As an important structural and functional component of the periodontal complex, cementum plays a crucial role in the regeneration of periodontal tissues1,2. Insufficient cementum regeneration may lead to an incomplete periodontal complex structure and the root resorption issues, because the newly formed alveolar bone may combine to the root dentin directly3,4. Cementum includes approximately 45% inorganic minerals and 55% organic components, while more than 90% organic components are collagen fibers. These collagen fibers provide a framework for the mineralization crystals, and the mineralized collagen fibrils constitute the basic building blocks of cementum5.

Bio-mineralization in collagen fibers is essential for the excellent mechanical and biological properties of cementum6. Cementum protein 1 (CEMP1) is a tissue-specific protein and has been considered as a regulator for cementum mineralization by modulating the nucleation, deposition rate, composition and morphology of hydroxyapatite (HA) crystals7,8. CEMP1 can promote the adhesion and differentiation of cementoblasts and progenitor cells, as well regulate cell phenotype from the non-mineralizing to the mineralizing9,10. Therefore, CEMP1 is considered to be a potential periodontal target to regulate the regeneration of cementum. However, instability of protein structure and the difficulty of extraction and purification limited the application of CEMP1(8).

Literature has reported that CEMP1’s short peptides can mimic its biological functions partly11,12,13,14. A peptide with 20 amino acids from N-terminus of CEMP1 (sequence: MGTSSTDSQQAGHRRCSTSN, N20) was reported to enhance the nucleation in vitro and the growth of HA crystals11. Interestingly, N20 was found to stimulate the proliferation and mineralization of periodontal ligament cells (PDL), leading to the formation of cementoid-like structures and contribution to the regeneration of periodontal defects12. It has been discovered that a peptide with15 amino acid located in the C-terminal domain of CEMP1 (sequence: QGQGDTEDGRMTLMG, C15) can regulate the Wnt signaling pathway to differentiate human oral mucosal stem cells (HOMSCs) into the mineralizing-like phenotype13. However, the discovery and selection of these two peptides seem arbitrary, with no concrete evidence to suggest that these sequences are key structural domains for CEMP1 to regulate biomineralization, and no evidence to suggest that these two peptides can replace CEMP1 for regulating mineralization within collagen fibers.

In this study, we constructed plasmids derived from the CEMP1 structure (CEMP1-△NC) by deleting the N-terminal 20 amino acids (N20) and the C-terminal 15 amino acids (C15). These plasmids were transfected into PDLCS cells to induce target protein expression and to investigate the functional significance of the N20 and C15 regions in CEMP1.Additionally, a novel biomimetic peptide (MGTSSTDSQQAGHRRCSTSNQ-GQGDTEDGRMTLMG), termed NC35, was synthesized based on the combined N20 and C15 sequences. This peptide was evaluated for its ability to bind functionally and to determine whether it could substitute for CEMP1 in promoting intrafibrillar mineralization and modulating biological signaling.

Methods and materials

Molecular docking

Autodock was primarily employed to forecast interactions between peptides and collagen proteins, as well as hydroxyapatite (HAP)15. Firstly, structural files for peptides, collagen proteins, and HA are prepared. Alphafold was utilized to predict three-dimensional structures of peptides with varying lengths, while protein structures of collagen were sourced from the PDB database, and the structure of HAP was obtained from the CAS database (CAS: 1306-06-5). These files were then converted into the PDBQT format using Autodock Tools. Next, regions on peptides or collagen proteins likely to engage in interactions were designated as three-dimensional search grids known as Grid Boxes. For peptide-HAP docking and collagen protein docking, the Grid Box dimensions were set to encompass the entire structures of the peptide and collagen protein, respectively, facilitating global search16. Autodock’s operational parameters were then configured, with the number of docking runs set to 200 times, while other parameters are kept at default conditions. Upon completion of docking, the output files from Autodock were analyzed, encompassing docking energies, binding conformations, interaction sites, and other pertinent information. Typically, conformations with lower docking energies were deemed the most stable binding conformations. PyMOL was employed to visualize the docking structures and interactions.

Cell culture

Teeth without caries or periodontal disease were extracted from 16 to 20 years old patients for orthodontic reason. hPDLSCs were isolated and cultures as previously described. Briefly, the periodontal ligament was scraped from the apical one-third of the tooth root and minced into approximately 1 mm × 1 mm fragments. The tissue was enzymatically digested with type I collagenase (3 mg/mL) at 37 °C and subsequently cultured in 3 mL of α-MEM (Gibco, USA) supplemented with 10% fetal bovine serum (FBS; Gibco, USA), 2 mM L-glutamine, and 1% penicillin/streptomycin (Gibco, USA) at 37 °C in a humidified atmosphere containing 5% CO₂. The culture medium was refreshed every 2–3 days until cell confluence reached approximately 80%, at which point cells were passaged using trypsin digestion. Cells from passages 3 to 5 were used for subsequent experiments and maintained in DMEM (Gibco, USA) supplemented with 10% FBS and 1% penicillin/streptomycin under standard culture conditions (37 °C, 5% CO₂, 95% air, 100% humidity). Informed consent was obtained from all donors, and the study protocol was approved by the Ethics Committee of the Affiliated Stomatology Hospital of Nanjing University (NJSH-2024NL-014).

Constructing pcDNA40-CEMP1 vector and transfecting them into PDLSCs

The open reading frame of CEMP1 was obtained from NCBI (GenBank Accession No. NM_001048212). The optimized coding sequences of CEMP1 and CEMP1 with N-terminal 20 amino acids, C-terminal 15 amino acids deletion (CEMP1-△NC) were synthesized (General Biosystems) and sub-cloned into the pcDNA3.1(+)-6×His vector (General Biosystems). The resultant cDNA construct were ligated into pcDNA3.1(+)-6×His vector, pcDNA3.1(+)-6×His-CEMP1 vector and pcDNA3.1(+)-6×His-CEMP1-△NC vector respectively. Then the plasmids were introduced into DH5α expression host Escherichia coli strain (Invitrogen) and were cultivated at 37 °C in Luria–Bertani medium containing 100 µg/mL ampimycin to an optical density at 600 nm (OD600) of 0.4. Subsequently, the plasmids were purified by GeneJET plasmids isolation kit (Thermo Fisher) and sequenced to confirm the insertion of targeted genes.

hPDLSCs were seeded on 24-well plates at 5 × 104 density and cultured in α-MEM medium (Gibco, USA), after all cells adhered to the wall, all groups were replaced with osteogenic induction medium (Gibco, USA) supplemented with 0.1µM dexamethasone (Gibco, USA),50 µg/mL ascorbic acid (Gibco, USA) and 10mM β-glycerophosphate (Gibco, USA).Thereafter, cells were randomly divided into three groups to be transfected with different plasmids. After 48-hour stable transduction, western blots expression was verified by fluorescence detection and western blot.

Mineralization activity

Cell cultures studies were undertaken after the hPDLSCs transfected with plasmids were incubated in osteogenic induction medium for several days (37℃, 5% CO2). Subsequent cytology experiments were repeated 3times in triplicate.

Alkaline phosphatase activity

After being cultivated in osteogenic medium for 7 days, the alkaline phosphatase activity was measured in PDLSCs using the BCIP/NBT alkaline phosphatase staining Kit (Beyotime Institute of Biotechnology, China) following the manufacture’s protocol and quantitated by measuring absorption at a wavelength of 562 nm.

Alizarin red staining (ARS)

After cultivated in osteogenic induction medium for 14 days, The PDLSCs were washed by de-ionised water and fixed with 70% ethanol for 1 h, then the cells were stained with 2% ARS (Sigma–Aldrich, USA) to detect calcium deposits and the CPC method was used to quantify calcium nodules as previously described17.

Real-time quantitative PCR (RT– qPCR)

RT-PCR was used to assess the expression of osteogenesis-related and cementogenic -related genes, including OPN, OCN, CEMP1 and CAP. After incubated in osteogenic medium for 7days, the total RNA of the cells was extracted with RNA quick purification Kit followed the synthesis of cDNA using the Hiscript Ш 1st Strand cDNA Synthesis Kit (R312-01, Vazyme, China). The gene expression was analyzed by iQTM SYBR Green Supermix (Q511-02, Vazyme, China). Relative level of different mRNA expressions was calculated by 2−ΔΔCt method. The primers are listed in S4.

Synthesis and characterization of peptides

The peptides were synthesized using standard solid-phase peptide synthesis, after purification, it was characterized by high-performance liquid chromatography (HPLC) and mass spectrometry. In the present study, the peptides were termed as P20, P15 and P35 respectively.

Mineralization assay of collagen

The assembly solution with 50 mM of glycine and 200 mM of KCl was prepared, and the pH value was adjusted to 9.2 by using 0.1 M Sodium. The solution was prepared and used immediately. Drop some amounts of collagen-I stock solution into the assembly solution to gain the final concentration of 50 mM and then incubated to gel at room temperature for 20 min. Afterwards, the solution was dropped onto the 200-mesh carbon-nickel TEM grid at 37℃ for 12 h. Then the grids were thoroughly washed by DW and randomly immersed in different remineralization solutions (RS)for 24–48 h: A, RS with P20; B, RS with P15; C, RS with P35; D, RS without peptides (negative control). Subsequently, grids were subjected to gradient dehydration and examined by TEM.

Molecular dynamical (MD) simulations

The small molecules were geometry-optimized using Gaussian16 at the B3LYP/def2-TZVP level of theory and subsequently parameterized with the GAFF2 force field using Sobtop. Initial configurations for the simulation systems were constructed using Packmol18with periodic box dimensions set to 10*10*10 nm³. Molecular dynamics simulations were performed using GROMACS 2022.419, each lasting 100 ns. System optimization was conducted with 5000 steps of steepest descent method followed by 5000 steps of conjugate gradient method to avoid unreasonable contacts. NPT ensemble was employed for system pre-equilibration, maintaining temperature at 298 K using V-rescale temperature coupling and pressure at 1 atm using Parrinello-Rahman pressure coupling. Non-bonded interactions were truncated at 1.2 nm, with an integration time step of 1 fs. Following equilibration, the Berendsen method was utilized20,21. Trajectory files were saved every 10.0 ps for subsequent analysis. The binding free energy of the simulated systems was calculated using the Molecular Mechanics/Poisson–Boltzmann Surface Area (MM/PBSA) and Molecular Mechanics/Generalized Born Surface Area (MM/GBSA) methods, implemented via the gmx_MMPBSA script (version 1.5) (https://pubs.acs.org/doi/https://doi.org/10.1021/acs.jctc.1c00645). The binding free energy (ΔG_bind) of protein–ligand complexes was computed using the following equation:

$${\rm \Delta G_{Bind} = \Delta G_{Complex} -\: \Delta G_{Protein}\:-\:\Delta G_{Ligand}}$$

where Gcomplex, Gprotein, and Gligand represent the free energies of the complex, the unbound protein, and the unbound ligand, respectively.

Characterization of interaction between P35 and collagen

FTIR examination

Peptide P35 was dissolved in deionized water to prepare a suspension at a concentration of 6 mg/mL. Type I collagen solution (3 mg/mL, derived from rat tail; Gibco, USA) was mixed with the peptide solution at a volumetric ratio of 5:1. The mixture was ultrasonically homogenized and incubated at 37 °C for 24 h. Following incubation, collagen suspensions, with or without the peptide, were freeze-dried and subsequently analyzed by Fourier-transform infrared (FTIR) spectroscopy.

Circular dichroism (CD) analysis

Circular dichroism (CD) spectroscopy was employed to assess changes in the secondary structure resulting from the interaction between collagen and peptide P35. Collagen solution was mixed with the P35 solution to achieve a final concentration of 0.6 mg/mL. The mixture was placed in a sealed container containing fresh ammonia water to gradually raise the pH to 9. CD measurements were performed at 25 °C using a quartz cuvette with a 0.01 cm path length, scanning from 190 to 260 nm at a rate of 100 nm/min. A collagen solution without P35 served as the control. Each sample was measured in eight independent replicates to ensure reproducibility.

Effect of P35 on the osteogenic and cementogenic differentiation of periodontal ligament stem cells

Biocompatibility of peptides

To investigate the biocompatibility of P15, P20 and P35 on PDLSCs, cells were seeded in 96-well microtiter plates at 5 × 104 density with a osteogenic medium overnight, then they were incubated with fresh culture media without peptides (control group) or containing 9ug/mL P15,P20 and P35 for 1,3,5 days, the cell viability (%) was analyzed using a Cell Counting Kit-8 (Dojindo, Japan), in brief, 10µL of CCK-8 solution was added to each well for 2 h, then the absorbance was measured at 450 nm by a microplate reader (Molecular Devices, Sunnyvale, USA). The cell viability was assessed by cell live/dead staining assays (Solarbio, China) after 5days of incubation in different conditioned medium.

ALP activity assay

PDLSCs were cultured for 14days with P15, P20 and P35 as previously described. Then the liquor-included culture medium was removed, the cells were washed with PBS thoroughly, ALP Assay Kits was used to measure the ALP activity and quantitatively assessment was done at 405 nm as mentioned above.

ARS

ARS was conducted to evaluate the formation of calcium nodules of PDLSCs after being co-incubated with P15, P20 and P35 for 14 days, the methos were as the same as the above.

Real-time reverse transcription-polymerase chain reaction (RT-PCR)

After cells were treated with different peptides for 7days or 14 days, the mRNA expression of the osteogenesis and cementogenesis-specific markers, including Runx-2, OCN, CAP as well as CEMP-1 were measured by real-time PCR (primers are listed in S5).

Western blot analysis

PDLSCs were treated with or without 9 ug/mL P15, P20 and P35 in 24-well culture plates for 14, 21 days, then cells were lysed by RIPA lysis buffer to obtain intracellar protein, afterwards the proteins were separated by 10% SDS-PAGE and then transferred onto PVDF membranes. After blocking with 5% skim milk solution overnight, the membranes were incubated with primary antibodies overnight at 4 °C with B-action used as an internal control. After the membranes were washed four-times and incubated with secondary antibodies for another one hour, western blots were visualized using a chemiluminescent reagent and a WB detection system.

Statistical analysis

Statistical comparisons between groups were performed using one-way analysis of variance (ANOVA), followed by Tukey’s post hoc test for pairwise comparisons. Data analysis was conducted using SPSS software (version 17.0), P value < 0.05 was considered statistically significant.

Results

Molecular docking

We first investigated and analyzed the molecular docking of full-length CEMP1 with collagen to determine the interaction between CEMP1 and collagen. As shown in Fig. 1A, collagen structure is depicted in red, while CEMP1 structure is represented in cyan, both displayed on protein surfaces. We found that their interactions primarily occurring at the α-helical ends and C-terminal residues of CEMP1, the hydrogen bonding interactions and distances between key residues is shown in Fig. 1B. Additionally, as depicted in Fig. 1C, besides hydrogen bond interactions, hydrophobic interactions also play a crucial role in the recognition and binding between the two protein molecules, particularly evident in the hydrophobic interactions between residues at the C-terminal of CEMP1 and collagen.

The interactions between CEMP1 with hydroxyapatite (HAP) are notable. The crystal structure of HAP has the alternating ion layers with calcium phosphate and hydroxide, while CEMP1 reportedly interacted with the phosphate groups. Autodock was performed to establish the molecular docking of CEMP1 with the HAP molecule. Figure 2A shows that the phosphate group interacts with CEMP1. And Fig. 2B presents by ligplot to display the residues that interact with phosphate groups, highlighting the importance of CEMP1 N-terminus in interacting with HAP.

Fig. 1
figure 1

The molecular docking results between full-length CEMP1 and collagen are presented as follows: (A) Visualization of the protein surface post-docking. (B) Hydrogen bonding interactions and distances between key residues. C)Depicts the interaction between full-length CEMP1 and collagen protein, as visualized using ligplot.

Full size image
Fig. 2
figure 2

Illustrates the molecular docking results between full-length CEMP1 and a single HA molecule. (A) Presentation of the position of HA within CEMP1 and distances to surrounding residues. (B) Visualization of interactions between HA’s phosphate groups and surrounding residues using ligplot.

Full size image

N20 and C15 of CEMP1 are essential in mediating the mineralization phenotype of PDLSCs

In order to verify the N20 and C15 sequences are the core regions of CEMP1 in regulating bio-mineralization and the cementogenic differentiation of PDLSCs, CEMP1-△NC plasmid that delete the sequences of N20 and C15, the full-length CEMP1 plasmid as well as the negative control (vector) were constructed and transfected into PDLSCs. As shown in Figure.3 A, all of them can be expressed in PDLSCs. Meanwhile, immunoblotting data labeled by His monoclonal antibody showed that CEMP1 and CEMP1-△NC were all well overexpressed after transfection (Figure.3B). As shown in Figure.3 C, PDLSCs in the CEMP1-△NC and control groups exhibited significantly lower ALP activity compared to the CEMP1 group (P < 0.001), with no significant difference observed between the CEMP1-△NC and control groups. A similar trend was observed in the quantitative analysis of calcium nodule formation (Figure.3D). Consistently, as shown in Figure.4E, the expression levels of osteogenic and cementogenic markers (CEMP1, CAP, OCN, and OPN) were significantly upregulated in the CEMP1 group compared to both the control and CEMP1-△NC groups. No significant differences were detected between the CEMP1-△NC and control groups. Control CEMP1 CEMP1-∆NC.

Fig. 3
figure 3

Different plasmids transduction and the effect on osteogenic/cementogenic differentiation of PDLSC. (A)Location of plasmids in PDLSCs by immune-fluorescence. (B) Detection of overexpression of target protein in PDLSC transfected with different plasmids. (C) ALP staining and. (D) ARS staining of PDLSC transfected with different plasmids for 14 Days. (E)belongs to the real-time PCR analysis of CAP, CEMP1, OCN and OPN expression of PDLSC transfected with different plasmids for 14 Days. Variances were compared by one-way ANOVA. *** P < 0.001.

Full size image

Influence of biomimetic polypeptide on collagen mineralization

After immersed in different remineralization solutions for 24 h, there were a large number of calcified clusters randomly distributed on the surface or between the collagen fibers in P15, P20 and P35 group, especially the aggregates were gradually distributed along the long axis of collagen in P35, diffraction analysis showed the mineral had an amorphous mineral phase (Fig. 4B, C,D, inset). In addition, there was no high contrast mineral aggregation on the surface of collagen which were exposed to m-SBF without the peptides (Fig. 4A). After 48 h, plate-shape crystallites were randomly aligned on the boundary or along the collagen fibril axis in all groups, however, there was few crystallites in control group, as shown in (Fig. 4E), compared with 24 h, the number of crystals increased significantly in P15 and P20 group and partial intrafibrillar mineralization was observed in some regions of collagen fibril (Fig. 4F, G, yellow arrow). As for P35, the collagen was fully wrapped by crystallites, the length and thickness of them were significantly enhanced than the other two groups and their c-axes aligned along the collagen fibril axis (Fig. 4H).

Fig. 4
figure 4

The TEM images of mineralized collagen fibrils after different treatments for 24 h（A-D）and 48h（E-H）(A,E in the control group; B,F in the P20 group; C,G in the P15 group; D,H in the P35 group).

Full size image

Molecular dynamics simulations of the interacting among biomimetic polypeptides, collagen and minerals

To further validate the rationality of this design and the mechanisms of biomimetic polypeptides in regulating biomineralization, molecular dynamics simulations were performed for P20, P15, and P35, respectively. These peptides were subjected to molecular dynamics simulations individually with minerals, collagen, and both simultaneously. We conducted molecular dynamics simulations of ternary complexes comprising peptides, minerals and type I collagen, and further analyzed the peptide structures. By tracking changes in secondary structure, we could assess alterations in peptide conformation during interaction processes, thereby analyzing the effects of different peptides in the system. Figure 5 illustrates a mixed system containing P20 jointly with minerals and type I collagen, we observed a gradual reduction in distance between P20 and both minerals and type I collagen. Structurally, residues 16–20 mostly adopted a turn conformation. This change in conformation may arise from spatial hindrance effects upon collagen binding and restrictions imposed by interaction forces, impacting subsequent reaction processes. As shown in Fig. 6, the mixed system involving P15 with minerals and type I collagen, residues 6–11 tended to form a 3-Helix. This could be attributed to P15’s shorter peptide chain, leading to conformational changes due to intramolecular hydrogen bonding, favoring subsequent reaction processes. As shown in Fig. 7, with an increasing number of amino acids, P35 in systems with all three components exhibited both better contact with minerals in disordered and turn structures, as well as maintaining a stable 3-Helix conformation, resulting in enhanced interaction and stability of the entire system.

Fig. 5
figure 5

The secondary structure analysis of the system where P15 binds jointly with minerals and type I collagen. (a)Conformations of P20 at various time points. (b) The temporal evolution of Test1’s secondary structure.

Full size image
Fig. 6
figure 6

The secondary structure analysis of the system where P15 binds jointly with minerals and type I collagen. (a)Conformations of P20 at various time points. (b) The temporal evolution of Test1’s secondary structure.

Full size image
Fig. 7
figure 7

The secondary structure analysis of the system where P35 binds jointly with minerals and type I collagen. (a)Conformations of P20 at various time points. (b) The temporal evolution of Test1’s secondary structure.

Full size image

Subsequently, binding free energy calculations were conducted to compare the affinity of the three peptides with minerals and collagen. Specific results are depicted in Fig. 8. It can be observed in Fig. 8A, B, C, the bonding energy of system with P20 and minerals is-2250 kJ/mol, while in the system where P20 individually acts with collagen, the average bonding energy is 0 kJ/mol. Compared with P15, the binding affinity of P20 with minerals significantly increases. In the systems where P20 interacts with both minerals and collagen, the average binding energy with minerals is approximately.

-1500 kJ/mol. There is also an increase in binding energy when P20 interacts solely with minerals. This suggests that P20 is a more favorable choice for promoting binding with minerals compared to P15.In addition, P15 shows a stronger binding ability with collagen compared to P20 in the system where P15 interacts with both minerals and collagen. Furthermore, compared to P15 and P20, P35 exhibits the lowest binding energy with minerals, regardless of interacting solely with minerals or in the presence of collagen, the average binding energy throughout the process is -2500 kJ/mol.

In order to explore which residues are responsible for the binding of P35 with type I collagen and minerals, the binding energy was decomposed and the binding free energy provided by each residue was averaged. It was found in Fig. 8D, in the system with only minerals, the main interacting residues were M1 (-12.96 kJ/mol), R15 (-13.165 kJ/mol), N20 (-14.919 kJ/mol), and Q23 (-18.849 kJ/mol), the binding energy provided by the residues at the C terminal were not very significant. Correspondingly, due to the addition of type I collagen, the residues that have little influence on the binding free energy before have an enhanced effect, namely R30 (-6.541 kJ/mol), M31 (-17.571 kJ/mol), T32 (-5.305 kJ/mol), L33 (-19.917 kJ/mol), and M34 (-16.649 kJ/mol).

Fig. 8
figure 8

A)The Bond energy of P20 with minerals or collagen. B) The Bond energy of P15with minerals or collagen. C) The Bond energy of P35 with minerals or collagen. Blue represents the binding energy between minerals and the peptide in the system containing only minerals. Orange indicates the binding energy between type I collagen and the peptide in the system with only type I collagen bound. Gray denotes the binding energy between minerals and the peptide in the mixed system of all three components. Yellow represents the binding energy between type I collagen and the peptide in the mixed system of all three components. D）Decomposition of combines free energy. Orange line: the system of P35 and minerals. Blue line: the system of P35, type I collagen and minerals.

Full size image

Characterization of interaction between P35 and collagen

The interaction between peptide and collagen is an important part in regulating biomineralization. As shown in Fig. 9A, the peaks at 1565 cm− 1、1553 cm− 1 and 1454 cm− 1 are attributed to the vibration of amide I、amide II and amide III respectively22. After P35 was added, the NH bending vibration peak showed slight movement, indicating the formation of hydrogen bond between collagen and the peptide23. And as depicted in Fig. 9B the peak at 1240 cm− 1 became broader and the tensity weakened. Moreover, a new peak appeared at 1133 cm− 1, suggesting the occurrence of a sesterification24.Circular dichroiam spectroscopy is a useful method to analyze the second structure of proteins25. Comparing the circular dichroiam spectrometer images of collagen before and after mixing with P35, it is obvious there is a negative peak at 200 nm and a positive peak at 215–227 nm, indicating the typical CD structural peaks of natural collagen fibers remain unchanged26. Additionally, the peak values range of triple helix structure shrinked after the addition of P35, suggesting the increase in intermolecular crosslinking which can stabilize the template of collagen to promote the nucleation and growth of minerals27as shown in Fig. 9C. Meanwhile, CD analysis result demonstrated 10.92%α-helix,52.95%β-fold,16.14%β-Turn and 19.99% irregular curl in the secondary structure of collagen fibers combined with P35.

Fig. 9
figure 9

Analysis of the interaction between type I collagen and peptide P35. (A) FTIR spectra showing the structural features of collagen before and after interaction with P35. (B) Magnified view of the selected region in (A), highlighting detailed spectral changes. (C) Circular dichroism spectra of collagen before and after binding with P35, illustrating changes in secondary structure.

Full size image

Effect of the biomimetic peptides on the osteogenic and cementogenic differentiation of periodontal ligament stem cells

The biocompatibility of the biomimetic peptides was evaluated by CCK-8 assay and live/dead staining. CCK-8 assay was used to evaluate the cell viability treated with different peptides for 1, 3, 7days.As shown in Fig. 10A, proliferation of cells in P35 group was more pronounced than P20 and P15 group at different time points, and no significant differences were found among the test groups, which indicated the polypeptides exhibited excellent biocompatibility and did not exert negative effects on cell proliferation, the live/dead staining imagines shown in Fig. 10B, demonstrated the polypeptides were of no cytotoxicity, consistent with the CCK8 results. Therefore, the polypeptides could be used for further osteogenic and osteogenic differentiation assessment.

Subsequently, the ALP staining displayed higher level of ALP activity in peptide groups than control group after 14 days of incubation, and a pronounced amount of ALP expression was found in P35 group compared with P15 and P20 group (Fig. 10C). The trend could be observed from the quantality analysis of ALP expression. Moreover, the formation of a mineralized nodule as the late-stage osteoblast activity marker, evaluated by ARS staining, displayed the peptide groups possessed the promoting effect on calcium deposition compared to control group, while P35 group exhibited the most pronounced ability to enhance mineralized nodules formation by compared with that of for P20 and for P15 (Fig. 10D).

After incubation with different polypeptides for 7 days and 14 days, the osteogenic and cementogenic differentiation of hPDLCs in different conditions was analyzed in mRNA and protein levels. As displayed in Fig. 10E and Fig. 10F, the mRNA expression levels of Runx2, OCN, CEMP1 and CAP were significantly upregulated under polypeptide treatment compared with that of control, and the P35 group demonstrated more excellent ability to promote the expression of osteogenic-related and cementogenic-related genes. Western bolt exhibited the expression of related genes(Runx2, CEMP1 and CAP)at the protein level was consistent with the trend of PCR. It has been found that CEMP1 transfected into nonmineralizing cells could increase the level of ALP and formation of mineralized nodules, the molecules associated with bone/cementum formation, were upregulated in mRNA and protein level.

Fig. 10
figure 10figure 10

The regulatory effect of three peptides on the differentiation of hPDLSCs. (A) The effects of peptides on hPDLSCs cell viability. (B)Live/Dead cell stain. (C) ALP staining and ALP activity detection. (D)Alizarin red staining and CPC quantitative analysis of calcium content. (E) (F) qRT-PCR analysis of hPDLSCs osteogenesis and cementogenesis related gene expression after incubation for 7days or 14days. (G)Western blotting analysis of hPDLSCs osteogenesis and cementogenesis related protein expression after incubation for 7days or 14days. Variances were compared by one-way ANOVA *P < 0.05, **P < 0.01, ***P < 0.001.

Full size image

Discussion

Cementum, a vital structural and functional component of the periodontal complex, plays a central role in periodontal maintenance and regeneration24,25. Current research on cementum regeneration primarily focuses on the role of non-collagenous proteins in regulating biomineralization28with CEMP1 being the most extensively studied5. In addition to analogs of naturally occurring non-collagenous proteins, short peptides derived from CEMP1, such as the N20 and C15 domains, have been developed to mimic its biological functions11,12,13.

Serine-rich sequence plays a crucial role in the process of mineralization, which can be attributed to their strong affinity for HAP, inducing its morphogenesis and serving as nucleation sites for the minerals5. CEMP1-p1 (MGTSSTDSQQAQHRRCSTSN), comprising the N-terminal 20 amino acids of CEMP1, contains 25% serine and possesses inherent properties stimulating HAP crystal nucleation and growth11,12. Arroyo R et al. found 15-amino acid peptide from CEMP1’s C-terminal domain (QGQGDTEDGRMTLMG) are located in CEMP1′s most intrinsic disordered region, while CEMP1 itself is an inherently disordered protein with a high proportion of random coils in its secondary structure. These characteristics endow the protein with flexibility and plasticity, enabling it to bind with other proteins and regulate crystal nucleation and growth29,30. Given the finding of previous researches and the results of docking, the N-terminal 20 amino acids from CEMP1 can effectively bind to HAP. Furthermore, the C-terminal amino acids of CEMP1 exhibit strong hydrogen bonding interactions with collagen. Therefore, we combined these two segments of amino acids to design a novel peptide to achieve functional complementarity and enhancing affinity with HAP and collagen.

ALP activity and the formation of mineralized nodules are considered as the early -stage marker and the final-stage marker of osteogenic differentiation, respectively31,32. Thus, ALP activity assays and ARS staining were carried out on the PDLSCs transfected with different plasmids. We found that CEMP1 overexpression led to the higher level of ALP, while the deficiency of N20 and C15 sequences caused the effect of CEMP1 to disappear. Moreover, the mRNA levels of osteogenic and cementogenic factors like OPN, OCN and CAP were remarkably upregulated after CEMP1 overexpressed, although CEMP1-△NC also counteracted this effect, indicating the regulatory function of CEMP1 on PDLSCs cells depends on the N20 and C15 sequences. Additionally, the deficiency of N20 and C15 sequences in CEMP1 also restricted the differentiation of PDLSCs into a mineralizing-like phenotype, further suggesting N20 and C15 are essential for CEMP1 in mediating the bio-mineralization and cementum formation.

CEMP1 has been widely studied as an NCP to regulate biomineralization and promote cementogenesis and osteogenesis33,34however, the researches on the effect and mechanism of biomimetic polypeptide derived from CEMP1 in inducing fibrillar mineralization are relatively scarce. In the process of biomineralization, NCP binds to special sites on the surface of collage, affecting the mineralization process in various ways35,36. In the present study, the effectiveness and mechanisms of P35 on induction fibrillar mineralization were evaluated by characterization in vitro and molecular dynamics simulations (MD).

After immersion in mineralization medium for 24 h, there was no obvious deposition of mineralized substance on the surface of the collagen fibers in the control group, indicating the interaction between biomimetic peptide and collagen is considered as the main factor for inducing the precursors of mineralization. Moreover, collagen fibrils alone did not affect the formation and distribution of the crystallites, which was consistent with the research findings of Niu et al.37. While, P35 group exhibited a remarkable capacity to attract minerals onto the surface of collagen compared with P20 and P15 group, and as the mineralization time continued, the length and thickness of the crystal increased, analysis of the diffraction patterns from mineralized fibrils showed that these minerals were completely crystalline and arranged with their c - axes oriented along the collagen fibril axis, which manifests the characteristic feature of intra mineralization14,21.

From the results of MD, it can be seen that P35, an intrinsic disorder protein is prone to change its orientation in accordance with the surrounding environment. When interacting with minerals and type I collagen, P35 is inclined to form 3-Helix to stabilize the structure, and it can increase the contact area with collagen and minerals to create more corner structures. Furthermore, D7 in the sequence of P35 is negative, therefore minerals preferentially accumulate in the N terminal of P355, while due to steric hindrance effect, the C terminal is apt to form different structural conformations when binding with type I collagen, resulting in different binding residues.

Compared with P15 and P20, P35 exhibited superior performance in mineralization regulation, the main reasons are as follows: (1) P35 not only binds tightly to minerals but also interacts well with type I collagen without affecting its interaction with minerals. From a binding strength perspective, this strongly supports that P35 is the optimal peptide for interacting with both minerals and type I collagen. (2) Although P15 and P20 could adjust their conformations according to the different systems, P35 has more flexibility to adapt to minerals and type I collagen, providing sufficient space for folding or stretching. Within the ternary mixed system, the residues at the N terminal of the polypeptide interact with minerals, and the C terminal residues bind to type I collagen, enabling a better balance of the system. The result further validated the conclusion of molecular docking.

Runx2 is considered as a regulator that participates in the differentiation and maturation of osteoblasts38. While CEMP 1 and CAP, the typical markers of cementogenesis, CAP is related with the mineralizing-tissue-forming progenitors in the periodontal ligament cells, it possesses the capacity to induce cementoblast progenitor cells to migrate and attach to the tooth root surface, and expression of CAP was upregulated during cementogenesis5. Cumulative researches have demonstrated that CEMP1 transfected into nonmineralizing cells could increase the level of ALP and formation of mineralized nodules, the molecules associated with bone/cementum formation were also upregulated in both mRNA and protein level. The results in the present research confirmed the active regulatory role of biomimetic peptide derived from CEMP1 during the biomineralization process of bone and cementum formation. As for P35, compared with P15 and P20, P35 displayed a more remarkable capacity to promote osteogenic and cementogenic differentiation of hPDLCs.

In summary, the present study demonstrates that the novel biomimetic peptide derived from CEMP1 possesses the capacity to induce collagen mineralization and enhance the proliferation and differentiation of hPDLCs toward a mineralizing-like phenotype. While peptide P35 shows promise as a potential candidate for cementum and bone regeneration and repair, several limitations must be acknowledged. First, the stability and degradation kinetics of the peptide under physiological conditions remain unexamined, which may significantly impact its functional performance in vivo. Second, as the current findings are based solely on in vitro experiments, the regenerative potential of P35 in the context of complex periodontal defects has yet to be validated in animal models. Third, although molecular docking suggested potential binding interactions between P35 and hydroxyapatite (HAP) or collagen, these static models do not account for dynamic conformational changes in physiological environments. Additionally, while molecular dynamics simulations offered insights into peptide-mineral and peptide-collagen interactions, the relatively short simulation timescale (in the nanosecond range) may limit the predictive accuracy regarding long-term structural stability. Further studies are required to comprehensively evaluate the biological functions and underlying mechanisms of P35 in regulating mineralization and to validate its efficacy in vivo.