The microorganisms used in this study included C. albicans SC5314, S. mutans 593, and liaS^–, which were obtained from the School of Stomatology, Fujian Medical University (Fuzhou, China). The liaS^– strain and its complemented strain (liaS^–-comp) were derived from S. mutans 593, using the construction methods and cultivation conditions described in previous studies (He et al., 2020; Alomeir et al., 2023). The dual-species coculture medium consisted of TYE broth containing 25 g/L tryptone (Oxoid, UK), 15 g/L yeast extract (Oxoid, UK), and 1% (w/v) sucrose (He et al., 2020). All the cultures were maintained at 37°C under 5% CO₂.

For biofilm formation, pretreated C. albicans (2 × 10⁴ CFU/ml) and S. mutans (2 × 10⁶ CFU/ml, including liaS^– and liaS^–-comp) were mixed at a 1:1 (v/v) ratio in TYE medium. The monospecies controls were grown under an equivalent system with 50% diluted medium (Rocha et al., 2018). The inoculated microplates were incubated for 24 h (37°C, 5% CO₂), and the biofilm coverage was dynamically monitored using a cell analysis system (Zen CELL Owl, Leica Microsystems).

A 24 h biofilm was constructed in a 96-well plate, and the crystal violet (CV) staining method was used to assess biofilm formation. After the supernatant was removed and the plates were washed with PBS, the biofilms were fixed in methanol for 15 min and stained with 0.1% (w/v) CV for 10 min. Following washing and drying, the biofilms were photographed using an optical microscope (ZEISS Stemi 508). The biofilms were then dissolved in 33.33% (v/v) acetic acid, and 100 μl of the solution was transferred to a new well plate for quantification on a microplate reader at OD575 nm (SpectraMax iD3). Biofilm samples were collected using sterile surgical blades, dissolved in PBS, and sonicated for 5 min to disperse the aggregates. Single-species and cocultured biofilm gradients were diluted and plated on sabouraud dextrose agar (SDA, Solarbio, China) and brain heart infusion (BHI, OXIOD, UK) agar to quantify C. albicans and S. mutans, respectively. After 48 h of incubation at 37°C, the biomass was evaluated. Each test was performed in triplicate, three times independently, and the average log₁₀ CFU was used to determine the fungal and bacterial biomass (Alomeir et al., 2023).

To evaluate the impact of the liaS gene on the acid tolerance response (ART) of S. mutans, modified protocols from previous studies were employed (Turner et al., 2020; Du et al., 2023). In brief, the overnight-cultured microorganisms of C. albicans and S. mutans were mixed at the aforementioned concentrations and ratios to prepare single-species and dual-species planktonic cultures. The test samples were exposed to acidic TYE culture medium (initial pH 5.5) adjusted with 0.1 M HCl for 24 h, whereas the control samples were maintained in a neutral environment (pH 7.0). The single-species and cocultured mixtures were inoculated into sterile centrifuge tubes and incubated at 37°C in a shaking incubator (150 rpm) to maintain a planktonic cell suspension. The pH of the culture medium was not artificially adjusted during the experiment to allow for natural acidification via microbial metabolism. Biomass quantification was performed by serial dilution plating on SDA and BHI agar. After 48 h of incubation at 37°C, the viable colonies (30–300 CFU/plate) were counted (Jafri et al., 2020; Guo et al., 2023). The results from three replicates are expressed as the log₁₀ CFU/ml.

The acid production capacity was assessed using a pH drop assay (Li et al., 2022; Lin et al., 2022a). The pH of the biofilm supernatants was measured using an Orion Dual Star pH/ISE benchtop meter (Thermo Scientific), which reflects the acid production from the glycolytic metabolism of carbohydrates by single and dual-species.

A multiparametric analysis of biofilms was conducted using confocal laser scanning microscopy (CLSM, Leica DM IRE2, Germany). The biofilms were developed on sterile glass slides (r = 0.6 cm) under anaerobic conditions. During initial cultivation, Alexa Fluor^® 647 (1.0 μM, ex/em: 650/668 nm; Molecular ProbesTM, Invitrogen) was added to label the EPS. After 24 h of incubation at 37°C, the planktonic cells were removed using PBS. Viable cells in biofilms were subsequently stained with SYTO^® 9 (2.5 μM, ex/em: 480/500 nm; Molecular ProbesTM, Invitrogen) for 15 min. Z-axis imaging was performed using a 40× oil immersion objective, with random field selection ensured through systematic sampling protocols. Biofilm thickness quantification was achieved using IMARIS V9.0 software (Bitplane AG, Zurich, Switzerland) that had built-in algorithms (Krzyściak et al., 2017; Liu et al., 2018; Sun et al., 2024).

The antioxidant stress capacity mediated by liaS in single and dual-species biofilms was assessed using time-H₂O₂ killing curves (Yadav et al., 2014; Graziano et al., 2015). Mature 24 h biofilms were washed twice with PBS and treated with 2× minimum inhibitory concentration (MIC). The MIC values are provided in Supplementary Table 2. After incubation at 37°C for 0, 15, 30, 45, or 60 min, the reactions were terminated using PBS containing 1.0 mg/ml catalase (Graziano et al., 2015). The biofilms were harvested for CFU counting as previously described. The survival rates of C. albicans and S. mutans were calculated to compare oxidative stress resistance across biofilm systems.

Intracellular ROS levels were quantified using 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA). The biofilms were loaded with 10 μM DCFH-DA probe and incubated at 37°C in the dark for 60 min. Cellular esterases convert the non-fluorescent probe to DCFH, which subsequently reacts with ROS to generate fluorescent 2′,7′-dichlorofluorescein (DCF; ex/em: 485/535 nm). The fluorescence intensity (FI) was measured using a microplate reader (SpectraMax i3x, Molecular Devices). Data were normalized to biofilm biomass (CFU counts) and expressed as FI/10³ cells to indicate ROS production per unit biomass.

Ribonucleic acid (RNA) was extracted from S. mutans and C. albicans in dual-species biofilms using TRIzolTM reagent (Invitrogen, USA) according to the manufacturer’s protocol. Next, 1.0 μg of RNA was reverse-transcribed with a PrimeScriptTM RT reagent kit (Takara Biotechnology, Japan). The primers (Sangon Biotech, Shanghai, China) used in this study are shown in Supplementary Table 1. RT-PCR was performed on cDNA using a StepOnePlusTM Real-Time PCR System (Applied Biosystems, USA) with TB GreenTM Premix Ex TaqTM II (Takara Biotechnology, Japan). The 16S rRNA and 18S rRNA were used as housekeeping genes, and gene expression was quantified via the 2^–ΔΔCt method (Nailis et al., 2006).

Cyclical liquid nitrogen freeze-thaw cycles were performed to increase the cAMP extraction efficiency. Following lyophilization and dry weight measurement, the samples were reconstituted in 200 μl cold PBS and subjected to 10 min of ultrasonication for cellular lysis. After centrifugation (5000 × g, 10 min, 4°C), the cAMP concentrations in the supernatants were quantified via ELISA according to the manufacturer’s protocols (Ruixin Biotechnology, China). The absorbance at 450 nm was measured with a microplate reader, and the biofilm cAMP content was normalized to the sample biomass using a standard curve-derived calibration.

To investigate the effects of cAMP signaling and the liaS gene on early fungal adhesion, dual-species biofilms of C. albicans and liaS^– were established through gradient supplementation of exogenous dibutyryl-cAMP (db-cAMP, Macklin Biochemical Co., Ltd., Shanghai, China) for functional complementation (Bernardoni et al., 2024). The fungal early adhesion model was developed under anaerobic conditions (37°C, 80 rpm, 8 h), with adherent fungal biomass quantified via CFU counting (Nabert-Georgi et al., 2018). Fungal morphology was assessed using scanning electron microscopy (SEM). Circular sterile glass slides (radius = 0.6 cm) served as carries for the biofilm, which was cultured in a 24-well plate. After rinsing thrice with PBS, the biofilm was fixed with a 2.5% glutaraldehyde solution (J&K Scientific, Co., Ltd., China). It was then dehydrated in a graded ethanol series, dried, and sputter-coated with gold-palladium. SEM images were randomly captured at 2,000× magnification using an EM8000 microscope (KYKY, China).

Thirty SPF-grade male Sprague–Dawley rats (3 weeks old, 50 ± 5 g) were randomly allocated to five groups (n = 6): WT-C. a, WT-S. m 593, liaS^–, WT-C. albicans + WT-S. m 593, and liaS^– + WT-C. a infection. Three days before infection, the rats received sterile water containing 0.1% (w/v) ampicillin, 0.1% (w/v) streptomycin, and 0.1% (w/w) carbenicillin to deplete the oral microbiota. Starting on day 5, the oral cavities were inoculated with 10⁸ CFU/ml S. mutans or 10⁶ CFU/ml C. albicans suspensions by swabbing for three consecutive days. Throughout the 21-days experiment, the animals were maintained on 5% sucrose water to promote cariogenesis, and their body weights were monitored. After euthanasia by cervical dislocation, the mandibles were aseptically dissected and homogenized in PBS using an ultrasonic processor. Serial dilutions were plated on mitis salivarius-bacitracin agar (MSBA, Difco, Biosciences) supplemented with 200 U/ml bacitracin and 1% potassium tellurite, followed by 48 h of anaerobic incubation at 37°C for S. mutans quantification. Caries progression was evaluated using the Keyes scoring system (Jafri et al., 2020; Chen et al., 2024). The rat teeth were stained with 0.4% ammonium purpurate for 12 h and graded under a stereomicroscope (ZEISS, Germany) for enamel (E), superficial dentin (Ds), and deep dentin (Dd) lesions.

The animal experiments were performed in strict accordance with the guidelines of the Animal Welfare Act by the ethics committee of Fujian Medical University and meet the ethical requirements (license number IACUC FJMU 2024-0378). This experiment met humane standards and minimizes the subject’s pain.

All the experiments were conducted in triplicate individually. Significant effects of the variables were determined by one-way ANOVA, followed by Tukey’s multiple comparison tests. Differences were considered statistically significant if P < 0.05. Statistical analysis was performed using SPSS 16.0 software (SPSS Inc., Chicago, IL, USA).

Biofilm quantification revealed that the liaS gene promotes biofilm formation in both single and dual species films (Figure 1A). Compared with those of the WT and liaS^–-comp strains, the percentages of liaS^– single-species decreased by 16.74% and 10.04%, respectively (P < 0.05). In the dual-species system, the combination of liaS-deficient samples showed significant reductions of 16.42% and 8.04% compared to the combination of WT-S. m and liaS^–-comp, respectively (P < 0.05). Biofilm formation was restored by supplementation with the liaS gene. There was no significant difference in fungal biomass between the single and dual-species biofilms of C. albicans. The bacterial biomasses in the single and dual-species biofilms of the liaS-deficient strain were 7.35 and 7.79 log₁₀ CFU/ml, respectively, which was significantly lower than those of the WT and liaS^–-comp strains (P < 0.05; Figure 1B).

Planktonic cultures demonstrated that C. albicans maintained stable biomass under both neutral (pH 7.0) and acidic (pH 5.5) conditions (Figures 2A, B). In contrast, S. mutans exhibited pH-specific growth patterns: while the WT and liaS^– strains showed comparable proliferation at neutral pH (6.10 ± 0.08 vs. 5.91 ± 0.25 log₁₀ CFU/mL; P > 0.05), acid stress (pH 5.5) significantly attenuated liaS^– growth in both mono- and dual-species cultures compared with the WT (P < 0.05). Additionally, in the 24-h planktonic culture system of the liaS gene knockout strain, the biomass was significantly lower than that of the WT-S. m and liaS^–-comp strains (P < 0.05). Metabolic profiling revealed impaired acidogenesis in liaS^– cultures: after 24 h of cultivation, liaS^– mono- and dual-species supernatants presented an attenuated pH reduction (10.07% and 8.00% less acidification, respectively) compared with the WT strain (Figure 2C). This metabolic deficit was rescued in the complemented strain (liaS^–-comp), and there was no significant difference in the pH value of the supernatant culture medium between single-species and dual-species biofilms.

The 3D reconstruction of biofilms and EPS distribution revealed significant structural alterations in liaS^–. The single-species liaS^– biofilms presented a 19.3% reduction in the average thickness (23.79 μm), with complementation restoring the thickness by 11.9% (25.97 μm) compared with that of WT-S. m (Figures 3B–D). In dual-species cultures, liaS^– biofilms were 41.1% thinner (21.96 μm) than WT biofilms were (P < 0.05), whereas liaS^–-comp biofilms presented minimal disparity (Figures 3E–G). Confocal imaging revealed that C. albicans and S. mutans (SYTO 9-green) formed dense aggregates with abundant EPS (red fluorescence) along fungal surfaces, whereas liaS^– displayed sparse bacterial clustering with reduced aggregation density and less EPS production (Figures 3C, F). Quantitative analysis of red fluorescence within the biofilm revealed that the liaS gene promotes EPS production (Figure 3I). Notably, the combination of liaS^– and C. albicans inhibited fungal hyphae dominance, suggesting an altered fungal-bacterial interplay. The complementation strain restored spatial organization and EPS synthesis patterns.

Biofilms exposed to 2× MIC H₂O₂ presented strain-dependent survival kinetics (0–60 min). After being exposed to H₂O₂ for 60 min, C. albicans demonstrated superior resistance in both single-species (10.32% reduction) and dual-species (16.17%–17.09% reduction) strains (Figures 4A, B). Compared with those of the WT and liaS^–-comp (P < 0.05) (Figure 4A), with exacerbation under dual-species conditions (21.39%–56.52% reduction) (Figure 4C). The liaS^– culture accumulated significantly higher intracellular ROS than WT-S. m and liaS^–-comp both in single and dual-species biofilms (P < 0.05). Cocultivation with C. albicans reduced the level of endogenous ROS by 46.54%, in contrast to the results in single-species cultures (Figure 4D). liaS displayed cross-kingdom transcriptional regulation in dual-species biofilm, accompanied by 63.85% downregulation of sodA (P < 0.05) and an 18.17% upregulation of perR (P < 0.05) in S. mutans. In C. albicans, 27.72% of the sod1 genes were upregulated (P < 0.05), 10.38% of the trr1 genes were downregulated (P > 0.05), and 101.83% of the cat1 genes were upregulated in C. albicans. This difference in gene expression regulation was reversed in liaS^–-comp (Figure 4E).

Quantitative cAMP analysis revealed that liaS had a regulatory effect on metabolites in the biofilms (Figure 5A). In S. mutans 593 single-species culture, the cAMP content (8.16 ng/ml) was 1.66-fold greater than in liaS^– (4.92 ng/ml) (P < 0.05). Compared with the single-species mixture, the liaS^– dual-species mixture did not present with significant differences in cAMP contents compared to the single species culture (P < 0.05). Early fungal adhesion in liaS^– and C. albicans biofilms depends on the dosage of db-cAMP. Compared with the combination of S. mutans 593 and C. albicans, exogenous supplementation with 1.56 mg/ml db-cAMP restored early adhesion, but the difference was not significant (Figure 5B). Notably, 1.56 mg/ml db-cAMP reversed these perturbations, significantly rescuing the early contribution of fungi in the liaS^– and C. albicans combination to the formation of complex biofilm structures (Figure 5C).

Regulatory profiling of the Ras1-cAMP-Efg1 signaling pathway in the liaS-deficient strain in dual-species biofilms revealed significant alterations in the expression of key genes governing the Ras1-cAMP-Efg1 cascade in C. albicans (Figure 5D). Specifically, the upstream regulator ras1 was downregulated by 10.4% compared with that of the WT (P < 0.05). The adenylate cyclase-encoding gene cry1 presented a 67.7% reduction in expression. The cAMP-dependent protein kinase (PKA) catalytic subunit genes tpk1 and tpk2 were downregulated by 34.6% and 35.8%, respectively. The core hyphal development transcription factor efg1 and the biofilm regulatory gene ume6 presented 68.5% and 27.9% reductions in transcript levels, respectively. Exogenous supplementation with 1.56 mg/ml db-cAMP restored the expression of this pathway in the liaS-deficient dual-species system, with statistically significant upregulation observed across pathway components (P < 0.05).

Compared with the WT-S. m, liaS^– exhibited significant transcriptional downregulation of virulence regulators (vicR: −26.93%; vicK: −49.64%) and (comD: −37.48%; comE: −47.66%). All expression deficits showed partial or full recovery after liaS complementation (Figure 6A).

Quantitative analysis using the Keyes scoring system in Sprague–Dawley rats demonstrated that liaS is a critical regulator of S. mutans-mediated cariogenesis. The single-species and dual-species colonization of S. mutans loads were 6.045, 4.085, 8.965, and 5.278 log₁₀ CFU/ml, respectively (Figure 7A). Keyes scores, which evaluate lesion depth and spread degree, revealed stark contrasts in caries severity (Figure 7B). The WT-S. m 593 infections induced dominant Dx lesions, with dual-species infections resulting in accelerated caries progression in the buccal surfaces, fissures, and proximal tooth regions (Figures 7C–E). In contrast, liaS^– infections significantly reduced the incidence of Dx lesions (P < 0.05) across both single-species and dual-species infections. Notably, the liaS^– and C. albicans combined infection displayed caries patterns that were phenotypically aligned with those of C. albicans single infection, which demonstrated minimal cariogenic activity.