Evaluation of antimicrobial and antibiofilm properties of proanthocyanidins from Chinese bayberry (Myrica rubra Sieb. et Zucc.) leaves against Staphylococcus epidermidis

Abstract Staphylococcus epidermidis has emerged in recent years as one of the most important opportunistic pathogens owing to its ability to attach to processing surfaces in the food industry. Demands of foodstuffs maintaining microbiological safety and stability enhance the need to develop natural antimicrobial agents as food preservatives. Proanthocyanidins from Chinese bayberry leaves (BLPs) belonging to the class of polyphenols promise to be a potential antibacterial material against bacterial adhesion and biofilm formation. The aim of the present study was to investigate the effects of BLPs on S. epidermidis growth and biofilm formation. BLPs possessed antimicrobial activity with MIC and MBC of 320 and 640 μg/ml, respectively. Scanning electron microscopy, transmission electron microscopy, and flow cytometry analysis revealed a loss of the cell structure and function after treatment of BLPs, evidenced by cell membrane hyperpolarization and changes in cellular morphology. BLPs inhibited the biofilm formation by S. epidermidis on polystyrene microplates. Atomic force microscopy analysis showed that BLPs could decrease the stiffness and adhesion force of the cell envelope, which might account for the inhibition of biofilm formation. In summary, this study indicated that BLPs have potential to be developed as natural preservatives to control S. epidermidis in foods.

concern about the uncontrolled use of chemical preservatives has grown louder in recent years (Witkowska, Hickey, Alonso-Gomez, & Wilkinson, 2013). Consequently, natural antimicrobial compounds are receiving a good deal of attention for the safety of food products, including extracts of some spices, herbs, and other plants (Tajkarimi et al., 2010).
Among phytochemicals, phenolic compounds have been extensively studied due to their diverse health benefits, mainly as antioxidants, anti-inflammatory, and antimicrobial agents (Daglia, 2012).
They are oligomeric flavonoids of catechin, epicatechin, and their gallic acid esters (Jagannathan & Viswanathan, 2018). Chinese bayberry (Myrica rubra Sieb. et Zucc.) has been cultivated in Southern China for more than 2000 years, and its leaves are luxuriant but always discarded (Zhang, Chen, Wei, Chen, & Ye, 2017). Studies showed that phenolic extracts from bayberry leaves exhibit antimicrobial properties (Li, Han, Chen, & Ye, 2012). However, the effects of phenolic extracts from bayberry leaves on biofilm formation are poorly recognized.
The structure of PAs depends on the nature of flavan-3-ol linkages substituted with hydroxyl groups along with aromatic and fused oxetane rings (Jagannathan & Viswanathan, 2018). A special type of PAs from Chinese bayberry leaves was previously identified by our group (Fu et al., 2014;Yang et al., 2011). In comparison with other types of PAs from plants, such as apple, cranberry, or grape seeds, proanthocyanidins from Chinese bayberry leaves (BLPs) contain a simple but potent bioactive unit, that is, epigallocatechin gallate (EGCG); thus, the structural-activity relationship for BLPs might be clearer than PAs from other plants (Zhang et al., 2017).
The current study investigated the ability of BLPs on inhibiting bacterial growth and biofilm formation by S. epidermidis, which was isolated from a milk powder processing factory. The mechanisms of antimicrobial and antibiofilm actions of BLPs were also investigated by a series of techniques, including scanning electron microscopy (SEM), transmission electron microscopy (TEM), flow cytometry analysis (FCM), and atomic force microscopy (AFM) nanoindentation.

| Polyphenols
The BLPs used in this study were obtained according to our previous studies (Fu et al., 2014;Yang et al., 2011;Zhang et al., 2016).
A stock solution was prepared by dissolving 25 mg BLPs powder in 1 ml sterile distilled water, and then, the solution was filtered through a 0.22 μm-pore membrane filter. EGCG was used as a positive control to evaluate the antibacterial and antibiofilm capabilities of BLPs. EGCG (Shanghai Yuanye Biotechnologies Co.; purity of approximately 90%) was also dissolved in sterile distilled water at a concentration of 5 mg/ml and was filter sterilized as above. The final concentrations of BLPs ranged from 20 to 2,560 μg/ml, and the final concentrations of EGCG ranged from 5 to 640 μg/ml.

| Bacteria and growth conditions
Staphylococcus epidermidis isolated from a milk powder processing factory located in northeastern China was cultured in tryptic soy broth (TSB, Hopebio) at 37°C for 18 hr to achieve the stationary phase (Zou & Liu, 2018). Cells were harvested by centrifugation (2,350 × g, 4°C for 10 min) and then washed twice by sterile phosphate-buffered saline (PBS, pH 7.4) solution and resuspended in fresh sterile TSB (a final concentration about 10 9 CFU/ml).

| Determination of MICs and MBCs
Minimal inhibitory concentrations (MICs) and minimal bactericidal concentrations MBCs of BLPs against S. epidermidis were determined using the twofold microplate dilution assay as described by Grenier, Chen, Ben Lagha, Fournier-Larente, and Morin (2015) with minor modifications. A suspension of stationary phase bacteria in TSB (approximately 10 6 CFU/ml) was added to each well of the 96well polystyrene microplates (Costar ® 3599, Corning Life Science).
MICs of BLPs were determined as the lowest concentration at which no turbidity can be observed after incubation at 37°C for 24 hr. To determine MBCs, 100 μl of each well showing no visible growth of S. epidermidis was spread on tryptone soya agar (TSA, Hopebio) plates. MBCs of compounds were determined as the lowest concentration at which no colony formation occurred after incubation at 37°C for 24 hr. All tests were performed in triplicate.

| Detection of viable and sublethally injured cells
To estimate the amounts of sublethally injured cells, S. epidermidis cells treated with BLPs at the MIC at 37°C for 24 hr were spread on nonselective and selective plates. Then, these plates were incubated at 37°C for 24 hr. In this study, the nonselective medium was TSA without additional NaCl, while the selective medium was TSA containing 5% NaCl. Stress-induced sublethal S. epidermidis were counted by obtaining the differences in colony counts between the nonselective and selective plates.

| Determination of MBICs
Minimal biofilm formation inhibitory concentrations (MBICs) were determined. S. epidermidis was subjected to a biofilm assay using 96-well polystyrene microplates for the examination of biofilm prevention, as described previously (LaPlante, Sarkisian, Woodmansee, Rowley, & Seeram, 2012). The plates were incubated statically for 24 hr at 37°C to allow biofilm formation, followed by the crystal violet staining method (Zou & Liu, 2018). MBICs of compounds were determined as the lowest concentration at which no detectable biofilm formation occurred.

| SEM assay
SEM was employed to confirm the changes in surface morphology according to Matijasevic et al. (2016). Briefly, the cell resuspensions were treated with polyphenols at the MIC level. Control (untreated bacteria) and treated samples were incubated for 2 hr at 37°C and then harvested by centrifugation at 2,350 × g, 4°C for 10 min, followed by immobilization with 2.5% glutaraldehyde overnight at 4°C. These samples were dehydrated by a graded series of ethanol (30%, 50%, 70%, 80%, 90%, 95%, and 100%), coated with gold-palladium, and observed in a Hitachi Model SU-8010 SEM (Hitachi, Ltd.).

| TEM assay
Ultrastructural damage of the bacterial cells was evaluated using TEM. Bacterial samples were prepared and dehydrated as mentioned in the SEM assay. Then, these samples were placed in a 1:1 acetone-Spurr resin mixture for 1 hr; a 1:3 acetone-Spurr resin mixture for 3 hr; and the absolute Spurr resin overnight at room temperature. Next, samples were sectioned in a LEICA EM UC7 ultratome.
Finally, these sections were stained with uranyl acetate and alkaline lead citrate prior to observation in a Hitachi Model H-7650 TEM (Hitachi, Ltd.).

| Flow cytometric analysis for cell membrane permeability
Bacterial samples were prepared as mentioned in the SEM assay, followed by measurement using the LIVE/DEAD BacLight™ Bacterial Viability Kit (L-7012, Invitrogen). Briefly, one milliliter of S. epidermidis resuspensions was incubated with BLPs or EGCG at the MIC level, respectively. Control (untreated bacteria) and treated samples were incubated for 2 hr at 37°C. After incubation, cells were harvested by centrifugation at 2,350 × g, 4°C for 10 min and washed thrice with PBS. Then, these samples were diluted 1:100 in filter-sterilized ddH 2 O to reach a final density of 1 × 10 6 CFU/ml. Last, one milliliter of each sample was mixed with 3 μl PI and 3 μl SYTO 9 and incubated at room temperature in the dark for 15 min. Stained samples were assayed in a Gallios flow cytometer equipped with a fully functional double laser and eight detectors (Beckman Coulter Inc.), and a total of 20,000 events were recorded. Data were analyzed using the Kaluza software package (Beckman Coulter Inc.). A heat-treated bacterial suspension (80°C/20 min) was analyzed as a positive control to check the applicability of the staining protocol for the sample analysis.

| Flow cytometric analysis for cell membrane potential
The effects of BLPs on the cell membrane potential of S. epidermidis were measured using the BacLight™ Bacterial Membrane Potential Kit (B34950, Invitrogen) as mentioned by Mora-Pale et al. (2015). Briefly, bacterial samples were prepared as mentioned in the SEM assay.
Meanwhile, an additional sample (untreated bacteria) was prepared as a depolarized control by mixing with 10 µl of 500 µM carbonyl cyanide 3-chlorophenylhydrazone (CCCP). One milliliter of each treated sample was mixed with 10 μl of 3 mM DiOC 2 (3) and incubated at room temperature for 15 min. Last, samples were assayed in the Gallios flow cytometer and a total of 20,000 events were recorded. Detection mode and data analysis were performed as described above.

| AFM analysis and force measurements
The effects of BLPs on the structural, adhesive, and mechanical properties of the cell envelope were measured by AFM based on Mularski, Wilksch, Hanssen, Strugnell, and Separovic (2016). Briefly, bacterial samples were prepared in the same manner as described in the SEM assay at a concentration of MBIC. A cell suspension (10 μl) was deposited onto a piece of mica plate (10 × 10 mm, Beijing Zhongjingkeyi Technology Co., Ltd) and dried naturally. The surface micrographs were imaged at room temperature in the air with a tapping mode by AFM (Cypher S, Oxford Instruments). A nominal resonance frequency of 300 kHz and a 50 N/m spring constant were applied. NanoScope analysis software (version 1.5, Bruker) was used for image manipulation. In addition, an XE-70 AFM (XE-70, Park Scientific Instruments) was used for adhesion force measurements using a cantilever (Si 3 N 4 ) with a scan rate of 1 Hz and a nominal spring constant of 0.08 N/m. XEI data processing software (version 1.8.0, Park Systems Corporation) was used for image manipulation.

| Statistical analysis
Unless indicated otherwise, all data were presented as means ± standard deviations (SD) based on three independent experiments. The data from all assays were compared using one-way analysis of variance (ANOVA) by applying Tukey's test with all calculations carried out using IBM SPSS Statistics 20.0 software (IBM Inc.), and the statistical significances were achieved when p < .05.

| RE SULTS AND D ISCUSS I ON
Staphylococcus epidermidis plays an influential role in creating the pathogenic biofilm. The present study describes antimicrobial and antibiofilm activities and mechanisms of BLPs against S. epidermidis.
Two polyphenols with a different mean degree of polymerization (mDP) were compared: BLPs (mDP at about 7.3 ± 0.1) and EGCG (positive control, monomer).

TA B L E 1
with EGCG at the MIC cannot be cultured on the selective media.
However, the number of sublethal cells after BLPs treatment was negligible; therefore, there was no sublethal damage caused by BLPs treatment.
Bacterial cells treated with EGCG at the MIC were unable to recover viability on the selective media, whereas those treated with BLPs at the MIC were able to recover viability overall on the same media.
Adaptation of bacteria by exposure to sublethal levels of some stresses was well documented (Fernández et al., 2018). Although the inhibitory effect of BLPs against S. epidermidis was less benign than that of cranberry extracts with a MIC of 160 μg/ml as reported by LaPlante et.al (2012), BLPs showed more favorable bactericidal activity with a MBC of 640 μg/ml (Table 1) than that of cranberry extracts with a MBC range from 1,250 to 5,000 μg/ml. As put forward by Mostafa et al. (2018), the difference in the MIC of plant extracts can be due to extensive variation in their method of extraction, constituents, and structural nature of their constituents.
These results suggested that BLPs were promising and effective compounds for antibacterial applications.

| Morphological and ultrastructural changes of S. epidermidis
The antibacterial mode of action of BLPs against S. epidermidis in the present was evidenced in two ways. Firstly, microscopies were applied to identify the morphological appearance, ultrastructure, and topography of S. epidermidis. Secondly, FCM was used to show that polyphenols appear to change the membrane potential, damage the integrity, and cause functional disorder of the bacterial cell membrane.

Morphological and ultrastructural changes of S. epidermidis
incubated with BLPs at the MIC were analyzed via SEM ( Figure 1) and TEM (Figure 2). Figure 1A-a Figure 1C-c), and particles of white precipitates can be identified on the surfaces of bacterial cells.
Moreover, compared with control cells (Figure 2A-a), the ultrastructure of S. epidermidis cells was obviously affected by BLPs treatment at the MIC ( Figure 2C-c). A significant increase in the cell wall thickness of S. epidermidis exposed to BLPs was found.
As is well known, the cell membrane is an active structure that acts as a barrier between the inner and outer portions of the cell and plays a key role in maintaining optimal internal conditions for metabolism and energy transduction. Many studies have suggested that phenolic compounds, such as syringic acid, primarily target the cytoplasmic membrane and retarded bacterial growth (Cui et al., 2012).
In the present study, BLPs contain unique EGCG as its terminal unit and most of its extension units, with an mDP of 7.3 ± 0.1 (Zhang et al., 2017). The previous studies revealed that PAs with lower mDP demonstrated better bioactivity properties (Zhang et al., 2016).
Thus, it is reasonable that BLPs showed less efficient antibacterial properties with a comparison to EGCG.
Changes in the morphology of S. epidermidis treated with BLPs at the MIC for 2 hr were slightly different from those caused by EGCG, see Figure 1B-b and C-c. The presence particles of white precipitates in association with the bacterial cells in Figure   The PI is widely used as a damage marker, which is excluded by cells with intact membranes but can enter cells with compromised membranes (Liu, Xia, Jiang, Yu, & Yue, 2018  . Combined with the plate count data, it is implied that BLPs at the MIC permeabilized the bacterial cytoplasmic membrane to retard the growth of bacteria.

| Effects of BLPs on biofilm formation, topography, and nanomechanical properties of the cell envelope of S. epidermidis
The inhibition of S. epidermidis biofilm formation by BLPs was evaluated by crystal violet staining following growth in a 96-well microplate. When S. epidermidis grew in the presence of BLPs, the specific antibiofilm effect was highly dose-dependent ( Figure 5B).
When added at a concentration of 160 μg/ml (MBIC), BLPs reduced biofilm formation by 83.72% without growth inhibition.
Biofilm formation decreased with increasing BLPs concentration (320 and 640 μg/ml), due to the growth inhibition effect caused by BLPs.
AFM was used to observe topological changes of the cell envelope of S. epidermidis when treated with BLPs at the MBIC. As shown in Figure 6A, untreated S. epidermidis cells had smooth surfaces with the average surface roughness of 0.83 ± 0.02 (n = 5, Table S2), which is in line with Gram-positive bacteria visualized by this technique (Yang et al., 2010). When treated with EGCG at the MBIC for 2 hr, cells became wrinkled leaving some debris around the cells ( Figure 6B), which resulted in a higher surface roughness value of 15.00 ± 7.20 (n = 5). Meanwhile, the cells' average height was decreased to 61.98 ± 16.14 nm (n = 5), which was significantly lower than that (146.59 ± 22.26 nm, n = 5) of untreated cells (p < .05). However, the treatment with BLPs at the MBIC after 2 hr induced the surface to aggregate with the highest surface roughness (45.53 ± 4.15, n = 5), while the cells' average height was increased to 152.39 ± 12.00 nm (n = 5) ( Figure 6C).
Since it is difficult to clarify the antibiofilm mode of action of The ability of S. epidermidis to form biofilms plays a key role in food contamination. As is well known, bacteria cells embedded in biofilms are more resistance to both mechanical removal and antimicrobial agents than planktonic cells (Donlan & Costerton, 2002 (Jin et al., 2010). As shown in Figure 6C and Table   S2, the value of the cell surface roughness increased induced by BLPs suggested accumulation of cross-linking between BLPs and the cell surface, which may change the permeability of the bacterial cell membrane and lead to the membrane damage. Thus, it was implied that BLPs could damage the structure of the cell wall and/or membrane, resulting in a decrease in Young's modulus. Additionally, there was a noticeable decrease in the adhesion force of the cell surface, which may result in the inhibition of bacterial biofilm formation. In the previous study, Young's modulus has been identified as an important factor in a broad range of biological processes, such as cell differentiation, cell deformability, and cell migration (Guo, Xia, Sandig, & Yang, 2012). Therefore, it was implied that the adhesion force of S. epidermidis cells decreased with the reduction of Young's modulus. Taken together, all the AFM morphological and mechanical data indicated that the antibiofilm mode of action of BLPs against S. epidermidis at the MBIC was related to the reduction of Young's modulus and cell adhesion force.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

E TH I C A L A PPROVA L
Human or animal testing is unnecessary in our study.