Unveiling structure-activity relationships of proanthocyanidins with dentin collagen

Abstract

Objectives:

To elucidate the structure-activity relationships (SARs) of proanthocyanidins (PACs) with type I collagen using sixteen chemically defined PACs with degree of polymerization (DP) 2–6.

Methods:

Under a dentin model, the biomimicry of PACs with type I collagen was investigated by dynamic mechanical analysis (DMA) and infrared spectroscopy. The dentin matrix was modified with PACs from Pinus massoniana [monomers (Mon-1 and Mon-2), dimers (Dim-1-Dim-4), trimers (Tri-1-Tri-4), tetramers (Tet-1-Tet-5), and hexamer (Hex-1)]. A strain sweep method in a 3-point bending submersion clamp was used to assess the viscoelastic properties [storage (E’), loss (E”), and complex moduli (E*) and tan δ] of the dentin matrix before and after biomodification. Biochemical analysis of the dentin matrix was assessed with FTIR spectroscopy. Data were statistically analyzed using one-way ANOVA and post-hoc tests (α = 0.05).

Results:

DP had a significant effect on modified dentin moduli (tetramers ≈ trimers > hexamers ≈ dimers > monomers ≈ control, p < 0.001). Trimers and tetramers yielded 6- to 8-fold increase in the mechanical properties of modified dentin and induced conformational changes to the secondary structure of collagen. Modifications to the tertiary structure of collagen was shown in all PAC modified-dentin matrix.

Significance:

Findings establish three key SARs: (i) increasing DP generally enhances biomimicry potential of PACs in modulating the mechanical and chemical properties of dentin (ii) the secondary structure of dentin collagen is affected by the position of B-type inter-flavanyl linkages (4β→6 vs. 4β →8); and (iii) the terminal monomeric flavan-3-ol unit plays a modulatory role in the viscoelasticity of dentin.

Graphical Abstract

1. Introduction

Biomimetic approaches entail the design and development of functional biomaterials, biological mechanisms, and processes aimed at mimicking the physicochemical, mechanical, and innate properties of natural materials [1,2]. Biomimicry drives improvements in tissue engineering and regeneration, particularly in cases when regeneration and repair of biological tissues is intrinsically limited, such as in dental tissues. Dentin biomodification is a recently established biomimetic approach that target modulation of biomechanics and biostability of the tissue for reparative and restorative dental interventions [3].

Dentin is a structurally and chemically complex mineralized tissue composed of a type I collagen-rich extracellular matrix that is impregnated by an inorganic phase of hydroxyapatite crystals. Dentin forms the bulk of a tooth and is susceptible to physiological and pathological induced alterations that impair mechanical functional features of the tooth. Prior efforts have identified plant-derived proanthocyanidins (PACs) as potent bioactive biopolymers that mimic native collagen cross-linking processes and, thereby, enhance the biomechanical properties of the dentin extracellular matrix [3–5]. Of particular interest is the biomimetic activity of PACs with proline-rich proteins such as collagen. PACs are structurally complex aromatic molecules with C15 monomer skeletons that contain one aliphatic and multiple phenolic hydroxyl groups. PAC oligomers form by condensation of the constituent flavan-3-ols units via interflavanyl C-C (B-type linkage) or ether (C-O-C) bonds (A-type linkage), or both [6,7].

While the detailed mechanism of PAC-collagen interaction remains to be elucidated, findings to-date indicate that PACs mimic collagen cross-linking by hydrogen bonding at the molecular, micro-fibrillar, and fibrillar scales of the collagen network [8,9]. This biomimicry process is mediated by the formation of PAC hydroxyl radicals and the reactivity of terminal flavan-3-ol C-rings towards C-C and C-O bond formation. Through inter(re)actions with hydroxyl, carboxyl, amino, and amide groups of collagen side chains, both hydrophilic and covalent interactions can occur between PACs and collagen [10,11].

The structural complexity of PACs derives from their significantly large number of configurational variations, i.e., the stereochemical variability within the monomers [catechin (C), ent-catechin (ent C), epicatechin (EC), and ent-epicatechin (ent EC)], differences in linkage positions and types, as well as the progressive degree of polymerization (DP), forming thousands or more distinct PAC structures in a single plant [3,12]. This intriguing variability contributes to the challenges associated with defining the precise chemistry of PACs and obtaining optimized and well-defined PAC-based biomaterials with demonstrated promise as sustainable biomimetic interventional materials.

Pinus massoniana is a renewable source rich in non-galloylated PACs that are potent mediators of dentin biomodification. Prior studies have demonstrated that P. massoniana reduce enzymatic degradation and improve the biomechanical properties of the dentin matrix [5,12–14]. Collectively, these studies point to both the DP (preferably ≥ 3) and specific structural features of PACs as being determinants of potency and modulation of dentin biomodification. However, the chemical complexity of the matrices in which these PACs occur, as well as general challenges associated with sourcing trimeric and tetrameric PACs in sufficient yield, represents a bottleneck for systematic and broader biomimetic investigations [13]. To overcome this challenge, a scaled-up phytochemical operation was recently accomplished that provided sufficient quantities of isolated medium-DP PACs with fully defined absolute stereochemistry from P. massoniana extracts, a plant source that has consistently shown to elicit high and stable dentin biomodification activity [13,15]. Following well-established drug discovery paradigms, mechanistic knowledge on the bioactivity of PACs could be advanced via the establishment of structure-activity relationships (SARs). Thus, the present study elucidated specific SARs of PACs with type I collagen using sixteen (16) chemically defined PACs with DPs 2 – 6, using methods to assess the viscoelastic properties and the chemical fingerprint of the dentin organic matrix.

2. Materials and methods

2.1. Sourcing and purification of PACs

The 16 individual PACs evaluated in this study were purified from extract of the inner bark of P. massoniana (Xi’an Chukang Biotechnology, China; batch PB120212). The phytochemical and spectroscopic procedures for the isolation and structural characterization have been reported in detail previously [14,15] and are summarized briefly as follows.

Isolation employed centrifugal partition chromatography (CPC), starting with 12 g of enriched tri- and tetrameric PACs and yielding 6.5 g and 5.6 g of fractions A and fraction B, respectively,[14] both of which were fractionated further on Sephadex LH-20 column (EtOH), affording subfractions A1−A6 and B1−B7, respectively. Fraction A1 (800 mg) contained mainly two pairs of racemates 1 and 2, identified as (±)-catechin (Mon-1) and (±)-epicatechin (Mon-2), respectively, via chiral phase HPLC. Fraction A2 (1.0 g) was fractionated over a RP-18 silica gel column (MeOH/H2O, 20−80%), and the major two subfractions A2b and A2c were then purified to afford 3 (Dim-1, 30 mg), 4 (Dim-2, 10 mg), 5 (Dim-3, 10 mg), and 6 (Dim-4, 20 mg), by semi-preparative HPLC (23% ACN in 0.1% formic acid H₂O, 2.5 mL/min). Similarly, after RP-18 silica gel column chromatography, 9 (Tri-3, 20 mg); 13 (Tet-3, 2.8 mg); as well as 10 (Tri-4, 12 mg) and 12 (Tet-2, 7.1 mg) were purified from their corresponding fractions A3, A4, and A5, respectively, by semi-preparative HPLC (23% ACN in 0.1% formic acid H₂O, 2.5 mL/min). Purification of A6 (1.5 g) via an RP-18 silica gel column (MeOH/H₂O, 25−30%) led to the isolation of the major tetramer, 11 (Tet-1, 900 mg). Fraction B5 (800 mg) was fractionated over a RP-18 silica gel column (MeOH/H₂O, 20−80%), and the major three subfractions B5a−B5c were further purified via semi-preparative HPLC (18% ACN in 0.1% formic acid H₂O, 2.5 mL/min) to afford compound 8 (Tri-2, 30 mg). Fraction B6 (1.6 g) mainly contained trimer 7 (Tri-1, 700 mg), which was purified by a RP-18 silica gel column. Compounds 14 (Tet-4, 2.0 mg), 15 (Tet-5, 25 mg), and 16 (Hex-1, 20 mg) were purified from fraction B7 by semi-preparative HPLC (20% ACN in 0.1% formic acid H₂O, 2.5 mL/min) after pre-fractionation with a RP-18 silica gel column.

Structure elucidation applied chiral (electronic circular dichroism) and non-chiral (mainly NMR, MS, UV) spectroscopy as well as chemical analyses (phloroglucinolysis, mass spectrometry, and chiral phase high-performance liquid chromatography). This characterized all isolates as condensed tannins, constituted of flavan-3-ol monomeric units that form PAC oligomers via interflavanyl linkages [B-type (4β→6 or 4β→8) and/or A-type (2β→O→7, 4β→8)] of different molecular weights, consistent with various DPs [15]. The 16 compounds included two monomers, Mon-1 (1) and Mon-2 (2) and 14 oligomeric PACs: four dimers, Dim-1-Dim-4 (3 – 6), four trimers, Tri-1-Tri-4 (7 – 10), five tetramers, Tet-1-Tet-5 (11 – 15), and one hexamer, Hex-1 (16) (Fig 1, S1, and S2). The recent PAC Block Arrays (PACBAR) nomenclature (Table 1), which includes the graphical PACBAR as well as macro and micro PACBAR representations [16], was instrumental in annotating the 16 PACs evaluated here in a way meaningful for the study of SARs.

Fig. 1.

Classical chemical structures of compounds 1−16 separated according to the degree of polymerization (DP). Hexamers (Hex-1, 16), tetramers (Tet-1 to Tet-5, 11–15), trimers (Tri-1 to Tri-4, 7–10), dimers (Dim-1 to Dim-4, 3–5) and monomers (Mon-1 and Mon-2, 1–2).

Table 1.

PACBAR graphs and annotations of the investigated panel of sixteen PAC oligomers (1–16) and two catechin-type monomers from P. massoniana, with their degree of polymerization (DP).

Compound	PAC code	PACBAR	Macro PACBAR	DP
1	Mon-1		C and eC	1
2	Mon-2		EC and eEC	1
3	Dim-1		EC=2b74b8=C	2
4	Dim-2		EC=2b74b8=eC	2
5	Dim-3		EC=2b74b8=eEC	2
6	Dim-4		EC=2b74b8=EC	2
7	Tri-1		EC-4b6-EC=2b74b8=C	3
8	Tri-2		EC-4b8-EC=2b74b8=C	3
9	Tri-3		EC-4b6-EC=2b74b8=EC	3
10	Tri-4		EC=2b74b8=eC-4b6-C	3
11	Tet-1		EC=2b74b8=EC-4b6-EC=2b74b8=C	4
12	Tet-2		EC=2b74b8=EC-4b6-EC=2b74b8=EC	4
13	Tet-3		EC=2b74b8=EC-4b8-EC=2b74b8=EC	4
14	Tet-4		EC=2b74b8=EC-4b6-EC=2b74b8=eEC	4
15	Tet-5		EC=2b74b8=EC-4b8-EC=2b74b8=C	4
16	Hex-1		EC=2b74b8=EC-4b6-EC=2b74b8=EC-4b6-EC=2b74b8=C	6

2.2. Biospecimen preparation and dentin biomodification

Dentin biospecimens were prepared using extracted human sound molars (IRB no. 2019–0416). Teeth were sectioned into mid-coronal dentin specimens [0.5 mm (thickness) × 1.7 mm (width) × 7.0 mm (length)]. Specimens were demineralized in 1.5 mL of 10% phosphoric acid (Ricca Chemical Company, Arlington, TX, USA) for 5 h at room temperature and under agitation [17]. Dentin biomodification was accomplished with 0.2 % compound (w/v) in 20 mM Hepes buffer (pH 7). Biospecimens (n = 3) were immersed in 100 μL of the compound solution for 1 h at room temperature and protected from light, subsequently washed, and rinsed with ultrapure water three times. For the control group, biospecimens were kept immersed in Hepes buffer under identical conditions.

2.3. Dynamic mechanical analysis of dentin

A dynamic mechanical analysis (DMA) method was used to determine the viscoelastic properties of the dentin extracellular matrix [18]. Briefly, dentin specimens were analyzed while immerse in ultrapure water, at room temperature and subjected to strain sweep using a 3-point bending submersion clamp (Q800 DMA, TA Instruments, New Castle, DE, USA) following the preset parameters: frequency of 1 Hz, amplitude varying from 1 to 100 μm and at a preload force of 0.01 N. The viscoelastic properties of dentin were determined by the storage (E’), loss (E”) and complex (E*) moduli, which represent the elastic, viscous and overall bulk viscoelastic properties of a material, respectively. The damping capacity was measured as the ratio of the viscous and elastic components (tan δ = E”/E’).

Data were statistically analyzed using one-way ANOVA and post-hoc tests Games-Howell or Tukey (α = 0.05, SPSS v.25, SPSS) to assess the effects of the DP and specific compound/structure on the mechanical properties of dentin ECM [complex (E*), storage (E’) and loss (E”) moduli, and damping capacity (tan δ)].

2.4. Biochemical analysis of the dentin extracellular matrix

The biochemical characterization of the functional groups of the dentin ECM was carried-out by Fourier-transform infrared spectroscopy (FTIR, Nicolet 6700, Thermo Fisher Scientific), equipped with an attenuated total reflectance (ATR) apparatus (single reflection diamond, Thermo Fisher Scientific). Spectra were collected over a range from 4,500 to 600 cm⁻¹ and with the accumulation of 128 scans per sample (n = 3). Background transmission spectra was collected prior to each analysis without a sample on the ATR crystal for computation of absorbance spectra. The same dentin specimens were pressed on the ATR plate surface and spectra was obtained. After acquisition, spectra were treated by a 14-point interpolated baseline subtraction, then normalized based on the mean values of all data sets.

The analysis of the secondary structure of collagen is possible by correlating the frequencies of amide bands with patterns of hydrogen bonding that represent the backbone of collagen polypeptide chains such as α-helices, β-sheet, etc. (Fig 3) [19,20]. Assessment focused on the predominant bands of type I collagen: amide I (C=O stretching mode at 1,630 cm⁻¹), amide II (N-H bending and C-N stretching modes at 1,550 cm⁻¹), amide III (C-N stretching modes and N-H bending at 1,240 cm⁻¹), and wagging vibrations of CH₂ groups at 1,450 cm⁻¹.

Fig. 3.

Biochemical characterization of type I collagen in dentin matrix. FTIR spectra of underivatized collagen dentin matrix, where the dashed vertical lines indicate peaks assignment for the collagen amide A at 3,290 cm⁻¹ (mainly N-H stretch), amide I at 1,630 cm⁻¹ (mainly C=O stretch), amide II at 1,550 cm⁻¹ (N-H bend coupled with C-N stretch), CH₂ scissoring at 1,450 cm⁻¹, and amide III at 1,240 cm⁻¹ (N-H bending coupled with C-N stretch). The vibrational modes of amide bands are represented in 3D structures (carbon, oxygen, nitrogen, and hydrogen atoms are represented by gray, red, blue and white colors, respectively).

The intensities and band area integrals of the assigned bands were fitted, calculated, and the structural variations in the collagen cross-linking and collagen triple helix investigated using indices calculated as ratios between each amide (II/III) and the CH₂ scissoring band. FTIR indices (amide II/CH₂ and amide III/CH₂) of the native and modified dentin matrices were statistically analyzed to assess the effects of the DP and specific compound/structure using one-way ANOVA and Games-Howell or Tukey as post-hoc tests (α = 0.05, SPSS v.25, SPSS). Spectra were analyzed with OMNIC Spectra (Thermo Fisher Scientific) and Origin Pro 8 (OriginLab, Northampton, MA) software.

3. Results

3.1. Dynamic mechanical analysis of the dentin extracellular matrix

The viscoelastic properties of the dentin matrix modified with P. massoniana compounds and control are shown in Fig 2. The degree of polymerization (DP) had a significant effect on the viscoelastic properties [storage (E’), loss (E”), and complex (E*) moduli] of modified dentin, in the following ranking order: tetramers ≈ trimers > hexamers ≈ dimers > monomers ≈ control (p < 0.001). Dentin modified with trimers and tetramers yielded the highest increase in E’ and E* modulus when compared to the control group (p < 0.001), with a remarkable 6- to 8-fold increase in their mechanical properties Fig 2A, 2D and Table S1. Moreover, the E” of trimer and tetramer modified dentin presented the highest values and fold variation of 11 to 16-fold increase (p < 0.001, Fig 2B). Hex-1 and dimer (Dim-1 to Dim-4) modified matrices presented intermediate E’, E”, and E* values (p < 0.001). There were no differences in all three moduli between the different tetramers (Tet-1 - Tet-5) and trimers (Tri-1 - Tri-4) when compared with PACs of the same DP (p > 0.05). Statistical differences were present between different dimers, where dentin modified with Dim-4 presented the highest E’ and E*, when compared to Dim-1 and Dim-3, and similar values to Dim-2 (E’, p = 0.017, E”, p = 0.043 and E*, p = 0.018). Dim-2 modified matrix depicted similar moduli when compared to Dim-1 and Dim-3 moduli. Monomers, however, did not alter the viscoelastic behavior of dentin, with the lowest E’ values of approximately 10 MPa (p > 0.05). There were no differences in E’, E”, and E* values between Mon-1 and Mon-2 (p > 0.05).

Fig. 2.

Results of the dynamic mechanical analysis of PAC-modified dentin matrices. The graphs depict means and standard deviation bars of (A) storage modulus (E’), (B) loss modulus (E”), (C) tan δ and (D) complex modulus (E*) of PAC-modified dentin as a function of their distinct degrees of polymerization (DP) and specific compound/structure. Different capital letters show statistically significant differences among groups of distinct DP values. Different Greek letters show statistical differences between isolates of the same DP, whereas bars depict the lack of statistical differences. Ctl = Control. Statistical difference set at p < 0.05.

The damping capacity (tan δ) of modified dentin also varied according to the DP. Dimer and hexamer modified dentin presented the highest damping capacity (p < 0.001), with up to a 2.3-fold increase in tan δ (Fig 2C). Tetramers and monomers groups presented similar damping capacity as the control. No differences existed in the damping capacity values of tetramers, trimers, and monomers, when compared to compounds of same DP (p > 0.05). However, significant tan δ differences were revealed between dimers, with Dim-2 effecting the highest damping capacity of all dimers, with values similar to Dim-1 (p = 0.001).

3.2. Biochemical analysis of biomodified dentin extracellular matrix

Fig 3 and 4 show representative FTIR spectra of native and PAC-modified collagen of dentin matrices, respectively, with band assignments for amide I at ~1,630 cm⁻¹, amide II at ~1,550 cm⁻¹, CH₂ scissoring at ~1,450 cm⁻¹, and amide III at ~1,240 cm⁻¹. All dentin matrix spectra displayed absorptions bands between 1,200 and 937 cm⁻¹, a region assigned to C–O bond stretching (Fig 4A and S3). Peaks located at 1,031 and 1,080 cm⁻¹ arise from C–O vibrations and C–O–C absorptions of carbohydrate moieties [21] and the symmetric extension of the C–O–C group in the pyran ring of condensed tannins [22]. Bands between 1,160 and 1,145 cm⁻¹ were assigned to the asymmetric stretching of C–O–C [23], the symmetric bonding of aliphatic C–H_, O–H or C–O stretch of different groups (Fig 4A) [24].

Fig. 4.

Biochemical characterization and structural modifications of PAC-modified dentin matrices. (A) FTIR spectra of dentin collagen (black, Control) and biomodified collagen by isolated PACs [two monomers (gray scale; Mon-1 and Mon-2), four dimers (blue scale; Dim-1 to Dim-4), four trimers (yellow scale; Tri-1 to Tri-4), five tetramers (pink scale; Tet-1 to Tet-5) and one hexamer (green; Hex-1)]. Peaks assigned to collagen are depicted by the dashed lines. FTIR spectra between 1,260 and 850 cm⁻¹ (right panel) depicts news bands in the range of trimer, tetramer and hexamer-modified dentin matrices. (B) and (C) The graphs depict means and standard deviation bars of the ratios calculated from the intensities and areas under the assigned peaks of amide II vs. CH₂ (1,550/1,450 cm⁻¹) and amide III vs. CH₂ scissoring (1,240/1,450 cm⁻¹). Different capital letters show statistically significant differences between native and modified collagen matrices of different degrees of polymerization (DP). Distinct Greek letters show statistical differences between compounds of same DP, whereas bars depict the lack of statistical differences. Ctl = Control. Statistical difference set at p < 0.05.

Identification of qualitative spectral differences of biomodified dentin collagen hinged on the following parameters: appearance of new bands, broadening and shifting of previously assigned bands, and increased absorbance compared to the unmodified control (Fig 4A and S3). The spectra of dentin collagen modified with monomers and dimers exhibited a discrete increase in the intensity of the amide I and II bands. Dim-1, −2 and −3 treatment spectra presented discrete shoulders between 1,135 and 1,091 cm⁻¹, however Dim-4 treatment did not present these variations (Fig 4A).

An increase in DP was associated with more pronounced modifications to the spectra of PAC-modified dentin, as shown by the presence of new peaks between 1,135 and 1,091 cm⁻¹, a shift or broadening of the bands between 1,160 and 1,145 cm⁻¹ with trimer, tetramer, and hexamer biomodifications (Fig 4A). Other noticeable spectral alterations were a higher absorbance and the broadening of peaks at 1,080 and 1,060 cm⁻¹ that resulted in one unified band for Hex-1, trimer, and tetramer treatments, most noticeably in the Tet-5 and Tri-2 treatment spectra (Fig 4A and S3). The increased adsorption behavior at 1,060 and 1,080 cm⁻¹ was attributed to the characteristic functional groups of high-molecular weight PACs [25] and were assigned to skeletal vibrations involving C–O bond stretching [26].

Interestingly, an increase in DP did not result in an increase of the intensity of collagen amide I/II/III bands. However, dentin biomodification significantly altered the areas and decreased the intensities under these assigned peaks (Fig 4A and 5B). These modifications were quantified via amide II and III to CH₂ band ratios (Fig 4B and 4C), which are considered a measure of the preservation of integrity of collagen [8]. The trimers, tetramers, and hexamer lowered the helical integrity measure of amide II/CH₂ ratio compared to the control, monomers, and dimer groups (p < 0.001). Accordingly, the groups with DP ≤2 and the control group exhibited no differences (p > 0.05). Regardless of DP, all amide III/CH₂ ratios decreased significantly compared to the control (p < 0.05). Compounds of identical DP exhibited no differences in their amide III/CH₂ ratios, except for the trimers: treatments with Tri-1 and Tri-3 resulted in a higher amide III/CH₂ ratio compared to Tri-2 and Tri-4 (p = 0.009 and p = 0.040, respectively; Fig 4C).

Fig. 5.

Collagen hierarchical structure and proposed structural collagen modifications and PAC-dentin matrix mechanisms of interactions. (A) Schematic of the primary, secondary, tertiary, and quaternary structures of collagen. (B) Typical spectra of dentin collagen (black) displaying the main amide I (blue), II (green) and III (orange) bands. PAC-modified dentin collagen amide bands are in evidence in dashed blue, green and orange respectively (PACs of degree of polymerization = 3, 4 or 6), and depict the differences in intensity and area when compared to native dentin collagen (black). The proposed structural collagen modifications associated to amide II and III bands are shown in the green and orange panels, respectively. The modification to amide II is associated to the secondary structure of collagen and transformation of free -NH₂ into N-H groups after PAC-dentin interaction. Alteration to amide III is linked to changes to the tertiary structure of collagen. (C) Postulated PACs-collagen interactions at different hierarchical levels of the structural organization.

4. Discussion

The overall study premise was that specific chemical features enable PACs to mimic physiological type I collagen cross-linking at the molecular, micro-fibrillar and fibrillar hierarchy of collagen, and, thereby, increase tissue strength. Understanding this biomimicry requires consideration of the molecular chemistry of type I collagen as the main organic component of the dentin extracellular matrix (ECM). The primary structure of collagen consists of a sequence of repeated amino acid motifs such as glycine and other residues (Gly-X-Y)_n [27]. The combination of hinge-like glycine and hydroxyproline units leads to helical macromolecules (i.e., α-helices and β-sheets), providing the 2D or secondary structure of collagen. The 3D or tertiary structure is formed by three polypeptide chains composed of repeating amino acid motifs with a triple-helical conformation coil arrangement [28]. The helical molecules are staggered and hierarchically organized in microfibrils to form collagen fibrils, known as the collagen 4D or quaternary structure (Fig 5A).

The stability and viscoelastic behavior of collagen is associated with the complex hierarchical structure, the length and exact amino acid sequence of a polypeptide chain, hydrogen bonding, and covalent inter-molecular cross-links [29]. An increase in cross-linking density leads to larger yield of stress and strain, as well as an enhancement of tissue strength [27]. In prior studies [13–15], P. massoniana was found to be a source for an extended and structurally distinct collection of PACs that mediate such biomimicry processes effectively. This justified this first SAR study of PACs.

Fig 2 shows that PACs with DP ranging from 2 to 6 increase the viscoelastic properties of the dentin matrix significantly. Trimeric (Tri-1 – 4) and tetrameric (Tet-1 – 5) PACs were the most bioactive molecules, eliciting increases to the dentin moduli (storage, loss, and complex modulus) by up to 8-fold (Table S1), while the interactions between the hexamer (Hex-1) and dimers (Dim-1 – 4) and dentin led to intermediate moduli (Fig 2A, B and D). A potential two-tailed role of the degree of polymerization (DP) can explain this increase. First, an increase of molecular polarity due to a large number of hydroxyl groups and aromatic rings increases hydrogen bonding. Secondly, as compound sizes increase, the spatial surface contacts between PAC and collagen are enhanced. Hence, higher PAC molecular weight interposes more intra-molecular cross-linking, which assures the collagen molecule stability, as well as induced additional inter-molecular and inter-fibrillar cross-linking, consequently improving the biomechanics of dentin [9,30]. Also, dentin biomodification by the hexamer, tetramers, and trimers resulted in conformational changes to the secondary collagen structure (Fig 4A and 5B) and revealed increased inter-molecular cross-links attributed to increased tissue stiffening. Thus, the DP is a key structural feature of PACs’ biomimicry of collagen cross-linking. Trimers and tetramers display the best features as dentin-bioactive compounds.

Fourier-transform infrared (FT-IR) spectroscopy analysis can provide essential information about chemical characteristics and provides means of characterizing protein conformation [19]. Herein, the IR analysis of biomodified dentin matrix showed prominent alterations to the intensities of amide II and III bands (Fig 4A and 5B). Furthermore, a specific decrease to the amide II/CH₂ and amide III/CH₂ ratios was observed in biomodified dentin in medium-DP PAC (Fig 4B) and all other PAC-modified dentin (Fig 4C).

The amide II absorption band is known to be proportional to the amount of -NH2 in collagen [31–33]. The decrease of the amide II/CH₂ ratio is suggestive of a transformation of free -NH2 into N-H groups as a result of collagen cross-linking [34,35]. Amide III band is associated to the integrity of the collagen triple helix [34] and modifications to amide III are linked to protein conformational changes. The observed decrease in amide III/CH₂ ratio suggests collagen denaturation for values ≤0.6 [19,36]. PAC-modified dentin exhibited lower amide III/CH₂ ratio than the control group, yet the values were above 1 (Fig 4C). All investigated PACs induced conformational changes to the dentin matrix, most likely due to increased hydrogen bonding and rearrangement of existing collagen hydrogen-bond patterns (Fig 5B) [35,37].

The amide III/CH₂ ratios between Tri-1 and Tri-3 were the only significant differences (Fig 4C); they could indicate a relationship between the type of IFL and the effectiveness of specific PAC hydroxyl groups in triggering protein conformational changes. For instance, in 5-deoxy PACs, the 4β→6 linkage is known to be more stable than the 4β→8 linkage [38,39]. As hydrogen bonds are essential for collagen stabilization and improved tissue biomechanics, the energetics of these bonds impact tissue hydration, substitutions of collagen residues, and molecular stability [38,40].

To expand the analyses of PAC SARs in dentin biomechanics, in a fashion that better matches the underlying biomolecular complexity, the viscoelastic properties of the dentin matrix were not only assessed by static biomechanical methods, but via dynamic mechanical analysis (DMA) [18]. Previously, assessed under uniaxial static forces, the mechanical properties of PAC-treated dentin tissues showed increases of the modulus of elasticity up to 30-fold [9,17,41]. In contrast, the DMA method enables measurements of multiple components of the viscoelastic properties of the dentin such as storage (E’), loss (E”), and complex (E*) moduli, as well as the damping capacity (tan δ). This allowed the identification of unique modifications to the elastic and viscous behavior of dentin collagen modified by groups of PACs of different DP, as well as between compounds of similar molecular size.

For instance, dentin matrix modified by dimers (Dim-1 to Dim-4) presented the most diverse viscoelastic behavior, with moderate storage modulus, but the highest damping capacity. Tissue viscoelasticity is impacted by changes in moduli, especially to loss modulus (E”), which affects the damping capacity (tan δ). An increase in damping capacity indicates an increased ability of the PAC dimer-modified dentin matrix to dissipate energy during oscillatory stress when compared to other groups. Dimer-modified dentin was the only group that exhibited viscoelastic differences between PACs of the same DP. Dim-1 to −4 contain the same interflavanyl linkage [A-type (2β→O→7, 4β→8)] between their epicatechin (EC) and variant terminal unit (Fig 1, S1 and Table 1). The findings suggest that the variability of the terminal unit elicits different viscoelastic behavior. Particularly the Dim-4 treatment group (epicatechin-(2β→O→7, 4β→8)-epicatechin) showed the highest viscoelastic behavior compared to other dimers (Fig 2). Catechin (C) and epicatechin (EC) are simple flavan-3-ols. While bioactivity differences between these epimers have been described for other rather general properties (“antioxidant” and radical scavenging [39,42] as well as pH stability [43]), the observed dentin-related properties of the monomer do not qualify them as viable biomodification candidates. However, as terminal units of PACs, minor difference in chemical structure and biological properties could explain the impact dentin matrix viscoelasticity outcomes as now observed. Apparently, terminal epicatechin units seem to slightly favor mediation and formation of intra-molecular and inter-molecular cross-linkages.

The viscoelastic behavior of Hex-1 modified dentin is similar to dimers and trimers with regard to the loss modulus and damping capacity during mechanical deformation (Fig 2). The decrease in the amide II/CH₂ ratio indicates that Hex-1 modifies the secondary structure of collagen in a similar fashion as the trimers and tetramers (Fig 4B). At the same time, the observed decrease in the amide III/CH₂ ratio revealed that Hex-1 induced conformational changes to the triple helix equal to the entire panel of PACs (Fig 4C). However, new IR bands arose between 1,135 and 1,091 cm⁻¹ only in trimer, tetramer, and hexamer modified dentin matrix. Specifically, the spectra of Hex-1 modified dentin showed the most intense absorption bands at 1,060 and 1,080 cm⁻¹. This was attributed to the characteristic functional groups of this large PAC (Fig 4A). In prior studies, PAC pentamers and polymers have shown bioactivity that was interpreted as being due to their increased number of hydroxyl groups, polarity, and size. The results suggest that, even though Hex-1 is a relatively large molecule (Fig 1 and S2), its overall flexibility resulting from multi-axis/-bond rotation might potentially lead to conformers that still maximize hydrogen bonding to the dentin collagen backbone (Fig 5).

Finally, the results provide insights SARs and dentin damping capacity. While all PAC-modified dentin matrix displayed tan δ values within a similar range, the damping capacity of dentin changed as a function of the molecular size and the chemical specificities of each compound.

Values of tan δ may vary according to the analytical methods and material, where values from 0.01 to 0.1 have been reported for plastic [44], <10⁻³ for metals [45], and 0.01 to 0.04 for “hard” biological tissues such as bone [46]. Native dentin shows tan δ values between 0.02 to 0.3 [18,47,48], but an ideal range has yet to be identified. Tetrameric PAC-modified dentin matrix exhibited the lowest damping capacity in the entire panel (Fig 2C). The tetramers consist of two A-type interflavanyl linkages (Fig 1 and Table 1), thereby offering a relatively high degree of conformational stability of their 3D structure [49], Fig S2, which likely affects their collagen interaction. Apart from the monomers, the PACs investigated herein all contain A-type interflavanyl linkages. The abundance of A-type IFLs in the investigated PAC species relates to their increased stability in solution and lower degradation rates [9]. This could explain why tetrameric PAC-modified dentin exhibited more stable hydrogen bonding and more rigid cross-links compared to other groups. It could also indicate that these tetramers leave less room for molecular sliding of the collagen molecules during vibrational stress.

The similar biomimicry outcome and lack of significant mechanical differences observed between some trimers and tetramers could reflect serendipitous similarities in their structural properties, such as the number of hydroxyl groups, spatial surface, and amphiphilicity. Hence, interactions of PACs with the dentin matrix that cannot be uncovered by IR spectroscopy could also contribute to biomechanical effects that, thereby, remain undetectable by current methods. The amphiphilic nature of generally more polar PACs can play a role in the extent of their interaction with dentin, as well as in the stability of hydrogen bonding and formation of hydrophobic pockets. This study provided the first available substantial set of SARs for PAC-dentin interactions. This data opens new opportunities for future investigations aimed at revealing the stability of dentin biomodification over time. In summary, the present investigation of 16 PACs purified from the highly active extract of P. massoniana led to the following set of SARs: (i) increasing degrees of polymerization (DP) generally enhance the biomimicry potential of PACs in modulating the mechanical and chemical properties of dentin; the optimum DP is 3 to 4. While hexamers are viable agents, their translational implementation remains challenging due to low yield and high purification effort; (ii) the secondary structure of dentin collagen is affected by the position of B-type inter-flavanyl linkages (4β→6 and 4β→8) as this can affect the precise position of OH-groups adjacent to the C-C inter-flavanyl bonds; and (iii) the terminal monomeric flavan-3-ol unit (catechin vs. epicatechin) plays a modulatory role in the viscoelasticity of biomodified dentin. The overall biomimicry potency imparts from a combination of the stereochemical structural features of PACs (molecular size, their type of IFLs, and their monomeric units) that collectively alter the viscoelastic behavior of dentin matrices.

5. Conclusions

The degree of polymerization (DP) is a key determinant feature of PAC SARs: it modulates the viscoelastic behavior and the chemical conformation of dentin collagen. Trimeric and tetrameric PACs showed the highest potency and overall, most promising biomechanical profiles as biomimetic agents. Hence, the observed variation in damping capacity of the dentin matrix induced by dimers and a hexamer revealed a substantial impact of a combined stereochemical features of PAC, that include the type, interflavanyl linkage, and specific monomeric units. The IR resonances revealed that mid-DP PACs increase the dentin matrix moduli by modifying the secondary and tertiary structures of collagen. In addition to marked alterations to the typical collagen amide II and III IR absorption bands, IR spectral features of dentin biomodified with PAC hexamer, tetramers, and trimers revealed the occurrence of additional absorbances which coincide with characteristic functional groups of PACs, and, thereby, benchmarks effective biomimicry action.

Highlights.

• Establishment of structure-activity relationships of Proanthocyanidins (PACs) with dentin collagen.

• The most favorable biomimetic activity is mediated by PAC trimers and tetramers.

• Mid-size PACs increase dentin moduli by modifying structural properties of collagen.

• B-type inter-flavanyl linkages position affects the secondary structure of collagen.

• Terminal monomeric unit of PACs plays a modulatory role in dentin viscoelasticity.

Abbreviations

PAC

proanthocyanidin

SAR

structure-activity relationship

DP

degree of polymerization

C

catechin

ent C

ent-catechin

EC

epicatechin

ent EC

ent-epicatechin

Mon

monomer

Dim

dimer

Tri

trimer

Tet

tetramer

Hex

hexamer

DMA

dynamic mechanical analysis

ECM

extracellular matrix

E*

complex modulus

E’

storage modulus

E”

loss modulus

FTIR

Fourier transform infrared spectroscopy