All solvents were of ACS or HPLC grade and were obtained from P.O.Ch. Company (Gliwice, Poland) or Merck (Darmstadt, Germany). Sephadex LH-20, BSA (V fraction), gelatin (B chain), tannic acid, (+)-catechin, vanillin, gallic acid, ferric chloride, and phosphate-buffered saline (PBS) were purchased from Sigma–Aldrich Co. Ltd. (Poznań, Poland). Pea protein isolate “Pisane” was obtained from Coscura Groupe Warcoing (Momalle, Belgium). All other reagents were of at least ACS grade and were acquired from P.O.Ch. Company (Gliwice, Poland). Tannase was kindly supplied by Mr. Nobouyshi Sato of Kikkoman Corporation (Elgin, IL, USA).
Broad bean seeds (Vicia faba Major) of Hangdown white cultivar were purchased in Olzans-CN LLC (Olsztyn, Poland). Coats covering seeds were manually separated from cotyledons, and the proportion between seed coat and cotyledon in the weight of the whole seed was determined. Then, seed coats were subjected to extraction of phenolic compounds, whereas proteins were isolated from cotyledons.
Isolation and characterization of tannin fraction
Extraction and fractionation
The broad bean coats were ground in a coffee mill (BSH Bosch & Siemens Hausgeräte GmbH, Munich, Germany) into fine powder (particle size <0.8 mm). A 60-g portion of broad bean coat powder was extracted using 80% (v/v) aqueous acetone at a solid-to-solvent ratio of 1:8 in a thermostatic shaking water bath (357 Elpan, Lubawa, Poland), at 60 °C for 15 min. Then, the supernate was filtered through filter paper and the extraction step was repeated twice more. The supernatants were combined, acetone was evaporated using Büchi Rotavapor R-200 (Büchi Labortechnik, Flawil, Switzerland) at 40 °C, and aqueous residue was lyophilized (Lyph Lock 6 freeze dry system, Labconco, Kansas City, MO, USA). One gram of crude extract of broad bean coat phenolic compounds was suspended in 10 mL of 96% (v/v) ethanol and applied onto a chromatographic column (30 mm i.d. × 230 mm l) packed with lipophilic Sephadex LH-20 gel (Sigma–Aldrich). Firstly, low molecular weight phenolics were eluted gravimetrically using 96% (v/v) ethanol, and then, solvent was changed over to 50% (v/v) acetone in order to elute tannins. Acetone from tannin fraction was evaporated, and remaining water was lyophilized.
Total phenolics content (TPC)
The TPC of broad bean coat crude extract and tannin fraction was determined using colorimetric assay with Folin-Ciocalteu phenol reagent according to Naczk and Shahidi . Briefly, 0.25 mL of methanolic solution of extract or tannin fraction was mixed with 0.25 mL of Folin-Ciocalteu reagent (diluted 1:1 with distilled water), and then, 0.5 mL of sodium carbonate saturated solution and 4 mL of water was added, and mixture was vortexed thoroughly (Genie2, Scientific Industries, Bohemia, NY, USA). Absorbance at 725 nm after 30 min color development was measured with Beckman DU-7500 spectrophotometer (Beckman Coulter, Fullerton, CA, USA) with prior centrifugation of samples. TPC was expressed as mg (+)-catechin equivalents per gram of extract or fraction from triplicate measurements.
Condensed tannins content
The method of Price et al.  was employed to determine condensed tannins content of crude extract and tannin fraction. Methanolic solutions of samples (1 mL) at a concentration of 0.25 mg/mL were mixed with 5 mL of vanillin reagent (obtained by dissolving of 0.5 g vanillin in 100 mL 4% (v/v) concentrated hydrochloric acid). The absorbance of the mixture was measured after a 20-min period of reaction development at 500 nm using Beckman DU-7500 spectrophotometer. Due to the lack of appropriate standard for this assay, the results were expressed as absorbance units per g of sample (A
Hydrolysable tannin content
In order to estimate hydrolysable tannins, content of broad bean coat tannin fraction was subjected to enzymatic hydrolysis with tannase . Four milliliters of tannin fraction solution in citric buffer (50 mM, pH 5.5), at a concentration of 4 mg/mL, was mixed with 1 mL of tannase solution (50 μg, 50,000 U/g; in the same buffer). Samples were incubated at 30 °C for 15 min, and then, pH was adjusted to 2 with 2 M HCl solution. Gallic acid liberated during hydrolysis was extracted into diethyl ether, solvent was evaporated, and solid residue was dissolved in 2 mL of methanol. The parallel control sample without tannase addition was incubated and extracted. Gallic acid content in the hydrolyzed and control samples was determined using a Shimadzu HPLC system (Shimadzu Co., Kyoto, Japan) consisting of LC 10ADVp pump, SPD-M10AVp photodiode detector, SCL-10AVp controller. Twenty microliters of methanolic solutions of sample was filtered through a 0.45-μm nylon filter and injected onto LiChrospher 100 RP-18 column (4.6 × 250 mm, 5 μm; Merck, Darmstadt, Germany). The chromatographic separation was performed in isocratic system of a mobile phase consisting of water/acetonitrile/acetic acid (88:10:2; v/v/v) with a flow rate of 1 mL/min. Gallic acid was identified and quantified with reference to gallic acid standard solution. Then, the difference in the content of gallic acid in the tannin fraction after tannase hydrolysis and in non-hydrolyzed sample was calculated.
Isolation and characterization of protein fractions
Extraction and fractionation
Ground broad bean cotyledons (10 g) were extracted with 100 mL of 20 mM phosphate buffer at pH 7.5 containing 1 M NaCl at room temperature for 2 h. Then, the suspension was centrifuged for 30 min at 5,000×g and at a temperature of 4 °C using MPW 350R centrifuge (MPW Med. Instruments, Warsaw, Poland). Precipitate was discarded, and 500 mL of deionized water was added to the supernate, and after 1 h, the suspension was centrifuged (5,000×g, 30 min). The precipitate obtained in that manner contained globulins, whereas albumins were present in the supernate. The globulins precipitate was dissolved in phosphate buffer and precipitated with 35% saturated ammonium sulfate, whereas to precipitate albumins from the supernate, 80% saturated ammonium sulfate was added and suspensions were centrifuged. Both globulins and albumins precipitates were dialyzed against deionized water using dialysis tubings (Sigma–Aldrich) with molecular cutoff of 12,400 Da at 4 °C for 48 h. Then, the protein extracts were lyophilized and chromatographically purified. Protein extract was dissolved in the mobile phase (50 mM phosphate buffer pH 6.9 containing 0.15 M NaCl and 0.1% sodium azide) at a concentration of 25 mg/mL and loaded onto the column (26 mm i.d. × 100 cm l) filled with Sephadex G-200 superfine gel (Sigma–Aldrich). The mobile phase was delivered to the column at a flow rate of 0.5 mL/min by peristaltic pump (Unipan 315, Warsaw, Poland). The fractions of 5 mL volume were collected using Redi Frac fraction collector (GE Healthcare). The elute was monitored at 220 and 280 nm with Beckman DU-7500 spectrophotometer. Once the chromatogram of protein separation was prepared, three individual fractions (I, II, III) were pooled and dialyzed. In order to estimate molecular weight of individual fractions, the column was calibrated with molecular weight marker consisting of the following: ferritin, aldolase, bovine serum albumin, chymotripsinogen A, cytochrome C (Sigma–Aldrich) with molecular weights of 440,000, 270,000, 67,000, 25,000, and 12,400 Da, respectively. The major fractions prepared according to above-mentioned procedure in five replicate processes were pooled once SDS–PAGE analysis confirmed the similarity of protein composition in the five preparations.
Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS–PAGE)
The molecular weight distribution of broad bean protein fractions was analyzed in Laemmli buffer system according to Hames  using Mini PROTEAN Tetra cell system (BioRad, Hercules, CA, USA). Fractions were dissolved in 62.5 mM Tris–HCl buffer at pH 6.8 with 2% SDS, 5% β-mercaptoethanol, 10% glycerol, and 0.002% bromophenol blue and heated for 5 min in boiling water. Cooled samples (20 μg) were loaded into each well of 12.5% polyacrylamide gel and separation was performed at 25 mA. Low Range Sigma Molecular Marker 6,500–66,000 Da (Sigma–Aldrich) was used to estimate molecular weights of individual protein subunits.
Bradford method  was employed to assess protein content of isolated fractions. The obtained results of protein content were utilized to calculate precipitating capacity per gram of protein in protein fraction.
Precipitating potential assay
The ability of broad bean coat tannins to precipitate isolated broad bean protein fractions as well as model proteins, i.e., BSA, gelatin, and pea proteins, was investigated employing the procedure described by Hagerman and Butler  with some modifications. Firstly, the effect of pH on the formation of phenolic protein insoluble complexes was investigated. Phenolic compounds solutions were prepared in 50% aqueous ethanol at a concentration of 1 mg/mL, whereas proteins were dissolved in McIvaine buffer at pH in the range from 2 to 8 and at a concentration of 1 mg/mL. The procedure of Hagerman and Butler  was scaled down: 200 μL of phenolic compounds solution was added to 400 μL of protein solution and mixed well. After 15 min of quiescent period, the reaction mixture was centrifuged at 4,000×g for 15 min (MPW Med. Instruments, Warsaw, Poland). The supernate was discarded, and the surface of the pellet and the walls of the tubes were rinsed with buffer to remove remained unbound phenolics. Then, the pellet was dissolved in 800 μL of sodium dodecyl sulfate (SDS)–triethanolamine solution (1% SDS and 5% triethanolamine), and 200 μL of 0.01 M ferric chloride solution (in 0.01 M HCl) was added. The reaction was developed for 15 min, and then, the absorbance at 510 nm was read against a reagent blank consisting of SDS–triethanolamine and ferric chloride mixture. Once the optimal pH for phenolic compounds–proteins complex formation was determined, the effect of phenolic compounds concentration on complex formation at optimal pH was assessed. The phenolic compound solutions at a concentration in the range from 0.1 to 3.0 mg/mL were mixed with individual proteins, and the amount of complexes formed was determined. The plot of phenolic compounds concentration expressed as mg catechin equivalents per mg of protein versus A
510 was prepared. The protein precipitation potential was expressed as the linear regression coefficient and presented in the comparison with values obtained for tannic acid.
Fluorescence quenching method
The interactions between tannin fraction and protein fraction yielding soluble complexes were investigated using fluorescence quenching method . The fluorescence quenching involves a reduction in fluorophore fluorescence in the presence of quencher. All measurements were taken in quartz cuvette (1.0 × 1.0 × 4.0 cm) using Perkin Elmer LS 50B fluorescence spectrometer (Beaconsfield, Great Britain). Fluorescence emission spectra were recorded in the wavelength range of 285–500 nm by exciting protein at excitation wavelength (λex) of 282 nm. The slit width for both excitation and emission was set to 5 nm. To determine the linear concentration range for protein fluorescence, a series of protein fractions and BSA solutions with increasing concentration were prepared in PBS buffer of pH 7.4. Suitable protein concentration was chosen for fluorescence quenching experiments. To 3 mL of protein solution, portions of 6 μL of tannin fraction solution (1 mg/mL) were added and the mixture was shaken. The changes of fluorescence intensity were measured within 30 s after addition. All fluorescence readings were corrected for protein dilution effect. The titration was performed in four replications. In order to avoid artifact quenching, the broad bean tannin fraction solution was checked for its intrinsic fluorescence. All measurements were taken at room temperature.
The results were expressed as mean values ± standard deviation from at least three replicates. The statistical analyses of data (linear regression analysis, standard errors of slopes) were performed using GraphPad Prism 5 (GraphPad Software Inc., San Diego, CA, USA). The determination of the regression coefficient (slope) values was based on the analysis of experimental data involving the precipitation of tannin–protein complexes at a minimum of five different tannin concentrations in the assay mixture.