INVESTIGATIONS OF WINE AND GRAPE SKIN TANNINS FROM THE OKANAGAN VALLEY. Dawn Michaela Visintainer. B.Sc., The University of British Columbia, 2010

Transcription

1 INVESTIGATIONS OF WINE AND GRAPE SKIN TANNINS FROM THE OKANAGAN VALLEY by Dawn Michaela Visintainer B.Sc., The University of British Columbia, 2010 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE COLLEGE OF GRADUATE STUDIES BIOCHEMISTRY AND MOLECULAR BIOLOGY THE UNIVERSITY OF BRITISH COLUMBIA (Okanagan) December 2013 Dawn Michaela Visintainer, 2013

2 Abstract A grape skin tannin profile was established throughout berry development in Oliver, Naramata, and Osoyoos, three sites in the Okanagan area of British Columbia, using both 70% acetone and 12% ethanol as extraction solvents. Tannins were analyzed by ultra-high pressure liquid chromatography (UHPLC) to determine how total tannins, mean degree of polymerization (mdp), and percent subunit composition varied. Tannins for each sample did vary by more than 20% with the exception of one sample set. All of the samples mdp did not vary by more than 20%, with the exception of two sample sets; most of the percent subunit composition also did not vary by more than 20% with the exception of one sample set. These same samples were then analyzed by methyl cellulose precipitation assay (MCP) to determine tannin content in relation to astringency. It was predicted that as more tannins bonded with polysaccharide over the course of berry development, the astringency would decrease. We found a low correlation between total tannins quantified by the MCP assay and by UHPLC (r 2 < 0.52), and between total tannins quantified by the MCP assay and mdp calculated via analysis by UHPLC (r 2 > 0.37). A new gelatin adsorption assay to determine tannin quantity in wines was also developed. This method allows tannins in wine to bind with gelatin. Then, by recording the absorption of phenolic compounds at 280 nm before and after binding, we ascertained a value related to astringency. There was a good correlation between the Gelatin Adsorption Assay and the established MCP assay: r 2 = 0.99 when using model wine with the addition of grape seed extract, and r 2 = 0.98 when using real red wine samples. This method will allow for a cheaper way to evaluate tannins in red wines. ii

3 Table of Contents Abstract... ii Table of Contents... iii List of Tables... v List of Figures... vii Acknowledgments... xv Chapter 1: Introduction Grape Vine and Berry Characteristics Phenolics and Their Biosynthesis Phenolic Accumulation During Ripening Extractability of Tannins from Grape Skin Thesis Aims and Hypotheses Chapter 2: Research Chapter 1: Polyphenol from Pre-Véraison to Harvest Synopsis Materials and Methods Results and Discussion Chapter 3: Research Chapter 2: Comparison of Tannin Content with their Reactivity as Measured by the MCP Assay Synopsis Materials and Methods iii

4 3.3 Results and Discussion Chapter 4: Research Chapter 3: New Gelatin Adsorption Assay Synopsis Materials and Methods Results and Discussion Chapter 5: Conclusion Conclusion Summary Limitations/ Variation Suggestions for Future Research References Appendices Appendix A: Calibration Curves for Subunits Appendix B: Calibration Curves for mdp Appendix C: Correlation Between total tannins measured by MCP and Total Tannins Measured by UHPLC of mdp iv

5 List of Tables Table 1.1: Total tannin véraison to harvest. CE= catechin equivalents Table 1.2: mdp véraison to harvest Table 1.3: Subunit composition in grapes skins as ripening proceeds Table 2.1: The weight of 50 grape berry skins from pre-véraison to harvest was analyzed by linear regression Table 2.2: The total amount of tannins per berry from pre-véraison to harvest was analyzed by linear regression. N=8 for Naramata and Osoyoos and N=7 for Oliver Table 2.3: The linear regression of total tannins measured by UHPLC from pre-véraison and post-véraison was analyzed for each site and solvent with an α value of Table 2.4: Average, standard deviation, and percent coefficient of variation for the total amount of tannins during the entire sampling period for each site and both extraction solvents Table 2.5: The mean degree of polymerization of tannins from pre-véraison to harvest was analyzed by linear regression for each site and each solvent Table 2.6: The linear regression of mpd of tannin molecules from grape skins measured by UHPLC during pre-véraison and post-véraison was analyzed for each site with an α value of Table 2.7: Average, standard deviation, and percent coefficient of variation for mdp over the entire course of berry development for each site and solvent Table 2.8: The linear regression for the subunit composition of tannin molecules from v

6 grape skins measured by UHPLC for pre-véraison and post-véraison were analyzed for each site with an α value of Table 2.9: Linear regression for tannin subunit composition during development for each site and both extraction solvents Table 2.10: Average, standard deviation, and percent coefficient of variation for tannin subunit composition during development for each site and both extraction solvents Table 3.1: Total tannins measured by MCP were analyzed using linear regression to access changes over the course of berry development for all sampling dates at each site with α of Table 3.2: Linear regression was analyzed for total tannins measured by MCP for different sampling dates at each site with an α value of Table 3.3: Linear regression was analyzed for total tannins measured by MCP compared to mean degree of polymerization (mdp) and for total tannins measured by MCP compared to those measured by UHPLC, for each sampling dates and site with an α value of Table 4.1: List of wine (winery and year) samples used to validate the new Gelatin Adsorption Assay vi

7 List of Figures Figure 1.1: Structure of a grape cluster and the cross section of a grape berry at maturity. Re-drawn from Ribéreau-Gayon et al., 2006a Figure 1.2: Berry weight changes throughout development from fruit set to ripeness. Redrawn and adapted from Coombe and Hale 1973: Berry growth; Phase I: early fruit development; Phase II: lag phase; Phase III: berry ripening Figure 1.3: Biosynthetic pathway of the major phenolic classes Figure 1.4: Basic monomer tannin structure found in grape skins can be linked by a carbon-carbon inter-flavan bond (C4-C6 or C4-C8) to form polymers. R1 can be a hydrogen group or a gallate group and R2 can be hydrogen or a hydroxyl group Figure 1.5: Structure of the flavanol-3-ol monomeric molecules that are subunits that make up tannins in grape skins Figure 1.6: Biosynthesis of (+)-catechin and (-)-epicatechin from caffeoyl CoA and malonyl-coa adapted and re-drawn from Adams (2006). CHS, chalcone synthase; CHI, chalcone isomerase; F3H, flavanone-3-hydroxylase; DFR, dihydroflavonol-4-reductase; LAR, leucoanthocyanidin reductase; ANS, anthocyanidin synthase; ANR, anthocyanidin reductase; UFGT, UDP glucose-flavonoid 3-O-glucosyl transferase Figure 1.7: Basic structure of anthocyanidins, the precursors to anthocyanins. Anthocyanins have a glucose molecule bonded to the oxygen which is bonded to the carbon at position three (R1). In certain varieties, vii

8 Cabernet franc and Merlot, but not Pinot noir, some esterification can occur on the OH function of the glucose with coumaric or acetic acid. R3 and R2 are substituted with either a hydrogen molecule (H), a hydroxyl molecule (OH), or a methoxy molecule (OCH3) Figure 1.8: Structure of the five anthocyanidins found in Pinot noir. Anthocyanidins are precursors to anthocyanins where anthocyanins have a glucose molecule bonded to the oxygen molecule at position Figure 1.9: Pathway showing anthocyanin biosynthesis Figure 1.10: Biosynthetic flavonoid pathway leading to the production of proanthocyanidin polymer (condensed tannin) and anthocyanidin re-drawn and adapted from Bogs et al. (2005); ANS, anthocyanidin synthase; UFGT, UDP glucose-flavonoid 3-O-glucosyl transferase; LAR, leucoanthocyanidin reducatase; ANR, anthocyanidin reductase Figure 2.1: Extraction kinetics of Pinot noir grape skin when using 20 ml of 70% acetone per g of grape skin as the solvent. Showing standard deviation error bars; N= Figure 2.2: Extraction kinetics of Pinot noir grape skin when using 20 ml of 12% ethanol per g of grape skin as the solvent. Showing standard deviation error bars; N= Figure 2.3: The weight of fifty berry skins at each sampling date. Showing standard deviation error bars; N= Figure 2.4: Grape skin tannins extracted from Oliver, Naramata, and Osoyoos over the course of berry development using 70% acetone as the solvent. viii

9 Showing standard deviation error bars; N= Figure 2.5: Grape skin tannins extracted from Oliver, Naramata, and Osoyoos over the course of berry development using 12% ethanol as the solvent. Showing standard deviation error bars; N= Figure 2.6: Grape skin tannins mean degree of polymerization extracted from Oliver (PN), Naramata (CF), and Osoyoos (CF) over the course of berry development using 70% acetone as the solvent. Showing standard deviation error bars; N= Figure 2.7: Grape skin tannins mean degree of polymerization extracted from Oliver (PN), Naramata (CF), and Osoyoos (CF) over the course of berry development using 12% ethanol as the solvent. Showing standard deviation error bars; N= Figure 2.8: Oliver (Pinot noir) tannin subunit composition over development extracted using 70% acetone. EGC: (-)-Epigallocatechin; ECG: (-)-Epicatechin-3-0-gallate; EC: (-)-Epicatechin; C: (+)-Catechin. N= Figure 2.9: Oliver (Pinot noir) tannin subunit composition over development extracted using 12% ethanol. EGC: (-)-Epigallocatechin; ECG: (-)-Epicatechin-3-0-gallate; EC: (-)-Epicatechin; C: (+)-Catechin. N= Figure 2.10: Naramata (Cabernet Franc) tannin subunit composition over development extracted using 70% acetone. EGC: (-)-Epigallocatechin; ECG: (-)-Epicatechin-3-0-gallate; EC: (-)-Epicatechin; C: (+)-Catechin. N= Figure 2.11: Naramata (Cabernet Franc) tannin subunit composition over ix

10 development extracted using 12% ethanol. EGC: (-)-Epigallocatechin; ECG: (-)-Epicatechin-3-0-gallate; EC: (-)-Epicatechin; C: (+)-Catechin. N= Figure 2.12: Osoyoos (Cabernet Franc) tannin subunit composition over development extracted using 70% acetone. EGC: Epigallocatechin; ECG: Epicatechin-3-O-gallate; EC: Epicatechin; C: Catechin. N= Figure 2.13: Osoyoos (Cabernet Franc) tannin subunit composition over development extracted using 12% ethanol. EGC: Epigallocatechin; ECG: Epicatechin-3-O-gallate; EC: Epicatechin; C: Catechin. N= Figure 3.1: Tannins quantified by MCP from Oliver (Pinot noir) throughout the course of berry development expressed in epicatechin equivalence. Showing standard deviation error bars; N= Figure 3.2: Tannins quantified by MCP from Naramata (Cabernet Franc) throughout the course of berry development expressed in epicatechin equivalence. Showing standard deviation error bars; N= Figure 3.3: Tannins quantified by MCP from Osoyoos (Cabernet Franc) throughout the course of berry development expressed in epicatechin equivalence. Showing standard deviation error bars; N= Figure 3.4: Correlation between total tannins measured by MCP and by UHPLC expressed in epicatechin equivalence. Samples are of grape skin extract from Oliver extracted with 70% acetone measured throughout berry development. N=9 at each sampling date Figure 4.1: (-)-Epicatechin calibration curve performed at 280nm. Showing standard x

11 deviation error bars; N= Figure 4.2: Reaction kinetics for tannin precipitating gelatin. Showing standard deviation error bars; N= Figure 4.3: Linearity of gelatin-tannin binding expressed in g L -1 (-)-epicatechin equivalence from varying concentrations of grape seed extract in model wine. Showing standard deviation error bars; N= Figure 4.4: Recovery of gelatin-tannin binding from a concentrated solution of grape seed extract added to wine or model wine in varying amounts to obtain different concentrations. Theoretical values represent 100% recovery. Data expressed in (-)-epicatechin equivalence. Showing standard deviation error bars; N= Figure 4.5: Correlation of MCP to the new Gelatin Adsorption Assay using grape seed extract added to model wine in different amounts. Data expressed in (-)-epicatechin equivalence. N= Figure 4.6: Correlation of MCP to the new Gelatin Adsorption Assay from twelve different wines. Data is expressed in (-)-epicatechin equivalence and each wine was analyzed in triplicate Figure A.1: Calibration curve for catechin and epicatechin with or without phloroglucinol. Showing standard deviation error bars; N= Figure A.2: Calibration curve for epigallocatechin (EGC) with phloroglucinol attached. Showing standard deviation error bars; N= Figure A.3: Calibration curve for epicatechin gallate (ECG) without phloroglucinol attached. Showing standard deviation error bars; N= xi

12 Figure A.4: Calibration curve for epicatechin gallate (ECG) with phloroglucinol attached. Showing standard deviation error bars; N= Figure B.1: Calibration curve for catechin (C) and epicatechin (EC) without phloroglucinol, concentration expressed in mmol/l to be used in calculation of mean degree of polymerization. Showing standard deviation error bars; N= Figure B.2: Calibration curve for catechin and epicatechin with phloroglucinol attached, concentration expressed in mmol/l to be used in calculation of mean degree of polymerization. Showing standard deviation error bars; N= Figure B.3: Calibration curve for epigallocatechin (EGC) with phloroglucinol attached, concentration expressed in mmol/l to be used in calculation of mean degree of polymerization. Showing standard deviation error bars; N= Figure B.4: Calibration curve for epicatechin gallate (ECG) without phloroglucinol, concentration expressed in mmol/l to be used in calculation of mean degree of polymerization. Showing standard deviation error bars; N= Figure B.5: Calibration curve for epicatechin gallate with phloroglucinol attached, concentration expressed in mmol/l to be used in calculation of mean degree of polymerization. Showing standard deviation error bars; N= Figure C.1: Correlation between total tannins measured by MCP and by UHPLC expressed in epicatechin equivalence. Samples are of grape skin extract from Oliver extracted with 12% ethanol measured throughout berry xii

13 development Figure C.2: Correlation between total tannin precipitated by MCP expressed in epicatechin equivalence and mdp calculated through analysis by UHPLC. Samples are of grape skin extract from Oliver extracted with 12% ethanol measured throughout berry development Figure C.3: Correlation between total tannins measured by MCP and by UHPLC expressed in epicatechin equivalence. Samples are of grape skin extract from Naramata extracted with 70% acetone measured throughout berry development Figure C.4: Correlation between total tannin precipitated by MCP expressed in epicatechin equivalence and mdp calculated through analysis by UHPLC. Samples are of grape skin extract from Naramata extracted with 70% acetone measured throughout berry development Figure C.5: Correlation between total tannins measured by MCP and by UHPLC expressed in epicatechin equivalence. Samples are of grape skin extract from Naramata extracted with 12% ethanol measured throughout berry development Figure C.6: Correlation between total tannin precipitated by MCP expressed in epicatechin equivalence and mdp calculated through analysis by UHPLC. Samples are of grape skin extract from Naramata extracted with 12% ethanol measured throughout berry development Figure C.7: Correlation between total tannins measured by MCP and by UHPLC expressed in epicatechin equivalence. Samples are of grape skin xiii

14 extract from Osoyoos extracted with 70% acetone measured throughout berry development Figure C.8: Correlation between total tannin precipitated by MCP expressed in epicatechin equivalence and mdp calculated through analysis by UHPLC. Samples are of grape skin extract from Osoyoos extracted with 70% acetone measured throughout berry development Figure C.9: Correlation between total tannins measured by MCP and by UHPLC expressed in epicatechin equivalence. Samples are of grape skin extract from Osoyoos extracted with 12% ethanol measured throughout berry development Figure C.10: Correlation between total tannin precipitated by MCP expressed in epicatechin equivalence and mdp calculated through analysis by UHPLC. Samples are of grape skin extract from Osoyoos extracted with 12% ethanol measured throughout berry development xiv

15 Acknowledgments This research project would not have been possible without the support of many people. I wish to express my gratitude to my two supervisors, Dr. Cédric Saucier and Dr. Melanie Jones who were tremendously helpful by providing useful comments, remarks, and engagement through the learning process of this masters thesis. I would also like to thank the members of my supervisory committee, Dr. Soheil Mahmoud and Dr. Miranda Hart; without their knowledge and guidance this study would not have been successful. A special thanks also to Dr. Adeline Delcambre because, without her knowledge and invaluable assistance, this study would not have been possible and also to Ryan Moss and Yann Andre for their guidance and support through the duration of this research project. I would like to thank my loved ones who have supported me throughout the entire process. I would also like to convey thanks to the sponsors who provided the funding which without this project would not have been possible; British Columbia Wine and Grape Council- Agriculture and Agri-Food Canada (Developing Innovative Agri-Products Grant), Canada Foundation for Innovation (Leader Opportunity Fund Grant), Agilent Technologies (Equipment Grant), and Natural Sciences and Engineering Research Council of Canada (Discovery Grant). xv

16 Chapter 1: Introduction 1.1 Grape Vine and Berry Characteristics The common wine and edible grapes are mostly woody or herbaceous lianas (treeclimbing plants) or shrubs with liana stems, from the genus Vitis in the family Vitaceae (Alleweldt and Possingham, 1988; Mullins et al., 1992). Inflorescences opposite to leaves and coiled tendrils characterize the Vitaceae, originating in Asia, but now grown primarily between 20 and 50 latitude, north or south of the equator (Alleweldt and Possingham, 1988; Mullins et al., 1992). Vitis contains approximately 60 species and is found natively in the temperate areas of the Northern hemisphere (Mullins et al., 1992). One of the most common wine grapes, Vitis vinifera, is native to the Mediterranean region, central Europe, and southwest Asia but has been spread throughout the world by human activity (Mullins et al., 1992). Vitis vinifera grape vines predominantly grow in drier, semi-arid climates where temperatures of the summer months, either July in the Northern hemisphere or January in the Southern hemisphere, vary between 15 C and 25 C (Conde et al., 2007). The acquisition of an adequate amount of water and nitrogen by vines of these regions is essential because of loamy sand or sandy loam soils in these areas, which are low in nitrogen and can become dry in the growing season (Winkel and Rambal, 1993; Bowen et al., 2005). Vine and Berry Physiology Vines produce clusters of grapes attached to the vine by the peduncle, which in turn extends into the rachis where individual berries are attached to the rachis by pedicels (Figure 1.1). The pedicel contains xylem vessels that extend into the berries, delivering water and nutrients to the berry from the vine (Ribéreau-Gayon et al., 2006a). During véraison, which is the onset of ripening, the flow of water switches from primarily xylem to primarily phloem transport 1

17 (Findlay et al., 1987; Lang and Thorpe, 1989; Lang and During, 1991; Coombe, 1992; Creasy and Lombard, 1993; Bondada et al., 2005). It is however, not agreed on how this switch occurs. Diurnal solute partitioning (Bondada et al., 2005), un-functional tracheids (Findlay et al., 1987), the formation of an embolism (Lang and Thorpe, 1989; Coombe, 1992; Creasy and Lombard, 1993), or the in-ability of xylem cells, during phase two of berry growth, to lengthen causing gaps in the xylem (Lang and During, 1991) are all current theories explaining how restricted flow to the xylem occurs. This restriction in xylem flow, during véraison, leads to decreased hydraulic conductivity in the berry. As a result of this process, sugary sap fills the berries. The sugar stays in the berries despite high water potentials in the vine (Creasy and Lombard, 1993). Figure 1.1: Structure of a grape cluster and the cross section of a grape berry at maturity. Re-drawn from Ribéreau-Gayon et al., 2006a. Vine development can be broken down into a number of growth stages associated with vine and berry maturity. During winter, the vines are dormant and have light or dark brown, 2

18 closed bud scales (Hellman, 2003). In the spring, the buds begin to swell but stay brown (Coombe, 1995; Hellman, 2003). When average daily temperatures reach 10 C, the shoots begin to grow, the buds burst, and green tips become visible (Hellman, 2003). The leaves then begin to unfold until nine or more leaves have unfolded (Hellman, 2003). At this time the inflorescences become visible and start to swell while the flowers get closely pressed together (Srinivasan and Mullins, 1981; Hellman, 2003). When the inflorescences are fully developed, the flowers separate. Flowering starts when the first caps are detached from the receptacle and continues until all the caps have fallen off, signifying the end of flowering and the start of fruit set (Hellman, 2003). Fruit set finishes two to three weeks after flowering starts (Mullins et al., 1992; Hellman, 2003). After fruit set, shoot growth comes to an end and the young fruits begin to swell and grow, clusters begin to hang, and the berries begin to touch (Hellman, 2003). Once the majority of berries are touching, the berries begin to soften, ripen and develop color, indicating the berries are in véraison (Coombe, 1992; Hellman, 2003). When all the berries have changed color and the sugar and acid levels are at the amounts sought after by the winemaker, harvest takes place. After harvest, the leaves begin to lose color and fall off and the vine enters dormancy again (Hellman, 2003). Growth Patterns and Compound Accumulation A double sigmoid growth pattern with two distinct growth periods (Figure 1.2) is observed in the berry, where changes in phenolic composition occur throughout maturation (Coombe and Hale, 1973; Considine and Knox, 1979; Kennedy and Waterhouse, 2000; Kennedy et al., 2001; Conde et al., 2007). Monomeric and polymeric flavanols accumulate in the skin during phase I, the first period of growth associated with the grape berries (Kennedy and Waterhouse, 2000; Kennedy et al., 2001). The first phase is associated with a rapid growth 3

19 period lasting 45 to 65 days where the berries have increased metabolic activity with the rapid accumulation of acids and elevated respiration (Harris et al., 1968; Pratt, 1971; Ribéreau-Gayon et al., 2006a). Phase II, or the lag period, occurs next with an absence of growth coinciding with véraison (Ribéreau-Gayon et al., 2006a; Conde et al., 2007). During véraison, there is a change in berry skin color with the berries becoming hard and acidic, with little sugar (Pratt, 1971; Conde et al., 2007). This phase lasts 8 to 15 days but can last longer if flowering is delayed (Ribéreau-Gayon et al., 2006a). During this time there is depletion in cytokinins and gibberellins required for growth and an increase in abscisic acid concentration (Hale, 1968; Ribéreau-Gayon et al., 2006a; Conde et al., 2007; Lacampagne et al., 2010). The next phase, phase III, coincides with a second period of growth (Ribéreau-Gayon et al., 2006a). The walls of the outer hypodermal cells swell, which correlates to an increase in fruit plasticity (Considine and Knox, 1979). This phase usually occurs in August (northern hemisphere) and lasts 35 to 55 days with the berries becoming larger, softer, sweeter, and less acidic, with robust flavour and color (Ribéreau-Gayon et al., 2006a; Conde et al., 2007). Total polyphenol contents follow a similar pattern as berry size, with a slow continuous increase from fruit set until véraison (Considine and Knox, 1979). Then, 28 to 35 days after véraison, a faster increase is seen (Considine and Knox, 1979). Towards harvest however, a decrease is observed in some cases (Pirie and Mulins, 1977; Considine and Knox, 1979; Delgado et al., 2004; Conde et al., 2007). 4

20 Figure 1.2: Berry weight changes throughout development from fruit set to ripeness. Redrawn and adapted from Coombe and Hale 1973: Berry growth; Phase I: early fruit development; Phase II: lag phase; Phase III: berry ripening. Grape Skin Physiology Grape skin consists of many phenolic compounds that influence the quality of wine. The accumulation of these compounds in the skin is influenced by many factors including varietal (Van Leeuwen et al., 2004; Chira et al., 2009; Revilla et al., 2010), vine vigor (Cortell et al., 2005), climate (Poni et al., 1994; Chira et al., 2011, Ferrer-Gallego et al., 2012), soil (Van Leeuwen et al., 2004; Morlat and Bodin, 2006; de Andres-de Prado et al., 2007), and agricultural practices such as irrigation (Van Zyl, 1987; Greenspan et al., 1996; Ojeda et al., 2002; Olle et al., 2011), fertilization (Spayd et al., 1994; Keller et al., 1999; Delgado et al., 2004; Bell and Henschke, 2005), and method used to train the vine (Jackson and Lombard, 1993). Grapes of the same variety, grown in different regions, can thus yield different amounts of polyphenols 5

21 (tannins and anthocyanins) and produce very different wines. A study by Chira et al. (2011) found that skin tannin mean degree of polymerization (mdp), percentage of galloylation (%G) and percentage of prodelphinidins (%P) were all affected by the varietal. Tannin monomer and dimer concentrations; however, were not affected by grape variety. Depending on varietal, the surface area of the skin at maturity can increase up to 640-fold from the size at fruit set (Considine and Knox, 1981). The skin makes up 5-10% of the total dry weight of the berry and can be divided into three layers consisting of two distinct cell types (Considine and Knox, 1979; Pinelo et al., 2006). The first cell type is a single layer of clear epidermal cells on the outside of the berry (Considine and Knox, 1979; Hardie et al., 1996). The second cell type forms the hypodermal layers beneath the epidermis and the cells are called collenchmatous cells (Considine and Knox, 1979; Hardie et al., 1996). This cell type acts as a hydrophobic barrier that protects the grapes (Considine and Knox, 1979). The three layers of the grape skin consist of the cuticle which is the outermost layer, the intermediate epidermis, and the hypodermis which is the innermost layer (Hardie et al., 1996; Pinelo et al., 2006). The hypodermal layers of the skin accumulate phenolics and their properties affect extraction into wine (Hanlin et al., 2010). Unlike other polyphenols found in grape skins, tannins have the unique ability to form complexes with proteins and polysaccharides (Haslam, 1974; Carvalho et al., 2006; Sarneckis et al., 2006; Escot et al., 2008). The ability of tannins to bind to proteins and polysaccharides in the cell wall has been correlated with a reduction in their release into wine (Downey et al., 2003; Hazak et al., 2005; Fournand et al., 2006; Cerpa-Calderon and Kennedy, 2008; McRae and Kennedy, 2011). Binding of tannins to proteins and polysaccharides involves hydrogen bonding and hydrophobic interactions (Oh et al., 1980; Haslam, 1988; McManus et al., 1985; McRae et al., 2010). Tannins are able to bind these molecules at different 6

22 sites on proteins and polysaccharides because they are amphipathic molecules, comprised of hydrophobic aromatic rings and hydrophilic hydroxyl groups (Haslam, 1988). There are many different proteins and polysaccharides in the cell wall that contain aromatic oxygen and glycosidic oxygen atoms and hydroxyl groups that form hydrogen bonds and hydrophobic interactions with tannin (McManus et al., 1985; Le Bourvellec et al., 2004; Ribéreau-Gayon et al., 2006b; Bindon et al., 2010a). As the grape berry ripens the cell wall loosens as a result of a decrease in the strength and elasticity of the grape skin after véraison and an increase in its extensibility (Huang et al., 2005). The epidermis and hypodermis swell after véraison and there is degradation of the middle lamella in the cell walls (Considine and Knox 1979; Hardie et al., 1996: Huang et al., 2005). Also as the grape berry ripens, grape skin tannins are more easily extracted because there is degradation of the cell wall caused from the hydrolysis of structural polysaccharides (Nunan et al., 2001; Canals et al., 2005; Huang et al., 2005). The apoplast also becomes acidified by expansins after véraison which breaks calcium bridges of the pectin molecules leading to loss of not only pectin but also calcium from the skin cell wall (Cosgrove, 2000; Huang et al., 2005). On the other hand, proteins that are covalently bound to grape skin cell walls increase with ripening (Huang et al., 2005). Wall-bound peroxidise, which catalyses the phenolic cross-link between polysaccharides and proteins in the cell walls, increases after véraison (Huang et al., 2005). Peroxidase is found ionically bound to all layers of the cell walls but is more concentrated on the outer cell layers where it is covalently bound (Calderón et al., 1993; Huang et al., 2005). Components that comprise the cell wall also change as the berry ripens where berry sizes and chemical composition of the skin is dependent upon cultivar (Kok and Celik, 2004). As ripening proceeds the quantity of cell wall material decreases (Ortega-Regules et al., 2008). This 7

23 is because as the grape ripens the cell volume increases which consequently leads to a decrease in cell wall thickness (Baravon et al., 2000; Ortega-Regules et al., 2008). The chemical composition of the cell wall also changes during ripening. Galactose levels decrease during ripening whereas glucose, arabinose, xylose, fucose, mannose, and rhamnose in the cell wall either increase, decrease, or stay the same depending upon variety (Baravon et al., 2000; Ortega- Regules et al., 2008; Vicens et al., 2009). There are many different components of the cell wall; however, in all varieties during ripening the quantity of cell wall material and galactose decreases (Barnavon et al., 2000; Ortega-Regules et al., 2008; Vicens et al., 2009). Overview of Polyphenols in Wine Wine is very complex consisting of many interacting molecules that contribute to its intricate nature. The largest component of wine is water which depending on the variety, can range from 75 to 90% (v/v). Varying amounts of phenolics, organic acids, mineral salts and pectins contribute to the 15% variation observed (Conde et al., 2007). Ethanol, produced through the alcoholic fermentation of yeast, is the second largest contributor to the composition of wine, ranging from 8 to 13% (v/v) (Ciani and Picciotti, 1995; Conde et al., 2007). This high concentration of alcohol and the natural acidity of wine during fermentation inhibit the growth of pathogenic and noxious microorganisms that spoil wine (Conde et al., 2007; Waite and Daeschel, 2007). The sugar content can vary tremendously, with dry wines having less than 2 g L -1 and sweet wines reaching as high as 200 g L -1 (Dominé et al., 2004 as cited in Conde et al., 2007). The sugar content, as well as various nutrients, provides the yeast in the wine with a medium for ideal growth, thus allowing fermentation to commence (Conde et al., 2007). Fermentation results in the production of ethanol, which allows for better extraction of phenolic compounds compared to an aqueous solution (Canals et al., 2005; Sacchi et al., 2005; Downey and Hanlin, 2010). 8

24 Phenolics, specifically anthocyanins and tannins, affect the taste, color, bouquet, and mouth-feel properties of wine (Soleas et al., 1997). Anthocyanins are only found in red grape varieties and contribute the color to red wine (Mazza et al., 1999; Winkel-Shirley, 2001; Fournand et al., 2006; Ribéreau-Gayon et al., 2006b; He et al., 2010). They are a group of phenolic compounds that are responsible for the purple, blue or red colors found in plants (Goto and Kondo, 1991). Anthocyanin pigments are found in the vacuoles of epidermal tissue of flowers and fruit, where ph influences the chemical form, and hence the color, of these compounds (Goto and Kondo, 1991; Manach et al., 2004). Tannins are the main phenolic component of red wine and consist of two different classes: condensed tannins and hydrolysable tannins. Hydrolysable tannins are extracted from oak barrels or added during winemaking, whereas condensed tannins are produced in the grape berry and thus are of importance for this study (Herderich and Smith, 2005). Tannins contribute to the astringency and bitterness experienced when drinking wine due to their ability to form complexes with proteins and polysaccharides (Haslam, 1974; Carvalho et al., 2006; Sarneckis et al., 2006; Conde et al., 2007; Escot et al., 2008; Chira et al., 2011). Bitterness and astringency are important sensory components in red wine. Bitterness is a taste mediated by sensory receptors (Chandrashekar et al., 2000; Vidal et al., 2004a). Astringency is a tactile sensation caused from the precipitation of salivary proteins leading to the loss of mouth lubrication (Green, 1993; Gawel, 1998; Vidal et al., 2004a). Astringency is one of the prime contributors to mouth-feel properties ascribed to wines and is commonly referred to as drying', 'roughing' and 'puckering (Gawel, 1998; McRae et al., 2013). The variation in mouth-feel and aging properties is attributed to the variation in tannin content, composition, and polymer length, with monomers being more bitter than astringent and larger molecular weight derivatives being 9

25 more astringent than bitter (Peleg et al., 1999; Vidal et al., 2003; Hanlin and Downey, 2009; Chira et al., 2011; McRae et al., 2013). 1.2 Phenolics and Their Biosynthesis Phenolics Overview Phenolics are a widespread group of more than 8000 specialized metabolites (Pereira et al., 2009; Dai and Mumper, 2010; Cartea et al., 2011), all of which contain an aromatic hydrocarbon ring with one or more hydroxyl groups associated with the ring structure directly (Soleas et al., 1997; Adams, 2006; Ribereau-Gayon et al., 2006b; Bowsher at al., 2008; Dai and Mumper, 2010). They account for approximately 40 percent of the organic carbon circulating in the biosphere (Croteau et al., 2000). Of these 8000 phenolics, more than 4000 come from a group known as the flavonoids (Harborne and Williams, 2000; Cheynier, 2005). Grape skin flavonoids consist of the tannins and the anthocyanins, which are secondary metabolites that aid in plant survival (Bennett and Wallsgrove, 1994). Flavonoids can be produced in response to a particular stress or during normal plant development (Treutter, 2006). For example, they are produced as a protective mechanism against predators (Bennett and Wallsgrove, 1994; Lev-Yadun et al., 2009; Barbehenn and Constabel, 2011), disease (Dai et al., 1995; Lima et al., 2012), and UV-B radiation (Solovenko and Schmitz-Eiberger, 2003; Berli et al., 2011). Plants also use anthocyanins to communicate with pollinators and seed dispersers by the use of color, thereby enhancing reproductive success of the plant (Weiss, 1995; Schaefer et al., 2004). This allows consumption of sphenolics to have either direct toxic effects or to provide barriers (Bennett and Wallsgrove, 1994). It has been shown in many cases that insects prefer cultivars low in phenolics (Larson and Berry, 1984; Leszczynski et al., 1989). Studies have also shown that phenolics produced by the plant cause resistance to other herbivores including birds and slugs (Bullard et al., 1980; Johnston and Pearce, 1994). 10

26 Tannins and anthocyanins along with other phenolic molecules have also been shown to have antioxidant properties in plants providing them with protection. The hydroxyl groups on phenolic compounds easily donate hydrogen to reactive oxygen and reactive nitrogen species (Paya et al., 1992; Porter et al., 2001; Choi et al., 2002; Valentao et al., 2002). If this donation occurs in a termination reaction then it breaks the cycle that would regenerate new oxygen or new nitrogen radicals (Pereira et al., 2009). The interaction of the phenolic compound with a reactive species generates a new radical from the antioxidant; however, this new radical is much more stable than the original reactive species (Parr and Bolwell, 2000; Nijveldt et al., 2001; Pereira et al., 2009). The stable phenolic radical then can either diffuse away or can be reduced back to what it was originally (Parr and Bolwell, 2000). Phenolic compounds are also effective at stopping the production of free radicals by chelating metal (Moran et al., 1997; Mira et al., 2002). Phenolics with 2,3-double bonds and both the catechol group in the B-ring and the 3- hydroxyl group are most effective at reducing iron, while phenolics with more hydroxyl groups are most effective at reducing copper (Mira et al., 2002). Phenolic compounds can also inhibit enzymes that are involved in the generation of radicals (Pereira et al., 2009). The benzenoid rings and hydrogen bonding potential of phenolic compounds, especially tannins, allow them to bind to proteins, such as lipoxygenase, and inactivate them (Adams, 2006; Pereira et al., 2009). Anthocyanins have also been shown to be directly involved in photoprotection either by shielding leaf tissue or by scavenging reactive oxygen species (Wang et al., 1997; Archetti et al., 2009; Zhang et al., 2012). They shield the leaf by absorbing light thereby reducing its penetration to the mesophyll and reducing the production of reactive oxygen species (Steyn et al., 2002; Manetas et al., 2003, Merzlyak et al., 2008). 11

27 Phenolic Structures and Properties Grape skin phenolics can be broken down into two classes: the flavonoids and the nonflavonoids. Flavonoids consist of the tannins and anthocyanins and all share the same basic C 6 - C 3 -C 6 structure containing two phenols (rings A and B) joined by a pyran ring (ring C) (Soleas et al., 1997). The non-flavonoids consist of the phenolic acids and the stilbenes. Phenolic acids have either a C6-C3 structure and are referred to as hydroxycinnamic acids or a C6-C1 structure and are referred to as the hydroxybenzoic acids (Ribéreau-Gayon et al., 2006b). This C6-C3 or C6-C1 linkage corresponds to the carboxyl linkage to the benzyl ring (Ribéreau-Gayon et al., 2006b). Stilbenes consist of two aromatic rings linked by a double bond and are found in oligomeric and polymeric forms (Monagas et al., 2005). Phenolics originate from the biochemical pathway beginning with phenylalanine (Figure 1.3) (Parr and Bolwell, 2000; Adams, 2006; Cartea et al., 2011). Phenylalanine is produced from shikimic acid which is a product of the shikimic pathway that combines products from the pentose phosphate pathway and glycolysis (Parr and Bolwell, 2000; Conde et al., 2007). Phenylalanine is converted into cinnamic acid by the enzyme phenylalanine ammonia lyase (Blount et al., 2000; Conde et al., 2007). Phenylalanine ammonia lyase is the key enzyme in secondary metabolite production. It regulates the flux into the production of secondary metabolites, the phenylpropanoid pathway, and not primary metabolites (Howles et al., 1996). Cinnamic acid then transforms into coumaric acid by the enzyme cinnamate-4-hydroxylase, then coumaric acid is transformed into 4-courmaroyl-CoA by the enzyme 4-hydroxycinnamate: CoA ligase (Croteau et al., 2000; Achnine et al., 2004; Conde et al., 2007). 4-courmaroyl-CoA combines with three molecules of malonyl-coa, produced from the acetate pathway from acetyl- CoA, starting the phenylpropanoid pathway (Adams, 2006; Conde et al., 2007). Coumaric acid can also be transformed, through a series of steps, into simple phenolics such as the phenolic 12

28 acids caftaric acid, coutaric acid, fertaric acid and quercetin (Conde et al., 2007). The phenylpropanoid pathway goes on to generate more complex phenolic compounds such as stilbenes, anthocyanins, monomeric flavanols, and polymeric flavanols, depending on which enzyme is utilized. For example at the start of the phenylpropanoid pathway, chalcone synthase leads to the production of the flavonoids, whereas stilbene synthase (SS) leads to the production of stilbenes, such as resveratrol (Conde et al., 2007). 13

29 Figure 1.3: Biosynthetic pathway of the major phenolic classes (Adams, 2006; Conde et al., 2007; He et al., 2010). Tannins Condensed tannins are polymeric flavan-3-ol molecules composed of flavanol monomer subunits and are found in the hypodermal layers in the skin (Adams, 2006; Ribéreau-Gayon et al., 2006b). All tannins share a basic structure (Figure 1.4) consisting of two benzene cycles 14

30 bonded by a saturated oxygenated heterocycle (Soleas et al., 1997; Adams, 2006; Ribéreau- Gayon et al., 2006; Dai and Mumper, 2010). The structure consists of the A-ring, a phloroglucinol derivative, which is linked to the C or pyran ring, which in turn is linked to the B or phenolic ring (Adams, 2006). The properties of the flavonoid monomers, their bonding, esterification to other compounds, and functional properties are the basis for tannin classification (Soleas et al., 1997). Figure 1.4: Basic monomer tannin structure found in grape skins can be linked by a carbon-carbon interflavan bond (C4-C6 or C4-C8) to form polymers. R1 can be a hydrogen group or a gallate group and R2 can be hydrogen or a hydroxyl group. The subunits (Figure 1.5) that make up tannins in grape skin are (-)-epicatechin, (-)- epigallocatechin, (+)-catechin, and (-)-epicatechin-3-o-gallate. These monomers can condense to form tannin polymers (proanthocyanins) with great complexity forming polymers with up to 80 subunits (Souquet et al., 1996; Adams, 2006). (-)-Epicatechin-3-O-gallate makes up a large portion of polymeric tannin subunits, this monomer consists of an epicatechin subunit condensed with a gallate group at the oxygen in position C3 of the pyran ring (Downey et al., 2003; Fournand et al., 2006). Tannin oligomers are mainly linked through C4-C8 (and sometimes C4- C6) known as B-type bonds (Fletcher et al., 1977; Foo and Porter, 1980; Adams, 2006; Hanlin et al., 2010). An additional bond can occur via an ether bond between C2-C5 or C2-C7, known as 15

31 A-type (Vidal et al., 2004b). It is currently unknown how this inter-flavan bond is formed between tannin monomers. It has been suggested however, that this polymerization occurs by the action of enzymatic or non-enzymatic oxidation in vacuoles through the use of vesicle or membrane transporters powered by tonoplast proton pumps (Abrahams et al., 2003; Baxter et al., 2005; Zhao et al., 2010). Figure 1.5: Structure of the flavanol-3-ol monomeric molecules that are subunits that make up tannins in grape skins. The production of monomeric tannins requires the action of several enzymes that join caffeoyl-coa and malonyl-coa from the shikimate pathway to produce tannins (Figure 1.6) (Adams, 2006). (+)-Catechin and (-)-epicatechin differ only by the stereochemistry of the hydroxyl group on C3 (pyran ring C) (Adams, 2006). The pyran ring is only found in R stereochemistry in wine and grapes and the two asymmetric centers C2 and C3 lead to four isomers (Adams, 2006). 2,3-trans-flavan-3-ol is produced from flavanonol by dihydroflavanol reductase, which can produce (+)-catechin directly by leucoanthocyanidin reducatase because it has the 2,3-trans configuration (Adams, 2006). (-)-Epicatechin has the 2,3-cis configuration and because of the stereochemistry cannot be directly converted from 2,3-trans-flavan-3-ol and has to be produced from cyanidin by anthocyanidin reductase (Xie et al., 2003; Adams, 2006). Anthocyanidin produces (-)-epigallocatechin from delphinidin (Xie et al., 2003). 16

32 Figure 1.6: Biosynthesis of (+)-catechin and (-)-epicatechin from caffeoyl CoA and malonyl-coa adapted and re-drawn from Adams (2006). CHS, chalcone synthase; CHI, chalcone isomerase; F3H, flavanone-3- hydroxylase; DFR, dihydroflavonol-4-reductase; LAR, leucoanthocyanidin reductase; ANS, anthocyanidin synthase; ANR, anthocyanidin reductase; UFGT, UDP glucose-flavonoid 3-O-glucosyl transferase. Anthocyanins Anthocyanins are also found in the hypodermal cells of the grape skin and are the second most abundant polyphenol (Adams, 2006). All anthocyanins share a basic structure, including two benzene rings bonded by an unsaturated cationic oxygenated heterocycle (Ribéreau-Gayon et al., 2006b). Like the tannins, anthocyanins consist of an A-ring, a C or pyran ring, and a B or phenolic ring (Adams, 2006). The oxygen molecule in the pyran ring is where the basic anthocyanin structure differs from the basic tannin structure. Anthocyanins also have an unsaturated cationic oxygenated heterocycle, whereas tannins have a saturated oxygenated 17

33 heterocycle (Ribéreau-Gayon et al., 2006b). They differentiate by the way they are substituted at the R groups (Figure 1.7). Figure 1.7: Basic structure of anthocyanidins, the precursors to anthocyanins. Anthocyanins have a glucose molecule bonded to the oxygen which is bonded to the carbon at position three (R1). In certain varieties, Cabernet franc and Merlot, but not Pinot noir, some esterification can occur on the OH function of the glucose with coumaric or acetic acid. R3 and R2 are substituted with either a hydrogen molecule (H), a hydroxyl molecule (OH), or a methoxy molecule (OCH3). Anthocyanins found in grape skin are delphinidin, cyanidin, petunidin, peonidin, and malvidin. Anthocyanins are classified as either glycoside or acylglycosides where acylation is made with p-coumaric, caffeic and acetic acids (Mazza and Francis, 1995; Ribéreau-Gayon et al., 2006b; He et al., 2010). In some varieties, such as Pinot noir, acylation is not present; therefore, they produce five (Figure 1.8) of the twenty-five anthocyanins (Cheynier et al., 2006; Adams, 2006). Anthocyanins can also be classified as monoglucoside or diglucosides. In V. vinifera species, such as Pinot noir and Cabernet franc, only monoglucoside anthocyanins can be produced (Ribéreau-Gayon et al., 2006b). In this case, glucose molecules can only be linked to anthocyanidins at position C3 to form 3-O-monoglucoside anthocyanins because they lack the ability to produce diglucosides (Ford et al., 1998; Ribéreau-Gayon et al., 2006b; Janvary et al., 2009). In other Vitis species, glucose molecules can be linked at both the C3 and C5 to produce diglucosides (Ribéreau-Gayon et al., 2006b; Janvary et al., 2009). 18

34 Figure 1.8: Structure of the five anthocyanidins found in Pinot noir. Anthocyanidins are precursors to anthocyanins where anthocyanins have a glucose molecule bonded to the oxygen molecule at position 3. Anthocyanins are synthesised in the flavonoid pathway (Figure 1.9) starting with the production of chalcone. Chalcone is produced through the condensation of three molecules of malonly-coa and one molecule of p-coumaroyl-coa by the action of chalcone synthase (Forkmann, 1991; Holton and Cornish, 1995; He et al., 2010). Chalcone is converted to naringenin by chalcone isomerase, and then naringenin is oxidized by flavanone 3β-hydroxylase to form dihydrokaempferol (Sparvoli et al., 1994; He et al., 2010). Dihydrokaempferol acts as a substrate for flavonoid 3 -hydroxylase and flavonoid 3,5 -hydroxylase, to produce the corresponding dihydroflavonols; dihydroquercetin, and dihydromyricetin, respectively (Sparvoli et al., 1994; He et al., 2010). These dihydroflavonols are sequentially reduced to their leucoanthocyanidins by dihydroflavonol 4-reductase (Boss et al., 1996; Bogs et al., 2006). Leucoanthocyanidins are then reduced to their corresponiding anthocyanidins by anthocyanidin 19

35 synthase, which can then be glycosylated to their corresponding anthocyanins by UDP glucoseflavonoid 3-O-glucosyl transferase (He et al., 2010). 20

36 Figure 1.9: Pathway showing anthocyanin biosynthesis (Matus et al., 2009; He et al., 2010). CHS, chalcone synthase;; CHI, chalcone isomerase;; F3H, flavanone 3β-hydroxylase;; F3 H, flavonoid 3 - hydroxylase; F3 5 H, flavonoid 3,5 -hydroxylase; DFR, dihydroflavonol 4-reductase; ANS, anthocyanidin synthase; UFGT, UDP glucose-flavonoid 3-O-glucosyl transferase. 21

37 The production of anthocyanins and condensed tannins (Figure 1.10) shares common steps. This involves the formation of flavan-3,4-diols which contributes to anthocyanin synthesis and to extension units of condensed tannins (Stafford as cited in Bogs et al. 2005). Leucocyanidin is reduced by leucoanthocyanidin reductase to catechin. Leucocyanidin can then be made into cyanidin by anthocyanin synthase. Anthocyanidin reductase can then reduce cyanidin to epicatechin, both of which can form the terminal subunit of the condensed tannin (Bogs et al., 2005). If UDP glucose-flavonoid 3-O-glucosyl transferase is used on cyanidin instead of anthocyanidin synthase, then the anthocyanin cyanidin-3-glucocide is produced (Bogs et al., 2005). Figure 1.10: Biosynthetic flavonoid pathway leading to the production of proanthocyanidin polymer (condensed tannin) and anthocyanidin re-drawn and adapted from Bogs et al. (2005); ANS, anthocyanidin synthase; UFGT, UDP glucose-flavonoid 3-O-glucosyl transferase; LAR, leucoanthocyanidin reducatase; ANR, anthocyanidin reductase. 22