Apramycin Sulfate was purchased from the China Institute of Veterinary Drug Control (Beijing, China). L-Citrulline and L-Glutamine (Biotechnology grade) were purchased from Shanghai Macklin Biochemical Technology Co., Ltd (Shanghai, China). Mueller Hinton (MH) Agar, MH Broth, Tryptic Soy Agar (TSA), Luria-Bertani (LB) broth, MacConkey Agar, and Xylose Lysine Deoxycholate (XLD) Agar were purchased from Guangdong Huankai Microbial Sci & Tech. Co., Ltd (Guangdong, China). Biotechnology grade cyclophosphamide was purchased from Shanghai Yuanye Bio-Technology Co., Ltd (Shanghai, China). Methanol and acetonitrile (high-performance liquid chromatography grade) were purchased from Thermo Fisher Scientific (Waltham, MA, United States). Carbonyl cyanide m-chlorophenylhydrazone (CCCP), phosphate buffer, and Sodium chloride were purchased from sigma Aldrich (Missouri, United States), GE Healthcare Life Sciences China (Beijing, China), and Shanghai Runjie Chemical Reagent Co., Ltd (Shanghai, China), respectively.

Standard strains of Escherichia coli (ATCC25922) and Salmonella choleraesuis (ATCC13312) were purchased from American Type Culture Collection (Manassas, VA, United States), Salmonella typhimurium (CICC21484) was purchased from the China Center of Industrial Culture Collection (Beijing, China), and Salmonella typhi suis (CVCC3783) was purchased from the China Veterinary Culture Collection Center (Beijing, China). Clinically resistant Salmonella strains (SP-JH-Y1407-52, SP-JH-Y1407-31, and SP-JH-Y1407-22) were donated by the Center for Agricultural Product Quality and Safety of Guangdong province (Guangdong, China).

The minimum inhibitory concentrations (MICs) of Apramycin against different Salmonella were determined using the microdilution method (Li et al., 2015). The increasing concentration method was used to induce drug-resistant strains. Salmonella strains (ATCC13312, CICC21484, and CVCC3783) were inoculated into the broth, and Apramycin at 1/2 MIC concentration was added. The MIC of the bacteria was determined after 1week of continuous culture at 37°C. The culture was continued with the new 1/2 MIC concentration until the MIC of the bacteria increased by more than 32 times, at which point the induction was stopped. The induced drug-resistant strains were passaged in a control medium without any drugs for three generations, and then, the MIC was determined to confirm that stable drug resistance was obtained.

Non-targeted metabonomics studies of Apramycin-resistant Salmonella ATCC13312 (Apr-R-ATCC13312) and a sensitive strain (ATCC13312) were performed. The resistant or sensitive Salmonella was cultured in broth to the late logarithmic growth stage. Then, 5ml of the bacterial suspension was taken and centrifuged at 16000g, at −10°C for 10min. The precipitate was washed twice with PBS. The bacteria were collected and quickly quenched in liquid nitrogen to stop their metabolism. Then, 1ml of methanol/acetonitrile/water (2:2:1, V/V) was added, vortexed for 60s, and extracted ultrasonically twice at a low temperature. The supernatant was placed at −20°C for 1h to precipitate proteins. The samples were centrifuged at 20000g at 4°C for 20min, and the supernatant was collected, lyophilized, and stored at −80°C. Before the test, 100μl of acetonitrile/water (1:1, V/V) was added to the lyophilized sample to dissolve the cell metabolites. The samples were separated using ultra-performance liquid chromatography with a HILIC column and analyzed using a triple TOF 5600 mass spectrometer (AB SCIEX, Framingham, MA, United States). Electrospray ionization (ESI) was used to monitor the samples in positive and negative ion modes. To avoid the influence of the fluctuation of the instrument detection signal, continuous analysis of the samples was carried out in random order. Quality Control (QC) samples were inserted into the sample queue to monitor and evaluate the stability of the system to ensure the reliability of the data. The mass spectrometry data were processed by Software Analyst 1.6. The total ion diagram and MRM multi peak diagram of the sample were obtained. According to the retention time and peak type of metabolites, the mass spectrum peaks detected by each metabolite in different samples were corrected again. The original data were converted into MzXML format by Proteo Wizard, and then, the XCMS program was used for peak alignment, retention time correction, and peak area extraction to form a data set for statistical analysis. Multi-dimensional and single-dimensional statistical analysis were performed, including unsupervised principal component analysis (PCA), supervised partial least squares discriminant analysis (PLS-DA), orthogonal partial least squares discriminant analysis (OPLS-DA), T-tests, and multiple of variation analysis.

Different Apramycin-induced resistant Salmonella was cultured in broth. In the same bacterial culture medium, different treatments were carried out, including adding the drug alone, adding the exogenous substances alone, and adding the drug and the exogenous substances at the same time. The bacteria of each group were cultured at 37°C for 8h. The cultures were sampled at 0, 2, 4, 6, and 8h for bacterial counting, and bacterial survival curves were drawn. The effects of different concentrations of exogenous substances on resistant bacteria were also investigated. In addition, the same method was used to verify the reversal effect of exogenous substances on clinically isolated resistant bacteria.

Apramycin; two exogenous substances (citrulline plus glutamine); Apramycin plus citrulline and glutamine; and Apramycin plus citrulline and glutamine plus 20μM CCCP were added into the culture medium of Apramycin-induced drug-resistant Salmonella (Apr-R-ATCC13312). The bacteria were counted at 0, 2, 4, 6, and 8h, and bacterial survival curves were drawn. The survival rate of each group was calculated in comparison with the bacteria in the culture medium without treatment.

Different exogenous substances were added into the culture medium of Apramycin-resistant Salmonella (Apr-R-ATCC13312 and Apr-R-CICC21484), which were then cultured at 37°C for 8h. Then, exactly 1ml of the bacterial solution was removed and centrifuged at 20000g for 5min at −10°C. The supernatant was discarded; the precipitate was washed twice with PBS and then diluted using PBS to approximately 1×10⁶ cells per ml. Then, the membrane potential was measured using a BacLight Bacterial Membrane Potential Kit (Life Technologies, Carlsbad, CA, United States) according to the manufacturer’s instructions.

Three induced drug-resistant Salmonella (Apr-R-ATCC13312, Apr-R-CICC21484, and Apr-R-CVCC3783) and a clinical drug-resistant Salmonella (SP-JH-Y1407-22) were chosen for the NADH measurement study. 300μl of Apramycin-resistant bacterial culture medium was sampled and added into MH broth with 120mm citrulline and 120mm glutamine, 120mm citrulline only, 120mm glutamine only, and MH broth without any exogenous substances. The bacterial culture of each group was then incubated at 37°C for 8h, at which point exactly 1ml of bacterial culture medium was removed, centrifuged at 20000g for 5min at −10°C, the supernatant was discarded, and the precipitate was washed twice with cold PBS. The bacterial precipitate was placed in a 1.5ml centrifuge tube, extracted, and detected using a NAD ⁺/NADH assay kit (BioAssay Systems, Hayward, CA, United States).

Apramycin-resistant Salmonella was inoculated into 10ml of MH broth containing Apramycin, Apramycin plus citrulline and glutamine, Apramycin plus citrulline, or Apramycin plus glutamine, respectively. After incubation at 37°C for 8h, 10ml of the culture was taken and centrifuged at 20000g for 5min at −10°C. The supernatant was discarded and the precipitate was washed with saline twice. The bacterial cells were added with 0.5ml saline, vortexed for 30s, and then treated with ultrasound for 10min. After centrifugation, the supernatant was collected, and the concentration of Apramycin was determined using an Apramycin ELISA kit (Shanghai Enzyme-linked Biotechnology Co., Ltd., Shanghai, China). The detection limit of the kit was 0.01ng/ml, the accuracy of the method was greater than 80%, the internal plate error of the absorbance of the kit was less than 8%, the inter plate error was less than 15%, and the standard curve ranged from 0.05ng/ml to 4.05ng/ml. This experiment was carried out on three induced drug-resistant strains (Apr-R-ATCC13312, Apr-R-CICC21484, and Apr-R-CVCC3783) and one clinical drug-resistant strain (SP-JH-Y1407-22).

Ninety healthy female mice were divided into nine groups, with 10 mice in each group. The groups comprised the negative control group without bacterial infection, the positive control group (mice infected with drug-resistant bacteria but without any treatment), and the other seven groups were treated with Apramycin, citrulline, glutamine, Apramycin plus citrulline, Apramycin plus glutamine, or Apramycin plus citrulline plus glutamine, respectively, after infection with drug-resistant bacteria. All animals were tested and found to be Salmonella typhimurium-free before the start of experiments.

Eighty mice were intraperitoneally injected with cyclophosphamide for 5days to establish a neutropenia model. After the establishment of the model, 0.1ml Apramycin-induced Salmonella typhimurium (Apr-R-CICC21484) suspension was injected intraperitoneally at a concentration of 10⁷ colony-forming units/ml. At 12h after infection, the mice in the experimental group were injected intraperitoneally with Apramycin or citrulline and glutamine. The doses of Apramycin, citrulline, and glutamine used in the experiment were 20mg/kg body weight (b.w.), 240mg/kg b.w., and 200mg/kg b.w., respectively. At 24h after treatment, all mice were euthanized. The liver, spleen, and blood of all mice were collected under sterile conditions and placed in saline. The tissue or blood of each mouse was homogenized separately, diluted with saline, and inoculated on different nutrient agar plates. Finally, the bacterial load of each mouse liver, spleen, and a blood sample was calculated to evaluate the effect of different treatment schemes.

Statistical analysis was performed on all the data from the metabolite supplement approach, measurement of the membrane potential, NADH measurements, determination of the intracellular drug concentration, and validation using a mouse infection model experiments using SPSS Statistics 26.0 software (IBM Corp., Armonk, NY, United States).

This study was approved by the Animal Ethics Committee of South China Agricultural University (application NO:2019B192) and was carried out in accordance with the ARRIVE guidelines. The mice in the study were euthanized by cervical dislocation.

After induction culture with Apramycin, three Apramycin-resistant Salmonella strains were obtained. The MIC value of the induced drug-resistant bacteria was up to 40 times higher than that before induction (Table 1). Three clinical isolates of Salmonella (SP-JH-Y1407-52, SP-JH-Y1407-31, and SP-JH-Y1407-22) were highly resistant to Apramycin and showed partial resistance to florfenicol, tetracycline, enrofloxacin, and other commonly used veterinary drugs (Table 2).

A non-targeted metabonomic experiment was conducted in Apramycin-resistant and sensitive Salmonella. The experimental bacteria were divided into two groups, with eight biological replicates in each group. The results showed that the response intensity and retention time of the chromatographic peaks of the QC samples overlapped, indicating that the variation caused by the instrument was small during the experiment. We first checked the data integrity and normalized the data between samples and metabolites to ensure that the samples and metabolites could be compared in parallel. The PLS-DA scores of the drug-resistant and sensitive groups are shown in Figures 1A,B. This multi-dimensional statistical analysis showed that the samples in each experimental group were gathered well, while the separation trend of samples between groups was obvious, indicating that the intracellular metabolites of the drug-resistant and sensitive bacteria were quite different.

According to the variable importance for prediction (VIP) analysis obtained using the OPLS-DA model, the influence intensity and explanatory ability of the expression pattern of each metabolite on the classification and discrimination of each group of samples were measured, and the biologically significant differential metabolites were mined. Metabolites with a VIP > 1 and a value of p <0.05 were identified as having significant differences. For a more comprehensive and intuitive evaluation of the rationality of candidate metabolites, and to determine the relationship between samples and the differences of metabolite expression patterns in different samples, we used the abundance of significantly different metabolites to cluster the samples. Figures 1C,D shows the hierarchical clustering results of significantly different metabolites under positive and negative ion modes. The same group of samples can appear in the same cluster. Volcano plots (Figures 1E,F) were constructed to analyze and display the changes of metabolites more intuitively. The results showed that 115 metabolites were significantly different between Apramycin-resistant and the sensitive strains, among which 34 increased and 81 decreased. These metabolites were mainly amino acids and their metabolites; products or intermediates of energy metabolism; and substances involved in the tricarboxylic acid cycle, such as citrulline, arginine, glutamine, lysine, Ornithine, ATP, ADP, phosphoenolpyruvate, and succinate.

Different concentrations of citrulline and glutamine were added to the culture medium containing a fixed concentration of Apramycin. Viable bacteria were counted at different times. The results showed that citrulline and glutamine promoted Apramycin killing of drug-resistant Salmonella, and over a certain range, the concentration of exogenous substances correlated positively with the intensity of the synergistic bactericidal effect. When 120mm citrulline, 120mm glutamine, and 1/2 MIC Apramycin were added, the killing rate of drug-resistant bacteria was 99.99% after 8h. The same effect was observed for the clinical drug-resistant Salmonella. The survival rate of drug-resistant Salmonella in the test group treated with citrulline, glutamine, and Apramycin was significantly lower than that in the test group treated with Apramycin alone from 2h after culture (p <0.05). Thus, citrulline and glutamine could help Apramycin to kill nearly all clinical drug-resistant bacteria after 8h of culture. The details are shown in Figure 2.

Drug-resistant Salmonella was cultured in the medium containing Apramycin, Apramycin plus citrulline, Apramycin plus glutamine, and Apramycin plus citrulline plus glutamine for 8h; then, the concentration of Apramycin in bacterial cells under different culture conditions was determined by an ELISA method. The results showed that when the drug-resistant Salmonella was cultured with Apramycin plus citrulline and/or glutamine, the intracellular drug contents of different serotypes of induced drug-resistant Salmonella and clinical drug-resistant Salmonella were significantly higher than those of the control group treated with Apramycin alone (p <0.05). However, when citrulline and glutamine were added at the same time, the intracellular drug concentration was the highest (Figure 3A).

Citrulline alone, glutamine alone, and the combination of citrulline and glutamine were added to the culture medium of different drug-resistant Salmonella. After 8h of culture, the intracellular concentration NADH in the bacteria was detected. The intracellular NADH content of all drug-resistant bacteria increased by more than three times when citrulline and glutamine were added together compared with that in the untreated control group (p <0.01). The NADH content in the resistant bacteria cultured with either citrulline or glutamine alone was also higher than that in the control group (Figure 3B).

Different exogenous substances were added to the culture medium of Apramycin-resistant Salmonella, which was then cultured at 37°C for 8h. The membrane potential of bacteria in the different experimental groups increased significantly with the addition of citrulline or glutamine. Moreover, the membrane potential of two Apramycin-resistant Salmonella (Apr-R-ATCC13312 and Apr-R-CICC21484) cultured with both citrulline and glutamine increased by 22.7 and 42.2%, respectively, compared with that in the untreated control group. Next, citrulline, glutamine, CCCP, (a chemical inhibitor of oxidative phosphorylation), and Apramycin were added to the culture medium of drug-resistant Salmonella at the same time, and the bacteria in the culture medium were counted at different times. The results showed that the drug-resistant bacteria grew normally, i.e., in the presence of CCCP, the exogenous substances lost their ability to reverse drug resistance (Figures 3C,D).

Mice were injected intraperitoneally with Apramycin-resistant Salmonella typhimurium (Apr-R-CICC21484) to establish a mouse model of drug-resistant bacterial infection. The infected mice showed depression, diarrhea, and death, and the Apramycin-resistant CICC21484 strain was isolated from the liver, spleen, and blood of the infected mice. The infected mice were treated with Apramycin and different exogenous substances, and the results showed that citrulline (240mg/kg b.w) and glutamine (200mg/kg b.w) combined with Apramycin (20mg/kg b.w) significantly improved the body condition of the infected mice. In addition, the bacterial levels in the liver, spleen, and blood of the combined treatment group were significantly lower than in those treated with Apramycin alone (p <0.01). Moreover, single citrulline or glutamine combined with Apramycin treatment also showed a good therapeutic effect, and the number of bacteria in the mice was significantly lower than that in mice treated with Apramycin alone (p <0.01; Figure 4). Thus, citrulline and glutamine combined with Apramycin could be used to treat an infection with Apramycin-resistant Salmonella.