2.1

Reagents

Unless otherwise stated, all chemicals were purchased from Sigma-Aldrich (St. Louis, MO, USA), including the 6 monomeric species (+)-catechin (pCT), (−)-catechin (CT), (−)-epicatechin (EC), (−)-epigallocatechin (EGC), (−)-epicatechin gallate (ECG) and (−)-epigallocatechin gallate (EGCG) ( Fig. 1 ). The representative EC–EC dimer, procyanidin B2 was purchased from Chromadex (Irvine, CA, USA). The representative pCT–pCT dimer ( Fig. 1 ) was synthesized from pCT following a published procedure . MegaNatural ^® Gold grape seed extract (Lot #: 05592502-01) was donated by the manufacturer (Polyphenolics, Madera, CA, USA). All reagents were used as received.

2.2

Preparation and characterization of PALM and PAHM

GSE was separated into a low molecular weight fraction (PALM) and high molecular weight fraction (PAHM) using adapted methods of extraction and preparative size exclusion chromatography (SEC) . In a typical procedure, 2.3 g of grape seed extract was dissolved in 200 mL of deionized water, which was subsequently extracted three times with ethyl acetate. The organic phase was lyophilized to remove the solvent, and approximately 0.27 g of residue was collected as PALM. The water phase was air-dried, re-dissolved in 20 mL of methanol, and loaded on a column (250 × 16 mm internal diameter) packed with Toyopearl TSK HW-40F resin (Tosoh, Japan). The column was sequentially eluted by methanol, water and acetone/water mixtures (20/80, 30/70, 40/60, 60/40, v/v). The resultant fractions were lyophilized and the last fraction (approximately 0.7 g) was designated as PAHM.

PALM, GSE and PAHM were subject to molecular weight analysis using gel permeation chromatography (GPC) following acetylation . Typically, the GSE-derived fraction (20 mg) was dissolved in a mixture of acetic anhydride (10 mL) and pyridine (10 mL, anhydrous) under N ₂ . The reaction mixture was stirred at room temperature for 72 h. Ice-cold water (∼300 mL) was then added to the reaction mixture leading to precipitation of acetylated product. The precipitates were collected by centrifugation and washed a few times with distilled water and methanol. GPC analysis were performed on a Tosoh Ecosec HLC-8320GPC system with three detectors used for measurements including a differential refractometer, a light scattering detector, and a UV detector . The mobile phase was tetrahydrofuran at 0.3 mL/min. Calibration of the system was performed with five different polystyrene standards (8000 to 90,000 Da).

2.3

Preparation of treatment and collagenase solutions

The treatment solutions were prepared by dissolving the 11 compounds to be tested, including pCT, CT, EC, EGC, ECG, EGCG, EC–EC dimer, pCT–pCT dimer, PALM, GSE and PAHM in ethanol/water (50/50, v/v) to a final concentration of 1 wt%. All treatment solutions had similar pH between 5 and 6 due to the weak acidity of phenol groups. For the collagenase solution, a TESCA buffer was first prepared by the addition of 11.5 g N -tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid, 50 mg sodium azide and 53 mg CaCl ₂ ·2H ₂ O into distilled water to a total volume of 1000 mL, followed by pH adjustment to 7.4. Then 1 g of collagenase with a molecular weight of ∼110 kDa (Collagenase Type I, Clostridiopeptidase A from Clostridium histolyticum, 125 U/mg) was dissolved in the TESCA buffer to a final concentration of 0.1% (w/v).

2.4

Preparation, demineralization and cross-linking of dentin films

Six non-carious human molars were collected after obtaining the patients’ informed consent under a protocol approved by the University of Missouri-Kansas City Adult Health Sciences IRB (IBC#12–14). Extracted teeth were stored at 4 °C in 0.96% (w/v) phosphate buffered saline containing 0.002% sodium azide. A water-cooled low-speed diamond saw (Buehler, Lake Bluff, IL, USA) was used to remove the occlusal one-third to one-half of the crown, followed by four additional cuts in the occlusal-apical direction to remove all side walls of the enamel. The resultant dentin block was sectioned in the mesial-distal direction, with a tungsten carbide knife mounted on an SM2500S microtome (Leica, Deerfield, IL, USA), into dentin films 6 μm thick. A total of 300 films (50 from each tooth) were obtained, and the final size of each film was uniform and approximately 5 mm × 5 mm.

Sixty dentin films were randomly assigned to 12 groups ( n = 5), including 11 treated groups and 1 control group. Each dentin film was demineralized with 35 wt% phosphoric acid for 15 s, rinsed in deionized water for 10 s, and then spread on a plastic cover slip (Fisher Scientific, Pittsburgh, PA, USA) with the aid of a fine paintbrush. After blotting away the excessive water, a small drop (approx. 0.15 mL) of selected treatment solution or deionized water (the control group) was immediately applied to cover the entire demineralized dentin film. After 1 min, the film was flushed with deionized water three times and further immersed in a copious amount of deionized water for 1 h to thoroughly remove any residual treatment solution. After being air dried overnight, the films were subject to the FTIR and collagenase digestion analyses.

2.5

FTIR spectroscopy

FTIR spectra of control and treated demineralized dentin films were collected at a resolution of 4 cm ⁻¹ and 128 scans per sample using a Fourier transformed infrared spectrometer equipped with an attenuated total reflectance (ATR) attachment (Spectrum One, Perkin-Elmer, Waltham, MA, USA). The ATR crystal was diamond with a transmission range between 650 and 4000 cm ⁻¹ , and a gauge force of 75 was applied to ensure a good contact between the films and the ATR top-plate. Area determination was performed for bands at 1235 cm ⁻¹ (amide III) and 1450 cm ⁻¹ (CH ₂ scissoring) using Spectrum software (Perkin Elmer, Waltham, MA, USA) following a two-point baseline correction and band area integration, and the band ratio (A1235/A1450) was calculated.

2.6

Resistance toward bacterial collagenase digestion

Following FTIR spectroscopy, the dentin collagen films were individually incubated in 30 μl of collagenase at 37 °C for 1 h, after which the remaining films were removed and the digest liquids were incubated for additional 24 h to ensure complete breakdown of solubilized collagen to tripeptides. The percentage of dentin film that was digested and released into the supernatants was gauged by MALDI–TOF mass spectroscopy following a published protocol , which is based on the signal strength of glycine–proline–arginine, a tripeptide resulted from the digestion of collagen by bacterial collagenase.

2.7

Preparation, demineralization and cross-linking of dentin beams

Like dentin films, dentin beams were obtained from non-carious human molars collected with patients’ consent. After removing the occlusal one-third to one-half of the crown, the same saw was used to make perpendicular cuts into the dentin surface at ∼1.1 mm increments. A single cut was then made parallel to and about 1.1 mm beneath the surface dentin, which freed the dentin beams from the remaining dentin block. The final cross-section of the beams was approximately 0.8 × 0.8 mm, and the length varied with respect to their position. All beams were carefully checked under light microscopy to make sure there was no remaining enamel or any defect. Eventually, a total of 11 teeth were processed into 63 dentin beams, which were pooled together and subject to complete demineralization in 10% phosphoric acid for 6 h . The resultant demineralized dentin beams were randomly assigned to 9 groups ( n = 7), including 1 control group and 8 treated groups. Beams in the treated groups were immersed in selected solutions that were found to increase dentin films’ resistance to biodegradation (see Section 3 ), including EGCG, PALM, GSE and PAHM for 2 h or 24 h. After treatment, the beams were thoroughly rinsed with deionized water and continued to be immersed in a large amount of deionized water for 24 h to remove residual treatment solution. The beams were stored in deionized water at 4 °C until tensile testing.

2.8

Tensile tests of dentin beams

The tensile properties of the beams were determined using an SSTM-5000 tensile tester (United Calibration Corporation, CA, USA) at a fixed gauge length of 2 mm. Specimens were mounted to the upper and lower grips of the tensile tester using a cyanoacrylate adhesive (Zapit, Dental Ventures of America, Corona, CA, USA). A constant strain rate of 0.5 mm/min was applied to stretch the dentin beams till failure. The direction of load was perpendicular to dentinal tubules owing to the manner of beam harvesting. During the test, specimens were kept fully hydrated with a water mist sprayer. Three tensile properties were retrieved from the stress–strain curves. First, the slope of the initial linear portion (up to strain of 1.5–5% depending upon treatment conditions) was calculated, representing the “initial modulus” ( E _i ). Second, the elastic modulus ( E ) was determined as the slope of the later linear portion, usually extending all to way to where failure occurred. Third, the maximum stress at the point of failure (correspondingly the peak value of the stress–strain curve) was recorded as the ultimate tensile strength (UTS).

2.9

Statistical analysis

Statistical analyses for FTIR band ratio and collagenase digestion were performed by one-way ANOVA and Tukey’s post hoc test. For tensile properties, the interaction between treatment agent and treatment time was analyzed by two-way ANOVA. The comparison of means across treatment agents was carried out by split file one-way ANOVA and either Tukey’s post hoc test (equal variance) or Dunnett’s T3 post hoc test (unequal variance). The comparison of means across treatment times was done by t -test. All statistical analyses were performed at 95% confidence level (differences were considered significant if p < 0.05).

2.1

Reagents

Unless otherwise stated, all chemicals were purchased from Sigma-Aldrich (St. Louis, MO, USA), including the 6 monomeric species (+)-catechin (pCT), (−)-catechin (CT), (−)-epicatechin (EC), (−)-epigallocatechin (EGC), (−)-epicatechin gallate (ECG) and (−)-epigallocatechin gallate (EGCG) ( Fig. 1 ). The representative EC–EC dimer, procyanidin B2 was purchased from Chromadex (Irvine, CA, USA). The representative pCT–pCT dimer ( Fig. 1 ) was synthesized from pCT following a published procedure . MegaNatural ^® Gold grape seed extract (Lot #: 05592502-01) was donated by the manufacturer (Polyphenolics, Madera, CA, USA). All reagents were used as received.

2.2

Preparation and characterization of PALM and PAHM

GSE was separated into a low molecular weight fraction (PALM) and high molecular weight fraction (PAHM) using adapted methods of extraction and preparative size exclusion chromatography (SEC) . In a typical procedure, 2.3 g of grape seed extract was dissolved in 200 mL of deionized water, which was subsequently extracted three times with ethyl acetate. The organic phase was lyophilized to remove the solvent, and approximately 0.27 g of residue was collected as PALM. The water phase was air-dried, re-dissolved in 20 mL of methanol, and loaded on a column (250 × 16 mm internal diameter) packed with Toyopearl TSK HW-40F resin (Tosoh, Japan). The column was sequentially eluted by methanol, water and acetone/water mixtures (20/80, 30/70, 40/60, 60/40, v/v). The resultant fractions were lyophilized and the last fraction (approximately 0.7 g) was designated as PAHM.

PALM, GSE and PAHM were subject to molecular weight analysis using gel permeation chromatography (GPC) following acetylation . Typically, the GSE-derived fraction (20 mg) was dissolved in a mixture of acetic anhydride (10 mL) and pyridine (10 mL, anhydrous) under N ₂ . The reaction mixture was stirred at room temperature for 72 h. Ice-cold water (∼300 mL) was then added to the reaction mixture leading to precipitation of acetylated product. The precipitates were collected by centrifugation and washed a few times with distilled water and methanol. GPC analysis were performed on a Tosoh Ecosec HLC-8320GPC system with three detectors used for measurements including a differential refractometer, a light scattering detector, and a UV detector . The mobile phase was tetrahydrofuran at 0.3 mL/min. Calibration of the system was performed with five different polystyrene standards (8000 to 90,000 Da).

2.3

Preparation of treatment and collagenase solutions

The treatment solutions were prepared by dissolving the 11 compounds to be tested, including pCT, CT, EC, EGC, ECG, EGCG, EC–EC dimer, pCT–pCT dimer, PALM, GSE and PAHM in ethanol/water (50/50, v/v) to a final concentration of 1 wt%. All treatment solutions had similar pH between 5 and 6 due to the weak acidity of phenol groups. For the collagenase solution, a TESCA buffer was first prepared by the addition of 11.5 g N -tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid, 50 mg sodium azide and 53 mg CaCl ₂ ·2H ₂ O into distilled water to a total volume of 1000 mL, followed by pH adjustment to 7.4. Then 1 g of collagenase with a molecular weight of ∼110 kDa (Collagenase Type I, Clostridiopeptidase A from Clostridium histolyticum, 125 U/mg) was dissolved in the TESCA buffer to a final concentration of 0.1% (w/v).

2.4

Preparation, demineralization and cross-linking of dentin films

Six non-carious human molars were collected after obtaining the patients’ informed consent under a protocol approved by the University of Missouri-Kansas City Adult Health Sciences IRB (IBC#12–14). Extracted teeth were stored at 4 °C in 0.96% (w/v) phosphate buffered saline containing 0.002% sodium azide. A water-cooled low-speed diamond saw (Buehler, Lake Bluff, IL, USA) was used to remove the occlusal one-third to one-half of the crown, followed by four additional cuts in the occlusal-apical direction to remove all side walls of the enamel. The resultant dentin block was sectioned in the mesial-distal direction, with a tungsten carbide knife mounted on an SM2500S microtome (Leica, Deerfield, IL, USA), into dentin films 6 μm thick. A total of 300 films (50 from each tooth) were obtained, and the final size of each film was uniform and approximately 5 mm × 5 mm.

Sixty dentin films were randomly assigned to 12 groups ( n = 5), including 11 treated groups and 1 control group. Each dentin film was demineralized with 35 wt% phosphoric acid for 15 s, rinsed in deionized water for 10 s, and then spread on a plastic cover slip (Fisher Scientific, Pittsburgh, PA, USA) with the aid of a fine paintbrush. After blotting away the excessive water, a small drop (approx. 0.15 mL) of selected treatment solution or deionized water (the control group) was immediately applied to cover the entire demineralized dentin film. After 1 min, the film was flushed with deionized water three times and further immersed in a copious amount of deionized water for 1 h to thoroughly remove any residual treatment solution. After being air dried overnight, the films were subject to the FTIR and collagenase digestion analyses.

2.5

FTIR spectroscopy

FTIR spectra of control and treated demineralized dentin films were collected at a resolution of 4 cm ⁻¹ and 128 scans per sample using a Fourier transformed infrared spectrometer equipped with an attenuated total reflectance (ATR) attachment (Spectrum One, Perkin-Elmer, Waltham, MA, USA). The ATR crystal was diamond with a transmission range between 650 and 4000 cm ⁻¹ , and a gauge force of 75 was applied to ensure a good contact between the films and the ATR top-plate. Area determination was performed for bands at 1235 cm ⁻¹ (amide III) and 1450 cm ⁻¹ (CH ₂ scissoring) using Spectrum software (Perkin Elmer, Waltham, MA, USA) following a two-point baseline correction and band area integration, and the band ratio (A1235/A1450) was calculated.

2.6

Resistance toward bacterial collagenase digestion

Following FTIR spectroscopy, the dentin collagen films were individually incubated in 30 μl of collagenase at 37 °C for 1 h, after which the remaining films were removed and the digest liquids were incubated for additional 24 h to ensure complete breakdown of solubilized collagen to tripeptides. The percentage of dentin film that was digested and released into the supernatants was gauged by MALDI–TOF mass spectroscopy following a published protocol , which is based on the signal strength of glycine–proline–arginine, a tripeptide resulted from the digestion of collagen by bacterial collagenase.

2.7

Preparation, demineralization and cross-linking of dentin beams

Like dentin films, dentin beams were obtained from non-carious human molars collected with patients’ consent. After removing the occlusal one-third to one-half of the crown, the same saw was used to make perpendicular cuts into the dentin surface at ∼1.1 mm increments. A single cut was then made parallel to and about 1.1 mm beneath the surface dentin, which freed the dentin beams from the remaining dentin block. The final cross-section of the beams was approximately 0.8 × 0.8 mm, and the length varied with respect to their position. All beams were carefully checked under light microscopy to make sure there was no remaining enamel or any defect. Eventually, a total of 11 teeth were processed into 63 dentin beams, which were pooled together and subject to complete demineralization in 10% phosphoric acid for 6 h . The resultant demineralized dentin beams were randomly assigned to 9 groups ( n = 7), including 1 control group and 8 treated groups. Beams in the treated groups were immersed in selected solutions that were found to increase dentin films’ resistance to biodegradation (see Section 3 ), including EGCG, PALM, GSE and PAHM for 2 h or 24 h. After treatment, the beams were thoroughly rinsed with deionized water and continued to be immersed in a large amount of deionized water for 24 h to remove residual treatment solution. The beams were stored in deionized water at 4 °C until tensile testing.

2.8

Tensile tests of dentin beams

The tensile properties of the beams were determined using an SSTM-5000 tensile tester (United Calibration Corporation, CA, USA) at a fixed gauge length of 2 mm. Specimens were mounted to the upper and lower grips of the tensile tester using a cyanoacrylate adhesive (Zapit, Dental Ventures of America, Corona, CA, USA). A constant strain rate of 0.5 mm/min was applied to stretch the dentin beams till failure. The direction of load was perpendicular to dentinal tubules owing to the manner of beam harvesting. During the test, specimens were kept fully hydrated with a water mist sprayer. Three tensile properties were retrieved from the stress–strain curves. First, the slope of the initial linear portion (up to strain of 1.5–5% depending upon treatment conditions) was calculated, representing the “initial modulus” ( E _i ). Second, the elastic modulus ( E ) was determined as the slope of the later linear portion, usually extending all to way to where failure occurred. Third, the maximum stress at the point of failure (correspondingly the peak value of the stress–strain curve) was recorded as the ultimate tensile strength (UTS).

2.9

Statistical analysis

Statistical analyses for FTIR band ratio and collagenase digestion were performed by one-way ANOVA and Tukey’s post hoc test. For tensile properties, the interaction between treatment agent and treatment time was analyzed by two-way ANOVA. The comparison of means across treatment agents was carried out by split file one-way ANOVA and either Tukey’s post hoc test (equal variance) or Dunnett’s T3 post hoc test (unequal variance). The comparison of means across treatment times was done by t -test. All statistical analyses were performed at 95% confidence level (differences were considered significant if p < 0.05).