A) water containing 0.1% formic acid and B) acetonitrile. The flow rate was set at 0.35 mL/min, and the injection volume was 5 μL. Mass spectrometric detection was conducted in both positive and negative ionization modes. Spray voltage was set to 3.8 kV (positive) and −3.0 kV (negative), respectively. Full-scan mass spectra were acquired over an m/z range of 70–1050 with a resolution of 70,000.

2.5. GC-MS-Based Metabolomics Analysis

For GC-MS analysis, the dried extracts were derivatized sequentially. First, 80 μL of methoxyamine hydrochloride in pyridine (15 mg/mL) was added to each vial, followed by incubation at 37 °C for 60 min with shaking to allow oximation. Afterwards, 50 μL of N,N-Bis(trimethylsilyl)trifluoroacetamide (BSTFA) derivatization reagent and 20 μL of n-hexane were added, followed by 10 μL of mixed fatty acid internal standards (C8–C24 in chloroform). The derivatization reaction was carried out at 70 °C for 60 min. After completion, samples were cooled and maintained at room temperature for 30 min prior to GC-MS metabolomics analysis.

GC-MS measurements were performed on an Agilent 8890 gas chromatograph coupled to a 5977B mass selective detector. Metabolites were separated using a DB-5MS fused-silica capillary column (30 m × 0.25 mm × 0.25 μm, Agilent J&W Scientific, Folsom, CA, USA). Helium (≥99.999% purity) was used as the carrier gas at a constant flow of 1.0 mL/min. Samples (1 μL) were injected in splitless mode at an injector temperature of 260 °C with a solvent delay of 5 min. The oven temperature program was as follows: 60 °C held for 0.5 min, increased to 125 °C at 8 °C/min, then to 210 °C at 8 °C/min, to 270 °C at 15 °C/min, and finally to 305 °C at 20 °C/min with a 5 min hold. Electron ionization (70 eV) was employed, with the ion source maintained at 230 °C, and the quadrupole temperature at 150 °C. Data acquisition was performed in full-scan mode (SCAN) across a mass range of m/z 50–500.

2.6. Data Processing and Statistical Analysis

Raw LC-MS data were analyzed using Progenesis QI V2.3 (Nonlinear, Dynamics, Newcastle, UK). The data processing workflow included baseline filtering, peak identification, integrals, retention time correction, peak alignment, and normalization. The processing parameters were defined as precursor tolerance at 5 ppm, product tolerance at 10 ppm, and product ion threshold at 5%. Metabolite annotation was performed by matching precise mass-to-charge ratios (M/z), secondary fragments, and isotopic distribution against entries in The Human Metabolome Database (HMDB), Lipidmaps (V2.3), Metlin, and LuMet-Animal3.0 databases. The extracted data were then further processed by removing any peaks with a missing value (ion intensity = 0) in more than 50% of groups, by replacing zero values with half of the minimum value, and by screening according to the qualitative results of the compound. Compounds with resulting scores below 36 (out of 60) points were also deemed to be inaccurate and removed. A data matrix was combined by merging datasets acquired in positive and negative ionization modes.

Raw GC/MS data were firstly converted to .abf format using Analysis Base File Converter prior to processing in MS-DIAL with the following analytical work flow, which includes peak detection, peak identification, MS2Dec deconvolution, characterization, peak alignment, wave filtering, and missing value interpolation. Metabolite identification was based on the LuMet-GC 5.0 database. Peak signal intensities were segmented and normalized according to the internal standards with RSD < 0.1 after screening. After the data was normalized, redundancy removal and peak merging were conducted to obtain the data matrix.

Multivariate statistical analyses were conducted in R. Principal component analysis (PCA) was applied to evaluate sample clustering patterns and analytical reproducibility. Orthogonal partial least-squares-discriminant analysis (OPLS-DA) was performed to identify the metabolites that differ between groups. Model robustness was assessed using 7-fold cross-validation and 200 Response Permutation Testing (RPT). Variable importance of projection (VIP) values obtained from the OPLS-DA model were used to rank the overall contribution of each variable to group discrimination. The statistical significance of differential metabolites was further examined using two-sided Student’s t-tests. Differential metabolites were selected with VIP values greater than 1.0 and p-values less than 0.05.

Following independent validation, metabolites confidently identified from both platforms were integrated at the biological interpretation level. Integration was performed using metabolite identity and pathway mapping rather than raw feature intensities. This approach enabled complementary metabolite coverage while avoiding inappropriate quantitative comparisons across platforms. Metabolic pathway enrichment analysis based on the KEGG database (http://www.genome.jp/kegg/, accessed on 13 February 2026) was used as a unifying framework to interpret coordinated metabolic perturbations captured by LC–MS and GC–MS.

All experiments were conducted using six biological replicates. Results are expressed as the mean ± standard deviation (SD). Statistical differences (p < 0.05) among groups were evaluated by one-way analysis of variance (ANOVA) followed by Duncan’s multiple range test using SPSS software (version 16.0). Data visualization and figure preparation were conducted using Origin 2021 and GraphPad Prism 9.

3. Results

3.1. Identification of Sub-Lethal Photodynamic Treatment Conditions

To investigate metabolic perturbations induced by curcumin-based PDT without causing extensive bacterial lysis, a sub-lethal PDT condition was first established. Bacterial viability following exposure to different photosensitizer concentrations and light doses was evaluated using a colony-forming unit (CFU) assay. As shown in Figure 1, blue light treatment (at an intensity of 0.198 J cm^−2^) showed a negligible bactericidal effect, which is consistent with a previous report [18]. Meanwhile, curcumin-based PDT induced a dose-dependent reduction in bacterial viability. According to the previous literature, sub-lethal bacterial inactivation is defined as treatment conditions that result in a significant but incomplete reduction in cell viability, corresponding to a decrease of approximately 1–3 log_10_ units [19]. Thus, 40 μM was selected as the sub-lethal dosage (CFU decreasing from 8.30 ± 0.03 to 6.03 ± 0.06 log) for subsequent metabolomics analysis. This condition ensured sufficient metabolic stress while preserving an adequate number of viable cells for reliable intracellular metabolite profiling.

3.2. LC–MS/MS-Based Metabolomics Analysis

Untargeted LC–MS/MS-based metabolomics was employed to characterize metabolic alterations induced by sub-lethal PDT. Principal component analysis (PCA) of the LC–MS/MS dataset further revealed tight clustering of the six biological replicates within each group, indicating good analytical reproducibility and metabolic consistency (Figure 2A). In contrast, a clear separation was observed between the control (CK) and PDT-treated groups, suggesting that LC–MS/MS-detected metabolites were strongly perturbed by PDT-induced oxidative stress and membrane-associated damage. To further enhance group discrimination and identify key contributing metabolites, OPLS-DA was performed, which also confirmed pronounced separation between the CK and PDT groups along the predictive component (Figure 2B). Model validation using 7-fold cross-validation yielded high R^2^Y and Q^2^ values (0.948 and 0.869, respectively) which confirmed the absence of overfitting and demonstrated robust model performance.

Based on the OPLS-DA model, differential metabolites were screened using the criteria of VIP > 1, p < 0.05. Volcano plot analysis revealed 91 significantly altered metabolites following PDT (Figure 3 and Supplementary Table S1), including 42 up-regulated and 49 down-regulated metabolites.

3.3. GC–MS-Based Metabolomics Analysis

Similarly, GC–MS-based metabolomics was also applied to profile changes in metabolites in S. aureus following PDT. As shown in Figure 4A, the PCA score plots of the GC–MS dataset clearly showed clear intra-group clustering and inter-group separation between the CK and PDT groups, suggesting that PDT markedly altered the intracellular metabolic profile. Consistent with PCA results, OPLS-DA analysis demonstrated distinct separation between the two groups along the predictive component (Figure 4B). Model parameters (R^2^Y and Q^2^) obtained from 7-fold cross-validation also showed high explanatory power and good predictive ability (0.876 and 0.731 respectively), indicating the robustness of the GC–MS-based analysis.

Using the same statistical thresholds (VIP > 1, p < 0.05), GC–MS analysis further identified 6 significantly altered metabolites following PDT (shown in Figure 5 and Supplementary Table S2).

3.4. Integrated LC–MS/MS and GC–MS Metabolomics Analysis

By integrating differential metabolites identified from both LC–MS/MS and GC–MS analyses, a total of 97 significantly altered metabolites were identified between the PDT and CK groups, including 49 up-regulated and 48 down-regulated metabolites. Notably, among the 97 significantly differential metabolites, more than half (55) were putatively identified at MSI levels 1–3, with 11 at level 1, 23 at level 2, and 21 at level 3. In addition, nearly 90% of identified significantly different metabolites (88 of 97) retained statistical significance after FDR correction, with a q-value less than 0.05, indicating that the majority of identified metabolic alterations remain robust after controlling for multiple comparisons.

To visualize the most influential metabolic alterations upon curcumin-mediated PDT, a lollipop plot was constructed based on the top 10 up-regulated and top 10 down-regulated metabolites ranked by VIP values (Figure 6). Among the down-regulated metabolites, several compounds associated with energy metabolism, nucleotide metabolism, and redox balance were significantly decreased, including betaine, 5-aminopentanoic acid, L-carnitine, adenine, NAD-related metabolites, and D-glycerate-3-phosphate. Notably, betaine and L-carnitine exhibited both large absolute fold changes and high VIP scores, suggesting a pronounced suppression of methyl-group metabolism and fatty acid transport-related pathways. Conversely, the up-regulated metabolite set was dominated by compounds related to lipid metabolism and amino acid derivatives, such as myristic acid (d3), arachidic acid (d3), phospholipid species (e.g., PE-related metabolites), and several bioactive or aromatic compounds. Overall, the lollipop plot highlights a clear metabolic disturbance in S. aureus upon PDT, characterized by the down-regulation of central energy- and osmolyte-related metabolites and the accumulation of lipid-associated metabolites.

To further elucidate the relationships among the most discriminative metabolites, Pearson correlation analysis was performed on the top 20 differential metabolites with the highest VIP values (Figure 7). The correlation analysis revealed two distinct and highly coherent metabolic clusters. One cluster consisted predominantly of down-regulated metabolites, including betaine, L-carnitine, 5-aminopentanoic acid, choline, NAD, adenine, and 2-aminopurine, which exhibited strong positive correlations with each other (r > 0.95 in most cases). This suggests that these metabolites are tightly co-regulated and likely participate in interconnected pathways related to one-carbon metabolism, amino acid catabolism, and cellular energy homeostasis. In contrast, the up-regulated metabolites formed a separate cluster, showing strong positive correlations among lipid-related metabolites (e.g., phospholipids and fatty acid derivatives) and aromatic or secondary metabolites. Importantly, strong negative correlations (r < −0.80) were observed between the up-regulated and down-regulated metabolite clusters, indicating that the observed metabolic perturbation involves a systematic redistribution of metabolic flux, potentially reflecting altered membrane composition, stress adaptation, or changes in metabolic demand.

Interestingly, curcumin, as the identified top differential metabolite, was also found to exhibit strong positive correlations with several up-regulated metabolites, including isoketocamphoric acid, integrin-related metabolites, benzoic acid derivatives, ambisentan, and aromatic aldehydes, forming a tightly connected subnetwork. Meanwhile, curcumin also showed pronounced negative correlations with multiple down-regulated metabolites, including betaine, L-carnitine, 5-aminopentanoic acid, choline, NAD, adenine, and 2-aminopurine. This antagonistic correlation pattern positions curcumin as a central metabolic driver within the correlation network. This observation further supported the hypothesis that curcumin-mediated PDT could induce a systemic metabolic shift.

KEGG pathway enrichment analysis (Figure 8) further revealed that oxidative phosphorylation was the most significantly enriched pathway, with the highest enrichment score, indicating that bacterial respiratory energy metabolism may be a primary target of PDT-induced perturbation. This finding is consistent with the observed down-regulation of NAD-, adenine-, and carnitine-related metabolites, which are essential for electron transport, ATP production, and fatty acid oxidation. In addition, pathways related to one-carbon metabolism, including one carbon pool by folate and folate transport and metabolism, were significantly enriched. These pathways are closely linked to betaine, choline, glycine, and serine metabolism, providing mechanistic support for the marked reduction in methyl-donor metabolites observed in the lollipop plot. Disruption of one-carbon metabolism can impair nucleotide synthesis, redox balance, and stress tolerance. Several amino acid metabolism pathways, such as glycine, serine and threonine metabolism, arginine and proline metabolism, histidine metabolism, and alanine, aspartate, and glutamate metabolism, were also enriched, indicating widespread remodeling of nitrogen metabolism. These changes are consistent with the strong correlations observed among amino acid-related metabolites in the correlation analysis. Notably, glycerophospholipid metabolism and ether lipid metabolism were significantly enriched, corroborating the accumulation of phospholipid- and fatty acid-related metabolites. This suggests enhanced membrane lipid remodeling, which may represent an adaptive response to PDT-induced oxidative or structural stress. Furthermore, enrichment of ABC transporters and two-component systems implies alterations in membrane transport and signal transduction, potentially reflecting changes in nutrient uptake, metabolite efflux, and stress signaling following PDT.

4. Discussion

Foodborne contamination by S. aureus remains a persistent challenge for the food industry due to its strong environmental adaptability, biofilm-forming capacity, and increasing tolerance to conventional chemical sanitizers and antibiotics [20]. Photodynamic treatment (PDT), as a novel non-thermal antimicrobial approach, has attracted growing interest and demonstrated high efficacy against a range of foodborne bacteria including S. aureus [21].

PDT mainly involves the light-mediated activation of a photosensitizer to generate reactive oxygen species (ROS), which subsequently damage multiple cellular components, including membranes, proteins, and nucleic acids. Due to this non-specific oxidative mode of action, classical PDT is thought to have a lower chance of resistance development compared with conventional antibiotics [22]. However, accumulating evidence suggests that certain bacterial species, particularly S. aureus, can exhibit tolerance or adaptive responses to PDT-induced stress [19], highlighting the need for a deeper mechanistic understanding of bacterial responses beyond simple viability assays.

In this study, integrated LC–MS/MS- and GC–MS-based metabolomics revealed that sub-lethal curcumin-mediated PDT induced extensive and coordinated metabolic perturbations in S. aureus. The most prominently affected pathways included amino acid metabolism, one-carbon metabolism, lipid metabolism, nucleotide metabolism, and energy-related processes.

Specifically, the observed depletion of key precursor metabolites involved in lysine and histidine biosynthesis, including L-glutamic acid, 5-aminopentanoic acid, and sarcosine, indicates a substantial impairment of amino acid biosynthetic capacity. These metabolites serve as critical nodes linking nitrogen metabolism to central carbon metabolism, particularly the tricarboxylic acid (TCA) cycle [23]. This observation is consistent with previous metabolomic studies of S. aureus under sub-lethal antibiotic stress, in which depletion of glutamate family amino acids was reported as a characteristic feature of disrupted metabolic homeostasis and compromised central metabolism [24]. Their depletion suggests that PDT disrupts not only anabolic pathways but also simultaneously weakens metabolic flexibility required for stress recovery. Similar coupling between amino acid metabolism and energy production has been reported in S. aureus during exposure to oxidative and antibiotic stress, underscoring the vulnerability of these pathways to ROS-mediated damage [25].

Perturbation of arginine- and proline-related metabolism further supports the notion that PDT compromises bacterial stress adaptation mechanisms. Arginine plays multifunctional roles in bacterial physiology, contributing to protein synthesis, polyamine biosynthesis, redox regulation, and acid resistance [26]. Disruption of arginine availability may therefore limit translational capacity while simultaneously weakening stress defense systems. Given that arginine metabolism has been implicated in bacterial persistence and tolerance to antimicrobial treatments, its perturbation following PDT likely contributes to the reduced survival observed under sub-lethal photodynamic conditions.

In addition to biosynthetic pathways, lipid metabolism emerged as another major target of PDT. KEGG enrichment highlighted glycerophospholipid metabolism and ether lipid metabolism, consistent with metabolite-level changes, such as the decreased levels of glycerol-3-phosphate and glycerophosphocholine, accompanied by an increase in 1,3-propanediol. Notably, glycerol-3-phosphate is a central precursor for phospholipid biosynthesis, and its depletion suggests impaired membrane renewal and repair [27]. Given that bacterial membranes are primary targets of singlet oxygen generated during curcumin-mediated PDT, perturbation of lipid biosynthesis and turnover likely reflects oxidative membrane damage and impaired membrane repair capacity. These findings provide metabolic-level evidence supporting the widely accepted notion that oxidative membrane damage and increased membrane permeability are key contributors to PDT-induced bacterial inactivation [28].

Furthermore, PDT profoundly disrupted nucleotide metabolism, as evidenced by decreased levels of adenine-, AMP-, GMP-, and NAD-related metabolites and significant enrichment of purine metabolism pathways. Nucleotides are indispensable for DNA replication, RNA transcription, intracellular signaling and DNA repair. This depletion aligns with transcriptomic findings by Rapacka-Zdonczyk et al. [19], who reported that S. aureus up-regulates DNA repair pathways (SOS response) following PDT exposure, suggesting an increased demand for nucleotide pools. From a functional perspective, nucleotide scarcity would severely constrain DNA repair and replication processes, thereby limiting the ability of S. aureus to recover from PDT-induced oxidative lesions.

Collectively, these results demonstrate that the antimicrobial efficacy of PDT against S. aureus involves profound metabolic disturbances, rather than relying solely on direct oxidative lethality. Our findings indicate that disruption of amino acid biosynthesis, lipid metabolism, and nucleotide metabolism significantly constrains the ability of S. aureus, at least in the majority of the bacterial population, to recover following sub-lethal PDT exposure. Interestingly, in a proteomic study investigating the erythrosine-mediated PDT in S. aureus, 17 differentially expressed proteins were identified, predominantly associated with oxidative stress defense, energy metabolism, translation, and protein biosynthesis [29]. Here, our metabolomic analysis revealed significant alterations in metabolites involved in amino acid biosynthesis, lipid metabolism, and nucleotide metabolism. The convergence of proteomic and metabolomic evidence may indicate that PDT mainly perturbs metabolic pathways, likely affecting stress adaptation mechanisms, membrane integrity, and cellular energy homeostasis.

Admittedly, this study still possesses several limitations. For instance, the investigation was only conducted using a single strain (S. aureus ATCC 29213). Although this strain is a well-established reference strain widely used for antimicrobial susceptibility testing and photodynamic inactivation studies [30], bacterial responses to aPDT may vary among strains. Therefore, future studies should validate the present findings using a broader range of isolates, including clinical and multidrug-resistant strains, as well as strains subjected to repeated PDT exposure, to enhance the translational and clinical relevance. Furhtermore, although the curcumin concentration (40 µM) and light dose (0.198 J cm^−2^) employed in the current study were far below the reported MIC of curcumin (594–678 µM) and lethal blue light intensity (3.24 J/cm^2^) against S. aureus ATCC 29213, the absence of a dark control (curcumin-only) and irradiation control (light-only) in our metabolomic profiling still limits our ability to differentiate metabolic perturbations caused by curcumin or light individually from those arising from their synergistic photodynamic interaction. Additionally, the observed metabolic changes in S. aureus upon PDT in the current study likely reflect a combination of bacterial passive responses resulting from ROS-mediated damage and active adaptive metabolic responses to support bacterial survival. We acknowledge that the current metabolomic data cannot definitively distinguish direct oxidative damage from adaptive metabolic remodeling. Further investigations are required to clarify the relationships between PDT-induced oxidative stress, metabolic disruption, and long-term phenotypic adaptation. Last but not least, previous studies have suggested that food matrices can profoundly influence the efficiency of photodynamic inactivation. Factors such as turbidity, light scattering, endogenous antioxidants, and interactions between curcumin and food components may affect ROS generation and alter bacterial responses compared with model systems [21,31]. Therefore, the findings of this study should be interpreted with caution when considering their applicability to real food processing conditions. Further validation using representative food matrices is required to determine whether similar metabolic effects occur under practical applications.

5. Conclusions

In conclusion, this study demonstrates that curcumin-mediated PDT induces extensive metabolic dysregulation in S. aureus, affecting interconnected pathways essential for biosynthesis, energy production, stress resistance, and cellular integrity. These findings provide mechanistic insight into PDT-induced bacterial inactivation. From an applied perspective, the findings in the current study highlight that targeting metabolic vulnerabilities may represent an effective strategy to enhance PDT efficacy, optimize photosensitizer selection, and design synergistic combination strategies for improved antimicrobial control.