The peptides used in this study were produced by Synpeptide Co., Ltd. (Nanjing, China) using conventional solid-phase peptide synthesis, then purified and identified using HPLC and ESI-MS. They correspond to the amino acid sequences of porcine amelogenin P173. In Figure 1, studied peptide 1 was synthesized from the N- and C-termini of porcine amelogenin; studied peptide 2 is the peptide TRAP synthesized. Protein was first dissolved in deionized water to prepare a peptide stock solution (Chu et al., 2018). The peptide stock solution was stirred for 24 h on a stirrer at 4°C and then was placed in a refrigerator for at least 24 h.

Fresh bovine incisor teeth after extraction (approved by the First Affiliated Hospital of Zhengzhou University’s Ethics Committee: 2021-KY-1050-002) were employed. The crowns were separated from the roots and prepared into enamel blocks (3 mm × 3 mm × 2 mm) using a diamond-coated band saw under continuous water cooling (Struers Minitom; Struers, Copenhagen, Denmark). The study design is illustrated in Figure 2. Forty-eight enamel samples, free of cracks, cavities, chalky lesions, and enamel dysplasia, were selected under a stereological microscope. The enamel surfaces were then ground flat, polished with silicon carbide sandpaper of various grits (320, 400, 600, 800, 1200, 1500, 2000, and 2500 grades; Buehler Ltd.), and water-cooled to remove the enamel layer with a thickness of about 150 μm. Subsequently, the labial side of the enamel sample was divided into three regions, each with a size of 1 mm × 3 mm (a, b, and c; Figure 3). The reference window with intact enamel was window a, the baseline for caries was window b, and the test window was window c.

The surfaces of the samples were all covered with acid-resistant nail varnish, except for windows b and c. To create artificial carious lesions, forty-eight blocks were immersed in 192 ml of demineralization solution (5.0 mM NaN₃, 50 mM acetic acid, 2.2 mM Ca (NO₃)₂, 2.2 mM KH₂PO₄, and 0.2 ppm NaF; pH was adjusted to 4.5 using 5 mM KOH), at 37°C for 3 days with 100 RPM continuous magnetic stirring (Ten and Duijsters, 1983). The demineralization solution was replenished every 24 h. After demineralization, all samples were rinsed with deionized water and dried naturally.

All enamel specimens were divided at random into four different remineralizing treatments (12 specimens per group): group A, remineralizing medium; group B, 100 μg/ml studied peptide 1 + remineralizing medium; group C, 100 μg/ml studied peptide 2 + remineralizing medium; group D, 2 ppm fluoride + remineralizing medium. Specimens from each group were immersed in 18 ml of the corresponding remineralizing medium (0.9 mM KH₂PO₄, 1.5 mM CaCl₂, 20 mM HEPES, 130 mM KCl, and 1 mM NaN₃; the pH was adjusted to 7.0 with 5 mM KOH), at 37 °C for 12 days with 100 RPM continuous magnetic stirring (Almqvist and Lagerlof, 1993). The remineralizing medium was replenished once per day. The enamel samples were washed with deionized water and air-dried after remineralization treatment.

Each specimen’s mineral density (MD) and lesion depth (LD) were assessed using a micro-CT system (Skyscan 1272, Bruker, Germany) after remineralization. The parameters were set to a voltage of 60 kV, a current of 160 μA, and a resolution of 8.5 μm. After scanning, the profiles of all samples were reconstructed using the NRecon software in Skyscan 1272. For MD calibration, a collection of mineral reference phantoms, including three HA disks (50, 250, and 750 mg/cm³), was scanned.

Each specimen’s sound, baseline, and test windows were all examined at their respective centers in a volume of interest (VOI) of 300 μm × 300 µm. By assuming a sound enamel with a maximum MD of 100 vol%, the MD profile was transformed into a relative MD. The data for the reconstruction of the three-dimensional image of the enamel blocks were collected with a resolution of 1024 × 1024 and voxel sizes of 8.5 μm isotropic. The average MD value was calculated from the MD of the sample. In addition, the three-dimensional VOI of 100 μm × 100 μm × 100 μm was randomly selected at the center of the three regions. The average densities were expressed as MDa, MDb, and MDc, respectively. The remineralization rate (%R) was calculated using the following formula: % R = ∆ Z ∆ Z d e × 100

ΔZde (MDa-MDb) is mineral loss after demineralization; ΔZre (MDa-MDc) is mineral loss after remineralization; and ∆Z (∆Zde-ΔZre) is mineral gain after remineralization.

A coronal image of each sample was obtained. To obtain the lesion depths of the baseline and test windows, the assessments were randomly selected at the center of each coronal image of the sample. The average values were LDb and LDc, and the changes in lesion depth (ΔLD) before and after remineralization can be obtained, ΔLD = LDb-LDc.

Enamel samples were first embedded in acrylic resin and then longitudinally cut in half and polished to achieve a flat surface and a sample thickness of 100 µm. These sections were analyzed using an X-ray micro-diffractometer (D8 Discover; Bruker, Germany). The data were collected in the scanning range of 20°–65° under the conditions of Cu target Kα radiation, tube voltage of 40 Kv, and tube current of 40 mA. The data were analyzed using Jade 6.0 software. Then the obtained XRD pattern was compared with a standard diffraction card (JCPDS 09-0432). (n = 4 for each group).

Enamel samples were first embedded in acrylic resin and then transversely cut in half. Then the samples were dried and sputter-coated with gold. The cross-section morphology was observed by SEM (Sigama, 500; Zeiss, Germany). (n = 4 for each group).

The enamel specimens were embedded in acrylic resin. From the center of each specimen, thin, plane, parallel pieces that were about 100 μm thick were cut and polished. Glass microscope slides were used to mount the slices, which were assessed using a polarized light microscope (Axio Scope Al, Zeiss, Germany). (n = 4 for each group).

The thermodynamics of Ca²⁺ binding to peptide 2 were studied using isothermal titration calorimetry (TA Instruments, USA). At 37°C and 350 rpm, ITC measurements were carried out in a standard volume of nanoITC. Water was put into the reference cell. A 10 mM HEPES solution containing peptide 2 was created (pH = 7.0). A solution of peptide 2 in HEPES at 0.42 mM received an injection of 12.5 mM CaCl₂ (10 mM). Twenty-five separate injections of 2 µL each were used to titrate CaCl₂ into the peptide solution at intervals of 180 s. The heat generated by calcium ions binding to TRAP per injection was measured and was displayed as differential power (µcal s⁻¹) vs. time (min). The area under each injection peak was integrated and presented as kcal mol⁻¹ of the injectant vs the molar ratio of [Ca²⁺] ÷ [TRAP]. In a control experiment, the same amount of CaCl₂ was titrated into HEPES (10 mM), and the results were then deducted from the raw heat of the reaction to get the effective heat of binding. ITCRun data collection software used the ITC one-set of sites model to calculate the thermodynamic values of calcium binding to peptide 2. Calcium binding properties are expressed as affinity constant, dissociation constant, enthalpy, and entropy.

Statistical analysis of the experimental data was performed using IBM SPSS 21.0 (IBM, Chicago, IL, United States). All data are presented as the mean and standard deviation. The ∆Z, ΔLD, and %R were determined by one-way ANOVA followed by post hoc LSD. Multiple comparisons were used to analyze the micro-CT data, and the test level was α = 0.05.

Figure 4 presents the micro-CT 2D image of the enamel sample (a,c,e,g) and the profiles of mineral density against depth from the objects’ surfaces (b,d,f,h). Images (a,c,e,g) include the sound, baseline, and test windows (from top to bottom). Subsurface lesions were confirmed using baseline profiles. The mineral density increased throughout the test profiles. The only medium (group A) that remineralized minerals showed mostly enhanced mineral density at 0–60 µm. At a depth greater than 60 µm for the baseline and test windows in the remineralizing media, the mineral density gradually converged. Except for the group that received only the remineralizing media, the other groups affected both the surface layer and the deeper layer of the lesion.

Figure 5 shows the %R of all groups. The mean %R ± SD values were 23.96% ± 6.16, 48.96% ± 6.06, 48.80% ± 8.95, and 61.29% ± 7.40 for the remineralizing medium, peptide 1, peptide 2 and fluoridate remineralizing medium groups, respectively. One-way ANOVA statistical analysis showed significant differences between the treatment groups (p < 0.05). Additionally, the LSD post hoc test for multiple comparisons indicated significant differences in %R within the treatment groups. Except for the peptide 1 and peptide 2 groups (p > 0.05), all groups had significant differences (p < 0.05).

Table 1 displays the ∆Z from the MD profiles of all groups after remineralization. The data are described using the mean ± standard deviation. Results of the one-way ANOVA revealed that all differences in the medium treatments were statistically significant (p < 0.05). Additionally, the MD between the treatment groups varied significantly according to multiple comparisons using the LSD post hoc test. However, there was no significant difference between the studied peptide 1 and peptide 2 groups (p > 0.05).

Table 2 illustrates the ΔLD in the profiles of each group after demineralization and remineralization. The data are described using the mean ± standard deviation. The one-way ANOVA identified a statistically significant difference in ∆LD among the test groups (p < 0.05). The LSD post hoc multiple comparison analysis showed that the ∆LD were greatest for the fluoridate remineralizing medium group and the least for the mineralizing medium group. There was no significant difference in the ∆LD between the studied peptide 1 and peptide 2 groups (p > 0.05).

Figure 6 shows the polarized light profiles of all groups. All demineralized regions with increased tissue porosity appeared to be positively birefringent when the slices were analyzed by PLM after imbibition in water (RI = 1.33). The enamel lesions had an intact surface layer (the area of negative birefringence) above positively birefringent lesion bodies. In all groups, except for the remineralizing medium group, the thickness of the surface layer and density clearly increased after remineralization. For the remineralizing medium group, a small increase in the density of the surface layer was observed.

Figure 7 presents the cross-section SEM images. Well-aligned enamel rods can be obversed in sound enamel 1). After 3 days of demineralization, the enamel rods of sound enamel were significantly destroyed 2). The remineralizing medium group formed mineral layers on its cross-section 3). The imprints of enamel rods are visible as shallow depressions in the cross-section image of peptide 1 group 4). The cross-section image of the peptide 2 group also shows similar inter-rod gaps. In addition, some minerals adsorb on the surface 5). The fluoridate remineralizing medium group obtained the most obvious enamel rods 6).

Figure 8 shows the XRD spectra of the surfaces of the natural enamel and enamel disks after remineralization. The characteristic peaks displayed on the remineralization surface of each group were compared with those of the standard card (JCPDS 09-0432). The diffraction peaks at approximately 29.0° (210), 31.8° (211), 32.9° (300), and 39.8° (310) are consistent with the characteristic crystalline peaks of HA, indicating that the substance generated after remineralization is HA. The (210), (211), and (300) peaks of XRD that were found on the remineralizing media indicate that crystalline HA started to form, but only in a small amount. In comparison with the remineralizing medium group, the peaks of (210) and (300) were enhanced, indicating that the remineralization of other groups was suppressed.

Figure 9 displays the heat generation spectrogram and effective binding isotherms between the calcium ions and peptide 2. The downward curve indicated an endothermic reaction. The change in the heat value of the binding isotherm was stable and abnormal in the heat generation spectrogram from the fifth drop in the binding isotherm graph, indicating binding saturation. The different thermodynamic values of the affinity constant indicated two distinct types of binding events between Ca²⁺ and peptide 2. A thermodynamic value analysis revealed that the second affinity constant (Ka₂ = 5.548 × 10⁵ M⁻¹) of the Ca²⁺ and peptide was higher than the first affinity constant (Ka₁ = 10 M⁻¹), and the first dissociation constant (Kd₁ = 1.000 × 10^–1 M) was lower than the second dissociation constant (Kd₂ = 1.803 × 10^–6 M), suggesting that Ca²⁺ binding to peptide 2 may be easier for the second type of binding event. The values of enthalpy (ΔH₁ and ΔH₂) were 106.4 kcal mol and -106.6 kcal mol, respectively, and entropy (ΔS₁ and ΔS₂) were 3.760 × 10² cal mol⁻¹ °C⁻¹ and -2.474 × 10² cal mol⁻¹ °C⁻¹, respectively, for Ca²⁺ and TRAP.