Strains of S. mutans (ATCC 25175) and L. salivarius (ATCC 11741) were sourced from the China General Microbiological Culture Collection Center (CGMCC; Beijing, China). L. salivarius was grown in de Man–Rogosa–Sharpe (MRS) broth and S. mutans was cultured with brain–heart infusion (BHI) broth at 37°C under aerobic and microaerophilic conditions, with all strains stored in the broth containing 30% glycerol at −80°C routinely and were subsequently incubated on appropriate agar plates for 24 h, followed by 2% (v/v) inoculation in the relevant broth at 37°C for an additional 18 h before experimental use.

The CFS of L. salivarius was manufactured following a modified protocol established by Liang (Liang et al., 2023). In summary, L. salivarius was regulated to a concentration of 1×10⁷ CFU/mL during the late logarithmic growth phase and subsequently incubated for 24 h at 37°C. After the bacterial culture, the spent culture underwent centrifugation (5,000×g, 10 min, 4°C), and the obtained supernatant was further filtered with a 0.22-µm filter to acquire the CFS.

An overnight culture of S. mutans was diluted to a predetermined final concentration of 1.0×10⁶ CFU/mL in BHI broth enriched with 1% sucrose. This diluted culture was then dispensed into a 96-well microtiter plate at a volume of 200 μL, either with or without the increase of CFS, and incubated for 24 h. The final CFS concentrations, both treated and untreated, ranged from 12.5% to 100% (v/v). For different time treatment, a certain amount of CFS was added into the 96-well microtiter plate at 0, 6, and 12 h, respectively, at 37°C for 24 h. The method of mediating the biofilm by CFS at 24 h was as follows: 200 μL of suspension of S. mutans was added to each well of a 96-well plate and then cultured for 24 h, the biofilm was washed with phosphate-buffered saline (PBS) two times, and then 100 μL of CFS was added and the culture was continued for another 24 h. A crystal violet staining assay was conducted to evaluate how effectively the CFS inhibited biofilm formation by S. mutans (Wasfi et al., 2018). Specifically, after removing the culture supernatant, the biofilm was washed with PBS and fixed in methanol for 30 min. Afterward, the wells were stained with 0.1% crystal violet solution for 30 min, and then dissolved in 33% glacial acetic acid over 30 min until fully solubilized. Finally, optical densities were recorded at 575 nm using a microplate reader (the amount of biofilm formation). As a negative control, CFS was replaced by MRS broth.

An overnight culture of S. mutans was maintained in BHI broth and subsequently diluted to 1.0×10⁷ CFU/mL (with 1% sucrose). A sterile cover slide was placed into the wells of a 24-well plate. In each well, 800 μL of the S. mutans suspension was combined with 160 μL of CFS or MRS broth and incubated under anaerobic conditions for 24 h. The cover slides were carefully rinsed three times using PBS, fixed, and prepared for scanning electron microscopy (SEM) observation (Hitachi, SU8100) following an established protocol (Liu et al., 2021).

Suspensions (1.75 mL) and 350 μL of CFS were added into the confocal dishes and cultured at 37°C for 24 h. The biofilm was subjected to staining for 30 min and subsequently observed using a confocal laser scanning microscope (CLSM) according to the BBcell Probe™ live/dead bacterial staining kit. In the control conditions, MRS broth served as a substitute for the CFS.

S. mutans were cultured statically in six-well polystyrene plates for 24 h, allowing loosely attached bacterial cells to be gently rinsed off and then the biofilms were scraped off and approximately 50 mg of biofilm mass was collected by centrifugation and frozen at −80°C until used. Library construction, sequencing, and analysis services were provided by GENEWIZ Life Sciences (Suzhou, China). Each group for transcriptomics had three replicates. The data presented in the study are deposited in the NCBI repository, accession number PRJNA1219341.

The biofilm formation assay adopted the same preparation as before. Following incubation, the culture suspension from the wells was discarded. The plate wells underwent two washes with sterile saline, after which biofilm was scraped and suspended in saline for transfer to a centrifuge tube. The RNAprep pure Cell/Bacteria Kit (Tiangen Biotech; Beijing; China) was used for total RNA extraction according to the instructions of this kit. RNA concentration and purity were assessed using the ND-1000 spectrophotometer. Primer [Sangon Biotech Company (Shanghai, China)] sequences are listed in Table 1 . The RNA was reversely transcribed into cDNA with the NovoScript Plus All-in-one 1st Strand cDNA Synthesis SuperMix (gDNA Purge) kit (Novoprotein, E047, Suzhou, China), subsequently utilizing the NovoStart SYBR qPCR SuperMix Plus reagent (Novoprotein, E096, Suzhou, China). The internal reference was 16S rRNA. The gene transcription level was evaluated using the 2^−ΔΔCT methodology.

Each experiment was performed three times, and all results were expressed as mean ± standard deviation. One-way analysis of variance (ANOVA) was used for statistical analysis along with Dunnett’s test through GraphPad Prism version 8.0.1. Subsequently, all pairs of mean comparisons were assessed using the post-hoc Tukey method. p < 0.05 was considered statistically significant, while p < 0.01 was considered highly statistically significant.

As illustrated in Figure 2A , all CFSs at varying concentrations demonstrated efficacy in suppressing the biofilm formation of S. mutans. Specifically, CFS could inhibit 99% of S. mutans biofilm formation when used at concentrations of 50% (v/v). When the CFS was diluted to the concentration of 25%, the effect of biofilm formation decreased to 93.4% (p < 0.05). As the CFS was diluted to 12.5%, the inhibition decreased to 28.6% (p < 0.01).

The biofilm formation process included four essential time points: 0 h, where bacteria initially adhered; 6 h, marking the initial colonization of bacteria; 12 h, indicating early biofilm development; and 24 h, reflecting mature biofilm formation. As Figure 2B shows, the amount of biofilm mediated by each time point decreased significantly (p < 0. 01); the 24-h-mediated biofilm showed a certain reduction in biofilm volume compared with the 24-h control biofilm (p < 0.01). The results showed that the CFS of L. salivarius had a strong inhibitory effect in the early stage of biofilm formation, but had a slightly destructive effect on middle and mature stage of biofilm formation.

At a dilution of 16.7% for the CFS, the SEM micrograph presented in Figure 2C demonstrated that a compact biofilm typical for S. mutans displayed a network-like composition identified as EPS. However, the biofilm architecture of S. mutans added with CFS appeared significantly more dispersed, exhibiting fewer micro-colonies on the surface compared to that of S. mutans alone, with a decrease in the quantity of EPS. Biofilms were created in the presence of CFS following 24 h of cultivation and analyzed by a confocal laser scanning microscope ( Figure 2D ). The images demonstrate fluorescence intensities in green (live bacteria) and red (dead bacteria). The control group revealed that the biofilm was distributed uniformly, accompanied by a comparatively dense structure and total surface coverage. As shown in the CFS treatment group, the biofilms seemed significantly more dispersed and notably looser. The surface area that the biofilm occupied decreased because of CFS, leading to a marked reduction of biofilm biomass.

Literature indicated that the antimicrobial components found in the CFS from Lactobacillus consist of organic acids, hydrogen peroxide, and bacteriocin (Vahedi Shahandashti et al., 2016; Liu et al., 2020). Bacteriocin is a protein synthesized by ribosomes, a metabolite secreted by bacteria during their reproductive process, possessing antibacterial properties (Kumariya et al., 2019). To preliminarily investigate the characteristics of the primary antibacterial agent, three portions of CFS were exposed to catalase and proteinase K, and pH was adjusted to neutral levels. After being treated with catalase and proteinase K, S. mutans showed a reduced but not lost antibiofilm effect. These findings suggested that hydrogen peroxide and a protein-like substance could serve as the principal antibiofilm agent in the CFS of L. salivarius. Additionally, this research revealed that the pH of the obtained CFS measured at 3.9. After adjusting the pH to 6.5, the CFS lost its antibiofilm capability; meanwhile, when the pH returned to the original value, the antibiofilm effect of CFS had not been fully restored ( Figure 3A ). As a result, we conducted further examinations to identify the main antibiofilm constituents of CFS. After the CFS was treated at different temperatures, there were no notable changes in its ability to inhibit biofilm formation. Even after 30-min treatment at 100°C, it still had the effect of inhibiting biofilm, and interestingly, it was slightly increased compared to that without treatment after 80°C ( Figure 3B ). This indicates that the active components of the CFS were thermally stable.

To analyze the effective components of CFS that enforce antibacterial properties, the extracellular metabolite profiles were compared between MRS and CFS through metabolomic analysis following a 24-h incubation period. The principal component analysis (PCA) and PLS-DA models highlighted significant metabolite differences between the two sample groups, all within the 95% confidence interval ( Figures 3C, D ). In the presence of the CFS, 436 metabolites were upregulated, while 694 were downregulated (p < 0.05, VIP > 1) compared with those in the presence of MRS broth ( Figure 3E ). There was a significant difference between the fermentation broths before and after cultivation, indicating that effective substances that may inhibit biofilm formation have been produced during the fermentation process. The VIP plots ( Figure 3F ) indicated that certain identified metabolites contributed to class differentiation. KEGG compound analysis of the upregulated metabolites was conducted, mainly including fatty acids, amino acids, nucleotides, and carboxylic acids ( Figure 3G ). The upregulated metabolites predominantly included various compounds including organic acids and derivatives, lipids and lipid-like substances, organoheterocyclic entities, organic oxygen compounds, and benzenoids were produced, as indicated by the HMDB compound classification analysis ( Figure 3H ). This section explored the characteristics of the effective components in CFS, which were thermally stable and possible protein and hydrogen peroxide-like substances and have a relatively narrow pH tolerance range, possibly being a type of organic acid or a substance that functions under acidic conditions. Furthermore, by combining with the non-targeted metabolomics method based on LC-MS to identify the differential metabolites between the CFS and the initial culture medium, we found that they were mainly fatty acids, amino acids, nucleotides, and carboxylic acids.

Given that CFS led to notable inhibition of biofilm formation by S. mutans, we conducted transcriptome analysis on both S. mutans treated with CFS and those under control conditions to monitor genome-wide gene expression alterations caused by CFS.

PCA showed that all biological replicates clustered closely, suggesting that gene expression underwent significant alteration due to CFS treatment compared to untreated controls, accounting for the 81.7% variance observed in the overall dataset ( Figure 4A ). The differentially expressed genes (DEGs) were identified using a modified p-value threshold of < 0.01 observed through the control and CFS groups. As depicted in the volcano plot and heatmap in Figure 4B , 487 genes were found to be upregulated, while 519 exhibited downregulation. DEGs were annotated, functionally utilizing KEGG enrichment analysis. The GO enrichment assessment concerning molecular functions, cellular components, and biological processes revealed the involvement of these DEGs in DNA binding, oxidoreductase activity, membrane, de novo IMP biosynthetic process, and carbohydrate transport ( Figure 4C ). The functional annotation of DEGs was carried out along with KEGG pathway and enrichment analysis ( Figures 4D, E ). Furthermore, 30 significant KEGG pathways were revealed, demonstrating potential links to the metabolic processes of S. mutans, including pathways related to pyruvate metabolism, purine metabolism, amino sugar and nucleotide sugar metabolism, galactose metabolism, and the TCA cycle. Certain pathways, especially pyruvate metabolism, the phosphotransferase system, and ATP-binding cassette (ABC) transporters, play vital roles in biofilm formation.

S. mutans possesses 14 phosphoenolpyruvate-dependent phosphotransferase systems (PTSs) and two ABC transporters, indicating that it can transport various sugars to meet its metabolic needs and adapt to complex and changing environments. Transcriptome analysis indicated a significant reduction in the key gene expression associated with carbohydrate metabolism because of CFS ( Tables 2 , 3 ). Selected genes related to two-component signal transduction systems (ciaH, ciaR, lytS, and lytR), polyketides/non-ribosomal peptides (mub gene clust), acid stress response (aguA and TreR), quorum sensing gene (LuxS), and exopolysaccharide-formation gene (dexA) were significantly downregulated and are presented in Table 4 , which could inhibit oxidative stress and attenuate the virulence of S. mutans.

To confirm the RNA-Seq findings, randomly selected genes underwent quantification through quantitative real-time polymerase chain reaction (qRT-PCR) to assess their transcription levels. As indicated in Figure 4F , the genes lrgB, LuxS, dexA, and ldh showed significant downregulation following treatment with the supernatant, which conformed with the results derived from RNA-seq analysis.

Throughout the entire experimental duration, the rats maintained stable health conditions. Weight gain among all groups showed no statistically significant differences ( Table 5 ). As shown in Figure 1B , in the caries-free group, S. mutans was not detected, while the levels of S. mutans in the caries model, L. salivarius, and CFS groups were approximately 4.0 × 10⁴ CFU/mL after infection for 5 days, demonstrating successful colonization of S. mutans within the oral cavities.

In order to improve the visibility of caries site on rat molars, micro-CT was used for 3D reconstructions of the mandibular molars, and the enamel was isolated from the complete mandible with a predetermined threshold. Additionally, the relevant sagittal slice of the homorganic molar was extracted for comparative analysis ( Figure 1C ). From the complete 3D reconstruction of dental hard tissue, the sagittal slice images of the caries model group were compared with those of the caries-free group, in which it was clear that the enamel (green) areas were discontinuous in the presence of caries. Additionally, to quantitatively assess the results from micro-CT, we calculated and analyzed the enamel volume and mineral density of molar teeth across the various experimental groups. The smaller the enamel volume, the more enamel loss and the more severe the caries. As shown in Figures 1D, E , the enamel volume and mineral density of the caries model group were obviously less than those in the caries-free group, indicating that the rat caries model was successfully established. The enamel volume and mineral density of the molars treated with CFS were higher than those in the caries-model group and lower than those in the caries-free group (p < 0.01). In addition, no significant differences were found in enamel volume and mineral density between the CFS and L. salivarius groups.