The peptides were synthesized by Ji’er Biochemical (Shanghai, China) with a purity greater than 95%. Before use, the peptides were dissolved in sterile deionized water at the concentration of 2 mM as stock solution, and diluted with phosphate buffered saline (PBS) to work concentration. Streptococcus mutans UA159 was purchased from Guangdong Microbial Culture Collection Center (Guangzhou, China) and cultured anaerobically in brain heart infusion broth (BHI; Thermo Fisher, CN) at 37°C under mixed gas conditions of 5% CO₂, 10% H₂, and 85% N₂ to logarithmic growth phase before experiments. The colony forming units (CFUs) and the optical density at 600 nm (OD₆₀₀) of the bacteria were plotted. OD₆₀₀ = 0.1 is equal to the density of S. mutans at 2 × 10⁶ CFU/mL in BHI.

Streptococcus sanguis ATCC 10556 and Porphyromonas gingivalis ATCC 33277 were cultured anaerobically in BHI. Bifidobacterium adolescentis ATCC 15703 and B. breve ATCC 15700 were cultured anaerobically in MRS broth (MRS; Thermo Fisher, CN) containing 0.05% L-cysteine hydrochloride. Lactobacillus acidophilus ATCC 4356 was cultured anaerobically in MRS broth.

The peptides (1 mM) were dissolved in 10 mM PBS, 30 mM sodium dodecyl sulfate (SDS) and 50% (v/v) trifluoroethanol (TFE), respectively. The peptide spectrum was scanned by a CD spectrometer (Brighttime Chirascan, Applied Photophysics Limited, UK) at wavelengths of 190 nm to 260 nm with a speed of 10 nm/min.

Equal volumes (50 μL) of bacterial suspension (2 × 10⁶ CFU/mL) and the peptide solution were mixed in wells on polystyrene plates (Corning, USA). After the plates were incubated anaerobically for 24 h, the OD₆₀₀ was measured. In negative control group, the peptides were replaced with PBS only, and chlorhexidine (CHX) was used as a positive control. The minimum inhibitory concentration (MIC) is defined as the minimum concentration that inhibits the visible growth of bacteria within 24 h. The minimum bactericidal concentration (MBC) is the minimum concentration that kills 99% of bacterial cells determined by colony counting (Cao et al., 2020).

The bacterial suspension (2 × 10⁶ CFU/mL, 100 μL) in BHI was mixed with the same volume of the peptide solution (3.1–12.5 μM). The peptides were replaced with PBS only in the negative control. Aliquots of the mixtures were pipetted at 0, 15, 30, 45, 60, 90, 120, and 180 min, and plated on BHI agar after proper dilution. The plates were incubated anaerobically for 48 h, followed by colony counting.

After exposing to the peptides (4 × MIC) for 1 h, the bacteria (1 × 10⁹ CFU/mL) were collected, fixed with 2.5% glutaraldehyde and dehydrated in serially diluted ethanol solutions at room temperature. The peptides were replaced with PBS only in the negative control. After embedding in resin for 48 h at 60°C, the bacteria were sliced into 60–80 nm sections and loaded on 150 mesh copper mesh. After staining with 2% uranium acetate, the sections were washed with 70% ethanol solution and ultra-pure water, respectively, then re-stained with 2.6% lead citrate solution. After washing and drying overnight, the sections were observed by TEM (HT7800, Hitachi, Japan) (Zhang et al., 2022).

The cell membrane integrity was studied by fluorescent staining (Zhang et al., 2020). Streptococcus mutans (1 × 10⁹CFU/mL) pre-treated with the peptides (1 × MIC) for 1 h was centrifuged and collected to remove the peptides. The bacteria were fixed with 4% paraformaldehyde for 1 h, followed by staining with PI solution (10 μg/mL) for 15 min in the dark. Then the bacteria were collected by centrifugation and stained with DAPI solution (10 μg/mL) for 15 min. The bacteria were washed, collected, and observed under an inverted fluorescence microscope with S. mutans treated with PBS as a negative control.

To confirm the impact of the peptides on S. mutans cell membranes, S. mutans treated with the peptides and stained with PI as described above were detected by flow cytometry (BF32366, Beckman Coulter, Brea, CA, USA).

Influence of the peptides on acid-producing capacity of S. mutans has been assessed (Wang et al., 2018). The bacterial suspension was adjusted to 1 × 10⁸ CFU/mL in BHI broth (the initial pH was 7.2) containing the peptides (1/2 × MIC, at which the bacterial growth was not affected), and 1% (w/v) glucose was added to trigger glycolysis. In negative control, the peptides were replaced with PBS only. After incubating for 180 min, an aliquot of the suspension was sampled every 30 min for detecting the pH value using a glass electrode pH meter.

Acid tolerance of S. mutans was assessed (Svensater et al., 1997). Bacteria (1 × 10⁶ CFU/mL) grown in tryptone-yeast extract medium (TYEM) broth (pH 7.5, containing 20 mM glucose) were transferred to BHI culture medium (pH 5.5) with or without the peptides (1/2 × MIC). The same volume of suspension was diluted and applied to BHI agar for determining the number of viable bacteria. After the remaining suspension was incubated for 2 h, the pH was immediately reduced to the killing pH (3.0) for S. mutans by adding 20% hydrochloric acid and further incubated for 3 h. The suspensions were also plated. After incubating all plates anaerobically for 48 h, the total CFUs were calculated.

The bacterial suspensions (1 × 10⁶ CFU/mL, 2 mL) were added to each well containing a coverslip on a 24-well microtiter plate, and incubated anaerobically for 24 h. After discarding planktonic bacteria, the coverslips covered by biofilms were removed to a new 24-well plate, and treated by the peptides (1/2 × MIC, containing 2% sucrose) in buffered peptone water (BPW). In negative control, the peptides were replaced with PBS only. After incubating anaerobically for 180 min, the supernatant was collected, and the production of lactic acid was measured by lactic acid measurement kit (Solarbio, China). OD₅₇₀ was recorded and the concentration of lactic acid was calculated (Chen et al., 2020).

The permeabilized cells were prepared and the activity of F₁-F₀ ATPase was determined (Xu et al., 2011). In general, S. mutans was collected and dispersed in Tris–HCl buffer (pH 7.0) containing 10 mM MgSO₄ and 10% (v/v) toluene. After swirling for 1 min, the bacterial suspension was incubated for 5 min at 37°C and rapidly frozen in liquid nitrogen. The frozen cells were quickly thawed and re-frozen twice. After collecting and exposing to the peptides (1/2 × MIC) for 120 min, the bacteria were collected by centrifugation and resuscitated in Tris-maleate buffer (pH 7.0, containing 10 mM MgCl₂). Finally, 5 mM ATP was added to trigger F₁-F₀ ATPase activity. In negative control, the peptides were replaced with PBS only. The released phosphate was identified (Bencini et al., 1983).

The effect of the derived peptides on the biofilm formation of S. mutans was evaluated (de Breij et al., 2018). Briefly, an equal volume of bacterial suspension (2 × 10⁶ CFU/mL) in BHI broth containing 1% sucrose (BHIs) was mixed with the peptide solution (0.8 to 25 μM) on a 96-well polystyrene plate (Corning, USA), and incubated anaerobically for 24 h. In negative control, the peptides were replaced with PBS only. After discarding the supernatant, the remaining biofilm was rinsed twice with PBS, fixed by methanol solution and thoroughly dried. CV solution (0.1%) was added to stain the biofilm for 15 min. After the excess dye was thoroughly washed with PBS, 200 μL of ethanol was added and OD₅₉₅ was determined.

The activity of the derived peptides on the bacterial adhesion ability was evaluated both on 96-well polystyrene plate surfaces and saliva-coated surfaces (Zhang et al., 2020). Bacterial suspension (1 × 10⁹ CFU/mL) was treated with the peptides (4 × MIC) for 15 min and the unbound peptides were removed by centrifugation (3,000 × g, 3 min). In negative control, the peptides were replaced with PBS only. After resuspending in BHI broth, an aliquot of the suspension was pipetted, plated on BHI agar, and cultured anaerobically for 48 h. The CFUs were calculated to determine the number of viable bacteria. A 96-well polystyrene plate was precoated with 100 μL of artificial saliva at 37°C for 15 min. After the saliva was removed, the saliva-coated plate was prepared. The remaining suspension (100 μL) was added to a polystyrene plate or a saliva-coated plate and incubated for 1 h under static conditions. After discarding the supernatant, the plates were gently washed twice with PBS to remove unadhered bacteria. After adding 200 μL of PBS to each well, all the plates were sonicated three times at a power of 120 W and a frequency of 40 kHz on ice for 90 s with an interval of 30 s. The sonication fluid was plated and the CFUs were numbered. The bacterial adhesion rate was expressed as the CFUs of adhesive bacteria divided by the CFUs of the viable bacteria.

The bacterial suspension (1 × 10⁹ CFU/mL) was incubated with the derived peptides (0.8–25 μM) in Phosphate urea magnesium (PUM) buffer for 20 min, and the OD₅₅₀ of the bacterial suspension was measured. Cetane (200 μL) was added to the remaining bacterial suspension, vortexed, and settled for 30 min. In negative control, the peptides were replaced with PUM only. The supernatant was pipetted to determine the OD₅₅₀. The degree of changes in surface hydrophobicity of S. mutans was shown as a percentage decrease of the OD₅₅₀ (Kim et al., 2019).

The production of water-insoluble EPS in biofilms was evaluated by phenol-sulfuric acid method (Abdel-Aziz et al., 2020). Briefly, equal volumes of bacterial suspension (1 × 10⁶ CFU/mL) and the peptides (1 × MIC) were mixed in a 24-well polystyrene plate and incubated anaerobically for 24 h to generate biofilms. In negative control, the peptides were replaced with PBS only. After discarding planktonic bacteria, the biofilms were collected and washed twice to clean out water-soluble EPS. After centrifugation, the pellet was re-suspended in 1 M NaOH solution (200 μL) for 2 h to extract water-insoluble EPS. Finally, 5% frozen phenol and sulfuric acid were added to the mixture at 1:1:5 (volume ratio), incubated for 1 h, followed by measuring OD₆₂₅.

The EPS distribution in biofilms was observed by CLSM (Deng et al., 2021). Equal volumes (1 mL) of the bacterial suspension (2 × 10⁶ CFU/mL) in BHIs broth and the peptides solution (1 × MIC) were mixed in each well containing a coverslip on a 24-well polystyrene plate. In negative control, the peptides were replaced with PBS only. Alexa Fluor 647 glucan (1 μmol/L) was used to label the dextran conjugate for 24 h. After that, S. mutans nucleic acids were labeled with SYTO 9 (2.5 μmol/L). CLSM (TCS SP8, LEICA, Germany) was used to observe.

The expression levels of virulence genes in S. mutans were measured by qRT-PCR with 16S rRNA serving as a quantified internal control (Xiong et al., 2020; Aqawi et al., 2021). Streptococcus mutans in the logarithmic growth phase was diluted to 1 × 10⁶ CFU/mL in BHIs broth, pretreated with the peptides (1/2 × MIC, diluted with BHIs) for 8 h and total RNA was purified. In negative control, the peptides were replaced with PBS only. The first-strand cDNA was synthesized using RT kit (TES201, Tsingke, China). The tested genes and primers are listed (Table 1). The reaction was performed on a CFX96 real-time system (CFX Connect, Bio-Rad, USA). Gene expression was presented by using the 2^{–Δ Δ Ct} method and normalized to the 16S rRNA level.

The cytotoxicity of the peptides was measured by the Cell Counting Kit-8 (CCK-8) (Utheim et al., 2016). Briefly, HOK cells (5,000 cells/well) were inoculated in a 96-well plate. After reaching 80% confluence, the cells were exposed to the peptides (3.1, 6.2, 12.5, 25, 50, and 100 μM) for 2 h. In negative control, the peptides were replaced with PBS only. Then CCK-8 solution (10 μL) was added and incubated for 2 h. OD₄₅₀ was measured.

The anti-caries effect of GHaR and GHa11R was detected in a caries rat model infected by S. mutans (Wang et al., 2018; Tian et al., 2022). Twenty-seven Sprague Dawley (SD) male rats aged 18 days (Slyke Jingda Experimental Animal Co., Ltd., China) were fed food (supplemented with 0.1% ampicillin) and water (containing 4,000 U/mL penicillin) for 3 days to sterilize the oral endogenous microorganisms. At the age of 22–24 days, 24 rats were anesthetized by isoflurane and infected orally with 200 μL of S. mutant (1 × 10⁹CFU/mL) once daily. Twenty-four hours after the last infection, a sterile cotton swab was used to scratch the oral cavity of the rats, which was then rinsed with PBS. The solution was plated on Mitis Salivarius agar (MSA) containing 20 U/mL bacitracin to confirm colonization. The infected rats were randomly divided into 4 groups with 6 rats in each, including the negative control group (UA159 + PBS), the positive control group (UA159 + CHX), and two peptide-treated groups (UA159 + GHaR and UA159 + GHa11R). Uninfected rats served as the blank group. The rat teeth were evenly brushed with a sterilized cotton swab soaked with 200 μL of PBS, CHX, GHaR or GHa11R three times a week for 8 weeks. All rats were fed Keyes 2,000 # cariogenic diet and purified water containing 5% sucrose. The health status of the rats was recorded. At the age of 81 days, after suffocation by CO₂, the left and right mandibles were removed. After autoclaving at 121°C for 10 min, the mandibles were cleaned by peeling off the remaining tissue, washed with PBS, and stained overnight with 0.4% murexide. The main groove of the tooth was exposed by a proximal sagittal semi-section and observed under a stereo microscope (M205 FA, LEICA, Germany) to assess the caries level according to a modified Keyes score. An “E” score of 1 indicates that the caries involved only the enamel. A “D” score of 2 indicates that the caries did not exceed 1/4 of the dentin thickness. A “Dm” score of 3 indicates that the range of caries was 1/4–3/4 of the dentin thickness. A “Dx” score of 4 indicates that the caries was more than 3/4 of the dentin thickness. The score for each mandible was the sum of the scores for the first and second molars.

The experiments were performed independently three times in triplicates. The data in the tables and graphs was a representation of one biological replicate and presented as the mean ± standard deviation (SD). The data were analyzed statistically using t-test with the GraphPad Prism, with a P-value of less than 0.05 considered statistically significant when comparing the treated groups to the control groups.

In our previous study, we performed sequence alignment on different AMPs and found that the positively charged amino acid lysine (Lys, K) in temporin peptides occurs more frequently than arginine (Arg, R) and histidine (His, H), while in the antibiofilm peptides, R and K are similarly chosen over H (Xie et al., 2019). In order to obtain derived peptides with antibiofilm efficacy, we used R to replace the H at GHa to design GHaR and GHa11R. Compared with GHa, the charge, amphiphilic index (AI) and Boman index (BI) of GHaR and GHa11R increased, and the grand average of hydropathy value (GRAVY) decreased (Table 2). Generally, α-helical AMPs with high amphiphilicity show strong activity; meanwhile their GRAVY is positive, and BI is negative or close to 0 (Saporito et al., 2018). The changes in the physiochemical characteristics of the derived peptides may enhance their antibacterial activities.

The structures of the peptides are shown in Figure 1. The spiral wheel of the peptides presented that on the one side the hydrophilic amino acid residues were located, while the hydrophobic residues were arranged on the opposite side, forming the hydrophilic and hydrophobic surfaces, respectively. The CD spectrum revealed that the peptides showed a random-coiled structure in aqueous conditions, while dissolved in a solvent of SDS (30 mM) or TFE (50%), they shown typical α-helical structures, indicating that GHa and the derived peptides had an amphiphile α-helical structure in simulated bacterial membranes. The predicted 3D structure of the peptides is also an α-helix structure. Meanwhile, the increased amphiphilicity of the derived peptides indicated that they have a more stable amphipathic α-helix structure than that of the parent peptide (Abbassi et al., 2013).

GHa has mild antibacterial activity against S. mutans with MIC of 12.5 μM and MBC of 25 μM. By mutating the amino acid residues of GHa, the ability of the derived peptides GHa11R and GHaR to inhibit S. mutans growth was improved, with the MIC decreased by 4 times and 8 times, respectively. The antibacterial efficacy of the peptides against several oral etiologic bacteria and probiotics was detected. As shown in Table 3, the peptides also showed good inhibitory activity against the oral pathogenic bacteria S. sanguinis and P. gingivalis, which are associated with oral infections and cardiovascular diseases. Comparatively, GHa and its derived peptides against pathogenic bacteria had lower MIC/MBC values; however, they showed higher antibacterial concentrations against the tested probiotics. CHX exhibits strong antibacterial efficacy against both pathogenic bacteria and probiotics.

The bacteriostatic activity of GHa, GHaR, and GHa11R against planktonic S. mutans was investigated (Figure 2A). GHaR inhibited S. mutans growth at 3.1 μM and GHa11R at 6.2 μM, similar to CHX. The time-killing kinetics demonstrated that the killing effect of these peptides on S. mutans was concentration- and time-dependent. After treatment with 6.2 μM GHaR and GHa11R, all bacteria were killed within 90 ∼ 120 min (Figure 2B). The microscopic morphology of S. mutans was observed under TEM. As shown in Figure 2C, the bacterial cell membrane was destroyed, resulting in leakage of cellular contents.

As shown in Figure 2D, after the bacteria were treated with the peptides (1 × MIC), PI penetrated cell membranes, showing red fluorescence, which indicated that the structure of S. mutans cell membranes was destroyed. Flow cytometry was used to further detect the membrane destruction efficiency of the peptides on S. mutans. In comparison with the untreated bacteria, after 1 h of treatment with the peptides (1 × MIC), a large number of S. mutans stained by PI was detected (Figure 2E).

As shown in Figure 3A, the pH value of the suspension in the control group declined sharply from 7.4 to 4.5 within 180 min. After exposure to the peptides, the drop in pH was alleviated and decreased to 5.4, indicating that the peptides inhibited the acidogenicity of S. mutans, especially GHaR and GHa11R. The effect of the peptides on the acid production of S. mutans was further determined (Figure 3B). In comparison with GHa, GHaR and GHa11R reduced lactic acid production by 50% within 180 min. After 3 h incubation under killing pH conditions, 10% of the S. mutans bacteria without the peptide treatment still survived, while the treated bacteria decreased to less than 1% of their original population (Figure 3C). Meanwhile, the activity of F₁-F₀-ATPase related to acid tolerance of S. mutans decreased by 50–70% after being treated with 1/2 × MIC of the peptides for 120 min (Figure 3D).

As shown in Figure 4A, when S. mutans was incubated with sub-MICs of GHaR and GHa11R for 24 h, more than 25–30% of the bacterial biomass in the S. mutans biofilms has been inhibited. Furthermore, we explored the attachment ability of the bacteria to polystyrene plates and saliva-coated polystyrene plates (Figure 4B). After treatment with GHaR and GHa11R, more than 60% of the S. mutans attached to polystyrene plates was reduced, indicating that the peptides disturbed the initial adhesion of S. mutans, thus inhibiting biofilm formation. Applying saliva to the surface of the polystyrene plate significantly enhanced bacterial attachment, with doubling biofilm formation in the negative control. However, both GHaR and GHa11R still showed effective efficacy in preventing the initial attachment of biofilms.

To explain how the peptides impact the adhesion ability of S. mutans, the hydrophobicity of bacterial surfaces exposed to the peptides was investigated. Nearly 40% of S. mutans in the negative control was transferred from the aqueous phase to the organic phase. After treatment, approximately 20% of S. mutans was transferred, and the percentage decreased dose-dependently (Figure 4C), confirming that the peptides decreased the hydrophobicity of S. mutans.

The glucan labeled by Alexa Fluor 647 can be embedded into EPS during the synthesis of the S. mutans biofilm matrix, showing red fluorescence under CLSM. The biofilms in the blank control were tightly packed, and EPS was highly expressed (Figure 4D). Although GHa (1/2 × MIC) showed no significant efficacy on the S. mutans biofilms, GHaR and GHa11R destroyed the confluent, tightly packed layers of the biofilms, resulting in scattered clusters of bacteria in the biofilms. As shown in Figure 4E, the EPS in biofilms decreased significantly after exposure to the derived peptides. The inhibitory activity of the peptides on EPS production has been explored (Figure 4F). The water-insoluble EPS production had been decreased by 40–50% after the bacteria were exposed to GHaR and GHa11R.

As shown in Figure 5A, the tested genes were downregulated after treatment with sub-MICs of the peptides, especially the gtfB and gtfC, which encode water-insoluble EPS, and the ldh. GHaR and GHa11R reduced the expression of the gtfB and gtfC genes by 70 to 80%. GHa11R reduced the expression of the gtfD gene by more than 99%. For the ldh gene, both GHaR and GHa11R reduced the expression of the gtfD gene by more than 95%.

After HOK cells were cocultured with GHa and its derived peptides, cell viability was detected. The results showed that GHa and GHa11R had no toxic effect on HOK cells at a concentration of 100 μM (Figure 5B). GHaR showed a weak inhibitory effect on cell viability at concentrations greater than 12.5 μM. However, the cytotoxicity was not dose dependent, and at least 80% of HOK cells treated with 100 μM GHaR were still alive.

The derived peptides exhibited better antibacterial efficacy than that of GHa, so we investigated whether GHaR and GHa11R reduce the incidence and severity of tooth decay in a rat caries model (Figure 6A). As shown in Figure 6B, all molars of the rats inoculated with S. mutans showed progressive development of dental caries. As indicated by the arrows, the lesions on the sulci surface were significantly more extensive in the control group. The total Keyes scores in mandibular first molar and second molar of the rats were evaluated. The severity of dental caries was decreased by treated with 25 μM GHaR and GHa11R (Figure 6C). GHaR showed effective anti-caries activity as good as CHX.