Determination of free and bound phenolic compounds and their antioxidant activity in buckwheat bread loaf, crust and crumb

When citing, please refer to the published version. This is the final peer-reviewed accepted manuscript of: Vito Verardo, Virginia Glicerina, Emiliano Cocci, Antonia Garrido Frenich, Santina Romani, Maria Fiorenza Caboni, Determination of free and bound phenolic compounds and their antioxidant activity in buckwheat bread loaf, crust and crumb, LWT, Volume 87, 2018, Pages 217-224, ISSN 0023-6438, https://www.sciencedirect.com/science/article/pii/S0023643817306333


Introduction 44
Wheat bread is considered to be a good source of energy for the human body. It is known that 45 bread obtained with natural raw ingredients such as cereals and seeds, spices, herbs and parts 46 of green plants, fruit or vegetable products and waste products from the food industry can be 47 M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT 6 Grandi Cucine S.p.A, Carpi, Italy). After baking, bread samples were left to cool at room 119 temperature for 2 hours, before performing analysis. The volume of the resulting bread (mL/g) 120 was 3.3±0.9 for the control; 3.2±0.5 for the bread obtained with the addition of 10% of 121 buckwheat flour; 2.9±0.1 for the product obtained with the addition of 20% of buckwheat 122 flour and 2.5±0.2 for the bread with the 30% of buckwheat flour. 123 Each type of bread was produced in triplicate. After baking, crust and crumb were separated 124 from each bread sample, frozen in encoded plastic bags at -20° C and then freeze-dried 125 (Thermo HETO, powerdry LYOLAB 3000; Waltham, USA). Dried samples were ground to a 126 fine powder in a blender mixer (Ika-Werke M20; Staufen, Germany) and used for the analyses. 127 128

Extraction of free and bound phenolic compounds from control and buckwheat bread 136
To determine the free and bound phenolic fraction of bread samples, the method developed by 137 Verardo et al. (2011) was applied. 138 Briefly, two grams of bread were extracted twice in an ultrasonic bath with a solution of 139 ethanol/water (4:1 mL/mL). The supernatants were collected, evaporated and reconstituted 140 with 2 mL of methanol/water (1:1 mL/mL). The extracts were stored at -18 °C until use. 141 To obtain the bound phenolic fraction, residues of free phenolic extraction were digested with 142 M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT 7 hydrolyzed solution was acified to pH 2-3 by adding 10 mol/L hydrochloric acid in a cooling 144 ice bath and extracted with 500 mL of hexane to remove the lipids. The final solution was 145 extracted five times with 100 mL of 1/1 diethyl ether/ethyl acetate (mL/mL). The organic 146 fractions were pooled and evaporated to dryness. The phenolic compounds were reconstituted 147 in 2 mL of methanol/water (1:1 mL/mL). 148 149 2.4. HPLC-ESI -MS analysis of phenolic compounds 150 HPLC analysis was performed by an Agilent 1100 series LC system (Agilent Technologies, 151 CA, USA) consisting of a vacuum degasser, autosampler, and a binary pump equipped with a 152 reversed-phase Kinetex TM C18 (100 mm × 4.6 mm, 2.6 µm) column (Phenomenex Inc,153 Torrance, CA, USA). The mobile phase and gradient program were used as previously 154 described by Gomez-Caravaca, Verardo, Berardinelli, Marconi and Caboni (2014). All 155 solvents were filtered with a 0.45 µm filter disk. The RP-HPLC system was coupled to a HP-156 Mass Spectrometer Detector (MSD, model G1946A) equipped with an ESI interface peak 157 integration and data elaboration were performed using Chemstation software (Hewlett 158 Packard, Wilmington, DE, USA). Parameters for analysis were set using negative ion mode 159 with spectra acquired over a mass range from m/z 50-1300. 160 161 2.5. DPPH radical scavenging activity 162 The free radical scavenging activity of extracts (FRSA) was determined using the DPPH 163 assay according to Parejo, Codina, Petrakisand and Kefalas (2000). Briefly, 0.1 mL of extract 164 was added to 2.9 mL of 100 µmol/L DPPH methanol solution. The absorbance was 165 determined at 517 nm after 30 minutes (at 25 °C). To assess the FRSA a Trolox calibration 166 curve was performed and the results were expressed as µmoles of Trolox equivalent/100 g of The spectrophotometric analyses were performed using a UV-1601 spectrophotometer from 169 Shimadzu (Duisburg, Germany). 170 171 2.6. ABTS Radical Cation Decolorization Assay 172 The ABTS assay, was performed by using the method previously described by Re et al. (1999) 173 where the radical monocation ABTS •+ is generated by oxidation of ABTS with potassium 174 persulfate and is reduced in the presence of hydrogen-donating antioxidants or with a standard. 175 Briefly, ABTS •+ was obtained by reaction of 7.0 mmol/L ABTS and 2.45 mmol/L potassium 176 persulfate (stand in the dark at room temperature for 16 h). The ABTS •+ stock solution was 177 diluted with water in order to obtain an absorbance of 0.700 ± 0.02 (λ = 734 nm). After that, 178 0.01 mL of sample extract was added to 1 mL of ABTS •+ and stored in a dark room for 10 179 minutes. The absorbance was measured at 734 nm (at 30 °C). A Trolox calibration curve was 180 performed and the results were expressed as µmoles of Trolox equivalent/100 g of bread d.w. 181  As expected, control bread formulated with refined wheat flour showed the apigenin-6-C-202 arabinoside-8-C-hexoside ((iso)-shaftoside) as principal free phenolic compound followed by 203 trans-ferulic acid. These data agreed with those reported by Gianotti et al. (2011). The content 204 of ferulic acid and other phenolic acids was in the same order of magnitude of that reported 205 by several authors (Abdel-Aal & Rabalski, 2013;Lu et al., 2014;Yu & Beta, 2015). 206 Buckwheat enriched breads, as control sample, showed the apigenin-6-C-arabinoside-8-C-207 hexoside isomers, however their content decreased when buckwheat flour ratio increased. 208 This trend is justified by the presence of these compounds in wheat flour but not in buckwheat 209 flour. Similar trend has been observed for the ferulic acid isomers (cis and trans). To better evaluate the distribution of phenolic compounds in control and buckwheat enriched 249 breads, the crumb was separated from the crust in each loaf samples; free phenolic fraction of 250 the two section of loaf is reported in Table 3. 251 Control and enriched samples showed that crust has the higher phenolic content compared to 252 the crumb. These data totally fitted with the results reported in other works (Balestra et al., 253 2011;Lu et al. 2014;Yu & Beta, 2015). According to Vitali, Dragojevic, and Sebecic (2009)  254 the high content of phenolic compounds in crust should be due to the effect of baking 255 temperature that probably hydrolyzes some complex phenols resulting in an increase of 256 extractable phenolic content. Anyway, Gélinas and McKinnon (2006) hypothesized that also 257 Maillard reactions are involved to some extent in the content of phenolic compounds in bread 258 crust. 259 Only four phenolic compounds were determined in control crumb; basically, apigenin-6-C-260 arabinoside-8-C-hexoside isomers represented more than 95 % of its total phenolic content; 261 minor phenolics were trans ferulic acid and p-hydroxybenzaldehyde. Twelve phenolic 262 compounds were quantified in control crust. Apigenin-6-C-arabinoside-8-C-hexoside was the 263 principal phenolic compound followed by the trans ferulic acid and their sum corresponds to 264 81.9% of total phenolic compounds. 265 To our knowledge, no studies on the distribution of phenolic compounds in buckwheat crumb M A N U S C R I P T

A C C E P T E D ACCEPTED MANUSCRIPT
The total phenolic content in the two loaf sections increased when higher buckwheat flour 268 ratio were used for formulation. 269 The first phenolic class in bread crumbs was the flavone; it represented the 69, 46 and 39 % 270 of total phenolic compounds in 10, 20 and 30 % buckwheat enriched bread, respectively. It is 271 important to underline that swertiamacroside was determined only in buckwheat bread 272 crumbs; probably, the high temperature in crust caused its degradation. The main phenolic class in crust was the phenolic acid group that increased with the increase 286 of buckwheat flour quantity. The same trend was reported by flavan-3-ols that was the second 287 phenolic class ranged from 4.7 to 11.2 % of total phenolic fraction. 288 It is important to underline that some compounds, such as quercetin and propelargonidins, 308 were not detected in bound form, probably due to their degradation and/or hydrolysis. 309 The results of the bound phenolic compounds determined in crumb and crust are given in 310 Table 5. 311 The substitution of wheat flour with buckwheat one caused, in general, significant increase of 312 phenolic acids content in bread crust; contrary, flavonoids content decreased. 313 Apigenin-6-C-arabinoside-8-C-hexoside was the main phenolic compounds in crust and 314 crumb, and its content decreased when higher amounts of buckwheat flour have been added in M A N U S C R I P T

A C C E P T E D ACCEPTED MANUSCRIPT
Trans ferulic acid was the second phenolic compound in bread crust; its content in crumb was 317 from 16 to 25 times lower than in crust. Moreover, the ferulic content in buckwheat breads 318 was lower than control bread. Similar trend has been showed by p-hydroxybenzaldehyde. 319 Crumbs of breads formulated with buckwheat flours contained increasing amounts of flavan-320 3-ols (from 12 to 29 mg/kg bread d.w.) according to the buckwheat flour ratio added during 321 bread formulation. 322 Rutin was also detected in bound form and its content was higher in buckwheat bread crumb 323 than crusts confirming the low thermal stability of this compound as previously reported for 324 its free form. 325 326

Antioxidant activity of bread samples 327
The results of antioxidant activity measured by DPPH and ABTS radicals, and total phenolic 328 content, expressed as sum of each phenolic compound determined by HPLC-ESI-MS, are 329 reported in Table 6. 330 Total free phenolic content in bread loaf varied between 109 and 235 mg/kg bread d.w., 331 buckwheat bread samples showed higher amounts of these compounds compared to control. 332 These data confirm the results showed in previous works (Angioloni, & Collar, 2011;333 Szawara-Nowak, Bączek & Zieliński, 2016) that demonstrate as the multigrain bread 334 exhibited increased polyphenol content, higher polyphenol bioaccessibility and higher 335 antioxidant power than bread obtained with single grains. 336 Significant correlations have been found between total phenolic content and DPPH and ABTS 337 assay results. Briefly, DPPH assay showed a positive correlation (r = 0.9354, p < 0.001) with 338 total phenolic amounts, and according to Yu et al. (2013) and Yu and Beta (2015), crusts 339 showed the highest scavenging activity due to the high phenolic content and probably to the M A N U S C R I P T

A C C E P T E D ACCEPTED MANUSCRIPT
Significant differences (p<0.05) were found among the buckwheat bread samples; in fact high 342 ratio of buckwheat flour correspond to high phenolic content and radical scavenging activity. 343 Similarly, ABTS assay reported a positive correlation (r = 0.9338, p < 0.001) with free total 344 phenolic content in breads. ABTS scavenging capacity was comprised between 118 and 899 345 µmoles TEAC/100 g bread d.w.; according to results obtained by Yu et al. (2013) and Yu and 346 Beta (2015), crust samples showed the highest antioxidant capacity, and buckwheat breads 347 scavenging capacity increased when high buckwheat ratio was used during the bread 348 formulation. 349 High correlation was found between DPPH and ABTS assay results (r = 0.9950, p< 0.001). 350 Total bound phenolic compounds content ranged between 77 and 213 mg/kg bread d.w. These 351 values are apparently in contrast with the data of other authors (Abdel-Aal & Rabalski, 2013;352 Yu & Beta, 2015) which showed that bound phenolic content was higher than free phenolic 353 content in bread obtained with wheat flour; according to Yu et al. (2013), the bound phenolic 354 content in refined flour could be more than ten times lower than in whole wheat flours. 355 Antioxidant activity, measured by DPPH and ABTS assays, decreased with the diminution of 356 buckwheat flour quantity. As noticed for free phenolic fraction, in each kind of bread crust 357 samples showed higher antioxidant activity than in loaf, and the last one presented higher 358 antioxidant activity than crumb. These results could be justified by a strong Maillard reaction 359 development in crust than in crumb. 360 Positive correlations were found between total bound phenolic content and DPPH (r = 0.7765, 361 p < 0.05) and between total bound phenolic content and ABTS (r = 0.8361, p < 0.05). 362 363

Conclusions 364
The phenolic composition of bread samples enriched with buckwheat flour was compared M A N U S C R I P T

A C C E P T E D ACCEPTED MANUSCRIPT
16 flavan-3-ols, propelargonidins and flavonols (i.e. rutin among others) in bread. In this way, 367 the buckwheat breads could represent "functional" breads, permitting to introduce these 368 bioactive phenolic classes in the diet. 369 Bread crust contains higher amounts of phenolic compounds and higher antioxidant activity 370 than bread crumb; however, the phenolic classes distribution varied between the two zone of 371 bread loaf. In fact, flavonoids were more concentrated in crumb than crust probably due to 372 their low thermal stability. 373 Finally, this work improves the information about the phenolic content of buckwheat 374 enhanced wheat bread and, to our knowledge, the phenolic composition of crust and crumb of 375 buckwheat breads has been showed for the first time. However, further researches are needed 376 to explore the neo-formation/degradation/hydrolysis reactions of phenolic compounds during 377 baking process in order to clarify the effect of temperature on single phenolic compound.

M A N U S C R I P T
A C C E P T E D ACCEPTED MANUSCRIPT Table 2. Content of single free phenolic compounds and relative classes in control (white bread) and buckwheat enriched bread loaf samples expressed as mg/kg d.w.