The small molecules used in this study were obtained from Chengdu Purechem-Standard Co., Ltd. Penicillin G (PG), ampicillin (Amp), ceftriaxone sodium (Cef), kanamycin, and nitrocefin were purchased from Dalian Meilun Biotechnology Co., Ltd. Dulbecco’s modified Eagle’s medium (DMEM), fetal bovine serum (FBS), trypsin (0.25%), and penicillin−streptomycin solution were purchased from Beijing Solarbio Science & Technology Co., Ltd. The substrate peptide of sortase A (Dabcyl-QALPETGEE-Edans) was purchased from GL Biochem (Shanghai), Ltd. S. aureus USA300, which expresses β-lactamase, was stored in our laboratory and cultured in Luria−Bertani (LB) media at 37°C with shaking or under static conditions.

The S. aureus sortase A gene was cloned and inserted into the pET-28a (+) vector, which was subsequently transformed into Escherichia coli BL21 (DE3) (Beyotime, Shanghai, China). The E. coli was cultured in LB medium supplemented with 0.2 mM isopropyl-beta-D-thiogalactopyranoside (Beyotime, Shanghai, China) at 18°C for 18 h. Then, the bacteria were harvested and lysed to obtain the supernatant. The supernatant was incubated with Ni-NTA agarose resin, and the purified protein was obtained after elution with 250 mM imidazole. To construct the sortase A mutants, primers designed for specific mutations (V168A, K173A, and L169A) were obtained using the QuikChange Site-Directed Mutagenesis Kit. The pET28a plasmid carrying the sortase A gene was used as a template, and the polymerase chain reaction (PCR) products were introduced into E. coli BL21 (DE3) after treatment with DpnI (Beyotime, Shanghai, China). The mutated proteins were obtained via the same purification method described above. Cloning, expression, and purification of β-lactamase and its mutants (Y96A, I158A, and I230A) were performed via the same methods. The mutants of sortase A and β-lactamase were used to identify the residues critical for interaction with the compounds. The primers used for this assay are shown in Table 1 .

The sortase A protein (10 µg) was coincubated with various concentrations of ECG and its analogues (0, 8, 16, and 32 µg/mL) at 37°C for 0.5 h. The substrate peptide (10 µg, final volume of 100 µL) was subsequently added, and the samples were incubated for 1 h in the dark. Then, the fluorescence intensity was detected by using a microplate reader (excitation 490 nm, emission 520 nm), and the inhibitory effects of ECG and its analogues against sortase A were evaluated via a previously described method (Song et al., 2017). The inhibitory effects of ECG against the mutants of sortase A (V168A, K173A and L169A) were determined via the same method.

AutoDock Vina software (Trott and Olson, 2010) was used to perform docking calculations after the compounds (set as the ligands) were docked into proteins (β-lactamase PDB ID: 6WGR; sortase A PDB ID: 6R1V) with AutoDock Tools on the basis of methods described previously (Gao et al., 2020). Molecular dynamics simulations of the candidates were carried out using GROMACS 2020.6 and previously reported methods (Liu TT. et al., 2019; Gao et al., 2020). Briefly, RESP charges were assigned to the ligands using Gaussian and amber, and topology files of the ligands were obtained by using AmberTools. The Amber 14sb force field and SPC/E water model were also used. The root mean square deviation (RMSD) was used to evaluate the equilibrium of the system, and key protein residues involved in the interactions with the ligands were identified via the molecular mechanics Poisson−Boltzmann surface area (MMPBSA) method.

Different concentrations of ECG and its analogues (0, 8, and 32 µg/mL) were used to treat S. aureus USA300, which were seeded into 96-well plates (5 × 10⁶ CFUs/well). The samples were cultured at 37°C until static. At 24 h later, the medium was removed, and the samples were washed with sterile phosphate-buffered saline (PBS) and stained with 0.1% crystal violet. The optical density at 570 nm (OD₅₇₀) of each well was measured after adding 33% glacial acetic acid to dissolve the solids to evaluate the biofilm formation inhibitory effects of the compounds (Grossman et al., 2021). The free LB medium was defined as negative control (NC).

S. aureus USA300 β-lactamase protein activity inhibition assays were performed according to methods reported previously (Zhou et al., 2020). Briefly, different concentrations of ECG and its derivatives (0, 32, 64, and 128 µg/mL) were used to treat the purified β-lactamase protein (5 µg) for 30 min at 37°C, and nitrocefin (5 µg) was added for the final 10 min of incubation. The OD₄₉₂ of each well was subsequently measured to determine whether these compounds affect the activity of β-lactamase (Zhou et al., 2020). The inhibitory effects of ECG against the β-lactamase mutants (Y96A, I158A, and I230A) were determined via the same method.

The minimum inhibitory concentrations (MICs) of PG, Amp, Cef, and kanamycin combined with or without the test compounds were determined on the basis of Clinical and Laboratory Standards Institute (CLSI) methods. Briefly, LB culture medium containing a series of concentrations of PG, Cef, Amp, or kanamycin (0–256 µg/mL) and ECG derivatives (0–32 µg/mL) was added to a 96-well plate, followed by the addition of S. aureus USA300 strains (5 × 10⁵ CFUs/mL). Samples were cultured at 37°C for 24 h, and the MIC values were defined as the concentration of compound at which there was no visible bacteria growth. The antimicrobial properties of the ECG and its analogues and their synergistic effects with PG, Cef, Amp, and kanamycin were analyzed on the basis of the MIC values (Guo et al., 2024).

On the basis of the synergistic effect between the ECG and its analogues and the tested antibiotics, Amp combined with ECG (32 µg/mL) was used to carry out a time-dependent bacterial killing assay. S. aureus USA300 in the logarithmic growth phase (approximately 2 × 10⁶ CFUs/mL) was treated with ECG (32 µg/mL), Amp (64 µg/mL), or their combination and cultured with shaking. Samples from each group were collected every 2 h and plated on LB agar medium after dilution. The colonies were analyzed to evaluate the ability of ECG to improve the bactericidal ability of Amp (Liu S. et al., 2019). The persister assay was performed on the basis of a previously described method with some modifications (Keren et al., 2004). Briefly, S. aureus USA300 was treated with 10 µg/mL gentamicin, then ECG (32 µg/mL) was added, and the mixture was cultured with shaking. Finally, bacterial colonies were harvested at the indicated time points to analyze the formation of persisters (Keren et al., 2004).

Mouse mononuclear macrophages (RAW264.7 cells) and human lung cancer epithelial cells (A549 cells) were cultured in DMEM supplemented with 10% FBS at 37°C with 5% CO₂. The RAW264.7 cells seeded in 96-well plates (2 × 10⁴ cells/well) were treated with S. aureus USA300 at a multiplicity of infection (MOI) of 60, after different concentrations of ECG (0, 16, or 32 µg/mL) were added, and the samples were cocultured for 5 h. Then, the samples were centrifuged (1,000 rpm for 10 min), the supernatant was added to an equal volume of lactate dehydrogenase (LDH) reagent (Solarbio, Beijing, China), and the mixture was incubated in the dark for 30 min. The level of LDH in each sample was determined by measuring the OD₄₉₀. After having been washed with sterile PBS, the cells were treated with ethidium bromide (EB, 50 µg/mL) to observe the inhibitory effect of ECG on the toxicity of S. aureus USA300 by using fluorescence microscopy (OLYMPUS, IX83). The cytotoxicities of the tested compounds were determined by detecting the LDH levels. For the adhesion assay, A549 cells in 24-well plates (1 × 10⁵ cells/well) were treated with S. aureus USA300 (MOI = 30) containing ECG (0, 8, or 32 µg/mL). At 1 h later, the medium was discarded, the cells were washed with sterile PBS, and the samples were plated onto LB agar medium after dilution. The number of colonies was used to assess the inhibitory effect of ECG on bacterial adhesion.

S. aureus USA300 was cocultured with different concentrations of ECG (0, 16, or 32 µg/mL) for 8 h. The bacteria and supernatants were subsequently collected separately after centrifugation (12,000 rpm for 5 min). The expression levels of sortase A and β-lactamase in the bacteria were detected by western blotting. The secretion of β-lactamase into the supernatant was evaluated via Coomassie brilliant blue staining. The antibodies used in these experiments were stored in our laboratory (mouse sortase A, 1:1,000; mouse β-lactamase, 1:1,000; goat anti-mouse secondary antibodies, 1:10,000).

Male C57BL/6J mice (weighing approximately 20 g) were purchased from Liaoning Changsheng Biotechnology Co., Ltd. All animal experiments were performed in accordance with the requirements of Jilin University. The mice had free access to food and water. The mice were intranasally inoculated with 30 μL of S. aureus USA300 suspension (5 × 10⁸ CFUs) to establish a pneumonia model. At 2 h later, ECG (50 mg/kg), Amp (50 mg/kg), or their combination was administered to the infected mice. Mice in the positive control (PC) group were infected with S. aureus USA300 and administered an equal volume of solvent instead of compound, and the mice in the negative control (NC) group were administered an equal volume of sterile PBS only. A total of 10 mice were assigned to each group for the survival assay, and the survival status of the mice was observed every 12 h. For other analyses, each group contained eight mice. After 48 hours of treatment, the mice were anaesthetized by injection of pentobarbital sodium (50 mg/kg) and then euthanized by cervical dislocation. The lung tissues of the mice were harvested, and pathological damage was observed after haematoxylin−eosin staining. Additionally, the tissue bacterial load was analyzed after homogenization and culture on LB agar, the ratio of the weight of each lung before and after drying was analyzed, and the levels of inflammatory factors in the alveolar lavage fluid was detected via enzyme-linked immunosorbent assays (ELISAs; Sangon Biotech, Shanghai, China).

The data are presented as the means with the standard deviations (SDs) or standard errors of the means (SEMs) from three independent experiments. Statistical analysis was performed via the unpaired t test in GraphPad Prism 9.5.0. Statistical significance was defined as p ≤ 0.05.

The relative activity of sortase A was defined as 100% at a test compound concentration of 0 µg/mL. Sortase A activity was reduced to 79.11%, 64.14% and 44.52% upon treatment with 8, 16 and 32 µg/mL ECG, respectively ( Figure 1A ). Similarly, reductions in sortase A activity were observed after treatment with 8, 16 and 32 µg/mL EGCG (13.68%, 36.35% and 61.22%, respectively) ( Figure 1B ), EGC (28.23%, 29.87% and 48.07%, respectively) ( Figure 1C ), C (10%, 14.32% and 30.88%, respectively) ( Figure 1D ) and EC (2.94%, 16.12% and 40.14%, respectively) ( Figure 1E ). These results showed that ECG and its analogues could significantly inhibit the activity of sortase A and that the gallic acid moiety played a role in this activity.

Molecular docking was carried out, and these compounds were found to bind to the catalytic pocket of sortase A ( Figure 2A ). The binding energies of ECG, EGCG, EGC, EC, C to sortase A were -7.2, -7.0, -6.7, -6.97 and -6.9 kcal/mol, respectively. Analysis of the sortase A binding site revealed that Thr180, Ser116, Glu108, Asn107 and Asn114 are involved in the binding of EC to sortase A ( Figure 2B ); that Gln178 and Asn107 of sortase A interact with C ( Figure 2C ); that EGCG interacts with residues Gln178, Leu169, Asp170 and Asn107 of sortase A ( Figure 2D ); and that residues Gln178, Asn114, Leu169, Thr180, Asp170 ( Figure 2E ), Asn107 and Gln178 are involved in the interaction between EGC and sortase A ( Figure 2F ). These results indicate that the residues involved in the binding of each compound were not exactly the same even though all of the interacting residues were located in the same catalytic pocket.

The RMSD values of sortase A and ECG fluctuated around 0.38 nm and 0.1 nm, indicating that sortase A and the ECG complex reached an equilibrium state during the molecular dynamics simulations ( Figure 3A ). Analysis of the binding free energies from the molecular dynamics simulations revealed that electrostatic interactions (ele), van der Waals forces (vdw) and solvation energy (sol) participated in compound binding ( Figure 3B ). Residue energy decomposition analysis revealed that Val168, Leu169, Ala104, Thr180, Asp165, Ile182, Asp170 and Lys173 have greater contributions to the binding energy ( Figure 3C ), which was confirmed by the closer proximity of these residues to ECG ( Figure 3D ). Site-directed mutagenesis was performed to construct the V168A, K173A and L169A mutants, and the inhibitory effects of ECG on these mutants were significantly reduced compared with inhibition of the WT protein ( Figure 3E ), indicating that these residues are important for ECG and sortase A binding.

sortase A is closely related to biofilm formation. Here, we quantitatively analyzed the effects of ECG and its analogues on S. aureus USA300 biofilm formation via crystal violet staining. S. aureus USA300 biofilm formation was defined as 100% in the absence of test compounds. When the bacteria were treated with 8 or 32 µg/mL C, biofilm formation was 103.40% and 85.84%, respectively ( Figure 4A ). Similarly, biofilm formation was reduced to 43.48% and 34.83% when the bacteria were treated with 8 and 32 µg/mL EGC, respectively ( Figure 4B ), and these values were 33.83% and 18.44%, respectively, upon treatment with ECG ( Figure 4C ); 13.26% and 10.92%, respectively, upon treatment with EGCG ( Figure 4D ); and 82.33% and 44.19%, respectively, upon treatment with EC ( Figure 4E ). These results showed that ECG and its analogues inhibited S. aureus USA300 biofilm formation to different degrees.

Adhesion to host cells is the first step in S. aureus infection. The virulence factor anchored by sortase A to bacterial cell walls is closely related to bacterial adhesion. Therefore, in this study, the effects of ECG and its analogues on the adhesion of S. aureus USA300 to human lung epithelial cells were analyzed. The number of S. aureus USA300 colonies resulting from the treatment of A549 cells in the absence of test compounds was defined as 100%. When S. aureus USA300-treated A549 cells received 32 µg/mL of the test compounds, the relative adhesion of the bacteria was 54.67% for C ( Figure 5A ), 14.89% for EGCG ( Figure 5B ), 46.90% for EGC ( Figure 5C ), 17.83% for ECG ( Figure 5D ) and 49.91% for EC ( Figure 5E ). ECG and EGCG showed significantly better antiadhesion effects than the other compounds, which is consistent with the above results.

The β-lactamase secreted by S. aureus can decompose β-lactam antibiotics and induce bacterial tolerance. Nitrocefin is a kind of cephalosporin that is usually used to detect the activity of β-lactamases (Zhou et al., 2020), almost all β-lactamases can hydrolyze the amide bond between the carbonyl carbon and the nitrogen group on the β-lactam ring of nitrocefin, which could be judged by the color changing or detecting the OD₄₉₂ to analyze the inhibitory effect of inhibitors on β-lactamase activity. In this assay, the samples that were not treated with ECG or its analogues were defined as 100%. When treated with 128 µg/mL test compounds, β-lactamase activity was reduced to 29.01% for ECG ( Figure 6A ), 40.85% for EGCG ( Figure 6B ), 73.24% for EGC ( Figure 6C ), 63.94% for C ( Figure 6D ) and 69.58% for EC ( Figure 6E ), indicating that ECG and its analogues inhibited β-lactamase activity.

This study evaluated the effects of ECG and its analogues on the tolerance of S. aureus USA300 to PG, Cef, Amp and kanamycin. When the concentrations of EGC, ECG and EGCG were 4 µg/mL, the MICs of PG and Cef were reduced by 2-, 8- and 2-fold, respectively ( Figures 7A, B ). The MIC values of Amp against S. aureus USA300 decreased by 16- and 2-fold when it was combined with ECG and EGCG (4 µg/mL), respectively, but cotreatment with EGC (4 µg/mL) did not change the MIC value of Amp ( Figure 7C ). The MIC value of kanamycin against S. aureus USA300 was greater than 256 µg/mL when kanamycin and ECG, EGCG, or EGC (32 µg/mL) was coadministered to S. aureus USA300 and the MIC value did not change significantly ( Table 2 ), suggesting that these compounds did not synergize with kanamycin. These results indicate that ECG has excellent potential for use in combination with β-lactam antibiotics against infections caused by S. aureus USA300. For the synergistic sterilization assay, S. aureus USA300 showed similar growth trends with and without ECG treatment, but in the Amp and Amp combined with ECG treatment groups, the bacterial density gradually decreased over time. In the combination group, a more rapid downward trend was observed, and bacteria were not detected in the medium after 10 h ( Figure 7D ). The persister formation assay revealed that the bacterial density in the gentamicin treatment group decreased sharply, suggesting that the bacteria were killed quickly, but the downward trend gradually flattened, indicating that some bacteria survived as persisters; however, bacteria were not detected in the combination group after 2 h of treatment ( Figure 7E ). Taken together, these results indicate that ECG enhanced the bactericidal ability of Amp and inhibited persister formation synergistically.

S. aureus induces cytotoxic effects that promote the development of infection. In this study, ECG was used to evaluate the cytotoxicity mediated by S. aureus USA300. Red fluorescence was abundant in the S. aureus USA300 treatment group, indicating that most of the cells died, but the intensity of red fluorescence decreased gradually upon treatment with 16 or 32 µg/mL ECG ( Figure 8A ). In the LDH assay, the amount of LDH secreted by RAW264.7 cells treated with S. aureus USA300 alone was defined as 100%. When 16 or 32 µg/mL ECG was added to the samples, the amount of LDH decreased to 69.14% and 44.48%, respectively ( Figure 8B ), indicating that ECG significantly alleviated the toxicity mediated by S. aureus USA300 to the cells. The amounts of LDH secreted from cells treated with 32 or 64 µg/mL ECG analogues were similar to that in the DMEM treatment group ( Supplementary Table S1 ), suggesting that these analogues did not exhibit cytotoxicity. Furthermore, the expression levels of sortase A and β-lactamase did not change significantly when the samples were treated with 16 or 32 µg/mL ECG ( Figures 8C, D ), and the secretion of β-lactamases in the bacterial culture supernatant was similar ( Figure 8E ), indicating that ECG did not affect the expression or secretion of β-lactamase or sortase A.

The β-lactamase–ECG complex was determined to be in equilibrium when the RMSD fluctuated around approximately 0.2 and 0.1 nm ( Figure 9A ). In this complex, ECG was located in the active center of the enzyme ( Figure 9B ), and binding site analysis revealed that Ile158, Ile230, Ala229, Pro265, Ala95, and Tyr96 of β-lactamase interacted with ECG, and Asn123 and Gln228 of β-lactamase formed H-bonds with ECG ( Figure 9C ). The binding free energy results revealed that Tyr96, Ile230, and Ile158 contributed more energy to this interaction, and the total free energy contributions of these residues were -8.73, -7.81, and -4.03 kJ/mol, respectively ( Figure 9D ). To identify key residues involved in ECG binding, site-directed mutagenesis of β-lactamase was performed, and the results revealed that the inhibitory effects of ECG against the Y96A, I158A, and I230A mutants were much lower than that against the WT protein ( Figure 9E ), suggesting that these residues are critical for ECG binding.

Mice in the PC and Amp treatment groups died after 12 h, and the survival ratios of the two groups were 6.67% and 13.33%, respectively. However, in the ECG and combination groups, the mice died at 24 or 36 h, and the survival ratios reached 46.67% and 76.67%, respectively ( Figure 10A ). In addition, the number of S. aureus USA300 colonies in the ECG and ECG combined with the Amp treatment groups was significantly lower than that in the PC and Amp treatment groups ( Figure 10B ). Additionally, the lung tissues from the PC and Amp treatment groups exhibited inflammatory factor infiltration and the alveolar structures were destroyed compared with those in the NC group, whereas these manifestations were alleviated in the samples from the ECG and ECG combined with Amp groups ( Figure 10C ). Compared with that in the PC group, the wet/dry lung weight ratios from the ECG and ECG combined with Amp treatment groups were significantly lower, but Amp treatment alone did not affect this ratio ( Figure 10D ). As important indicators of inflammation level, TNF-α and IL-1β are often used to quantitatively evaluate the inflammation level of cells or tissue (Li et al., 2019; Wang et al., 2022). S. aureus infection is usually accompanied by an excessive inflammatory response, accompanied by an excessive release of TNF-α and IL-1β (Ghosh et al., 2023; Ghosh and Bishayi, 2024); the levels of TNF-α and IL-1β here in the samples from the ECG and ECG with Amp groups were significantly lower than those in the other groups, and Amp treatment alone slightly reduced the levels of these cytokines ( Figures 10E, F ). These results suggest that ECG alone or in combination with Amp could delay mouse death, improve mouse survival, weaken the colonization ability of bacteria, and relieve edema and inflammation in the lungs.