Impact of alternative protein fining agents on the phenolic composition and color of Syrah red wines from warm climate

*REVISED Manuscript (text UNmarked) 1 Impact of alternative protein fining agents on the phenolic composition and color 2 of Syrah red wines from warm climate 3 Belén Gordillo, Francisco Chamizo-González, M. Lourdes González-Miret*, and 4 Francisco J. Heredia 5 6 Food Colour and Quality Laboratory, Facultad de Farmacia, Universidad de Sevilla, 7 41012 Sevilla, Spain 8 9 Belén Gordillo: [email protected] 10 Francisco Chamizo-González: [email protected] 11 Francisco J. Heredia: hered[email protected] 12 M. Lourdes González-Miret: [email protected] 13 14 * Corresponding author: 15 M. Lourdes González-Miret 16 Food Colour & Quality Lab., Área de Nutrición y Bromatología. Facultad de Farmacia. 17 Universidad de Sevilla. 41012-Sevilla, Spain 18 Tel.: +34 955420938 19 e-mail: mire[email protected] 1 2 20 ABSTRACT 21 Currently, the wine industry has an increasing interest in developing alternative solutions to traditional 22 animal proteins fining agents. In this study, the impact of different protein fining agents on the turbidity, 23 phenolic composition and color of 2-month and 12-month Syrah red wines was assessed. Wines fined 24 with egg albumin and plant-based proteins from potato, pea, and grape seed as recent alternative, were 25 compared to unfined control wines. Changes on turbidity, phenolic composition and color (by 26 Differential Colorimetry) showed that animal and plant proteins differed in their clarifying efficiency 27 and ability to interact with colorless phenolics and anthocyanins, depending on the age of wine, with 28 important consequences on color quality and stability. Plant proteins showed lower effectiveness to 29 reduce wine turbidity than egg albumin but modified in different way the phenolic composition, 30 inducing lower color differences with respect to control wine and similar stability, especially potato and 31 grape seed proteins. 32 33 Keywords: Protein fining agents; grape seed protein; phenolic composition; color; red wine. 34 35 3 36 1. Introduction 37 Phenolic compounds are the main chemical substances responsible for the organoleptic characteristics of 38 wines such as color, bitterness and astringency, especially in red wines. The color of red wine is one of 39 the most important parameters to decide its quality since it is the first attribute perceived, so it influences 40 consumer’s preferences and market decisions. Color is directly related to the anthocyanin composition 41 of wine and the interactions among them or with other wine components (mainly colorless phenolics) by 42 copigmentation, which contributes to the color evolution and stabilization ( ordillo Ce udoastante 43 odr ue - ulido on le -Miret  Heredia, 2013; Escribano-Bailón  Santos-Buelga, 2012). 44 On the other hand, the bitterness of wine is mainly due to flavan-3-ols, and also caused by some 45 flavonols, hydroxycinnamates and benzoic acid derivatives (Ferrer-Gallego, Hernández-Hierro, Rivas46 Gonzalo,  Escribano-Bailón, 2014; Ferrer-Gallego, Brás, García-Estévez, Mateus, Rivas-Gonzalo  47 De Freitas, 2016). Moreover, low molecular weight flavanols and their oligomeric/polymeric derivatives 48 (procyanidins or tannins) are the main phenolics contributing to the astringency, which tend to decrease 49 during red wine maturation and aging by polymerization and precipitation reactions (Quijada-Morín, 50 Williams, Rivas-Gonzalo, Doco,  Escribano-Bailón, 2014; Ramos-Pineda, García-Estévez, Brás, 51 Martín del Valle, Dueñas,  Escribano Bailón, 2017). Thus, to elaborate full-bodies red wines with 52 stable deep colors and equilibrated taste sensation of bitterness and astringency it is crucial to control 53 and modulate the phenolic composition during vinification. 54 After fermentative processes, wines are turbid and unstable media mainly due to the presence of 55 microorganisms (yeast and bacteria), tartrate crystals, rests of grape skin and pulp, and aggregates of 56 macromolecules (mainly pectin and protein materials) formed during the fermentative maceration 57 (Vernhet, 2018). In advanced stages of vinification, the presence of colloidal unstable species is related 58 to the formation of less soluble phenolic species that tend to co-aggregate progressively during wine 59 aging, which determine its natural slow precipitation and sedimentation (González-Neves, Favre  Gil, 60 2014). 4 61 This diversity of particles, responsible for hazes and deposits, can indistinctly aggregate the coloring 62 matter and colorless phenolics affecting the sensory quality of wines, so they need to be removed or 63 stabilized to prevent alterations of taste, flavor, or color previous to bottling and consumption. 64 Clarification with fining agents is very common operation in oenology, which consists of adding an 65 exogenous substance in a turbid wine that drags down other suspended particles by flocculation or 66 adsorption (Gambuti, Rinaldi, Romano, Manzo  Moio, 2016; Vernhet, 2018). The main benefits rely 67 on increasing the wine limpidity, color stability and modulating mouthfeel perception by eliminating or 68 reducing some phenolic compounds of colloidal nature implicated on oxidation processes or aggressive 69 taste sensations (Marangon, Vincenzi  Curioni, 2019). However, this is a major challenge for red 70 wines having insufficient levels of phenolics since clarifying in excess can negatively affect the 71 stabilization processes related to small solutes and to macromolecules affecting especially the color, as 72 typically occurs in warm climate regions. 73 Among clarifying substances, the protein fining agents are of great interest for wine fining because they 74 have good ability to interact with wine phenolics and have demonstrated different affinity to diverse 75 phenolic classes (Maury, Sarni-Manchado, Poinsaut, Cheynier,  Moutounet, 2016; Maury, Sarni76 Manchado,  Cheynier, 2019; Río Segade, Paissoni, Vilanova de la Torre, Gerbi, Rolle  Giacosa, 77 2020). However, despite its effectiveness, traditional animal-derived fining agents like milk and egg 78 proteins has been subjected in the last decade to increased regulation by the European Union, Australia, 79 or New Zealand because of their potential allergenic risk or food intolerance (Tschiersch, Nikfardjam, 80 Schmidt  Schwack, 2010; Marangon et al., 2019). For this reason, the use of plant-derived 81 macromolecules such as proteins, cell wall material, or fiber from different vegetal sources have been 82 recently proposed as alternative solutions for the clarification of white, rose, and red wines (Guerrero, 83 Smith & Bindon, 2013; Cosme, Capão, Filipe-Ribeiro, Bennett  Mendes-Faia, 2012; Bautista-Ortín, 84 Cano-Lechuga, Ruiz-García  Gomez-Plaza, 2014; Gambuti et al., 2016; Marangon et al., 2019; 85 Jiménez-Martínez, Bautista-Ortín, Gil-Muñoz  Gómez-Plaza, 2019). At this respect, special attention 5 86 has been paid toward the use of protein fining agents endogenous to grapes and byproducts as proteins 87 from seeds (Vincenzi et al., 2013; Gazzola, Vincenzi, Marangon, Pasini  Curioni, 2017). 88 Notwithstanding, the effectiveness of different protein fining sources strongly differs depending on the 89 type and age of wines in which are applied (Gonzalez-Neves et al., 2014; Martínez-Lapuente, 90 Guadalupe  Ayestarán, 2017; Marangon et al., 2018; Río-Segade et al., 2020). Moreover, several 91 studies have also demonstrated controversial effects on wine sensory characteristics, stability, and 92 composition depending on the origin of the protein source and their hydrolysis grade, the dose applied, 93 or the contact time (Tschiersch et al., 2010; Oberholster, Carstens  Du Toit, 2013; Ghanem et al., 94 2017; Kang, Niimi  Bastian, 2018; Jimenez-Martinez et al., 2019). In this sense, further studies about 95 the effectiveness of alternative protein agents are still needed, which could be of great interest especially 96 in warm climate vinifications. 97 Thus, the main objective of this work was to assess the impact of different vegetal protein fining agents 98 (potato, pea, and grape seed proteins) in the turbidity, phenolic composition, and color of Syrah wines 99 from warm climate (2 and 12-month from the end of fermentation), and compare them as potential 100 alternatives to traditional animal-derived proteins such as egg albumin. Special attention was focused on 101 color quality and stability by Differential Colorimetry, which provides relevant color information related 102 to visual perception of qualitative and quantitative color variations. 103 2. Material and methods 104 2.1. Fining agents and preparation of grape seed protein concentrate 105 Fining agents used in the clarification trials were commercial powdered protein isolates from egg 106 (OVOVIN, egg albumin), potato (PROVEGET FINE, Solanum tuberosum) and pea (PROVEGET 100, 107 Pisum sativum), all of them provided from Agrovin S.A. (Ciudad Real, Spain) and approved by the 108 International Oenological Codex and EC 606/2009 Regulation. 109 With the aim of comparing the fining efficiency of the selected protein agents, a grape seed protein 110 concentrate (GSP) was experimentally prepared and included in this study as alternative protein-fining 6 111 source. GSP was obtained from defatted grape seed flour (industrial wine by-product, ALVINESA 112 Natural Ingredients, Ciudad Real, Spain) by alkaline solubilisation with isoelectric precipitation, 113 according to Gazzola et al., (2017). The alkaline solubilisation of grape seed proteins was carried out at 114 pH 10.5 in a Bioreactor Bio Bundle Microbial System (Applikon Biotechnology®, Holland) using the 115 pH-stat method. For this purpose, 1 kg of defatted grape seed flour was re-dissolved in 5 L of distilled 116 water (20% p/v) during 12 h in agitation (300 rpm). After that, the alkaline aqueous solution was 117 centrifuged (9000 rpm, 15 min, 10 ºC), decanted 24 h in refrigeration (5 ºC), and then acidified to pH 3.0 118 with 6 M HCl. The precipitated material was recovered by centrifugation (9000 rpm, 30 min, 10 °C) and 119 lyophilized to obtain a fine powder (GSP concentrate), which was stored at -20 ºC until its use. The 120 chemical composition of the GSP extract was determined according to the AOAC standard protocols 121 (AOAC, 1990) and shown in Table S1. 122 Protein content (%) of fining agents were: egg albumin (82.3 % ± 0.3), potato isolate (76.9 % ± 0.3), pea 123 isolate (68.7 % ± 0.5) and GSP (32.1 % ± 0.6). Protein content was determined using a LECO TruSpec® 124 CHNS MICRO microsample elemental analyzer (Leco Instrumentos S.L., Madrid, Spain) based on the 125 nitrogen determination by thermal conductivity detection system, after sample combustion. Protein 126 content was computed using a nitrogen-to-protein conversion factor of 6.25 for animal protein and of 127 5.53 for plant based proteins (Zhou et al., 2011). 128 2.2. SDS-PAGE analysis 129 The range of molecular weight of protein fining agents was determined by sodium dodecyl sulphate 130 polyacrylamide gel electrophoresis (SDS-PAGE) usin the Invitro en™ system (Thermo Fisher 131 Scientific Inc.). Fining agents were re-dissolved in Tris-HCl buffer pH 7.5 at 2 mg/mL. Aliquots of these 132 concentrates (20 μL) were mixed with 5 µL of 0.02 M Tris-HCl loading buffer pH 6.8 (containing 40% 133 w/v glycerol, 4.6% w/v SDS, 20% v/v 2-mercaptoethanol, and 0.01% Bromophenol Blue, all from 134 Sigma-Aldrich, Milan, Italy), and then denatured at 100°C for 5 minutes. Denatured protein samples 135 were loaded into 12% polyacrylamide gel 1.5 mm (acrylamide: bisacrylamide, 30:1) and electrophoresis 7 136 was run with constant voltage mode (160 V) at room temperature for 50-60 min until the tracking dye 137 Bromophenol Blue ran off the gel. Gels were stained with Quick Coomassie Stain (Quimigen S.L., 138 Madrid) and then de-stained with water. The molecular weight standard proteins (ranging from 10 to 139 200 kDa) used were the a e uler™ Unstained rotein Ladder (Thermo Fisher Scientific Inc.). 140 2.3. Winemaking protocols and fining treatments 141 Two Syrah red wines elaborated in warm climate (“Condado de Huelva” Desi nation of Ori in 142 southwestern Spain) at different stages of vinification (2-months (W2) and 12-month (W12) from the 143 end of fermentation) were used for clarification assays. W2 corresponded to a very recent wine, and 144 W12 corresponded to a more stabilized wine in which some reactions (aggregations or sedimentation) 145 have occurred along the winemaking period. These are two stages of the vinification process in which 146 clarification could be performed: when wine is ready to initiate the stabilization (around 2 months) 147 process and when is ready to be bottled (around 12 months). 148 W2 and W12 red wines were made from grapes Vitis vinifera var. Syrah by traditional fermentative on149 skin maceration for 6 days. Healthy grapes were harvested at optimum technological maturity (average 150 13.5 ºBé and 13.9 ºBé in W2 and W12, respectively), destemmed and crushed, and distributed into 151 stainless steel tanks for maceration, where alcoholic fermentation (20-25 ºC) was induced by inoculating 152 selected yeast (Saccharomyces cerevisiae 25 g/hL, Viniferm BY, Agrovin, Ciudad Real, Spain). 153 Fermentative maceration occurred along 5 days (fermentation caps were punched down once a day); 154 after this, the mash was drawn off to remove the solid parts, and the free run wine was racked to 155 stainless steel tanks to finish the fermentation. Malolactic fermentation was induced by inoculating 156 selected lactic acid bacteria (Oenococcus oeni VINIFERM Oe 104, 14 mL/hL, Agrovin, Ciudad Real, 157 Spain). When fermentative processes finished, sulfur dioxide and total acidity were adjusted at the same 158 levels for all wines. W2 and W12 wines were maintained in 50 L stainless-steel tanks until fining trials. 159 The mean conventional oenological parameters of wines, assessed according to the Official Methods 160 established by European Union (UE, 2003) were: W2 (alcohol content = 13.5 v/v; reducing sugars = 8 161 1.88 g/L; titratable acidity = 6.07 g/L as tartaric acid; pH = 3.38; volatile acidity = 0.40 g/L as acetic 162 acid; free sulfur dioxide = 13.3 mg/L; total sulfur dioxide = 119.3 mg/L) and W12 (alcohol content = 163 13.8 v/v; reducing sugars= 1.75 g/L; titratable acidity = 5.63 g/L as tartaric acid; pH = 3.35; volatile 164 acidity = 0.60 g/L as acetic acid; free sulfur dioxide = 25.8 mg/L; total sulfur dioxide = 112.5 mg/L). 165 The clarification assays were performed in triplicate, in 200 mL glass containers, for each wine (W2 and 166 W12) and each fining agents, which were prepared following the recommendation of manufacturers and 167 the doses applied of each product were the maximum recommended, being 10 g/hL for egg albumin and 168 potato protein isolates, and 30 g/hL for pea protein isolate. For comparative purposes, GSP was applied 169 at two doses: 10 and 30 g/hL (GSP10 and GSP30, respectively). The fining agents were added and 170 homogenized with the wines, the containers completely filled were then closed and maintained at room 171 temperature (10 ± 2 ºC, in the dark) for 6 days. These are the mean conditions for the normal treatment 172 in wineries. Moreover, triplicates of 200 mL of untreated W2 and W12 wines were used as control. 173 After 6 days of clarification, wine samples were separated from lees and stored closed in glass 174 containers at room Tª (10 ± 2 ºC in dark) during 1 month. 175 Wine samples were taken for chemical and colorimetric analysis at day 1 (before clarification), day 6 176 (end of clarification), and at the end of storage (30 days after clarification). 177 178 2.4. Analysis of turbidity Turbidity of wines was measured using a 2100P Portable Turbidimeter (HACH® Be RightTM, Loveland, 179 CO, USA) before treatment and after 6 days of fining. Wine samples were analyzed inside a glass vial 180 and measured four times rotating the vial after each measurement. Results were expressed in NTU 181 (Nephelometric Turbidity Unit). 182 2.5. HPLC-DAD analysis of phenolic compounds 183 High performance liquid chromatography (HPLC) was applied for the determination of the monomeric 184 anthocyanins, phenolic acids, monomeric flavan-3-ols and flavonols by direct injection of the samples, 185 previously filtered through a 0.45 m Nylon filter. An Agilent 1200 (Palo Alto, CA), equipped with 9 186 quaternary pump, UV-Vis diode-array detector, automatic injector, and the ChemStation software was 187 used for the analyses. All analyses were made in triplicate. The separation, identification and 188 quantification of compounds was performed following a modification of the method described in 189 Gordillo et al. (2013). Phenolic compounds were separated on a Zorbax C18 column (250 x 4.6 mm, 5 190 m particle size) maintained at 38 ºC. Acetonitrile-formic acid-water (3:10:87) as solvent A, and 191 acetonitrile-formic acid-water (50:10:40) as solvent B were used. Acetonitrile and formic acid were 192 HPLC grade (Merck, Darmstadt, Germany) and purified water was obtained from a NANOpure 193 Diamond system (Barnsted Inc., Dubuque, IA, USA). The elution profile was as follows: 0-10 min with 194 6% B; 10-15 min with 30% B; 15-25 min with 40% B; 25-30 min with 45% B; 30-33 min with 50% B; 195 33-34 min with 60% B; 34-35 min with 6% B. The flow-rate was 0.8 mL/min and the injection volume 196 was 50 l. UV-Vis spectra were recorded from 200 to 800 nm with a bandwidth of 2.0 nm. The 197 wavelengths of detection were 525 (monomeric anthocyanins), 280 nm (benzoic acids and monomeric 198 flavanols), 320 nm (hydroxycinnamic acids and their tartaric esters) and 360 nm (flavonols). 199 Identification of phenolics was performed according to the spectra features and retention times with 200 those of the available pure standards and our data library of the standards. The quantification was made 201 by external calibration comparing the areas with the following commercial standards: malvidin 3-O202 glucoside ( 97%, Extrasynthese, Genay, France), catechin, p-coumaric acid, and quercetin ( 96%, 203 Sigma-Aldrich, Madrid, Spain). The concentration of compounds was expressed as mg/L. 204 The determination of procyanidins (dimeric and oligomeric flavan-3-ols) were performed, in triplicate, 205 according to Jara-Palacios, Gordillo, Gonzalez-Miret, Hernanz, Escudero-Gilete & Heredia (2014) by 206 rapid resolution liquid chromatography (RRLC). After filtration through a 0.45 m Nylon filter, samples 207 were injected (0.5 L injection volume) in an Agilent 1290 chromatographic system, equipped with 208 quaternary pump, UV-VIS diode-array detector, automatic injector, and ChemStation software (Agilent 209 Technologies, Palo Alto, USA). A C18 Poroshell 120 column (2.7 m, 5 cm x 4.6 mm) was used. The 210 solvents were formic acid and water (1:999 mL:mL) as solvent A, and acetonitrile as solvent B at the 16 361 grape seed proteins with wine phenolics (Table S1). Indeed, further studies focused on increasing the 362 purity of grape protein extracts are still needed in order to optimize their efficacy as fining agents in 363 relation with the dose applied and contact time (Jiménez-Martínez et al., 2019; Marangon et al., 2019). 364 As in W2 wines, the content of some individual phenolic acids was affected by fining treatments in W12 365 wines, although this effect was not reflected in the total content. In particular, wines fined with pea 366 proteins and GSP30 showed significant (p<0.05) lowest contents of gallic acid than control wines (90.2, 367 92.7, and 94.1 mg/L, respectively; Table S2), which proved its higher ability to bind such as small 368 colorless phenolics than the other fining sources. 369 As shown in Table 2, most of the individual anthocyanin monomers were not affected in W12 wines by 370 any of the fining treatments, as well the total monoglucosides and acetylated derivatives. Consequently, 371 although the global levels of Total Monomeric Anthocyanins (Figure 1B) were slightly decreased after 372 clarification, the differences in relation to the content of control W12 wines were not significant. These 373 results agreed with Granato et al. (2018), who showed that different fining agents had higher impact in 374 the anthocyanin composition of young wines than in one-year older ones. Nevertheless, a more 375 comprehensive assessment of the impact of fining treatments in the phenolic composition of aged wines 376 should include the evaluation of the anthocyanin-derived pigments, which are progressively formed 377 during vinification contributing in a higher extent to the total pigment content in advanced stages of 378 vinification (González-Neves et al., 2014; De Freitas  Mateus, 2011). 379 3.3. Impact on wine color by Tristimulus Differential Colorimetry 380 Table 3 shows the effect of the clarification treatments on the colorimetric characteristics (mean±SD, 381 n=3) of 2-month and 12-month Syrah wines. In W2, all the fining treatments significantly affected most 382 of the CIELAB (L*, a*, b*, C*ab, hab) and CIELUV (suv) color parameters compared to those of control 383 wines (CW, unfined). In general, the lightness (L*) and hue (hab) values slightly increased in fined wines 384 while the chroma (C*ab) and saturation (suv) values decreased. These trends indicate that clarification 385 induced both quantitative and qualitative color changes in young Syrah wines. From a quantitative point 17 386 of view (L*, C*ab, and suv), fined wines showed clearer, less intense and less saturated colors than 387 control wines, which was also confirmed by the decreases of C*ab values. Likewise, the variations of the 388 qualitative attribute of color (hab) denote a slight reduction of the bluish component of the red tonality in 389 fined wines regarding control ones, as observed by the increases of the negative b* values. 390 Notwithstanding, the rate of the quantitative and qualitative color changes varied between the fining 391 sources. Egg albumin and pea proteins produced the highest impact on the CIELAB quantitative 392 attributes of color by decreasing the wine color intensity by 5-4% (C*ab=48.6 and 49.1 versus 51.2 in 393 CW, p<0.05) and increasing the lightness by 5% (L*=55.4 and 54.8 versus 52.5 in CW, p<0.05). Even 394 so, these fining sources had not significant effect on the hue values with respect to CW (hab= -1.30° and 395 -1.40° versus -1.45° in CW), which indicate that maintained better the bluish-red tonality of young 396 Syrah wines. On the other hand, fining with GSP had the lowest impact in color intensity and lightness 397 (C*ab values decreases and L* increases by 2% with respect to CW), but the influence on the tonality of 398 wines was higher in comparison to the other fining sources (hab increases of +0.9 with respect to CW). 399 In this case, grape seed proteins at the higher dose applied (GSP30) showed no significant differences 400 for any of the color parameter with respect to CW, except for the lightness L*. Similarly, although the 401 CIELUV saturation (suv) values decreased in all fined wines, those treated with GSP showed the lowest 402 reductions in relation to CW making the wine to keep the purity and intensity of their original color. The 403 higher effect of egg albumin in reducing the color intensity in young wines with respect to the most 404 plant-based fining agents tested (mainly potato and GSP proteins) agree with the results reported by 405 Gazzola et al. (2017) and Gambutti et al. (2012).The differences found in the quantitative and qualitative 406 color effects between the fining sources could be explained by their selectivity to remove specific 407 anthocyanins compounds and families, which influenced not only the global pigment contents of wines 408 but also the proportions of individual anthocyanins, as reported by Granato et al. (2018). Moreover, it 409 was observed different rates on the reduction of copigments families such as flavanols, flavonols and 410 phenolic acids (González-Neves et al., 2014). 18 411 On the other hand, most of the CIELAB and CIELUV parameters were not significantly different 412 between fined wines and control ones (CW) when applied in advanced stages of vinification (W12 413 wines). Quantitatively, only wines fined with pea proteins had significant (p<0.05) higher values of L* 414 and lower of C*ab than CW (L*=60.6 versus 59.2; C*ab= 41.2 versus 42.6), which meant lighter and less 415 intense colors (L* increased by 2% and chroma decreased by 3%). Similarly, these wines showed less 416 saturated colors than CW (lower values of suv; p<0.05). Regarding the hue, W12 wines showed higher 417 values than W2 wines (hab=8-9° versus -1.45°/-0.59°) corresponding to the redness region of the 418 CIELAB space (between 0°-10° and positive values of b*), which indicates an important reduction of 419 the bluish tonalities typical of the earlier stages of vinification. Qualitatively, although wines fined with 420 egg albumin, potato and pea proteins had significant differences for the hue values compared to those of 421 control wines, the changes in the tonality between wines were quite small and thus, could be considered 422 negligible. 423 The impact of clarification observed in the color of W12 wines agree with the changes found in the 424 phenolic composition. On the one hand, all the fining treatments did not affect the content of most of the 425 anthocyanin compounds and the total levels of monomeric pigments. In fact, pea proteins were the 426 fining source that most reduced the content of copigments such as monomeric flavanols and flavonols 427 which could lead to higher loses on the color intensity of wines. 428 To quantify the color changes due to the fining treatment, we have used the Color Difference (ΔE*ab) 429 defined by CIELAB, which provide relevant color information related to visual perception (Gordillo et 430 al., 2015). For each ΔE*ab, it is possible to calculate the relative contributions of color attributes 431 changes: %ΔL (relative difference of lightness), %ΔC (relative difference of chroma) and %ΔH (relative 432 difference of hue). These contributions allow comparing objectively the quantitative and qualitative 433 effects of the different treatments on color. 434 The mean color differences (E*ab) between each fining treatments and its corresponding control wine 435 (not-treated) were calculated for W2 and for W12 wines, to compare the magnitude of the clarification 19 436 effects on the color of wines when this treatment is applied at different stages of vinification. As 437 observed in Figure 2, the E*ab values between control and fined treatments were comparatively higher 438 for W2 than for W12 wines (E*ab = 1.6-3.9 and 0.38-2.1, respectively) confirming that clarification had 439 more impact on color when applied in earlier stages of vinification (2 months versus 12 months). In the 440 case of W2, the highest color differences with respect to control wines (treated-W2 vs not-treated-W2) 441 was found in wines fined with egg albumin (E*ab= 3.9) compared to the E*ab values obtained for the 442 rest of the fining agents (ranging from 1.6 to 3.2). These results indicate that the selected plant-based 443 proteins had less impact on the color of young Syrah wines than traditional fining agents based on 444 animal proteins, which confirm those found by Gonzalez-Neves et al. (2014). However, taking into 445 account that E*ab around or higher than 3 CIELAB units indicates that color differences can be 446 perceived by the human eyes ( art ne el osa re , Hita  Negueruela, 2001), the effects of egg 447 albumin and pea proteins on wine color could be considered visually discernible. In both cases, the 448 differences were mainly due to a higher contribution of changes on the quantitative attributes of color 449 (%L=56% and 51%; %C= 44% and 47%, respectively). These results suggest that excessive fining 450 can be detrimental to the sensory quality of red wines. In contrast, the lowest color differences in 451 relation to control wines were found in wines fined with GSP and potato proteins (E*ab <2.5, not 452 clearly perceptible), in agreement with Gazzola et al. (2017). When fining treatments were applied in 453 advances stages of vinification (W12), wines fined with pea proteins led to the highest color differences 454 with respect to unfined wines mainly due to lightness and chroma changes (higher weight of %L and 455 %C), which agree with the effect observed in the individual color attributes. Nevertheless, in all cases 456 the E*ab values were lower than 2 units, and thus, considered not visually appreciable. 457 458 459 The assessment of the global color differences occurring during 30 days of storage after clarification allowed evaluating the color variation of each wine over time (ΔE*ab= [(L30-L02 + (a*30-a*0)2 + (b*30b*0)2]1/2, and compare the impact of fining treatments on color stability in W2 and W12 wines (Table 4). 460 In W2, all the fined wines showed significant lower values of color differences with respect to control 20 461 wines (ΔE*ab ranging from 7.2 to 9.1 versus 9.8 in CW), which indicate lower color variation and thus, 462 higher color stability. Among fining treatments, the higher stabilizing effect was found in wines fined 463 with e albumin (lower ΔE*ab) but the differences in relation to the plant-based protein agents were not 464 significant, except for S at the hi her dose (ΔE*ab =7.2 versus 9.1, respectively). In fact, wines fined 465 with GSP at the two doses (GSP10 and GSP30) showed higher color variation (lower stability) 466 compared to wines fined with potato and pea proteins. Probably the presence of other plant components 467 co-extracted with grape seed proteins in GSP extract could exert a negative impact on the color stability 468 of wines when used as fining agents (Table S1). The trend of the color attributes during storage (positive 469 values of L* and Δhab, negative of C*ab) indicates that all wines increased the lightness and hue 470 values but decreased the chroma, that is, showed clearer and less intense red-orange colors with the 471 time. The magnitude of the changes reflects that the most affected parameters were chroma and hue. 472 Among fining treatments, wines fined with egg albumin maintained better the color intensity while those 473 fined with potato proteins showed more stable hues (significant lower decreases of C*ab and increases 474 of Δhab, respectively). 475 In the case of the older Syrah wines (W12), the global color variations were comparatively lower than in 476 youngest ones (W2), mainly due to increases of hue during storage while the changes on chroma and 477 lightness were in most cases negligible (slight increases or decreases of C*ab and L*). At this regard, the 478 higher color stability (lower E*ab values) was found in wines fined with pea proteins and GSP at the 479 lowest dose (GSP10), but the differences with respect to control wines were only significant for the 480 formers. In both cases, the higher color stability achieved was due to their lower increases of hue, that is, 481 these wines showed more stables hues. 482 4. Conclusions 483 Results showed than plant-based proteins (such as potato and pea, and grape seed proteins as recent 484 proposal) are suitable sources to be used as clarifying agents both in earlier and advanced stages of 485 vinification of red wines from warm climate as alternative to traditional fining agents from animal origin 21 486 such as egg albumin. The effectiveness at reducing wine turbidity, and the impact on the phenolic 487 composition (colorless phenolics and anthocyanin pigments) and color quality and stability, have 488 demonstrated in most cases sensory benefits in fined wines in relation to unfined ones, compared to egg 489 albumin. Therefore, the exploitation for their potential use in the wine industry could be of great interest. 490 Although the use of seed proteins from grape by-products has the advantages of being endogenous to 491 grape, not allergenic, and adds value to this byproduct, further studies are still needed to optimize their 492 efficacy as fining agent in relation with the protein purity of the extracts, the dose applied and contact 493 494 time, and hence, the wine industry application. 495 Acknowledgments 496 Authors thank the assistance of the technical staff of Biology Service (SGI, Universidad de Sevilla, 497 Spain). 498 Funding sources 499 This research was financially supported by the Ministerio de Economía y Competitividad, Spain, 500 501 Gobierno de España (Project AGL2017-84793-C2). 22 502 Conflict of interest statement 503 504 Authors declare no conflict of interest 505 References 506 Bautista-Ortín, A.B., Cano-Lechuga, M., Ruiz-García, Y.,  Gómez-Plaza. E. Interactions between 507 grape skin cell wall material and commercial enological tannins. Practical implications. (2014). 508 Food Chemistry, 152, 558-565. 509 CIE. Technical eport Colorimetry; Commission Internationale de l’Eclaira e Central ureau: Vienna 510 Austria, 2004. 511 Cosme, F., Capão, I., Filipe-Ribeiro, L., Bennett, R.N.,  Mendes-Faia, A. Evaluating potential 512 alternatives to potassium caseinate for white wine fining: Effects on physicochemical and sensory 513 characteristics. (2012). LWT-Food Science and Technology, 46, 382-387. 514 De Freitas, V.  Mateus, N. Formation of pyranoanthocyanins in red wines: a new and diverse class of 515 anthocyanin derivatives (2011). Analytical and Bioanalytical Chemistry, 401, 1467-1477. 516 Escribano-Bailón, M.T.,  Santos-Buelga, C. Anthocyanin copigmentation - Evaluation, mechanisms 517 and implications for the colour of red wines. (2012). Current Organic Chemistry, 16, 715-723. 518 Ferrer-Gallego, R., Brás, N.F., García-Estévez, I., Mateus, N., Rivas-Gonzalo, J.C., De Freitas, V.,  519 Escribano-Bailón, M.T. Effect of flavonols on wine astringency and their interaction with human 520 saliva. (2016). Food Chemistry, 209, 358-364. 521 Ferrer-Gallego, R., Hernández-Hierro, J.M., Rivas-Gonzalo, J.C.,  Escribano-Bailón, M.T. Sensory 522 evaluation of bitterness and astringency sub-qualities of wine phenolic compounds: Synergistic 523 effect and modulation by aromas. (2014). Food Research International, 62, 1100-1107. 524 Gambuti, A., Rinaldi, A., Moio, L. Use of patatin, a protein extracted from potato, as alternative to 525 animal proteins in fining of red wine. (2012). European Food Research and Technology, 235, 753526 765. 23 527 Gambuti, A., Rinaldi, A., Romano, R., Manzo, N.,  Moio, L. Performance of a protein extracted from 528 potatoes for fining of white musts. (2016). Food Chemistry, 190, 237-243. 529 Gazzola, D., Vincenzi, S., Marangon, M., Pasini, G.,  Curioni, A. Grape seed extract: the first protein530 based fining agent endogenous to grapes. (2017). Australian Journal of Grape Wine Research, 23, 531 215-225. 532 Ghanem, C., Taillandier, P., Rizk, M., Rizk, Z., Nehme., N., Souchard, J.P.,  El Rayess, Y. Analysis of 533 the impact of fining agents types, oenological tannins and mannoproteins and their concentrations 534 on the phenolic composition of red wine. (2017). LWT-Food Science and Technology, 83, 101535 109. 536 Gómez-Míguez, M., González-Miret, M.L.,  Heredia, F.J. Evolution of colour and anthocyanin 537 composition of Syrah wines elaborated with pre-fermentative cold maceration. (2007). Journal of 538 Food Engineering, 79, 271-278. 539 González-Neves, G., Favre, G.,  Gil, G. Effect of fining on the colour and pigment composition of 540 young red wines. (2014). Food Chemistry, 157, 385-392. 541 Gordillo, B., Lopez-Infante, M.I., Ramírez-Perez, P., González-Miret, M.L., & Heredia, F.J. Influence 542 of prefermentative cold maceration on the color and anthocyanic copigmentation of organic 543 tempranillo wines elaborated in a warm climate. (2010). Journal of Agricultural and Food 544 Chemistry, 58, 6797-6803. 545 ordillo . Ce udoastasnte . . odr ue - ulido F. . on le -Miret, M.L.,  Heredia, F.J. 546 Application of the differential colorimetry and polyphenolic profile to the evaluation of the 547 chromatic quality of Tempranillo red wines elaborated in warm climate. Influence of the presence 548 of oak wood chips during fermentation. (2013). Food Chemistry, 141, 2184-2190. 549 Gordillo, B., Rodríguez-Pulido, F. J., González-Miret, M. L., Quijada-Morín, N., RivasGonzalo, J. C., 550 García-Estévez, I., Heredia, F.J.,  Escribano-Bailón, M. T. Application of differential 24 551 colorimetry to evaluate anthocyanin–flavonol–flavanol ternary copigmentation interactions in 552 model solutions. (2015). Journal of Agriculture and Food Chemistry, 63, 7645-7653. 553 Granato, T.M., Ferranti, P., Iametti, S.,  Bonomi, F. Affinity and selectivity of plant proteins for red 554 wine components relevant to color and aroma traits. (2018). Food Chemistry, 256, 235-243. 555 Guerrero, R.F., Smith, P.,  Bindon, K.A. Application of Insoluble Fibers in the Fining of Wine 556 Phenolics. (2013). Journal of Agricultural and Food Chemistry, 61, 4424-4432. 557 Heredia F. . lvare C. on le -Miret, M.L.,  Ramírez, A. CromaLab n lisis de Color; 558 Registro General de la Propiedad Intelectual, 2004. 559 Jara-Palacios, M. J., Hernanz, D., González-Manzano, S., Santos-Buelga, C., Escudero-Gilete, M. L., & 560 Heredia, F. J. (2014). Detailed phenolic composition of white grape by-products by RRLC/MS and 561 measurement of the antioxidant activity. Talanta, 125, 51-57. 562 Jiménez-Martínez, M.D., Bautista-Ortín, A.B., Gil-Muñoz, R.,  Gómez-Plaza, E. Fining with purified 563 grape pomace. Effect of dose, contact time and varietal origin on the final wine phenolic 564 composition. (2019). Food Chemistry, 271, 570-576. 565 Kang, W., Niimi, J.,  Bastian, S.E.P. Reduction of red wine astringency perception using vegetable 566 protein fining agents. (2018). American Journal of Enology and Viticulture, 69, 22-3. 567 Marangon, M., Vincenzi, S.,  Curioni, A. (2019). Wine fining with plant proteins. Molecules, 24, 1-16. 568 art ne . . el osa . re , M.M., Hita, E.,  Negueruela, A. I. Visual and instrumental color 569 evaluation in red wines. (2001). Food Science and Technology International, 7, 439-444. 570 Martínez-Lapuente, L., Guadalupe, Z.,  Ayestarán, B. Effect of egg albumin fining, progressive 571 clarification and cross-flow microfiltration on the polysaccharide and proanthocyanidin 572 composition of red varietal wines. (2017). Food Research International, 96, 235-243. 573 Maury, C., Sarni-Manchado, P., Poinsaut, P., Cheynier, V.,  Moutounet, M. Influence of 574 polysaccharides and glycerol on proanthocyanidin precipitation by protein fining agents. (2016). 575 Food Hydrocolloids, 60, 598-605. 25 576 Maury, C., Sarni-Manchado, P.,  Cheynier, V. Highlighting protein fining residues in a model red 577 wine. (2019). Food Chemistry, 279, 272-278. 578 Oberholster, A., Carstens, L.M.,  Du Toit, W.J. Investigation of the effect of gelatine, egg albumin and 579 cross-flow microfiltration on the phenolic composition of Pinotage wine. (2013). Food Chemistry, 580 138, 1275-1281. 581 Quijada-Morín, N., Williams, P., Rivas-Gonzalo, J.C., Doco, T.,  Escribano-Bailón, M.T. 582 Polyphenolic, polysaccharide and oligosaccharide composition of Tempranillo red wines and their 583 relationship with the perceived astringency. (2014). Food Chemistry, 154, 44-51. 584 Ramos-Pineda, A.M., García-Estévez, I., Brás, N.F., Martín del Valle, E.M., Dueñas, M.,  Escribano 585 Bailón, M.T. Molecular approach to the synergistic effect on astringency elicited by mixtures of 586 flavanols. (2017). Journal of Agricultural and Food Chemistry, 65, 6425-6433. 587 Río Segade, S., Paissoni, M.A., Vilanova de la Torre, M.M., Gerbi, V., Rolle, L.,  Giacosa, S. Phenolic 588 composition influences the effectiveness of fining agents in vegan-friendly red wine production. 589 (2020). Molecules, 25, 1-20. 590 Tschiersch, C., Nikfardjam, M.P., Schmidt, O.,  Schwack, W. Degree of hydrolysis of some 591 vegetable proteins used as fining agents and its influence on polyphenol removal from red wine. 592 (2010). European Food Research and Technology, 231, 65-74. 593 Vernhet. A. Red wine clarification and stabilization. Chapter 16.2 in Red wine technology, Academic 594 press, Elsevier, 2018, 237-251. 595 Vincenzi, S., Dinnella, C., Recchia, A., Monteleone, E., Gazzola, D., Pasini, G.,  Curioni, A. Grape 596 seed proteins: A new fining agent for astringency reduction in red wine. (2013). Australian 597 Journal of Grape and Wine Research, 19, 153-160. 598 Zhou, T., Zhang, T., Liu, W., & Zhao, G. Physicochemical characteristics and functional properties of 599 grape (Vitis vinifera L.) seeds protein. (2011). International Journal of Food Science and 600 Technology, 46, 635-641. Table 3 Table 3. Colorimetric parameters (mean±SD, n=3) of 2-month and 12-month Syrah wines (W2 and W12, respectively) after clarification (6 days) with protein fining agents. W2 wine Control Egg albumin Potato protein Pea protein GSP10 GSP30 L* 52.53 ± 0.06 a 55.42 ± 0.40 b 54.67 ± 0.07 b 54.81 ± 0.16 b 53.63 ± 0.83 b 53.66 ± 0.68 b a* 51.15 ± 0.21 a 48.56 ± 0.19 b 49.46 ± 0.44 b 49.05 ± 0.40 b 50.02 ± 1.02 b 50.30 ± 0.83 a b* -1.29 ± 0.08 a -1.10 ± 0.13 a -1.21 ± 0.06 a -0.95 ± 0.13 b -0.52 ± 0.55 b -0.58 ± 0.45 a C*ab 51.16 ± 0.21 a 48.57 ± 0.19 b 49.47 ± 0.45 b 49.06 ± 0.39 b 50.02 ± 1.01 b 50.29 ± 0.83 a hab -1.45° ± 0.08 a -1.29° ± 0.16 a -1.40° ± 0.04 a -1.11° ± 0.15 b -0.59° ± 0.65 b -0.67° ± 0.53 a suv 1.45 ± 0.01 a 1.31 ± 0.01 b 1.35 ± 0.01 b 1.35 ± 0.02 b 1.40 ± 0.06 b 1.41 ± 0.02 a W12 wine L* 59.21 ± 0.02 a 59.68 ± 0.53 a 59.61 ± 0.45 a 60.62 ± 0.02 b 58.82 ± 0.30 a 59.02 ± 0.04 b a* 42.12 ± 0.71 a 41.32 ± 0.75 a 41.80 ± 0.47 a 40.69 ± 0.75 b 42.17 ± 0.20 a 41.81 ± 0.12 a b* 6.50 ± 0.15 a 5.97 ± 0.02 b 5.90 ± 0.10 b 6.43 ± 0.19 b 6.55 ± 0.21 a 6.54 ± 0.08 a C*ab 42.62 ± 0.69 a 41.75 ± 0.75 a 42.21 ± 0.48 a 41.20 ± 0.72 b 42.68 ± 0.23 a 42.32 ± 0.13 a hab 8.78° ± 0.33 a 8.23° ± 0.14 b 8.04° ± 0.07 b 8.98° ± 0.41 b 8.82° ± 0.24 a 8.89° ± 0.09 a suv 1.15 ± 0.03 a 1.11 ± 0.03 a 1.12 ± 0.02 a 1.08 ± 0.03 b 1.16 ± 0.01 a 1.14 ± 0.01 a Different letters in the same row indicate significant differences (p<0.05) for each fining treatment respect to control wine. Table 4 Table 4. Color variations (∆E*ab, ∆L*, ∆C*ab, ∆h*ab, mean±SD, n=3) of control wines and wines treated with protein fining agents after 30 days of clarification. Abbreviations: W2 and W12 wine: 2-month and 12-month Syrah wines. Control Egg albumin Potato protein Pea protein GSP10 GSP30 W2 wine E*ab 9.81 ± 0.06 a 7.21 ± 0.48 b 7.70 ± 0.44 b 7.88 ± 0.34 b 8.64 ± 0.12 b 9.13 ± 0.07 a L* +1.65 ± 0.02 a +1.06 ± 0.51 a +0.73 ± 0.54 a +0.89 ± 0.12 a +0.53 ± 0.11 a +0.72 ± 0.27 a C*ab -5.10 ± 0.14 a -1.67 ± 0.79 b -3.77 ± 0.88 a -2.96 ± 0.24 b -4.88 ± 0.68 a -4.77 ± 1.19 a hab +9.68 ± 0.02 a +8.25 ± 0.90 a +7.99 ± 0.01 b +8.76 ± 0.37 b +8.51 ± 0.06 b +9.21 ± 0.79 a W12 wine E*ab 5.09 ± 0.15 a 5.27 ± 0.28 a 5.67 ± 0.20 b 4.62 ± 0.09 b 4.76 ± 0.25 a 5.05 ± 0.27 a L* -1.23 ± 0.43 a +0.64 ± 0.33 b +0.62 ± 0.36 b -1.39 ± 0.72 a -1.44 ± 0.10 a -0.02 ± 0.01 b C*ab +0.63 ± 0.03 a -0.19 ± 0.26 a -0.64 ± 0.05 a +1.55 ± 0.04 a +1.00 ± 0.57 a +0.20 ± 0.63 a hab +6.47 ± 0.25 a +7.18 ± 0.20 b +7.65 ± 0.14 b +5.66 ± 0.59 a +5.84 ± 0.16 b +6.75 ± 0.35 a ΔE*ab= [(L30-L0)2 + (a*30-a*0)2 + (b*30-b*0)2]1/2, L* =L30-L0, C*ab= C*ab30-C*ab0, hab = hab30-hab0. Different letters in the same row indicate significant differences (p<0.05) for each fining treatment respect to control wine. 1 Concentration (mg/L) Figure 1 Click here to download Figure(s): Gordillo_AlterProtFining_Figure 1_REV2.doc Figure 1. A. 140 120 a b b a 100 Total Anthocyanins Total Phenolic acids Total Monom eric Flavan-3-ols Total Procyanidins Total Flavonols b b b b b b b b 80 60 40 a b b b b b 20 0 CW EA PT PE GSP10 GSP30 W2 wine 2 Concentration (mg/L) B. 140 120 100 a Total Anthocyanins Total Phenolic acids Total Monomeric Flavan-3-ols Total Procyanidins Total Flavnonols b b b b 80 b 60 40 a a b a b b 20 0 CW EA PT PE GSP10 GSP30 W12 wine 1 %H %C %L  E*ab Figure 2 Click here to download Figure(s): Gordillo_AlterProtFining_Figure 2.doc Figure 2. 5 4 3 2 1 0 EA PT PE GSP10 GSP30 EA PT PE GSP10 GSP30 W2 wine W12 wine 1 Figure S1 Click here to download Supplementary Material: Gordillo_AlterProtFining_Suplementary Material_Figure S1.doc Figure S1. SDS-PAGE analyses of the fining agents (MMP, molecular weight protein marker; EA: Egg albumin; PT: potato protein; PE: Pea protein; GSP: grape seed protein). 1 Figure S2 Click here to download Supplementary Material: Gordillo_AlterProtFining_Suplementary Material_Figure S2.doc Figure S2. HPLC chromatograms recorded at 280, 525, 320 and 360 nm of W2 and W12 wines. Peaks: A) 1, gallic acid; 2, (+)-catechin; 3, (-)-epicatechin; B) 1, delphinidin 3-glucoside; 2, cyanidin 3glucoside; 3, petunidin 3-glucoside; 4, peonidin 3-glucoside; 5, malvidin 3-glucoside; 6, petunidin 3acetylglucoside; 7, peonidin 3-acetylglucoside; 8, malvidin 3-acetylglucoside; 9, petunidin 3-pcoumaroylglucoside; 10, peonidin 3-p-coumaroylglucoside; 11) malvidin 3-p-coumaroylglucoside; C) 1) t-caftaric acid; 2) t-coutaric acid; 3) p-coumaric acid; D) 1) myricetin-3-glucuronide; 2) myricetin-3glucoside; 3) quercetin-3-glucuronide; 4) quercetin-3-glucoside; 5) laricitrin-3-glucoside; 6) kaempferol-3-glucoside; 7) isorhamnetin-3-glucoside; 8) syringetin-3-glucoside. A. 1 280 nm W2 wine W12 wine 3 2 min B. 5 525 nm W2 wine W12 wine 1 3 4 2 6 7 8 9 10 11 min 2 C. 1 320 nm W2 wine W12 wine 2 3 min D. 360 nm 2 W2 wine 4 W12 wine 3 5 8 1 7 6 min 1 2.5 5 7.5 10 12.5 15 17.5 20 min min mAU 4 Figure S3 Click here to download Supplementary Material: Gordillo_AlterProtFining_Suplementary Material_Figure S3.doc Figure S3. RRLC chromatogram recorded at 280 nm of W2 and W12 wines. Peaks: 1, Procyanidin B1; 2, Tetramer 1; 3, Procyanidin B2; 4, Procyanidin B2 3-O-gallate; 5, Procyanidin B7; 6, EC Gallate; 7, Tetramer 2. mAU 10 280 nm W2 wine 6 W12 wine 5 4 3 6 7 2 1 2 0 2.5 5 7.5 10 12.5 15 17.5 20 min Table S1 Click here to download Supplementary Material: Gordillo_AlterProtFining_Suplementary Material_Table S1.doc Table S1. Chemical composition the GSP extract (mean values ±SD; n=3). % w/w GSP Protein 32.16 ± 0.60 Fat 0 Carbohydrates 53.45± 1.33 Ash 5.76 ± 0.18 Total Fiber 8.03 ± 0.5 Total Phenolics (Folin Ciocalteau) 0.70 ± 0.10 1