Escherichia coli K-12 strain ATCC 25404 was used as a model biofilm-forming strain. To generate a fluorescent variant, E. coli was transformed with the GFP plasmid pCM18 (Hansen et al., 2001), which conferred resistance to erythromycin. E. coli were grown overnight in Luria Bertani (LB, Land Bridge Technology, Beijing, China) broth containing 100 μg/mL ampicillin and IPTG with shaking (250 rpm) at 37°C. V. parahaemolyticus S36 and L. monocytogenes WaX12 used this study were isolated and stored in our laboratory. V. parahaemolyticus S36 and L. monocytogenes WaX12 were isolated from shrimp and pork samples by using specific selective media, species-specific gene and API system tests (BioMérieux, Marcyl’Etoile, France). V. parahaemolyticus S36 was cultured in tryptic soy broth (TSB, Beijing Land Bridge Technology Company Ltd, Beijing, China) plus 3% NaCl. L. monocytogenes WaX12 serotype 1/2a was grown in brain heart infusion (BHI, Land Bridge Technology, Beijing, China). The cultures were diluted to obtain a bacteria population of 9 log CFU/mL.

Biofilm formation experiments were carried out as described previously (Krom and Willems, 2016; Song et al., 2016) with minor modifications. Static biofilms were grown in 24 well polystyrene microtiter plates (Sangon Biotech Co., Ltd, Shanghai, China). The test bacteria cultures were diluted in fresh culture medium (1:100) and aliquoted into wells. E. coli was incubated statically to form biofilms for various time (2, 4, 6, 8, 10, 12, 24, and 48 h). L. monocytogenes and V. parahaemolyticus were incubated statically to form biofilms for 48 h (Ayebah et al., 2006; Song et al., 2016).

Acidic electrolyzed water was prepared according to Wang et al. (2014). The AEW generator model FW-200 (AMANO Corporation, Kanagawa, Japan) was ran for 15 min with the amperage set as 10 A before collection of a sample for testing. The pH and ORP were determined using a pH/ORP meter (model pH 430, Corning Life Sciences, New York, United States). The ACC in AEW was determined by a colorimetric method using a digital chlorine test kit (RC-2Z, Kasahara Chemical Instruments Corp., Saitama, Japan). All measurements were carried out in triplicate. The physicochemical properties of each AEW are shown in Table 1.

After incubation, biofilm production was quantified using a crystal violet staining method as described previously by Antoniani et al. (2010). Biofilms in the wells of the polystyrene microtiter plates were air-dried for 10 min, then stained with 1 mL of 0.1% (w/v) crystal violet (Sangon Biotech Co., Ltd, Shanghai, China) for 30 min. The wells were then washed three times with 0.1 M phosphate-buffered saline (PBS) (Sangon Biotech Co., Ltd, Shanghai, China). Biofilm was solubilized using 1 mL of 95% ethanol (Sinopharm Chemical Reagent Co., Ltd., Shanghai, China) for 30 min. The optical density of each well was measured at wavelength of 600 nm.

The viability of the biofilm cells was measured by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay which has been attributed to Mshana et al. (1998) and Krom et al. (2007). One milliliter of culture medium and 0.1 mL of 5 mg/ml MTT solution were added to each well, then incubated at 37°C for 2 h. The culture supernatant was then discarded and 1 mL of dimethyl sulfoxide was added to each well to solubilize the MTT for 2 h. The optical density of each well was measured at wavelength of 570 nm. All measurements were carried out in triplicate and the mean data are presented.

Following biofilm formation, the suspension was gently aspirated from the plate and the wells were rinsed three times with PBS to remove non-adherent cells. The biofilms were then exposed to 1 mL sterile deionized water (SDW) or AEW produced by different NaCl concentration (0.1, 1, 3, and 5 g/L) marked AEW-1, AEW-2, AEW-3, AEW-4, respectively, at room temperature (25 ± 1°C). Subsequently, 1 mL neutralizing agent (PBS containing 0.8% Na₂S₂O₃) was added to stop the bactericidal effects of AEW after a 30 s treatment (Luppens et al., 2002; Wang et al., 2014). Surviving cells were collected by vortexing and scraping of the wells and transferred to tubes containing sterile 0.85% NaCl solution. Serially dilution of the bacterial population was plated onto correspondent agar and incubated at 37°C for 24 h. Three replicates were tested for each treatment. According to Wang et al. (2016), the percentage of reduction biofilm cells (%) = (the cells numbers in control group - the cells numbers in treatment group)/the cells numbers in control group × 100. The percentage of reduction biofilm cells represents the removal efficiency.

Biofilms which were treated with AEW-3 and SDW and untreated control, were then fixed with 4% glutaraldehyde overnight, and dehydrated in an ascending acetonitrile series (30, 50, 70, 80, 90, and 100% twice for 10 min each). Samples were sputtered with gold and observed with a Nova 450 scanning electron microscope (FEI, Hillsboro, OR, United States). Epifluorescence visualizations were carried out without previous fixation or dehydration and directly observed in a EVOS^® FL Auto Cell Imaging System (AMG, Thermo Fisher Scientific, Waltham, MA, United States). We initially observed different areas in one sample at low magnification. Then we choose one area in each sample based on similarity of all images and use high magnification of this area for analysis. The images from three independent experiments with three replications were used for analysis. Pictures were obtained using the same settings for each picture. The fluorescent density in the biofilm cells were quantified by the ImageJ software (National Institutes of Health, Bethesda, MD, United States)¹.

Extracellular polymeric substance in a biofilm was extracted using the sonication method (Liu et al., 2007; Gong et al., 2009; D’Abzac et al., 2010). The density of suspended cultures was initially measured at OD_{595 nm}. The biofilm cells were then collected by vortexing and scraping in 1 mL 0.01 M KCl solution. The cells were pretreated with a sonicator (VCX 500, SONICS, Newtown, CT, United States) for four cycles of 5 s of operation and 5 s of pause at a power level of 3.5 Hz. The sonicated suspension was centrifuged (4,000 rcf, 20 min, 4°C), and the supernatant was then filtered through a 0.22 μm membrane filter (Sangon Biotech Co., Ltd., Shanghai, China). The amounts of protein and carbohydrate in the filtrate were analyzed. The amounts of carbohydrate and protein were quantified by the phenol–sulfuric acid method and Lowry method (Kim and Park, 2013; Nakamura et al., 2013). The amount of protein and carbohydrate were quantified by OD_{750 nm}/OD_{595 nm} and OD_{490 nm}/OD_{595 nm}, respectively. Each experiment was carried out at least three times.

Extracellular polymeric substance was extracted as described in EPS chemical analysis. All Raman spectrum were obtained by a Senterra R200-L Dispersive Raman Microscope (Bruker Optics, Ettlingen, Germany) at room temperature. Raman spectrum of each sample was determined as the average of fifteen measurements at different random sites on the biofilm. The Raman measurements were recorded with an accumulation time of 60 s in the 500–1250 cm^-1 range. Raman spectral acquisition and preprocessing of preliminary data were carried out using the Bruker OPUS software.

The percentage of reduction biofilm cells, carbohydrate and protein content in EPS and Raman intensity were analyzed by analysis of one-way ANOVA using the SPSS statistical software (version 19.0; SPSS Inc., Chicago, IL, United States). The level of statistical significance was p < 0.05.

Bacterial biofilm formation is a dynamic process with distinct phases of development (Hall-Stoodley et al., 2004). In principle, there are three stages of biofilm development: an initial attachment and growth phase, a mature phase and a dispersal phase. In our model system, stage one was evident when the number of small homogeneous green-fluorescent colonies of biofilm cells gradually increased (Figures 1BI–III). Figure 1A shows the corresponding increase in cell viability and biomass of the biofilm from 2 h to 6 h. Stage two of the biofilm developmental process could be seen from 8 h to 24 h where we observed the aggregation of cells into mature biofilm (Figures 1A,BIV–VII). The mature biofilm showed green fluorescence exclusively in their border regions, resulting in multilayer films of bacterial cells. This indicated that the biofilm had been established and the biomass within the biofilm was relatively constant. After 24 h of cultivation, stage three commenced and the biomass appeared disaggregated as transient motility of the biofilm cells led to dispersal (Figures 1BVIII,IX).

The numbers of E. coli cells in the biofilm after 24 h were approximately 6.77 log CFU/mL as confirmed by the plate count method. Figure 2A showed the effect of the different AEW treatments to cell numbers in the mature biofilms in our model system. AEW-1, AEW-2, AEW-3, and AEW-4 were produced using increasing concentrations of NaCl and the reduction in cell number was positively correlated to the NaCl concentration. The bactericidal activity of the AEW-3 treatment was optimal and the trend in reduction of biofilm cell number was confirmed by the visualization of E. coli biofilms using epifluorescence microscopy (Figure 2C). These images showed that only a few scattered viable cell aggregates were observed in the biofilm after 2 min exposure to AEW. The fluorescent intensity, an indication of cell density and viability, was still at maximum intensity after the SDW treatment and then decreased progressively after the AEW treatments. The fluorescent intensity was minimal after AEW-3 treatment and we selected this treatment for further analysis.

Increasing the contact time with AEW-3 from 2 to 10 min increased cell death and removal, minimally (Figures 2B,D). Taken together, it was observed that increasing the potency of AEW using increasing NaCl concentration was more important than treatment time for the eradication of E. coli biofilm cells. In addition, the number of viable cells which escaped from the biofilm after AEW-3 treatment was below the limit of detection (<1.4 log CFU/mL). However, the residual viable cells after SDW treatment were as much as 6 log CFU/mL. We chose AEW-3 exposure for 5 min for further experimentation.

To test the effect of AEW on EPS production, we analyzed the total carbohydrate and protein content of EPS in a 24 h E. coli biofilm. As shown in Figures 3A,B, both total carbohydrate and total protein were reduced after exposure to SDW and AEW-3 for 5 min. However, total protein was reduced more than total carbohydrate. The total protein with AEW was 65% of the control, whereas total carbohydrate with AEW was 72% of the control. However, SDW treatment resulted in only little reduction of carbohydrate and protein, and had no significant difference compared to control.

Representative Raman spectra of EPS in the spectral fingerprint range of 500–1250 cm^-1 are presented in Figure 3C. The tentative peak assignments of the bands are summarized in Table 2. The prominent Raman bands of EPS belonged to carbohydrates: 561 cm^-1 (C-O-C glycosidic ring def polysaccharide) and 1090–1095 cm^-1 (C-O-C glycosidic link). The band intensity weakened dramatically with AEW treatment. Besides, EPS after AEW treated showed a shift in 561 cm^-1 toward 581 cm^-1, 1095 cm^-1 toward 1106 cm^-1. Two broad Raman bands at 1020–1085 cm^-1 and 855–899 cm^-1 were assigned to C-C stretching of carbohydrates (polysaccharides). After AEW-3 treatment, these intensities were significantly weakened as shown in Figure 3D.

Extracellular polymeric substance showed a decrease in the magnitude of Raman intensity at 637–695 cm^-1, 830–850 cm^-1, 1003 cm^-1, and 1005 cm^-1, the bands which corresponds to proteins. For example, bands at 1005 cm^-1 could be observed in the spectrum of the control group and SDW group. However, these bands were not present in the spectra of EPS after AEW treatment.

The reduction of Raman intensity corresponding to DNA, such as the bands at 788 cm^-1 and 830–850 cm^-1, arises from the destruction of the ring structure, indicating degradation of the DNA. This reduction provides further evidence for cell death or removal.

To gain further insight into the mode of action of AEW in eradicating biofilms, SEM images of E. coli biofilm treated with AEW were performed. Representative SEM images of E. coli biofilm are shown in Figure 4. Overall, the untreated samples revealed biofilms had well organized network structures encased in a protective EPS. After SDW treatment, there were still aggregates of cells held together by EPS. However, there were significant differences after AEW treatment, where only a few cells were scattered sporadically and cell lysis was evident when compared to control. The SEM images revealed that most of the biofilm cells were detached from the biofilm matrix and suggested that EPS was disrupted after AEW treatment.

While AEW demonstrated activity against E. coli biofilms, it was of interest to explore whether biofilms formed by other foodborne pathogens such as L. monocytogenes and V. parahaemolyticus, were also sensitive to AEW.

The efficacy of AEW in eradicating biofilm cells of L. monocytogenes and V. parahaemolyticus are showed in Figure 5. The populations of L. monocytogenes and V. parahaemolyticus biofilm cells were decreased by 82 and 52% after AEW treatment. There were significant differences obviously between the reduction of biofilm cell number after AEW treatment and those after SDW treatment (p < 0.05). SDW treatment marginally reduced the population of biofilm cells of L. monocytogenes and V. parahaemolyticus. In contrast, the L. monocytogenes biofilm was more susceptible to AEW treatment than the V. parahaemolyticus biofilm. Additionally, the detachment of residual viable cells from a biofilm after AEW treatment was markedly less when compared to biofilms treated with SDW (data were not shown). After AEW treatment, the detached cells which were a potential source of secondary pollution were lower than detecting value.

Extracellular polymeric substance analysis of V. parahaemolyticus and L. monocytogenes biofilm is shown in Figures 6A,B. SDW treatment only result in 3–6% EPS reduction and there was no significant difference between SDW treatment and control. The carbohydrate and protein content of EPS in V. parahaemolyticus were reduced by 34 and 44%. Comparatively, there was a reduction of 53% carbohydrate and 75% protein content in EPS of L. monocytogenes biofilm. AEW treatment therefore is more effective in reducing protein than carbohydrate content in EPS of L. monocytogenes and V. parahaemolyticus. Representative Raman spectra of EPS in the spectral fingerprint range of 500–1250 cm^-1 are presented in Figures 6C–F. The Raman bands of EPS in L. monocytogenes biofilm at 1090 cm^-1 (C-O-C glycosidic link) were decreased after AEW treatment, indicating the destruction of the carbohydrate structure (Figures 6C,D). Similar trends were observed for carbohydrates at band 565 cm^-1 (C-O-C glycosidic ring def polysaccharide) and band 1095 cm^-1 (C-O-C glycosidic link).

The band intensity was also weakened by a big margin after AEW treatment as shown in Figures 6E,F. Raman intensity in V. parahaemolyticus and L. monocytogenes biofilm was also reduced at bands 637–695 cm^-1 and 1003 cm^-1 (proteins) after AEW treatment. The Raman bands at 1,003 cm^-1 could be observed in the spectra of the control and SDW group, but these bands were not present in the spectra of EPS in L. monocytogenes biofilm after AEW treatment. The decrease of Raman intensity at band 788 cm^-1 which corresponds to DNA, arises from the destruction of the ring structure and indicated the degradation of DNA.

The eradication effect of AEW on biofilm formed by L. monocytogenes and V. parahaemolyticus was further evaluated using SEM. Representative SEM images (Figures 7A,B) show that untreated biofilm had nearly uniform and dense mature architecture. L. monocytogenes biofilm forms a much stronger structure than V. parahaemolyticus biofilm. SDW treatment had no obvious effect in biofilms. In contrast, AEW removed the biofilm cells and destroyed the high ordered structures, resulting in much less dense, and individually formed colonies when compared to the control (Figures 7A-III,B-III).