Paradoxical effects of galloyl motifs in the interactions of proanthocyanidins with collagen-rich dentin

Abstract

Plant-derived proanthocyanidins (PACs) mediate physicochemical modifications to the dentin extracellular matrix (ECM). The structure-activity relationships of PACs remain largely unknown, mostly due to the varied complex composition of crude extracts, as well as the challenges of purification and mechanistic assessment. To assess the role of galloylated PACs as significant contributors to high yet unstable biomodification activity to the dentin ECM, we removed the galloyl moieties (de-galloylation) via enzymatic hydrolysis from three galloyl-rich PAC-containing extracts (Camellia sinensis, Vitis vinifera, and Hamamelis virginiana). The biomechanical and biological properties of dentin were assessed upon treatment with these extracts vs. their de-galloylated counterparts. An increase in the complex modulus of the dentin matrix was found with all extracts, however, the crude extract was significantly higher when compared to the de-galloylated version. Exhibiting the highest content of galloylated PACs among the investigated plants, Camellia sinensis crude extract also exhibited the biggest relapse in mechanical properties after one-month incubation. De-galloylation did not modify the damping capacity of dentin ECM. Moreover, PAC-mediated protection against proteolytic degradation was unaffected by de-galloylation. The de-galloylation experiments confirmed that gallic acid in galloylated rich-PAC extracts drive a stronger yet significantly less sustained mechanical effect of in dentin ECM.

1. Introduction

Pitfalls to current resin-based dental restorative therapies include the development of secondary caries [1,2]. Biomimetic strategies targeting the dentin have shown increased stability and strength of the resin-dentin interface in adhesive restorations. One biomimetic strategy utilizes plant-based proanthocyanidins (PACs) for biomodification of the dentin extracellular matrix (ECM) and enhancement of their physicochemical properties [3–6]. By mediating exogenous collagen crosslinking, certain PACs increase the dentin biomechanical properties, provide resistance to proteolytic dentin degradation, and decelerate the deterioration of the dentin-resin interface [7–9]. Further in-depth evaluation of predictability and sustainability of these effects remains key for successful future clinical application.

The interaction of PACs with dentin depends on their structural features, such as degree of polymerization, presence of galloyl moieties and degree of galloylation, type of interflavan linkages, and overall stereochemistry [5,10,11]. Prior studies have shown that, by choosing appropriate plant sources and production/fractionation methods, these characteristics can be selected to tailor the desired biomechanical responses. Long-term findings revealed that PACs extracts with the most favorable biomodification potency may not necessarily sustain the elicited bioactivity over an extended time [5,10,11]. Particularly, PACs extracts containing galloyl groups (flavanol 3-O-gallic acid esters) are among those PAC species that exhibit a notable reduction of their strong initial bioactivities over time, before reaching a plateau [11,12]. Therefore, the study premise is that galloyl functionalities are major contributors to a relatively unstable form of PAC-dentin biomodification.

Numerous bioactivities have been attributed to galloylated phenolic natural products, with epigallocatechin gallate (EGCg) being the most extensively characterized molecule. Representing a galloylated PAC monomer, EGCg has recently been recognized as an invalid metabolic panacea and highly improbable drug lead [13]. Galloyl moieties are 3,4,5-trihydroxylated benzoic acids (i.e., gallic acid) that attach to the C-3 hydroxyl function of the PAC monomers via an ester bond. The number of hydroxyl groups in PAC monomers has shown a positive correlation with increased modulus of elasticity of dentin matrices [10]. As such, due to the 3,4,5-trihydroxyl functions of the gallocatechin monomer and the additional hydroxyl groups of the galloyl moiety, EGCg exhibits the highest dentin bioactivity of the monomeric catechins [10]. Previous reports highlight the role of the galloyl moiety to significantly enhance the binding affinity to proteins [14–18].

However, this simplistic correlation is misleading. First, the instability of the bioactivity is a result of several chemical processes, such as epimerization, oxidation, and ester hydrolysis, that can be catalyzed by increased temperature and pH [19–22]. Second, the apparent correlation of hydroxyl counts and bioactivity ignores time as an important factor. Therefore, in order to systematically investigate the dentin biomodification potential of galloylated PACs, the present study subjected three PAC-containing extracts with different contents of galloyl moieties and distinct PAC fingerprints to a two-step de-galloylation protocol to create matching pairs of PAC extracts that were galloylated vs. galloyl-free (“galloyl knock-out” extracts). These materials were then used to investigate the specific effect of galloylation on the activity and stability of PAC-induced dentin matrix biomodification. The null hypotheses tested were 2-fold: that knocking-out galloyl motifs in PACs will not affect (1) the mechanical properties of biomodified dentin, as assessed by immediate and long-term (up to 6-months) measurements; nor impact (2) the protective effect of PACs on dentin matrix biodegradation.

2. Materials and methods

2.1. Plant extracts selection, chemicals, and materials

The following extracts were selected due to their diverse content of galloylated compounds: Camellia sinensis leaf extract (CS, Sunphenon 90D, Taiyo International Inc. Minneapolis, MN, USA, lot.103281) and Vitis vinifera L. grape seed (VV, MegaNatural Gold grape seed extract, Polyphenolics, Madera, CA, USA, batch #06112508-01/122112505-01) represented extracts with high and low content of galloylated PACs, respectively. Hamamelis virginiana (HV, Mountain Rose Herbs Inc., Eugene, OR, USA) was selected due to its high content of galloylated compounds despite lower PAC concentration. All extracts were prepared in-house using 70% acetone extraction [11]. Tannase, an enzyme purified from Aspergillus ficuum (CAS: 9025-71-2, Lot# BCBS4299V) was purchased from Sigma-Aldrich (St. Louis, MO, USA). All solvents were of analytical grade obtained from Fisher Scientific (Hanover Park, IL, USA) or Sigma-Aldrich (St. Louis, MO, USA). Water was deionized to 18.2 MΩ cm at 25°C through the Milli-Q Synthetic A10, Millipore Water (Bedford, MA, USA) system using a Quantum Ultrapure Cartridge and fed through a double cartridge ion exchange system (Culligan, Northbrook, IL, USA).

2.2. Two-step de-galloylation protocol for the extracts

The applied two-step de-galloylation (“knock out” [KO]) protocol included the enzymatic hydrolysis of the galloyl moiety followed by countercurrent separation (TBE-300C, Shanghai Tauto Biotech Co., Ltd, Shanghai, China) to remove the released by-product, gallic acid (GA). A pilot experiment had confirmed the enzymatic-induced hydrolysis of galloyl moieties using isolated EGCg by a tannase purified from Aspergillus ficuum [23]. The extracts from CS and VV were hydrolyzed with tannase (20:1 ratio, w/w) in 8 mL Milli-Q water at 37°C for 20 hours and 2 weeks for the HV extract with tannase (10:1 ratio, w/w). The hydrolysis reaction was quenched by adding methanol, until no changes were observed in the high performance liquid chromatography (HPLC) analysis. The de-galloylated extracts were analyzed by HPLC and nuclear magnetic resonance (NMR) spectroscopy (ECZ 400 MHz spectrometer, Jeol Resonance Inc., Peabody, MA, USA) equipped with a liquid nitrogen cryogenic probe (SuperCOOL probe). Subsequently, countercurrent separation was applied to knock out the released GA from the extracts. A quaternary system of hexane/ethyl acetate/methanol/water (HEMWat, 1:9:1:9, v/v) was thoroughly equilibrated in a separatory funnel. The upper organic phase was used as a stationary phase by applying descending operation mode. Operating at a constant temperature (25°C), high-speed counter current chromatography (HSCCC) instrument was filled with stationary phase (3 column volumes, 300 rpm, 3 mL/min). Next, the coils were rotated to 800 rpm as the lower aqueous mobile phase was pumped at a flow rate of 1.3 mL/min. Once a hydrodynamic equilibrium was achieved after pumping mobile phase, the filtered sample solution (~200 mg dissolved in 2.0 mL of both phases) was injected into the column. The fraction collection was set at 11 min per test tube. The Sf value was 0.72. Elution ran in a head-to-tail mode for 32 test tubes, followed by extrusion. The partition coefficient (K value) of GA was determined as 0.697 by the “shake-flask” method [24]. Concentrated fractions were analyzed by thin layer chromatography (TLC), HPLC, and ¹H NMR.

2.3. Chemical Analysis of Extracts

The products from galloyl ester hydrolysis, GA and the corresponding de-galloylated PACs, were analyzed using LC-20AB HPLC (Shimadzu Kyoto, Japan) with a YMC-Pack ODS-AQ 250 × 10 mm, S-5 μm, 120 Å column equipped with a photodiode array (PDA) detector (SPD-20A, Shimadzu, model SPD-20A). The mobile phases consisted of A: acetonitrile, and B: 0.1% formic acid in H₂O. An isocratic 20% of A with a flow rate of 2.5 mL/min was applied. The PDA chromatograms were extracted at 280 nm.

The CS, VV, and HV hydrolysis products and their corresponding unmodified crude extracts were analyzed by Delta 600 HPLC (Waters, Milford, MA, USA) using a YMC-Pack ODS-AQ 250 × 4.6 mm, S-5 μM, 120 Å column equipped with a Waters 2996 PDA detector. The solvents were A: acetonitrile, and B: 0.1% formic acid in H₂O. A linear gradient from 5% to 100% B in 40 min was used with a flow rate of 1 mL/min.

2.4. Preparation and biomodification of dentin specimen

Fifteen human sound third molars were selected and extracted after approval from the University of Illinois at Chicago (IRB no. 2019-0416). Teeth were cut to obtain mid-coronal dentin specimens of 0.5 × 1.7 × 7.0 mm (thickness × width × length). A dimple was made at one end of each sample to allow for repeated measurements on the same surface. Specimens were demineralized in 10% phosphoric acid (Ricca Chemical Company, Arlington, TX, USA) for 5 hours [25]. Solutions containing 0.65% w/v of crude (C) extracts: VV_C, CS_C, and HV_C, and their correspondent de-galloylated counterpart (gallic acid knock out, GAKO): VV_GAKO, CS_GAKO, and HV_GAKO, were prepared in 20 mM HEPES buffer (pH 7.0). All dentin specimens (n = 8) were individually placed in 100 mL of the corresponding solution for 1 h, followed by ultrapure water rinse. A control group was treated with HEPES buffer only. All specimens were incubated in simulated body fluid (SBF) at 37°C. SBF consisted of 5 mM HEPES, 2.5 mM CaCl₂•2H₂O, 0.05 mM ZnCl₂, and 0.3 mM NaN₃ [26], adjusted to a pH of 7.4 with 1 M NaOH. The SBF was changed every two weeks.

2.5. Dynamic Mechanical Analysis (DMA) of dentin matrix

The response of a viscoelastic material such as dentin ECM, is measured as it is deformed over a range of strain. Here, we used a micro-DMA method, to determine the viscoelastic properties of bulk dentin matrix [27]. After treatment, specimens were placed in a three-point bending submersion clamp (Q800 DMA, TA Instruments, New Castle, DE, USA) and a strain sweep was conducted with predefined parameters (frequency: 1 Hz, amplitude: 1 to 100 μm and preload force: 0.01 N) [27]. Specimens were tested in distilled water immersion at room temperature. The storage modulus (E’), representing the elastic properties of a material and the damping capacity, describing the ability of the material to dissipate energy (tan δ = loss modulus [E”]/E’), were both calculated within the linear region of viscoelasticity. All demineralized dentin specimens with baseline E’ ranging between 6.5 – 10.5 MPa were included in the study and specimens from the same tooth allocated randomly into different experimental groups. Measurements were determined after biomodification and at 1-month, 3-month, and 6-month incubation in SBF.

2.6. Biodegradation of the dentin matrix by exogenous proteases

Mid-coronal dentin was sectioned into squares of 2.0 × 0.5 mm (length × thickness) and demineralized in 10% phosphoric acid for 5 h. Specimens were allocated into seven groups (n = 10) and treated as described in Section 2.4. After treatment, specimens were rinsed with ultrapure water and then dried in a vacuum desiccator containing anhydrous calcium sulfate for 24 h at room temperature. The dry weight (M1) was determined using an analytical balance (XPR Microbalance, Mettler Toledo Inc.) with a precision of 0.001 mg. After 1 h of rehydration, specimens were immersed in digestion medium (0.2 M ammonium bicarbonate buffer, pH 7.8) containing bacterial collagenase (150 μg/mL, from Clostridium histolyticum, Sigma-Aldrich, St. Louis, MI, USA) for 24 h at 37°C. A second dry mass weigh (M2) was recorded and the percentage of biodegradation was determined as follows: R (%) = 100 − ((M2 × 100)/M1) [25].

2.7. Statistical Analysis

The equality of variances intra-groups, determined by Levene’s test, was found not to meet the assumption of homogenous distribution (p < 0.001). The de-galloylation effect was determined using paired t-test for each extract (C vs. GAKO) using the immediate values of mechanical properties (E’, tan δ) at α = 0.05. The biodegradation (mass loss) was analyzed using one-way ANOVA and Games-Howell post-hoc tests (α = 0.05). The mechanical stability of biomodification was determined using two-way ANOVA repeated measures with a Greenhouse-Geisser correction and Bonferroni post-hoc test (α = 0.05). The investigated factors were dentin treatment (seven levels: control, CS_C, VV_C, HV_C, CS_GAKO, VV_GAKO, HV_GAKO) and evaluation time (four levels: biomodification, 1-month, 3-month, and 6-month storage). All data were analyzed using SPSS Statistics software package (SPSS V26, Chicago, IL, USA).

3. Results

3.1. Chemical analysis of tannase-induced hydrolysis products

A pilot study showed tannase can catalyze the de-galloylation reaction under mild conditions. This is exemplified in Figure 1 and Figure S1, depicting the conversion of EGCg into epigallocatechin (EGC) and GA. Moreover, a second step of knocking out the liberated GA by HSCCC was applied to hydrolyzed extracts. Analyses of the HPLC and ¹H NMR data confirmed that de-galloylation had occurred and the released gallic acid moieties had been removed by a two-step “knock out” process. A comparison of the HPLC chromatograms of the hydrolyzed extracts and their galloyl knock-out counterparts showed increased amounts of GA (Figure 2). The application of HSCCC yielded seven, eight, and six fractions from corresponding CS, VV, and HV hydrolyzed crudes, respectively (Table 1). A total of 31.8 (fraction 4), 7.2 (fraction 5), and 65.0 (fraction 4) mg of GA of ≥ 98% was assigned by HPLC analysis at 280 nm (Figure S2). A more detailed analysis of the hydrolyzed products by ¹H NMR is included as Supplementary Information (Figures S3 - S5).

Figure 1.

A. De-galloylation of EGCg by tannase from Aspergillus ficuum. B. HPLC analysis of EGCg enzymatic hydrolysis product.

Figure 2.

HPLC chromatographic comparison of the enzymatic hydrolysis products of three PAC-containing plant extracts (CS [A], VV [B], and HV [C]), each with their corresponding crude extract (injecting 10 μL of ~10 mg/mL solutions).

Table 1.

Basic structural assignments of fractions corresponding to CS, VV, and HV hydrolyzed crudes and their estimated GA content (wt %) before knock out. EGC: epigallocatechin, GC: gallocatechin, GA: gallic acid, EC: epicatechin, C: catechin.

	Fractions
	1	2	3	4	5	6	7	8
CS (200 mg) + tannase (10 mg) - estimated GA content 16.7 wt %
Weight (mg)	0.5	21.0	131.2	31.8	1.5	16.3	5.3
Major components	polar	A (A+B)-type PACs	EGC+GC	GA	GA+EC	EC	C+fatty
VV (200 mg) + tannase (10 mg) - estimated GA content 3.6 wt %
Weight (mg)	0.4	147.8	26.9	14.0	7.2	4.7	4.7	3.4
Major components	polar	B-type PACs	B-type PACs	B-type PACs	GA	EC	C	fatty
HV (200 mg) + tannase (20 mg) - estimated GA content 32.5 wt %
Weight (mg)	0.4	141.4	3.1	65.0	3.1	6.9
Major components	polar	hydrolysable tannins	EGC+GC	GA	C	fatty

3.2. Dynamic mechanical analysis of dentin matrix

Figure 3 summarizes the results of the mechanical properties, E’ and tan δ, of the dentin matrix after biomodification with the crude vs. the GAKO extracts as a function of time. The de-galloylation of crude extracts resulted in significant reduction of the storage modulus for the CS_GAKO (p < 0.0005), VV_GAKO (p < 0.014), and HV_GAKO groups (p < 0.0005). The most significant reduction (58%) was found for CS, followed by HV (50%) and VV (27%). Interestingly, the damping capacity (tan δ) that represents the ability of the dentin matrix to dissipate energy, was not affected by the de-galloylation [CS (p = 0.78), VV (p = 0.84), HV (p = 0.13)]. A significant statistical interaction occurred between the studied factors (treatment vs. time) for E’ (p < 0.001) and tan δ (p < 0.022). Extracts from both the C and the GAKO groups significantly increased the E’ of the dentin matrix, up to 9.6-fold (p < 0.001), when compared to the control group. All groups showed stable E’ values up to 6 months, except for the CS_C group, which significantly decreased after one month (43%, p < 0.002), before reaching a plateau (p = 1.0). Biomodification with GAKO extracts presented sustained E’ values (p > 0.174). The damping capacity of groups CS_C, CS_GAKO, VV_C, VV_GAKO, and HV_C decreased significantly after one month (p < 0.01) but remained stable thereafter for up to 6 months (p > 0.1). Conversely, the damping capacity of the HV_GAKO group was stable within the evaluated timeframe (p > 0.052).

Figure 3.

Viscoelastic mechanical properties of biomodified dentin matrices. Mean and standard deviation of (A) storage modulus (E’) and (B) Tan δ of PACs extracts and their GAKO analogues. Lack of statistically significant differences within groups is depicted by a horizontal bar. Symbol (*) above groups represents significant differences between crude groups and their GAKO counterparts (p < 0.05). CS: Camellia sinensis, VV: Vitis vinifera, HV: Hamamelis virginiana, C: crude, GAKO: gallic acid knock out.

3.3. Biodegradation of the dentin matrix by exogenous proteases

Biomodified dentin showed significantly lower biodegradation rates (15 to 24% mass loss) than the control group (78% mass loss, p < 0.001), regardless of the extract (Figure 4). De-galloylation did not affect the biodegradability of biomodified the dentin matrix (p > 0.8).

Figure 4.

Biodegradation rates (%) of dentin matrices upon PACs-mediated biomodification. Symbol (*) depicts statistical differences between groups (p < 0.05). CS: Camellia sinensis, VV: Vitis vinifera, HV: Hamamelis virginiana.

4. Discussion

The elucidation of the mechanism of interactions between PACs and dentin extracellular matrix (ECM) is fundamental to effectively tailor the clinical usage of this very promising class of bioactive agents and to assess their long-term effectiveness. Two key predictive outcomes are the enhancement of dentin’s biomechanical properties, often characterized by measurements of the modulus of elasticity, and the biodegradability of the dentin ECM. Herein, a two-step de-galloylation method using enzymatic hydrolysis and countercurrent separation effectively removed (“knocked out”) gallic acid after liberating the galloyl residues from galloyl-rich PAC-containing extracts. This new approach enabled the study of the role of galloylation on dentin biomodification and stability. As determined by the dynamic mechanical analysis of dentin, the de-galloylation (gallic acid knock out) reduced the biomodification potency of all investigated extracts (Camellia sinensis, Vitis vinifera, and Hamamelis virginiana). Camellia sinensis, with the highest content of galloyl moieties, showed the largest mechanical reduction (58%) after de-galloylation, closely followed by Vitis vinifera (50%) and Hamamelis virginiana (27%). Furthermore, long-term evaluations showed dentin ECM treated with CS_C crude extract exhibited a significant decrease in the mechanical properties within only 1-month incubation. However, the protection against biodegradation remained unchanged (Figure 4). Because the galloyl knock-out affected the immediate and long-term mechanical outcome with no influence on the dentin biodegradability, only the first null hypothesis was rejected.

Chemical analysis of CS_C extract by HPLC and NMR revealed a content of ~35% EGCg with smaller amounts of epicatechin gallate (ECg), gallocatechin gallate (GCg), and catechin gallate (Cg), in monomeric and dimeric forms. Given such a high galloyl content, CS_C extract met the expectation of being the most active (9.6-fold increase in storage modulus) among all investigated extracts. At the same time, CS_C also exhibited the largest drop in mechanical properties, as early as one-month post-treatment (43% reduction). This revealed the relative instability of interactions of galloylated PACs with the dentin ECM. The removal of galloyl moieties from the CS extract, yielding CS_GAKO, resulted in reduced initial bioactivity of CS (58% lower storage modulus compared to the crude analog), whereas the mechanical properties remained stable after 6-month incubation. The lack of stability found in the CS_C group can be explained by the intrinsic susceptibility of galloylated PACs to hydrolysis [23], even when bound to the dentin matrix. The mechanism involved is analogous to the process catalyzed by the tannase used in the de-galloylation process and can be both a chemical process and/or mediated by dentin ECM-bound hydrolases. Tannase is a naturally occurring enzyme that cleaves ester and depside linkages in hydrolysable and condensed tannins [28,29]. As a result, gallate esters of tannins (i.e., gallotannins) and phenolic compounds such as EGCg transform into their free OH counterparts (EGC from EGCg), releasing gallic acid [29]. This conversion decreases the binding ability of tannins with protein macromolecules [30]. Other events that could alter galloylated PAC-dentin interaction are epimerization of the catechin C-rings and, likely more prominent in gallates, oxidation of the trihydroxy motif on the B-ring of PAC monomers in the gallo- series (e.g., gallocatechin; notably distinct from galloyl PACs) [19–21].

Conversely, the instability of interactions between galloylated PACs and dentin was not observed in the VV and HV groups. Certain plant species and natural sources present prevalence of particular structural motifs. Here, Vitis vinifera extract showed a robust and stable enhancement of the dentin matrix mechanical properties, which was expected for two highly-contributing reasons: due to the small amount of galloyl groups (predominantly from monomeric ECg, not from the more potent oligomers) [3,6,11]; and, due to the higher content of non-galloylated oligomers PACs with higher-order degree of polymerization, known to mediate covalent and covalent-like bonds with type I collagen [3]. Given the lower content of GA, the mechanical reduction found in the VV_C group after 6 months (24.3%) may be attributed to the in-situ hydrolysis of galloylated PACs, although the decreased E’ values were not statistically significant.

Dentin biomodification mediated with HV_C showed a modest yet sustained activity resulting in a 4.8-fold increase in storage modulus. In contrast to VV and CS, HV exhibits a high content of hydrolysable tannins (80%), such as ellagitanins and gallotannins and a small amount (~3%) of condensed tannins (PACs like EGCg, GCg, and Cg) [31]. Two reasons can explain this finding: first, the amount of galloyl groups from PACs in HV is too small; and secondly, hydrolyzable tannins interact with proteins through hydrophilic or hydrophobic connections [32]. Hydrophobic interactions generate hydrophobic pockets that ultimately protects adjacent H-bonds from hydrolysis [4] and likely could account for the mechanical stability found in the HV_C group. Interestingly, higher enzyme concentrations and longer fermentation times were needed to achieve a complete hydrolysis of the HV_C extract. This points to a possible stabilizing mechanism resulting from the particular molecular configuration of the hydrolysable tannins in HV. Another reason may be that the enzyme does not have the structural affinity and capacity to hydrolyze several galloyl moieties from the same molecule, thus requiring more time and higher enzyme concentration. Further investigations are necessary to determine the nature of these potentially hydrolysis resistant molecular properties.

The viscoelastic properties of dentin can be influenced by the intrinsic water dynamics [33], the molecular structure of the PAC [5], and the type of molecular interaction between PACs and collagen [27]. After biomodification, de-galloylation did not change the damping capacity (tan δ) of any of the PAC-treated groups. Further, all groups except HV_GAKO and control followed the trend characterized by a drop after 1 month, before reaching a plateau. Lower tan δ has been associated with collagen crosslinking predominantly via covalent intermolecular linkages [27,33], due to their strong and highly stable nature [33]. After 6 months, the damping capacity of control group showed a clear trend leading to an increased tan δ, revealing a more viscous-like behavior. Variations in viscoelastic response, observed in control group are likely due to initial signs of degradation, as a result of the inherent endogenous activity of metalloproteinases [27,34].

Overall, the treatment with crude and GAKO extracts significantly reduced the biodegradation rates of dentin matrix from 78%, observed in control group, to 19% (Figure 4). The lack of significant differences among the extracts revealed that the protective mechanism of PACs does not relate to PAC galloylation. These non-specific protections stem from possible direct inhibition of enzymes [35] to non-specific collagen binding that could form a surface coating hiding protein cleavage sites [4].

In summary, while galloyl moieties in PACs participate in the interaction with the dentin matrix, the instability of their ester linkages leads to a marked loss of biomodification effect over a relatively short period of time. The short-term stable biomodification of the galloyl-rich hydrolyzable tannins of HV were attributed to the formation of protective hydrophobic pockets [36]. However, the biomechanical data obtained after selective and effective two-step de-galloylation demonstrated that HV relied greatly on the galloylated compounds for dentin biomodification, thereby raising uncertainty of their clinical predictability.

The hydrolysis of EGCg and galloylated PAC oligomers is associated with the release of gallic acid [23] and potentially other by-products. PAC leaching and loss of activity can result in negative side effects, including failure of therapy and unknown anticipated effects.

Thus, the 2-step de-galloylation method effectively removed the galloyl moieties from three different PAC extracts. The presented de-galloylation protocol offers a new methodology for the targeted elimination of galloyl groups from active extracts with high galloyl content. As this approach eliminates one element of chemical variability/instability for biological inquiries, it can be applied in future studies to produce biomaterials with a more predictable and/or controlled bioactive performance. By confirming a less favorable biomodification potential of galloylated PACs, the results spark the need for further probing of defined structural elements of PACs, via establishment of structure-activity-relationships for advancing their development as a tailored and controlled multi-functional biomaterial.