Escherichia coli DHB4, Acinetobacter baumannii ATCC 19606, Enterobacter cloacae ATCC 700323, Klebsiella pneumoniae ATCC 700603, and Pseudomonas aeruginosa ATCC 27853 strains were purchased from China General Microbiological Culture Collection Center (CGMCC) and were cultured in Luria Bertani (LB) and stored in the logarithmic phase with 20% glycerin at −80°C in our lab.

E. coli multidrug-resistant (MDR) strain BC1 was isolated from patients with UTI at First Clinical Hospital of Yichang (the First College of Clinical Medical Science, China Three Gorges University), with an approval and written informed consent for research from the Ethics Committee of the First Affiliated Hospital of China Three Gorges University and informed consent of the patient (Wang et al., 2020).

1,3-Dibenzyl-4,5-diphenyl-imidazol-2-ylidene silver (I) acetate was obtained from the School of Chemistry, University College Dublin, Ireland. 2-phenyl-1,2-benzisoselenazol-3(2H)-one was purchased from Selleck Chemicals. MEF cell line (embryonic fibroblast cells), and S-KN-SH (human neuroblastoma strain) were purchased from China National Collection of Authenticated Cell Cultures. Protease inhibitor cocktail and 2′,7′-Dichlorodihydrofluorescein diacetate (H₂DCF-DA) were purchased from MedChemExpress. Iodacetamide (IAM) and Penicillin-streptomycin solution were purchased form ThermoFisher. CellTiter-Glo^® for ATP assay was purchased from Promega. Bacterial RNA Extraction Kit, 2× Universal Blue SYBR Green qPCR Master Mix, and SweScript All-in-one First-Strand cDNA Synthesis SuperMix for qPCR were purchased from Vazyme. LB medium was purchased from EMD Millipore. E. coli DHB4 Trx protein and anti-E. coli Trx1 polyclonal antiserum were purchased from Santa Cruz, IgG2a mouse monoclonal antibody was obtained from VIROGEN, rabbit anti-sheep IgG-HRP and Anti-Dnak antibodies were from Santa Cruz, protease inhibitor cocktails were from Roche, and all the other reagents were from Sigma-Aldrich.

MEFs and S-KN-SH were cultured in DMEM medium supplemented with 10% (v/v) fetal bovine serum, 50 units/mL penicillin-streptomycin solution were added at 37°C in a 5% CO₂ incubator. Cells were seeded in 96 micro-well plates and grown till 80% confluency. Cells were further treated with a serial combination of 0–8 μg/ml SBC3 and Ebselen (0.55, 1.09, 2.19, and 2.72 μg/ml Ebselen equal to 2, 4, 8, 10 µM Ebselen, separately) for 48 h, and the cell toxicity was detected by ATP assay as described by manufacture.

There are five Gram-negative pathogen species: E. coli DHB4, uropathogenic E. coli BC1, A. baumannii ATCC 19606, E. cloacae ATCC 700323, K. pneumoniae ATCC 700603, and P. aeruginosa ATCC 27853. All cultures were grown until an A₆₀₀ of 0.4 and at 37°C, 180 rpm and further diluted in 1:1,000 times in LB medium and further treated by a serial combination of SBC3 and Ebselen for 24 h, and the A₆₀₀ of bacterial culture was detected to calculate bacterial viability. Bacterial viability was measured using Ultraviolet-visible spectrophotometry (UV-Vis) at A₆₀₀. The mean OD of four wells in the indicated groups was used to calculate the percentage of bacterial viability as follows: percentage of bacterial viability = (A_treatment − A_blank)/(A_control − A_blank) × 100% (where, A = absorbance). Values were plotted by averaging duplicate wells.

The synergism of SBC3 and Ebselen was determined using the Bliss Independence Model, which calculates a degree of synergy using the following formula: S = (f_X0/f₀₀) (f_0Y/f₀₀) − (f_XY/f₀₀), where f_XY refers to E. coli DHB4 growth rate in the presence of the SBC3 and Ebselen in combination at a concentration X, for SBC3, and Y for Ebselen; f_X0 and f_0Y refer to E. coli DHB4 growth rates in the presence of the individual SBC3 and Ebselen at a concentration of X and Y, respectively; f₀₀ refers to E. coli DHB4 growth rate in the absence of SBC3 and Ebselen; and S corresponds to the degree of synergy.

E. coli DHB4 was grown until an A₆₀₀ of 0.4 at 37°C, 180 rpm and treated with 6 μg/ml SBC3 and 21.6 μg/ml Ebselen for 30 min. Cells were obtained by 4°C, 12,000 rpm, 15 min centrifugation and further fixed with 2.5% glutaraldehyde. Bacterial cells were stained by 2% uranyl acetate and dropped to the surface of copper wire mesh for the sample preparation. The morphology of DHB4 cells was observed by transmission electron microscopy (TEM, Hitachi H-7500).

E. coli DHB4 cells were cultured until an A₆₀₀ of 0.4 at 37°C, 180 rpm and incubated with 6 μg/ml (10 μM) SBC3, 21.6 μg/ml (80 μM) Ebselen, and 6 μg/ml SBC3 and 21.6 μg/ml Ebselen for 30 min respectively. The DHB4 cells were stained at 37°C with 10 μmol/L H₂DCF-DA. After 30 min incubation, the intracellular ROS production level was quantified by flow cytometry (BECKMAN COULTER, CytoFLEX). Briefly, an FSC/SSC plot was draw with both axis are in log mode. Then the samples were run at low-rate settings to acquire at least 5,000 cells per sample. A gate was created on the unstained sample and was applied to the rest to acquire fluorescent data from the stained sample and calculate the statistics.

E. coli DHB4 cells were cultured until an A₆₀₀ of 0.4 at 37°C, 180 rpm and incubated with 6 μg/ml (10 μM) SBC3 and 21.6 μg/ml (80 μM) Ebselen, 6 μg/ml SBC3 and 24.3 μg/ml (90 μM) Ebselen, 6 μg/ml SBC3 and 27.0 μg/ml (100 μM) Ebselen for 30 min respectively. DHB4 cells were obtained by 4°C, 5,000 rpm, 3 min centrifugation, and treated by a protease inhibitor cocktail (in 50 mmol/L Tris-HcL-EDTA buffer, pH 7.4). The DHB4 cells were further disrupted by sonication (30 W, 10 min), and the supernatants were obtained by 4°C, 12,000 rpm, 20 min centrifugation. The total protein has been detected by BCA assay and normalized to 20 μg, and the Trx/TrxR activities and GSH amount assays were performed in a 96-well plate by DTNB assay as described previously (Zou et al., 2017). The activity of the untreated group was 100%.

DTNB experiments were carried out with 96 micro-well plates in the reaction mixture containing pH 7.5 50 mM Tris-HCl, 200 μM NADPH, 1 mM EDTA, 1 mM DTNB, in the presence of 5 μM E coli Trx. The UV-vis at A₄₁₂ was measured to represent TrxR activity. The Trx activity was detected by DTNB assay coupled with 100 nM E coli TrxR instead of 5 μM Trx in the solution. To measure GSH total amount, 25 μg of the cell lysates was added in the solution containing 50 nM GR, pH 7.5 50 mM Tris-HCl, 200 μM NADPH, 1 mM EDTA, 1 mM DTNB.

E. coli DHB4 cells were cultured until an A₆₀₀ of 0.4 at 37°C, 180 rpm and incubated with 6 μg/ml (10 μM) SBC3 and 21.6 μg/ml (80 μM) Ebselen, 6 μg/ml SBC3 and 24.3 μg/ml (90 μM) Ebselen, 6 μg/ml SBC3 and 27.0 μg/ml (100 μM) Ebselen for 30 min respectively and were detected by Western blotting assay. After lysis by sonication as described above, the cell lysates were obtained by centrifugation at 12,000 rpm for 20 min and re-suspended in lysis buffer containing 30 mM IAM (protease inhibitor) only for glutathione-protein complexes. Western blotting assay was performed with anti-E coli Trx1 polyclonal antiserum (for Trx1) or IgG2a mouse monoclonal antibody (for glutathione-protein complexes). The normalized Trx1 level with respect to the reference DnaK has been presented.

E. coli DHB4 cells were cultured until an A₆₀₀ of 0.4 at 37°C, 180 rpm and incubated with 6 μg/ml (10 μM) SBC3 and 21.6 μg/ml (80 μM) Ebselen, 6 μg/ml SBC3 and 24.3 μg/ml (90 μM) Ebselen, 6 μg/ml SBC3 and 27.0 μg/ml (100 μM) Ebselen for 30 min respectively. The total RNA was extracted, and further reverse transcribed to cDNA. Real-time fluorescence PCR (qPCR) analysis was performed with a StepOne Plus system. The qPCR primers for the genes of interest, trxa/trxb/grxa/rnr and a housekeeping gene rrsA were used as the reference and are listed in Supplementary Table S1, and the cycling protocol was briefly listed as follows. An initial denaturation at 95°C for 30 s, 40 cycles at 95°C for 5 s, 61°C for 30 s. A corresponding melting curve was constructed to ensure that there was no contamination. qPCR assay was performed in triplicate for each sample and a no template control was included to rule out contamination and primer-dimers’ formation. The expression fold changes of trxa/trxb/grxa/rnr were calculated based on the comparison with rrsA.

All animal experiments were performed in accordance with the relevant guidelines and regulations. Sixty healthy male BALB/c mice were randomly divided into four groups (n = 15). Mice were injected i.p. with DMSO, 2 mg/kg SBC3, 25 mg/kg Ebselen, 2 mg/kg SBC3 and 25 mg/kg Ebselen in combination on day 3 and 1 before infection. On day 0, mice were infected i.p. with 100 μL 2 × 10⁸ CFU/mL E coli BC1 to construct an acute peritonitis model. Seven days post infection, the overall survival was calculated.

Forty healthy male BALB/c mice were divided into four groups randomly (n = 10). Mice were injected i.p. with DMSO, 2 mg/kg SBC3, 25 mg/kg Ebselen, 2 mg/kg SBC3 and 25 mg/kg Ebselen in combination on the day 3and 1 before infection. On day 0, mice were infected i.p. with 100 μL 2 × 10⁷ CFU/mL E coli BC1 to construct a mild peritonitis model. The bacterial load was calculated by counting the colony formation unity (CFU). Mice serum were collected on the day 2, 4, and 6- post-infection and centrifuged at 3,000 rpm for 10 min. The creatinine (Cre), serum urea (UA), alanine transaminase (ALT), and aspartate aminotransferase (AST) were determined by biochemical analyzer (HITACHI 7180) as described previously (Wang et al., 2020; Dong et al., 2022).

Statistical analyses were performed by GraphPad Prism 9.0 (GraphPad Software). Means of data between two groups were contrasted using Student’s t test. Data among groups were contrasted using ONE-WAY ANOVA. Sample rates between two groups were tested with chi-square analysis. Overall survival was analyzed by the log-rank (Mantel-Cox) test. p values of <0.05 were significant.

The synergistic effect of SBC3 in combination with Ebselen on the growth of E. coli DHB4 was detected by UV-Vis assay at A₆₀₀. The results showed that SBC3 alone inhibited E. coli DHB4 growth with a MIC of 25 μg/ml (Supplementary Figure S1A), while the addition of 2.16 μg/ml Ebselen decreased the MIC of SBC3 to 0.25 μg/ml (100 times, Figure 1A) with the S degree of 0.77 (Supplementary Table S2). SBC3 alone inhibited E. coli BC1 growth with a MIC of 16 μg/ml (Supplementary Figure S1B), while the addition of 0.54 μg/ml Ebselen decreased the MIC of SBC3 to 0.25 μg/ml (64 times, Figure 1B) with the S degree of 0.81 (Supplementary Table S2). SBC3 alone inhibited A. baumannii ATCC19606 growth with a MIC of 20 μg/ml (Supplementary Figure S1C), while the addition of 0.27 μg/ml Ebselen dramatically decreased the MIC of SBC3 to 0.25 μg/ml (80 times, Figure 1C) with the S degree of 0.24 (Supplementary Table S2). SBC3 alone inhibited E. cloacae ATCC700323 growth with a MIC of 8 μg/ml (Supplementary Figure S1D), while the addition of 0.27 μg/ml Ebselen decreased the MIC of SBC3 to 2 μg/ml (4 times, Figure 1D) with the S degree of 0.77 (Supplementary Table S2). SBC3 alone inhibited K. pneumoniae ATCC700603 growth with a MIC of 20 μg/ml (Supplementary Figure S1E), while the addition of 0.54 μg/ml Ebselen decreased the MIC of SBC3 to 4 μg/ml (5 times, Figure 1E) with the S degree of 0.62 (Supplementary Table S2). SBC3 alone inhibited P. aeruginosa ATCC27853 growth with a MIC of 25 μg/ml (Supplementary Figure S1F), while the addition of 1.08 μg/ml Ebselen decreased the MIC of SBC3 to 2 μg/ml (12.5 times, Figure 1F) with the S degree of 1.00 (Supplementary Table S2).

The toxicity of SBC3 and Ebselen on the growth of MEFs and S-KN-SH were also detected. Cells were treated with a serial concentration of SBC3 for 24 h and the cell toxicity was detected by ATP assay. The results showed that SBC3 possesses a dose-dependent toxicity on mammalian cells, the addition of Ebselen will effectively reduce its applicable concentration and will not increase the IC₅₀ values of SBC3 against MEFs and S-KN-SH (Figures 1G–J). Thus, the above results indicated that Ebselen and SBC3 in combination showed the synergy of antibacterial effect with the decreased application dose of SBC3.

Cell viability was represented by measuring A₆₀₀. The growth curve showed a synergistic antibacterial effect of SBC3 and Ebselen on E. coli DHB4. 6 μg/ml SBC3 and 21.6 μg/ml Ebselen inhibited E. coli growth 480 min post-treatment (Figure 2A and Supplementary Figure S2). The effect of SBC3 in combination with Ebselen on the morphology of E. coli DHB4 was observed by TEM. Normal E. coli DHB4 has a smooth surface and an integrated cell membrane and cell wall. By treatment with 6 μg/ml SBC3 and 21.6 μg/ml Ebselen for 30 min, E. coli DHB4 cells were deformed, cell membranes and cell walls were ruptured, and overflow of cell contents was detected; meanwhile, although 21.6 μg/ml Ebselen treated E. coli DHB4 cells showed morphological changes, there is no obvious rupture of bacterial cells (Figure 2B). All the results showed that the treatment of SBC3 in combination with Ebselen displays strong bactericidal activity against E. coli DHB4.

The effects of SBC3 in combination with Ebselen on bacterial TrxR and Trx activities and the GSH amount in E. coli DHB4 were measured by a DTNB assay. All the results showed that the combination of 6 μg/ml SBC3 and 27.0 μg/ml Ebselen could efficiently inhibit the activity of TrxR and deplete the amount of GSH when compared to the control or 27.0 μg/ml Ebselen alone; meanwhile, the activity of Trx1 showed no difference when compared to the control (Figures 3A–C).

The mean fluorescent intensity (MFI) of ROS in E. coli DHB4 cells was detected by flow cytometry after being stained by H₂DCF-DA (Ezraty et al., 2017; Zou et al., 2017). The result showed that the intracellular ROS level in E. coli DHB4 cells that were treated with 6 μg/ml SBC3 and 21.6 μg/ml Ebselen was significantly up-regulated when compared with the control (Figures 3D,E). The thiol-containing compound dithiothreitol (DTT) is an organic reducing agent that can remove ROS, and the protective effect of dithiothreitol (DTT) against SBC3 in combination with Ebselen treated E. coli DHB4 was evaluated by UV-Vis at A₆₀₀ and FCM, separately. Although 4 mmol/L DTT can effectively rescue E. coli DHB4 from the bactericidal effect of 6 μg/ml SBC3 and 21.6 μg/ml Ebselen, DTT at this concentration can be toxic. That is why with the addition of 4 mmol/L DTT, full recovery of E. coli DHB4 from SBC3 and Ebselen treatment was not possible (Figure 3F). Meanwhile, the results in Figure 3G showed that 6 μg/ml SBC3 and 21.6 μg/ml Ebselen could cause significantly high MFI of ROS, which also has been statistically reduced with the addition of 4 mmol/L DTT. Both results indicated that the death of E. coli DHB4 caused by SBC3 and Ebselen in combination was related to ROS production.

Whether SBC3 in combination with Ebselen could affect the expression levels of protein Trx1 and S-PSSG was further analyzed by Western blot assay. The results showed that the Trx1 protein expression level was not different by densitometric analysis, indicating the treatment of SBC3 in combination with Ebselen has no influence on Trx1 protein expression (Figures 4A,B). Meanwhile, protein S-PSSG was significantly reduced by 6 μg/ml SBC3 in combination with 27.0 μg/ml Ebselen treatment compared to 6 μg/ml SBC3 or 27.0 μg/ml Ebselen alone (Figures 4D,E), which further points to the loss of GSH. Furthermore, the mRNA expression levels of key gens in TDRS were also detected by qPCR. The results show that the relative expression level of trxa (encodes Trx1, Figure 4C), trxb (encodes TrxR, Supplementary Figure S3A), grxa (encodes GR1, Figure 4F) and rnr (encodes RNR, Supplementary Figure S3B) in SBC3 in combination with Ebselen treatment group were significantly up-regulated when compared with those for the control. These results show that SBC3 in combination with Ebselen affects the Trx system bioactivity rather than mRNA or protein expressions.

All the results suggest that SBC3 in combination with Ebselen could disrupt E. coli intracellular redox homeostasis to perform the bactericidal activity.

Mice were randomly divided into four groups, and i.p. injected with DMSO, 2 mg/kg SBC3, 25 mg/kg Ebselen, 2 mg/kg SBC3 and 25 mg/kg Ebselen in combination on the day 3 and 1 before infection. On day 0, mice were further infected with 100 μL 2 × 10⁸ CFU/ml or 100 μL 2 × 10⁷ CFU/ml to construct acute and mild peritonitis models, respectively. 90% of the mice with acute peritonitis treated with SBC3 in combination with Ebselen survived, compared with 30% in the DMSO group (Figure 5A). Meanwhile, SBC3 and Ebselen in combination led to a significant reduction in bacterial load compared with the DMSO treatment (Figure 5B). Mice treated with SBC3 and Ebselen in combination achieved a 76% reduction, followed by 54% in the SBC3 treatment and 26% in the Ebselen treatment (Figure 5B). All the findings demonstrated the highly effective antibacterial effect of SBC3 and Ebselen in combination against MDR E. coli BC1 in vivo.

The peripheral blood from different groups of mice were collected, Cre, UA, ALT, and AST in mouse blood serum were detected, and the results showed that SBC3 and Ebselen treatment had no influence on the Cre (Figure 5C), ALT (Figure 5D), and AST levels (Figure 5E). Although there was a slight increase of the UA, the results still fall into the normal level (1.7–8.3 mM, Figure 5F).

Overall, SBC3 and Ebselen treated-mice exhibited the greatest effects in reducing the MDR E. coli BC1-caused infection, demonstrating SBC3 and Ebselen may assist the recovery from MDR E. coli-caused infection.