All bacterial strains used in this study are listed in Table 1. Wild-type E. faecalis OG1RF and constructed E. faecalis OG1RF::p23cfp that expresses a constitutive CFP were obtained from the laboratory of Professor Gary M. Dunny (University of Minnesota, United States). All E. faecalis strains were statically cultured at 37°C in M9 medium [3g/L yeast extract, 10g/L casamino acids, 36g/L glucose, 0.12g/L MgSO₄, and 0.011g/L CaCl₂ (Bandyopadhyay et al., 2016)] or in brain heart infusion broth (BD Co., United States). If needed, the antibiotic rifampicin was added at a concentration of 200μg/ml (Bandyopadhyay et al., 2016).

Bacillus subtilis natto NTU-18 (BCRC 80390) isolated from a commercial product was maintained in our laboratory (Kuo et al., 2006; Kuo and Lee, 2008). B. subtilis natto was cultured in LB broth [10g/L tryptone, 5g/L yeast extract, and 10g/L sodium chloride (Atlas, 2010)] in orbital shakers at 37°C with shaking at 125rpm.

Bacillus subtilis natto supernatant was prepared using methods described in our previous study (Lin et al., 2021). In brief, overnight cultures of B. subtilis natto were diluted 1:100 in M9B medium (8.5g/L Na₂HPO₄·2H₂O, 3g/L KH₂PO₄, 11.5g/L sodium acetate, and 1ml/L acetic acid were added as buffering agents to the original M9 medium) and cultured aerobically in orbital shakers at 37°C with shaking at 125rpm. After 24h of cultivation, the cultures were centrifuged (4,000×g, 10min) to remove all cells. Then, the supernatant was filter-sterilized through 0.22-μm filters (Pall Co., United States) and stored at 4°C.

Caco-2, a colon adenocarcinoma cell line, was purchased from the Bioresource Collection and Research Center (BCRC, Taiwan). The cells were routinely maintained in Dulbecco’s modified Eagle medium (DMEM containing 4.5g/L glucose; GeneDireX Inc., Taiwan) supplemented with 10% fetal bovine serum (FBS; Gibco, United States) and 1x penicillin-streptomycin-glutamine (GeneDireX Inc., Taiwan) and incubated at 37°C in a 95% humidity and 5% CO₂ air atmosphere. The cells were passaged every 5–6days in 10cm² cell culture dishes.

An in vitro bacterial adhesion assay was performed as described by Letourneau et al. with some modifications (Letourneau et al., 2011). To prepare monolayers of Caco-2 cells for the in vitro bacterial adhesion assay, one milliliter of cell suspension (2×10⁵ cells/ml) was seeded in three sets of duplicate wells (one for each treatment) of a 24-well plate, and the plate was incubated in a cell culture incubator until the cells were fully confluent. The cells were then washed with phosphate-buffered saline (PBS), and the culture medium was replaced with 900μl of antibiotic-free DMEM supplemented with 10% FBS; then, 100μl of B. subtilis natto supernatant or M9B medium was added.

Overnight cultures of E. faecalis cells were centrifuged, washed twice with PBS containing 2mm EDTA, and resuspended in DMEM with 10% FBS. A volume of bacterial culture corresponding to 10⁶ E. faecalis cells was used to inoculate Caco-2 cells. The same volume of E. faecalis culture was also added to a medium mixture (90% DMEM containing 10% FBS; 10% M9B medium) without Caco-2 cells to determine the total number of bacterial cells in the inoculum. The E. faecalis and Caco-2 cells were then cocultured at 37°C with 5% CO₂ for 3h. After 3h of incubation, the culture medium was removed, and the infected Caco-2 cells were washed 3 times with PBS. All cells were then detached with 0.05% trypsin-EDTA for 20min at 37°C. Then, serial dilutions of these samples were plated on selective Todd Hewitt broth agar medium containing 30g/L Bacto Todd Hewitt Broth (Neogen Cor., United States), 15g/L agar and 50μg/ml rifampicin, and the adherent E. faecalis cells were counted.

An autoaggregation assay was performed as described by Baccouri et al. and Kaur et al. with some modifications (Kaur et al., 2018; Baccouri et al., 2019). Overnight cultures of E. faecalis OG1RF were diluted 1:100 in M9B medium with or without B. subtilis natto supernatant treatment (0, 10, and 50% v/v) and grown in culture tubes with screw caps and rubber liners (Kimble Inc., United States). All culture tubes were incubated anaerobically at 37°C for 24h. After 24h of cultivation, treated and untreated E. faecalis cultures were centrifuged, washed twice with PBS, and resuspended in PBS to give final OD₆₀₀ of 1. The E. faecalis suspensions were vortexed for 15s and then incubated at 37°C for 4h. After 0 and 4h of incubation without mixing, one milliliter of the suspension from the top of the tube was taken to measure the absorbance (A) at 600nm. Autoaggregation was then calculated as follows: autoaggregation (%)=[1−(A_4h/A₀)×100].

Overnight cultures of E. faecalis OG1RF were diluted 1:100 in a 24-well plate containing 1ml of medium and sterile cover glasses and anaerobically cultured at 37°C for 24 or 48h. The biomass that adhered to the cover glass was prefixed with 2.5% glutaraldehyde and 2.5% paraformaldehyde in 0.05M cacodylate at 25°C for 60min.

After prefixation, the samples were washed three times with 0.05M cacodylate and then postfixed with 1% osmium tetroxide in 0.05M cacodylate at 25°C for 60min. The fixed samples were chemically dehydrated using a graded ethanol series [30, 50, 70, 85, 90, 95, and 100% (two times)] and processed in a CO₂-based critical point dryer and ion coater. The dried samples were observed using an FEI Inspect S scanning electron microscope at a high voltage of 15kV and a magnification of 3,000×.

A ninety-six-well polystyrene plate-based bacterial biofilm formation assay was performed as described by Dale et al. with some modifications (Dale et al., 2015). In brief, overnight cultures of E. faecalis were diluted 1:100 in M9B medium containing 10% B. subtilis natto supernatant or not, and 100μl was dispensed into 8 wells per treatment. In addition, Oxyrase^® for Broth (Oxyrase Inc., United States) was added to the cultures to generate anaerobic conditions. The 96-well plates were then incubated in the chamber of a SpectraMax^® 190 microplate reader (Molecular Devices, LLC, United States) at 37°C for 24h. The OD₆₀₀ was measured every 2h to monitor cell growth. After 24h of cultivation, the culture medium was removed, and the biomass in the bottom of the 96-well plates was washed three times with double-distilled water (ddH₂O) and then air-dried for 2.5h. Next, the biomass was stained with 0.1% safranin solution for 20min, washed five times with ddH₂O, and air-dried. The safranin-stained biomass was quantified by measuring the OD₄₅₀ value. Biofilm formation is expressed as an index of the biomass stained with safranin (OD₄₅₀ value) normalized to the cell growth (OD₆₀₀ value at 24h). The relative biofilm biomass values were calculated by further normalizing the biofilm index values of the treated group to those of the negative control group to which no B. subtilis natto supernatant was added.

Overnight cultures of E. faecalis OG1RF::p23cfp were diluted 1:100 in M9B medium with or without B. subtilis natto supernatant treatment (0, 10, and 50% v/v) and cultured on glass coverslips in 35×12mm tissue culture dishes (Alpha Plus Scientific Co., Taiwan). All dishes were incubated anaerobically at 37°C for 24 or 48h. After 24 or 48h of cultivation, the culture medium in the tissue culture dishes was removed, and the biomass attached to the glass coverslips in the tissue culture dishes was washed twice with PBS to remove unattached cells and then fixed with 2% paraformaldehyde (PFA). Fixation with 2% PFA in PBS was performed at 4°C for 10min. After fixation, the biomass attached to the glass coverslip in the bottom of each petri dish was visualized using a white light laser confocal microscope Leica TCS SP8 X (Leica Microsystems, Ltd., Germany) and analyzed using Leica Application Suite X software.

The extraction and analysis of E. faecalis biofilm polysaccharides were conducted using the method described by Liu et al. (2020a, 2020b) with some modifications. Overnight cultures of E. faecalis were diluted 1:100 in M9B broth with or without B. subtilis natto supernatant (0, 10%, or 50% v/v). One milliliter of diluted E. faecalis culture was added to 24-well plates and incubated anaerobically at 37°C. After 24 or 48h of cultivation, the culture supernatant was removed, and the biomass that adhered onto the bottom of each well was washed with ddH₂O and then air-dried for 1h. Next, the adhered biomass was harvested by scraping the surface thoroughly with a sterile polyester-tipped swab after 1ml of PBS was added into each well. The cell suspensions of two wells corresponding to the same treatment were mixed together and centrifuged at 5,000×g for 30min at 4°C. The concentrated precipitates were resuspended in aqueous solution (2ml) containing 0.85% NaCl and 0.22% formaldehyde, and the E. faecalis biofilm polysaccharide was extracted at 80°C for 30min. The polysaccharide dissolved in the formaldehyde solution was recovered further via centrifugation at 15,000×g and 4°C for 30min. The polysaccharide concentrations were quantified using the phenol-sulfuric acid (PSA) method (Dubois et al., 1951). In brief, 100μl of polysaccharide solutions or standard (D-glucose solution) was mixed equally with 5% (w/w) phenol solution in microcentrifuge tubes. Immediately afterward, 1ml of concentrated sulfuric acid was added. The tubes were then incubated for 5min at room temperature, and 200μl of the reaction mixture was added to a 96-well plate. The absorbance was measured at 492nm using a Multiskan FC microplate photometer (Thermo Fisher Scientific Inc., United States).

Enterococcus faecalis cell envelope polysaccharides were prepared using the method described by Dale et al. (2015, 2017) with some modifications. In brief, overnight cultures of E. faecalis were diluted 1:100 in M9B broth without and with B. subtilis natto supernatant (0 and 10% v/v) and cultured anaerobically. After 24h of incubation, E. faecalis cells were collected by centrifugation, and pelleted cells were washed using sucrose solution [25% sucrose and 10mm Tris-HCl (pH 8)]. Cells were then resuspended in sucrose solution supplemented with lysozyme and mutanolysin and incubated overnight at 37°C with gentle agitation. Next, supernatant fractions were harvested by centrifugation, followed by treatment with RNase A, DNase to remove contaminating nucleic acids, and proteinase K to remove protein impurities. The remaining impurities in the supernatant fraction were further extracted with chloroform. The aqueous phase was transferred to a new tube, and polysaccharides were precipitated by the addition of ethanol to a final concentration of 75% and incubation at −80°C for 30min. Precipitated polysaccharides were washed with 75% ethanol and allowed to air dry.

The carbohydrate composition of E. faecalis cell envelope polysaccharides was analyzed following complete acid hydrolysis of the polysaccharides. Acid hydrolysis of E. faecalis cell envelope polysaccharides was carried out with 1.95N trifluoroacetic acid (TFA) at 80°C for 6h. The mixture was cooled, evaporated, and then resuspended in Milli-Q water. Monosaccharides were analyzed on an HPAEC system (Dionex BioLC) equipped with a gradient pump, a pulsed amperometric detector (PAD) using a gold working electrode, and an anion – exchange column (Carbopac PA −10, 4.6×250mm). The detection condition was an isocratic NaOH concentration of 18mm at ambient temperature. The flow rate was 1.0ml/min. Identification and quantification of the monosaccharides were carried out in comparison with established standards. Data were collected and integrated on a PRIME DAK system (HPLC Technology, Ltd., United Kingdom).

Overnight culture of E. faecalis OG1RF was diluted 1:100 in M9B medium with or without B. subtilis natto supernatant treatment and incubated anaerobically at 37°C for 24h. After 24h of cultivation, 600μl of bacterial culture was treated with 1,200μl of RNAprotect Bacteria Reagent (Qiagen Ltd., Germany) at room temperature for 5min. The cells were then collected by centrifugation for 10min at 4°C, flash-frozen in liquid nitrogen and stored at −80°C until RNA extraction. For RNA extraction, cells were lysed with lysozyme (30mg/ml) and mutanolysin (500U/ml) in Tris-EDTA (TE) buffer at 37°C for 10min (Bandyopadhyay et al., 2016; Manias and Dunny, 2018). Total RNA was then extracted using an RNeasy Mini Kit (Qiagen Ltd., Germany) according to the manufacturer’s instructions. Five micrograms of total RNA was subjected to DNase treatment with Turbo DNase (Ambion Co., United States) according to the manufacturer’s instructions. The RNA purity and concentration were measured using an ND-1000 spectrophotometer (NanoDrop Technologies, Inc., United States), and RNA integrity was validated using a Bioanalyzer 2,100 (Agilent Technologies, Inc., United States). Then, ribosomal RNA was removed using a RiboMinus^™ Transcriptome Isolation Kit (Invitrogen Co., United States). The cDNA library was constructed with purified mRNA with a SureSelect Strand Specific RNA Library Preparation Kit (Agilent Technologies, Inc., United States) according to the manufacturer’s instructions. RNA-seq using the Illumina NovaSeq 6,000 (paired end) platform was performed at Welgene Biotech (Taiwan).

The raw image data were converted to sequence data using bcl2fastq conversion software v2.20. After adaptor clipping and sequence quality trimming with Trimmomatic v0.36, the clean reads of the control and treated groups were mapped to the reference genome of the WT strain E. faecalis OG1RF (NC_017316.1) using HISAT2.

Differential expression analysis was performed using cuffdiff (cufflinks v2.2.1) with genome bias detection/correction and Welgene Biotech’s in-house pipeline. Genes with a value of p≤0.05 and an FC≥2 were considered significantly differentially expressed. Functional enrichment assays of the differentially expressed genes (DEGs) in each experiment were performed using clusterProfiler v3.6.

The RNA-seq data discussed in this study have been deposited in NCBI’s Gene Expression Omnibus database (Edgar et al., 2002; Barrett et al., 2013) and are accessible at GEO series accession number GSE184249.¹

Total RNA was prepared according to the methods above. Approximately 500ng of total RNA was used to synthesize cDNA with a Quant II Fast RT kit (BIOTOOLS Co., Ltd., Taiwan). Each reaction product was diluted 1:20 in ddH₂O, and 1μl was then used for quantitative real-time polymerase chain reaction (RT-qPCR) using SYBR^® Green Supermix (Bio-Rad Laboratories, Inc., United States) and a Bio-Rad CFX384 instrument (Bio-Rad Laboratories, Inc., United States). A total reaction volume of 10μl containing 2μl of gene-specific primer mixture at a concentration of 2μm was used in each well. Each reaction was performed in triplicate, and the threshold cycle (C_T) values were obtained. Relative quantification was performed using the 2^-ΔΔCT method (Livak and Schmittgen, 2001). Gene expression was normalized to that of the housekeeping gene gyrB, which is moderately expressed (Chatterjee et al., 2013). The sequences of the primers are listed in Table 2. Three biological replicates were performed to show repeatability.

Bacterial adhesion to and subsequent colonization of host tissues are key steps in the initial stage of biofilm formation (Baldassarri et al., 2001; Monte et al., 2014). To test our hypothesis that B. subtilis natto supernatant would interfere with E. faecalis biofilm formation, we first investigated whether the ability of E. faecalis to adhere onto the human intestinal epithelial cell line Caco-2 was inhibited by B. subtilis natto supernatant treatment. A certain number of E. faecalis cells were added to Caco-2 monolayer cultures in the presence or absence of B. subtilis natto supernatant. After 3h of cocultivation, the E. faecalis cells adhered to the Caco-2 cell monolayers were detached and quantified via plating of the entire cultures on selective medium. As shown in Figures 1A,C, the concentrations [colony-forming units (CFU)/ml] of adhered E. faecalis cells [both wild-type (WT) and cyan fluorescent protein (CFP)-expressing strains] decreased markedly due to B. subtilis natto supernatant treatment. The relative adhesion of E. faecalis OG1RF (WT) cells decreased from 1.55 to 1.02%; the relative adhesion of E. faecalis OG1RF::p23cfp cells decreased from 0.93 to 0.51% (Figures 1B,D). These results indicated that B. subtilis natto supernatant likely interferes with the ability of E. faecalis to adhere to host tissues, such as the human intestinal tract.

Previous studies (Kragh et al., 2016) have shown that bacterial cells tend to clump together in multicellular aggregates before they form biofilms. Herein, we investigated whether B. subtilis natto supernatant affects the formation of E. faecalis aggregates. Overnight cultures of E. faecalis were diluted in M9B medium in the presence or absence of the B. subtilis natto supernatant and incubated under anaerobic conditions. After 24h of cultivation without mixing, both 10 and 50% (v/v) B. subtilis natto supernatant-treated E. faecalis cultures remained turbid, whereas control (untreated) cultures settled at the bottom of the tube (Figure 2A). Treated and untreated E. faecalis cells were collected and used in 4-h autoaggregation assay. As shown in Figure 2B, percentage autoaggregation of both 10 and 50% (v/v) B. subtilis natto supernatant-treated E. faecalis was significantly lower than that of untreated E. faecalis (control group; 10%: p<0.05; 50%: p<0.01). In addition, the aggregates of E. faecalis cells formed on glass coverslips with or without B. subtilis natto supernatant were observed and analyzed using scanning electron microscopy (SEM). As shown in Figure 2C, E. faecalis cells in the control group formed dense aggregates after 24h of cultivation. However, the dense aggregation of E. faecalis cells was markedly reduced upon treatment with 10 and 50% v/v B. subtilis natto supernatant. Even though E. faecalis grew for 48h, no large cell clumps were observed in the B. subtilis natto supernatant-treated groups. Collectively, these results indicated that B. subtilis natto supernatant might inhibit E. faecalis autoaggregation, thus affecting biofilm initiation and development.

Ninety-six-well polystyrene plate-based biofilm assays devised by Dale et al. (2015) were performed here to investigate the effect of B. subtilis natto supernatant on E. faecalis biofilm growth over 24h. Overnight cultures of E. faecalis were diluted in M9B medium in the presence or absence of the B. subtilis natto supernatant (10% v/v) and incubated under anaerobic conditions for 24h. The optical density at 600nm (OD₆₀₀) values of the E. faecalis cultures were measured every 2h to monitor the growth status of E. faecalis within 24h. After 24h of cultivation, the biomass that adhered to the bottom of the 96-well plates was washed, dried, and stained with safranin. Biofilm production was expressed as an index of the biomass stained with safranin [optical density at 450nm (OD₄₅₀) value] normalized to the cell growth (OD₆₀₀ value at 24h). As shown in Figures 3A,B, the OD₆₀₀ values of treated E. faecalis cultures were higher than those of untreated E. faecalis cultures when E. faecalis grew to the exponential growth phase. These results indicated that B. subtilis natto supernatant might slightly promote rather than inhibit the growth of E. faecalis in 24h. However, the biofilm growth of E. faecalis at 24h was significantly inhibited by B. subtilis natto supernatant treatment. The relative biofilm production rates (%) of WT and CFP-expressing E. faecalis OG1RF strains decreased by 21.18 and 26.99%, respectively (Figure 3C). These results showed that B. subtilis natto supernatant could inhibit E. faecalis biofilm production without killing planktonic E. faecalis cells.

The difference between the B. subtilis natto supernatant-treated and untreated groups seemed to be more significant in SEM images (Figure 2C) than in the biofilm formation assay (Figure 3C). In SEM experiments, E. faecalis cells attached to glass coverslips that had no coating, whereas E. faecalis cells in the biofilm formation assay attached to the tissue culture-treated bottom of a 96 well polystyrene plate. We found that E. faecalis cells had relatively difficulty attaching to noncoating abiotic surfaces, such as glass coverslips. Furthermore, some E. faecalis cells or clumps that attached to glass coverslips may be removed due to multistep sample pretreatment before SEM observations. Therefore, we inferred that these factors might contribute to the deviation between the results of the SEM experiment and the 96-well plate-based biofilm formation assay (Figures 2C, 3C, respectively).

The 3D architecture of the biofilms of the CFP- expressing strain E. faecalis OG1RF::p23cfp that formed on the glass coverslips at the bottom of the tissue culture dishes with or without B. subtilis natto supernatant was visualized and analyzed using confocal laser scanning microscopy (CLSM). The results are shown in Figure 4. After 24h of cultivation, the E. faecalis in the control group formed dense and well-organized biofilms, whereas the E. faecalis in both B. subtilis natto supernatant-treated groups formed loose and even disorganized biofilms, especially those in the group with high-dose B. subtilis natto supernatant (50% v/v) treatment. In addition, the biofilms of both 10 and 50% (v/v) B. subtilis natto supernatant-treated E. faecalis were thinner (at almost 12 and 5μm, respectively) than the E. faecalis biofilm in the control group, which was 18μm thick. Although all the control and B. subtilis natto supernatant-treated biofilms grew thicker from 24 to 48h, the inhibitory effect of B. subtilis natto supernatant on E. faecalis biofilms still existed (approximate biofilm thickness: 25μm in the control group, 18μm in both treated groups). The CLSM results were consistent with the results shown by biofilm assays conducted in 96-well polystyrene plates. Taken together, these results indicated that B. subtilis natto supernatant interfered with E. faecalis biofilm growth and resulted in the formation of looser and thinner biofilms.

We found that E. faecalis cells can easily and effectively attach to the bottom of tissue culture-treated plates or dishes and form biofilms. Therefore, in the 96-well plate-based biofilm formation assay, we hypothesized that approximately the same number of E. faecalis cells attached to the tissue culture-treated surface in the presence or absence of B. subtilis natto supernatant. The results in Figure 3C show that E. faecalis biofilm production was inhibited by B. subtilis natto supernatant treatment. In addition, the E. faecalis biofilms in the B. subtilis natto supernatant-treated groups were thinner than those in the control (untreated) groups under CLSM (Figure 4). However, B. subtilis natto supernatant did not inhibit the growth of E. faecalis (Figures 3A,B). Based on these observations, we inferred that the secretion of extracellular polymeric substances (EPSs) in E. faecalis biofilms might be changed. When bacteria adhere to a solid surface, they continue to grow and secrete EPS, which comprises polysaccharides, proteins, nucleic acids, and fatty acids (Liu et al., 2020a,b). EPS plays a key role in forming the three-dimensional architecture of biofilms. As reported previously, EPS in most biofilms accounts for more than 90% of the dry mass (Flemming and Wingender, 2010). Furthermore, polysaccharides are the major components of EPS (Flemming and Wingender, 2010). Herein, we investigated whether the polysaccharide contents of E. faecalis biofilms were reduced. The quantification results in Figure 5 show that the polysaccharide contents of E. faecalis biofilms indeed decreased in both the 10 and 50% (v/v) B. subtilis natto supernatant-treated groups after 24h of cultivation. When E. faecalis grew for 24 to 48h, the biofilm exopolysaccharide contents in all the control and treated groups increased. However, the polysaccharide contents in both the 10 and 50% (v/v) B. subtilis natto supernatant-treated groups were significantly lower than those in the control groups. Thus, B. subtilis natto supernatant likely inhibited E. faecalis biofilm polysaccharide production.

Previous studies (Haussler et al., 2003; Kragh et al., 2016) have reported that greater cell surface “stickiness” may increase the tendency of bacterial cells to form aggregates and attach to a solid surface, thus increasing biofilm formation. The properties of the bacterial cell surface, including stickiness, may be associated with the composition and organization of the bacterial cell envelope, which comprises the inner cell membrane and the cell wall (Sengupta et al., 2013; Choi et al., 2015). Here, we investigated whether the carbohydrate composition of the E. faecalis cell envelope was impacted by B. subtilis natto supernatant treatment (10% v/v, 24h) using high-performance anion-exchange chromatography (HPAEC). As shown in Figure 6, several types of monosaccharides, including glucose, glucosamine, and galactosamine, were detected in the E. faecalis cell envelope. Glucose, which accounted for approximately 45mol % of all monosaccharides in the E. faecalis cell envelope, was the most abundant monosaccharide. Upon treatment with B. subtilis natto supernatant, the content of glucose increased by approximately 5mol %, whereas the content of other monosaccharides decreased. Restructuring of the monosaccharide composition in the cell envelope can likely result in changes in cell surface properties, thus affecting the biofilm formation ability of E. faecalis.

To further understand how E. faecalis responded to B. subtilis natto supernatant, we searched for DEGs between E. faecalis treated with or without B. subtilis natto supernatant using RNA-seq analysis. According to the RNA-seq results, approximately 95.94 and 97.98% of the clean reads from the treated and control (untreated) groups were mapped to the reference genome, respectively. Among the 2,657 genes detected by RNA-seq [2,658 genes present in E. faecalis OG1RF (Bourgogne et al., 2008)], 95 genes were identified as differentially expressed [value of p≤0.05 and fold change (FC)≥2] in the treated group (Figure 7 and Supplementary Table 1). Among these DEGs, 70 genes were identified as significantly downregulated, and 25 genes were found to be significantly upregulated. Furthermore, Gene Ontology (GO) annotation analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis were performed to identify DEGs at the biological function level. The GO annotation results showed that the genes associated with the ATP binding term were differentially expressed, and the KEGG pathway mapping results showed that some DEGs were involved in the peptidoglycan biosynthetic process (Supplementary Figures 1, 2). These results indicated that some genes encoding ATP binding protein or peptidoglycan biosynthetic genes might be differentially expressed due to B. subtilis natto supernatant treatment.

To verify the RNA-seq results, we selected a total of 13 DEGs, including some ATP binding protein-encoding genes and peptidoglycan biosynthetic genes, for further RT-qPCR analysis. As listed in Table 3, the up- or downregulation trends and relative expression of most selected DEGs, including three peptidoglycan biosynthetic genes [murD (OG1RF_RS03760), murN (OG1RF_RS08680), and uppS (OG1RF_RS09795)], bgsA (OG1RF_RS11220), and walK (OG1RF_RS05035), were consistent with the RNA-seq results. Furthermore, a correlation between the six downregulated DEGs listed above and E. faecalis cell wall homeostasis and biofilm formation has been reported previously (Dubrac et al., 2008; Vollmer et al., 2008; Theilacker et al., 2009; Watanabe et al., 2012; Bucher et al., 2015). Peptidoglycan, an essential component of the cell wall in almost all bacteria, plays a key role in maintaining cell shape and serves as a scaffold to anchor other cell envelope components, such as proteins and teichoic acids (Vollmer et al., 2008). Gene bgsA (OG1RF_RS11220) encodes the putative glucosyltransferase designated biofilm-associated glycolipid synthesis A and synthesizes diglucosyl–diacylglycerol (DGlcDAG) in E. faecalis (Theilacker et al., 2009). DGlcDAG, a precursor of glycolipid and lipoteichoic acid, is involved in E. faecalis biofilm production, adherence to host cells and virulence in vivo (Theilacker et al., 2009). Gene walK (OG1RF_RS05035) encodes the membrane-linked histidine kinase WalK. WalK is a key element in the WalK/WalR two-component signal transduction system, which regulates genes involved in cell wall metabolism, biofilm production, virulence regulation, oxidative stress resistance, and antibiotic resistance in low-G+C Gram-positive bacteria, including E. faecalis (Dubrac et al., 2008; Watanabe et al., 2012). Based on these findings, we focused on the six downregulated cell wall- and biofilm-related genes in E. faecalis in further investigations.

To investigate whether B. subtilis natto supernatant inhibits the expression of other genes in the peptidoglycan biosynthesis pathway and WalK/WalR regulatory system, the relative expression of UDP-GlcNAc enolpyruvyl transferase (murA; OG1RF_RS10140), murB (OG1RF_RS10725), murF (OG1RF_RS03015), and murG (OG1RF_RS03765) in the peptidoglycan biosynthesis pathway and the response regulator-encoding gene walR (OG1RF_RS05030) in the WalK/WalR two-component system were assessed further using RT-qPCR. As shown in Figure 8, the relative expression of several peptidoglycan biosynthetic genes except murF was indeed inhibited due to B. subtilis natto supernatant treatment. The relative expression of walR in the WalK/WalR two-component system was also inhibited. Taken together with the RT-qPCR results of the six cell wall- and biofilm-related genes identified from RNA-seq, these results indicate that B. subtilis natto supernatant might interfere with the E. faecalis peptidoglycan biosynthesis pathway, WalK/WalR two-component system and bgsA, thus inhibiting biofilm formation.