The diameters of the inhibition zone of CA against Y. enterocolitica were measured by the poured-plate method (Dai et al., 2018) with some modifications. Before starting the experiments, bacterial PBS (0.02 M, pH 7.4) suspensions with a concentration of 10⁸ CFU/mL were prepared. Briefly, 100 mL of sterilized LB agar (cool to 45°) was homogeneously mixed with 0.7 mL of diluted Y. enterocolitica to reach the final concentration of 10⁴ CFU/mL and then quickly poured into a sterile plate. After solidification, 6 holes with 8 mm diameters were made by sterile punch, and filled with 200 μL of sterile water, gentamicin (40 μg/mL), CA solutions (20, 40, 60, and 80 mg/mL CA), respectively. After incubation for 12 h in a 37° constant temperature incubator, the diameters of the inhibition zone were measured. Each experiment was performed in triplicate.

According to previous reports (Mitani et al., 2018), an acidic environment might contribute to the antibacterial effect to a certain degree. In order to exclude the possible effect of acidic characteristic of CA on its antibacterial function to Y. enterocolitica, the antibacterial effect of PBS with the same pH value as 80 mg/mL CA solution (pH 2.5) was assessed and compared with CA. The experiment was conducted as above.

The time-dependent bactericidal kinetics of CA was analyzed by growth curve (Shi et al., 2018a). Bacterial PBS suspensions were obtained as above. Briefly, 0.5 mL of 20, 40, 60, 80, and 100 mg/mL CA were added into 4.5 mL of LB broth containing 10⁶ CFU/mL bacteria to make the final concentrations of 2, 4, 6, 8, and 10 mg/mL CA, respectively. All the samples were then incubated at 37° in a shaking incubator at 150 rpm. During the incubation, 200 μL of the culture was pipetted out to a sterile 96-well plate (Corning) at 2 h intervals to determine OD₆₀₀ by a Multiskan FC Microplate Reader (Thermo Fisher Scientific, United States). The same volumes of sterile water and gentamicin (40 μg/mL) instead of CA were used as negative and positive controls, respectively. The absorbance at each time point was used to plot a growth curve.

To test the effect of CA on preformed biofilm of Y. enterocolitica, the biofilm of Y. enterocolitica was first formed and tested according to a previously reported method (Wang et al., 2020a) with some modifications. An overnight bacterial culture in LB broth was harvested and adjusted to 10⁶ CFU/mL using fresh LB broth. Aliquots of 200 μL bacterial suspension were added into a 96-well flat-bottomed polystyrene microtiter plate (Corning) and incubated for 24 h at 37° at the static condition to ensure the full adhesion of biofilm to the bottom of the plate. Following 24 h of bacterial adhesion and biofilm formation, the supernatants were decanted, planktonic cells were discarded, and the plate was rinsed three times with 200 μL PBS. Then, 200 μL 4, 6, 8, and 10 mg/mL CA and water were added into the wells, respectively, and incubated for 30 min at 37° with slight shaking. Afterward, the plate was washed three times using sterile water, fixed with methanol for 20 min, washed three times with water, stained with 0.1% (w/v) crystal violet for 10 min, and rinsed three times again. Finally, 200 μL of 95% ethanol was added to the crystal violet-stained wells, and the amount of biofilm formation was measured by the absorption at 595 nm using a Multiskan FC Microplate Reader (Thermo Fisher Scientific, United States). The background from un-inoculated media was subtracted.

The inhibition of biofilm formation of Y. enterocolitica by CA was performed according to a previously reported method (Maury et al., 2019) with some modifications. An overnight bacterial culture in LB broth was harvested. Then, aliquots of 100 μL of the above prepared bacterial suspension were added into a 96-well plate (Corning) respectively, followed by the addition of 100 μL 8, 12, 16, and 20 mg/mL CA into the wells, respectively, ensuring the final bacterial concentration of 10⁸ CFU/mL in each well and the final concentration of 4, 6, 8 and 10 mg/mL CA. The plate was then incubated for 24 h at 37° at static condition, after which the remaining operation procedures were performed as mentioned above, and the absorbance of each well at 595 nm was obtained. The same volume of sterile water instead of CA was used as a negative control.

The changes in microstructure and morphology of Y. enterocolitica after treatment with CA were imaged by Scanning electron microscopy (SEM) (Fang et al., 2018). Log-phase Y. enterocolitica with a concentration of 10⁸ CFU/mL was mixed with CA (final concentration was 10 mg/mL) and cultured for 30 min at 150 rpm at 37°. Sterile water was used as a control. The samples were treated according to the reference and finally observed via SEM (JSM-7500F, Hitachi, Japan).

The intracellular alterations of Y. enterocolitica after treatment with CA were visualized by Transmission electron microscopy (TEM) (Li et al., 2019b). The treatment method of Y. enterocolitica with CA was identical to SEM protocol. The samples were treated according to the reference and finally observed via TEM (JEM-1200EX, JEOL Ltd., Tokyo, Japan).

The Propidium iodide (PI) staining experiment was conducted using a previously reported method (Logue et al., 2018). Log-phase bacterial PBS suspensions were obtained as above. Specifically, Y. enterocolitica cells (10⁸ CFU/mL) were incubated with CA (final concentration was 10 mg/mL, 150 rpm, 37°) for 0.5, 3, 6, and 9 h, respectively. After incubation, bacteria were obtained by centrifugation at 6,000 rpm for 10 min, washed three times with 0.9% NaCl solution, re-suspended in NaCl solution, and diluted to approximate 10⁶CFU/mL. Then, 1 mL of the bacterial solution was mixed with 3 μL PI (20 mM in DMSO) in the dark for 15 min at room temperature. After that, the fluorescence intensity of samples was measured by a FACS AriaIII Flow cytometer (FC) (Becton Dickinson, Franklin Lakes, NJ, United States) at an excitation wavelength of 490 nm and an emission wavelength of 635 nm.

LIVE/DEAD BacLight Bacterial Viability Kit (Invitrogen, United States) was used to further elucidate the effect of CA on the cell membrane permeability (Watson et al., 2019). According to the specification, log-phase bacterial PBS suspensions (10⁸ CFU/mL) were incubated with CA (final concentration was 10 mg/mL, 150 rpm, at 37°) for 0.5, 3, 6, and 9 h, respectively. After incubation, bacteria were harvested by centrifugation at 6,000 rpm for 10 min, washed three times with 0.9% NaCl solution, resuspended in NaCl solution, and diluted to approximate 10⁸CFU/mL. Then, the bacterial solution was incubated with 3 μL mixed dye of PI: SYTO-9 (1:1) in the dark for 15 min at room temperature. After that, samples were dropped on a microscope slide and observed using a TCSsp5II Laser scanning confocal microscope (LSCM) (Agilent, United States) at the maximum excitation/emission wavelength 490/635 nm for PI and 480/500 nm for SYTO-9.

Quartz crystal microbalance (QCM) was used to test the binding of CA to Y. enterocolitica. Briefly, 50 μL of log-phase bacterial PBS suspension (10⁸ CFU/mL) was dropped on the surface of the Au chip and incubated for 24 h at 4°) at static condition to ensure the full adhesion of bacteria to the surface of the chip. Then, the chip was rinsed with slow water flow to remove the unattached bacteria and dried with nitrogen. Finally, the chip was placed into the sample room and analyzed by QCM (QE401-F1719, Q-sense, Biolin Scientific, AB, Finland). For the experiment, ultrapure water was first injected until the stable baseline. Then, 10 mg/mL CA and ultrapure water were sequentially injected until the baseline was stable. The equipment can examine the interaction through the frequency shifts [including resonance frequency (Δf) and energy dissipation (ΔD)] of the electrode if CA could bind to Y. enterocolitica. A chip without bacteria attachment was used as a control to exclude the non-specific binding of CA to the chip.

Chlorogenic acid can naturally emit fluorescence at 440 nm (using an excitation wavelength of 380 nm) (Wang et al., 2020b). In this study, the fluorescence detection method by FC was designed to test the binding of CA to Y. enterocolitica. Specifically, 200 μL of log-phase bacterial PBS suspension (10⁸ CFU/mL) was incubated with 400 μg/mL, 600 μg/mL, 800 μg/mL, 1 mg/mL, and 2 mg/mL CA for 30 min at 150 rpm, 37°), respectively. After incubation, the mixtures were centrifuged at 6,000 rpm for 10 min, washed three times with PBS to remove the unbound CA, and bacteria concentration was adjusted to approximately 10⁶ CFU/mL with PBS. Finally, the fluorescence intensity of samples was recorded by a FACS AriaIII flow cytometer (Becton Dickinson, Franklin Lakes, NJ, United States). The bacteria treated with ultrapure water instead of CA were used as a control.

The in vitro antibacterial results of CA against Y. enterocolitica are shown in Figure 1A, and the diameters of the inhibition zone are shown in Supplementary Table 2. As seen in Figure 1A, CA exerted a dramatic growth inhibition on Y. enterocolitica. The diameters of the inhibition zone increased in a concentration-dependent manner ranging from 40 to 80 mg/mL, demonstrating that CA exerted antibacterial activity dose-dependently. Yet, no obvious inhibition zone was observed at 20 mg/mL CA in comparison with the gentamicin group.

Meanwhile, the effect of the acidic characteristic of CA on Y. enterocolitica was conducted (Figure 1B). The diameters of the inhibition zone are shown in Supplementary Table 3. Compared with 80 mg/mL CA, which exerted distinct antibacterial activity (with a diameter of inhibition zone of 13.5 mm), PBS with the same pH value as 80 mg/mL CA solution (pH 2.5) had no obvious antibacterial activity (no inhibition zone). Hence, it was concluded that the antibacterial effect of CA was not caused by its acidic characteristic but by other mechanisms of action.

The growth curve was plotted to further test the antibacterial effect of CA. According to Figure 1C, compared with the negative control group, a significant decline of OD₆₀₀ was observed in CA groups. At CA concentrations of 2, 4, and 6 mg/mL, the growth of Y. enterocolitica was initially repressed, and such a tendency lasted for 4-8 h, indicating the occurrence of partial bacteria inhibition or lysis by CA. However, at 8 mg/mL and 10 mg/mL CA, the growth curves were horizontal, and the OD₆₀₀ values remained unchanged throughout the overall process, just as that of the positive control of gentamicin, indicating the complete bacterial inhibition or lysis by CA. Moreover, from the overall tendencies of growth curves, we concluded that the antibacterial activity of CA against Y. enterocolitica was dose- and time-dependent.

The SEM microstructural analysis of Y. enterocolitica is shown in Figure 3A. In the control group, normal Y. enterocolitica was a typical rod-shaped structure with a smooth surface, and the average diameter was about 1 μm. However, after CA treatment, the shape of bacteria became irregular and wrinkled, and the surface of bacteria became rough. Moreover, some severe damages and ruptures of bacteria were observed, and the integrity of bacteria was destroyed.

The destruction of the bacterial structure by CA was further observed by TEM (Figure 3B). The untreated bacteria were bacilliform and intact. However, after incubation with CA, the cell wall disappeared locally, and the bacterial integrity was ruptured. Moreover, the contents of bacteria flowed out, and the interior structure of bacteria was completely damaged.

To further demonstrate the antibacterial mechanism of CA against Y. enterocolitica, the bacteria treated with CA for a different time were stained with PI. As a kind of nucleic acid dye, PI can penetrate cells and bind to nucleic acid only when the cell membrane is disrupted. Thus, when the cell membrane was ruptured, PI entered the cell, and its fluorescence was reserved and could be detected by FC (Shi et al., 2018b). In short, the higher the fluorescence intensity, the more severe damage on the membrane. As shown in Figure 4, the fluorescence intensity of the bacteria after treatment with CA was significantly higher than that of negative control and increased with incubation time, indicating that CA could exert the bactericidal activity by disrupting the bacterial cell membrane, which was in accordance with the results of SEM and TEM.

To further investigate if CA could cause changes in the bacterial cell membrane permeability, a live/dead cell staining method was performed using two nucleic acid dyes of SYTO-9 and PI. When SYTO-9 is used alone, all bacteria are stained (those with intact membrane and those with damaged membrane). On the contrary, PI can only penetrate the damaged bacteria (Shi et al., 2018b). However, when both dyes are used together, PI can conceal the fluorescent color of SYTO-9. Thus, red fluorescence can be seen in damaged bacteria. In other words, the higher red fluorescence intensity reflects the higher membrane permeability.

As seen in Figure 5, bacteria with red fluorescence were observed after the treatment of CA, indicating that CA could increase the bacterial cell membrane permeability. Furthermore, with the extension of CA processing time, more bacteria with red fluorescence were observed, and bacteria with green fluorescence accordingly decreased. These results suggest that CA destroys the cell membrane of Y. enterocolitica and increases the permeability of the cell membrane in a time-dependent manner.

Quartz crystal microbalance is a biosensor platform that operates on the principle of mass detection due to the piezoelectric effect. When a mass is adsorbed on the surface of the QCM electrode, the oscillation frequency of the electrode shifts and the weight changes (Speight and Cooper, 2012; Liu et al., 2013; Yilmaz et al., 2015; Emir Diltemiz et al., 2017; Feng et al., 2017). When the CA solution flowed through the Au chip surface with Y. enterocolitica, the frequency drastically increased until reaching a stable plateau. After using ultrapure water to rinse away the non-specific interaction, the frequency drastically decreased until reaching another one stable plateau, which was under the zero baseline (with the f value of -1.7912 Hz) (Figure 6A). Similarly, the weight changes on the Au chip had the same tendency line as f changes (with the weight of 32.05 ng/cm²) (Figure 6B). On the contrary, for the negative control, when the CA solution flowed through the Au chip without bacteria attaching to, the frequency also drastically increased until reaching stable plateau; but after rinsing with ultrapure water, the frequency drastically decreased until reaching the zero baseline (Figure 6A), suggesting no interaction. Differently, there were almost no weight changes for the negative control over the whole process (Figure 6B). The results indicated that the changes of f and weight were on account of the possible interaction of CA with Y. enterocolitica, and not because of the non-specific binding. Therefore, it can be inferred that specific binding of CA to Y. enterocolitica occurred.

Based on the characteristic of CA, namely, emitting fluorescence at 440 nm at an excitation wavelength of 380 nm (Supplementary Figure 1), a fluorescence detection assay by FC was performed to further analyze the binding of CA to Y. enterocolitica. As shown in Figure 6C, after incubation with CA and removal of unbound CA, the fluorescence at 440 nm was detected by FC, indicating the binding of CA to Y. enterocolitica. In addition, the fluorescence intensities had a linear relation and a positive correlation with the concentrations of CA.