Following the synthetic route shown in Figure 2, CS-SS-LF was synthesized. ¹H NMR and FTIR confirmed the formation of the CS-SS-LF conjugate. As indicated in Figure 3, ¹H NMR of the newly synthesized compound was in full agreement with the proposed structure. The signals of the phenyl hydrogen group and the vinyl hydrogen of LF moiety were observed at 8.82 ppm and 7.66 ppm, respectively. The typical signals of the methylene hydrogen were detected around 3.23–3.00 ppm, which indicated the formation of ester moiety. Moreover, the construction of the acylamide moiety was confirmed by the signal around 5.59 ppm, which corresponded to the hydrogen of the-CONH-group. The typical skeletal signals of CS were appeared around 4.12–3.64 ppm. The above appeared chemical shifts indicated the successful formation of the CS-SS-LF conjugate. The connection was also verified using FTIR analysis. As shown in Figure 4, as compared to those of CS, the new peaks of CS-SS-LF appeared at ~1294 cm⁻¹ and 1620 cm⁻¹, which assigned to stretching of amines and aromatic C-C (Jalvandi et al., 2017). As measured by UV–vis at 295 nm, the drug content of CS-SS-LF was 21.5 wt%, indicating synthesis success.

The amphiphilic CS-SS-LF conjugates readily formed micelles by self-assembling in aqueous medium. The morphology and size distribution of CS-SS-LF micelles were studied using scanning electron microscopy (SEM) and dynamic light scattering (DLS). Based on DLS measurements, CS-SS-LF micelles had a diameter of approximately 130 nm (Figure 5A) and a surface charge of 11.7 mv (Figure 5B), indicating that they contained cationic amine groups. As demonstrated by SEM in Figure 5C, the CS-SS-LF micelles displayed spherical morphology with approximately 120 nm in diameter. The stability of colloids is one of the most important aspects of nanoscale drug delivery systems. DLS measurements showed that CS-SS-LF micelles remained unchanged in diameter and polydispersity index for 6 days (Figure 5D). These results demonstrated that amphiphilic CS-SS-LF conjugates are excellent drug carriers since they can self-assemble into micelles.

The trigger effect of H₂S was further investigated by incubating CS-SS-LF micelles with or without 10 mM Na₂S. As shown in Figure 6, about 55% drug was released in 2 h and reached 85%in 10 h with Na₂S treatment. In contrast, less than 20% of the conjugated LF was released in PBS from CS-SS-LF micelles. As the disulfide bond being cleaved by H₂S, rapid release occurs (Pal et al., 2016). The sustained retention is advantageous in micelles delivery because it prevents leakage of the drug prior to reaching the target site and ensures its delivery in larger quantities to be released at the infection site. According to these results, CS-SS-LF micelles exhibited high stableness under physiological conditions and can be used for targeted release.

To investigate the antimicrobial specificity of CS-SS-LF micelles, three strains (Salmonella, P. aeruginosa, and S. aureus) were selected as the model bacteria to qualitatively analyze the MIC of H₂S-sensitive micelles, and the results were presented in Supplementary Table S1. It is apparent that the MIC of CS-SS-LF micelles against P. aeruginosa and S. aureus were significantly higher compared with that of free LF, marking that the nanodrug basically lost its antimicrobial activity against these two strains. However, the MIC of CS-SS-LF micelles against Salmonella was only slightly elevated (0.4 μg/mL vs. 0.25 μg/mL of net LF), revealing the high sensitivity and specificity of the modified micelles against Salmonella.

The bactericidal activity of CS-SS-LF micelles and LF were tested with Salmonella at different concentrations. As shown in Figure 7, cultured Salmonella bacteria in TSB containing CS-SS-LF micelles, equivalents of CS or LF, were analyzed for their growth curves. According to the results, CS-SS-LF micelles (80 μg/mL) inhibited the growth of Salmonella cells effectively compared to the blank control. By contrast, CS treatment had a less effective effect on inhibiting bacterial growth. Meanwhile, 10 μL samples (10⁶ dilution) were cultured on Petri dishes for 10 h and then the CFU were counted. Based on these results, CS-SS-LF micelles demonstrated ability in suppressing planktonic Salmonella growth.

As an opportunistic human pathogen with Gram-negative status, Salmonella is commonly employed as a biofilm model. Salmonella biofilms from established cultures were treated with different concentrations of CS-SS-LF micelles (12.21, 24.42, 48.84, 97.and 68 μg/mL) for 24 h to evaluate CS-SS-LF micelles^, disruption properties. It was determined that CS-SS-LF micelles were more effective at destroying live cells at all concentrations tested than LF alone (Figures 8A, B).

As shown in Figure 8C, Salmonella biofilms exposed to CS-SS-LF micelles for 24 h exhibited fewer scattered cell aggregates than individual reagents, as evidenced by visualization of biofilms by fluorescence microscopy.

Salmonella biofilms were treated with different samples and their apparent morphologies were assessed using SEM. As indicated in Figure 8D, biofilms that were treated with TSB have clearly intact bacteria cells with well-defined shapes and well-organized architectures. While some bacteria on the surface of the biofilms were destroyed by the CS treatment, intact cells and obvious aggregates were still visible. Surface roughness, cellular deformation, and cytoplasm leakage of bacteria were observed in LF and CS-SS-LF micelles groups. Noun broken bacterial cells were found in treated biofilms containing CS-SS-LF micelles.

Taken together, these results demonstrated that polycationic properties enabled CS-SS-LF micelles to be more effective in eliminating Salmonella biofilms than single drug.

For the MTT assay (Figure 9A; Supplementary Figure S1), cells were incubated with various concentrations of CS-SS-LF micelles for 24 h. We found that over 85% of the incubated cells remained viable after 24 h incubation even under the highest concentration of CS-SS-LF micelles at 250 μg/mL. In addition, the blood compatibility of CS-SS-LF micelles was estimated via a red blood cell hemolysis assay in vitro. As shown in Figure 9B, 250 μg/mL concentration in CS-SS-LF micelles group display no obvious in hemolysis of RBCs. The biocompatibility of these micelles further supported their potential as effective targeting vehicle for drug delivery.

Test mice were infected with Salmonella (10⁶ CFU/mouse) and given therapy 24 h after infection. As indicated in Figure 10, liver, spleen, and kidney viable counts decreased when CS-SS-LF micelles (10 mg/Kg) and LF were treated by intraperitoneal injection. In addition, compared with LF, the CS-SS-LF micelles displayed better bacteria clearance in spleen and kidney. Study results showed that mice treated with CS-SS-LF micelles experienced decreased Salmonella infection in the intraperitoneal cavity. The strategy could serve as a useful tool in developing new therapies for Salmonella-associated infections.

Chitosan (~10 kDa), MTT 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyl tetrazolium bromide, Succinic anhydride, 4-(dimethylamino) pyridine (DMAP), 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (EDC), Levofloxacin (LF), Dichloromethane (DCM) 2-Hydroxyethyl disulfide, N-hydroxysuccinimide (NHS) were obtained from Sigma Co. (St. Louis, MO, United States). Gibco Ltd. (Grand Island, NY, United States) provided the Roswell Park Memorial Institute (RPMI) 1,640 medium, while Hyclone (Logan, UT, United States) provided the fetal bovine serum (FBS).

Prof. Wang (College of Life Sciences, Northwest A&F University) generously donated Salmonella typhimurium (SL1344) and Macrophages (RAW264.7). We obtained Kunming mice from Pengyue Experimental Animal Breeding (Jinan, China).

A total of 10 mg CS-SS-LF conjugation was dispersed in 10 mL DMSO. Stirring the solution for 5 h and dialyzing it against deionized water (MWCO 3.5 kDa) followed. After that, micelles solution was filtered through a microporous membrane with a pore size of 0.45 μM.

A Dynamic Light Scattering (DLS) method (Malvern Zatasizer NANOZS90, Malvern, United Kingdom) was used to determine the average particle size and zeta potential of CS-SS-LF micelles. The morphology of the CS-SS-LF micelles was studied using a field emission scanning electron microscope 163 with an accelerating voltage of 10 kV by Hitachi.

DLS was used to measure the diameter and PDI changes of CS-SS-LF micelles after incubation in PBS (pH 7.4) for 6 days.

Assaying drug release from LF micelles (CS-SS-LF) in the presence or absence of a 10 mM Na₂S was conducted in a dialysis tube (MWCO 3.5 kDa) under 100 rpm shaking at 37°C in PBS (10 mM, pH 7.4) containing 10% FBS. The release of drugs was measured using 1.0 mL of CS-SS-LF micelles (1 mg/mL) dispersion dialysis against 30 mL of a corresponding medium. For every desired time interval, 2.0 mL of the release medium was removed and replaced with equal quantities of fresh medium. UV–vis absorption spectrum was used to determine the drug concentration. The data are presented as mean ± SD (n = 3).

To investigate the antibacterial activity of this micelles in vitro, a two-fold dilution method was employed to determine the minimum inhibitory concentration (MIC) of CS-SS-LF micelles against Salmonella, P. aeruginosa, and S. aureus. To obtain working cultures, 50 μL of TSB medium was added in a 96-well plate, then 50 μL of bacterial suspensions (Salmonella, P. aeruginosa, and S. aureus, respectively) was inoculated in each well to adjust the final cell density to approximately 1.0 × 10⁵ CFU/mL. The bacteriostatic agents (CS-SS-LF micelles and LF) was diluted successively and then added to the prepared bacterial suspensions. The concentration of bacteriostatic agent corresponding to wells without bacterial precipitation was the MIC.

Similarly, with the addition of different concentrations of CS-S-S-LF micelles, the bacteria samples (0.4 OD₆₀₀, 0.1 mL) were well mixed with 3.9 mL tryptone soya broth (TSB). At intervals, the OD₆₀₀ was monitored after shaking the mixtures at 37°C.

For the formation of biofilms, 100 μL of bacterial TSB solutions (~10⁸ CFU/mL) were cultivated in 96-well plates for 24 h at 37°C. After removing non-adhered bacteria, the plate was washed with PBS three times. Afterward, 100 μL TSB containing CS-SS-LF micelles, LF and CS were incubated at 37°C with the existing biofilm. The biofilms were incubated with TSB only as blank control. Each treatment was divided into six parallel wells. As previously reported (Zhang et al., 2013), biofilm mass (Crystal violet staining assay) and viable cells (MTT assay) were evaluated.

As previously described (Mu et al., 2016), Salmonella biofilms were grown on glass coverslips placed at the bottom of 12-well plates. The coverslips were washed, and the residual biofilms treated as before.

In order to prepare the SEM sample, Salmonella strains were pre-incubated with TSB, CS, LF and CS-SS-LF micelles, followed by removal of the medium, followed by fixing with 4% glutaraldehyde in PBS for 2 h. After that, 40, 50, 70, 90, and 100% ethanol was used to dehydrate the samples. Supercritical CO₂ drying was then used to dry the bacterial biofilms. SEM analysis was performed on the dried samples after they were plated with platinum.

MTT assays were used to assess the cytotoxicity of CS-SS-LF micelles. RPMI 1640 medium containing 10% FBS was used to culture RAW 264.7 and buffalo rat liver 3A (BRL-3A) cells (6 × 10³ cells/well), which were incubated at 37°C with 5% CO₂ for 24 h. Following that, CS-SS-LF micelles were added to the cells for 24 h in different concentrations. Afterwards, cells were further incubated for 4 h with 20 μL MTT solution. Blue violet crystal formazan was dissolved in 100 μL DMSO. The absorbance was measured at 570 nm using a microplate reader with PBS as a blank control.

The hemolysis assay was performed with CS-SS-LF micelles to check its compatibility with blood. Briefly, red blood cells (RBC) solution (2% w/w) of 0.1 mL was added to 1.9 ml of the solution of CS-SS-LF micelles (250 μg/ml). Deionized water was used as positive controls (PC), while PBS was used as negative controls (NC). After 3 h incubation at 37°C, the samples were centrifuged at 6500 rpm for 10 min, and then 200 μL supernatant was removed carefully and the optical density was measured at 545 nm by a microplate reader. The hemolytic ratio (%) was calculated based on the formula:

The experiment was conducted using female Kunming mice (6 weeks old). Salmonella was injected intraperitoneally into mice at a dose of about 1 × 10⁶ CFU per animal (Zhi et al., 2016). Three times a day for 3 days, mice received subcutaneous injections of CS-SS-LF micelles, LF, CS, or PBS (10 mg/kg of CS-SS-LF micelles, n = 5). An aseptic procedure was used to remove the liver, spleen, and kidneys from the mice, which were homogenized with sterile saline solution in a final volume of 1 mL. By plating serial dilutions of the cultures onto TSB agar, CFUs were counted.

Plots and statistical analyses were carried out using the software GraphPad Prism 7.01. Experimental data were expressed as mean ± standard deviation (SD).