DEVELOPMENT AND VALIDATION OF METHODS FOR THE INVESTIGATION OF WINE

by RYAN KURTIS MOSS

B.Sc., University of British Columbia, 2011

A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF

MASTER OF SCIENCE

in

THE COLLEGE OF GRADUATE STUDIES

Chemistry

THE UNIVERSITY OF BRITISH COLUMBIA (Okanagan)

September 2014

© Ryan Kurtis Moss, 2014

Abstract

Stilbenoids are secondary plant metabolites responsible for the protection of vine from bacterial and fungal infection. Red wine has been shown to be a major source of these compounds in the human diet, where they display an array of health benefits.

The first goal of our study was to develop and validate a robust and selective method for quantification of the major stilbenoids in red and white wine using ultra-high-performance liquid chromatography coupled with electrospray ionization/quadrupole time-of-flight mass spectrometry (UHPLC-ESI-Q-TOF). Both isomers of and were quantified externally using authentic standards, while was quantified in trans-piceid equivalents. Due to the minimal amount of sample preparation and the short method runtime, results were obtained rapidly and with low expenditure of energy, chemicals, and labor.

The method was validated with respect to linearity, limit of detection (LOD), limit of quantification (LOQ), accuracy, intra- and inter-day precision and stability. All six stilbene monomers were quantified in 44 British Columbian wines; the highest total stilbene concentration in an individual wine was 28.81 mg/L in Pinot Noir, while the average across all wines was 8.49 ± 6.25 mg/L.

Another method was developed for separation, identification and semi-quantification of all derivatives of resveratrol that are present in wine.

ii

A total of 41 (both known and novel) stilbenoids were detected in extracted red wine. In addition to the well-known monomeric stilbenes, several resveratrol-resveratrol homodimers

(m/z 453.1344), resveratrol-piceatannol heterodimers (m/z 469.1293) and piceatannol- piceatannol homodimers (m/z 485.1236) were detected. Modified dimers of resveratrol were also detected. Multiple trimers of resveratrol (m/z 679.1978) were detected for the first time in red wine, as well as some known and some novel tetramers (m/z 905.2604).

A solid-phase extraction (SPE) method was developed for quantification of the stilbenoid oligomers in red wines. The monomers and oligomers in red wine from the Okanagan Valley

(Cabernet Sauvignon, Merlot and Pinot Noir) and Québec (Maréchal Foch, ,

Sabrevois, St.Croix, and ) wines were semi-quantified as using this method. The highest concentration of total stilbenoids was 10.67 mg/L in Pinot Noir with an overall average of 3.32 ± 2.86 mg/L in all wines.

iii

Preface

All of the optimization, sample preparation and analysis using ultra-high performance liquid chromatography, coupled with quadupole time-of-flight mass spectrometry were performed on the University of British Columbia, Okanagan campus (UBCO) by Ryan Moss. Many of the standards used in this study were synthesized at The University of Adelaide, School of

Agriculture, Food and Wine in Dr. Dennis Taylor’s lab, by Qinyong Mao. The study was designed by Dr. Cédric Saucier, Dr. Dennis Taylor, and Ryan Moss. A manuscript containing the identification of numerous stilbenoid derivatives, many newly discovered in wine, was prepared primarily by Ryan Moss with editorial assistance from Dr. Saucier. The manuscript was published in Rapid Communications in Mass Spectrometry in May 2013. The same results were presented in a poster presentation at the 2013 annual American Society of Mass Spectrometry in

June 2013.

iv

Table of Contents

Abstract ...... ii Preface ...... iv Table of Contents ...... v List of Tables ...... viii List of Figures ...... ix List of Abbreviations ...... xii Acknowledgements ...... xiv CHAPTER 1 – INTRODUCTION ...... 1 1.1 General ...... 1 1.1.1 Stilbenes ...... 5 1.1.2 Metabolism and Biosynthesis of Stilbenes in Vitis vinifera...... 6 1.1.3 Medical and Health Benefits of Stilbenes ...... 11 1.2 Instrumentation Background ...... 12 1.2.1 Ultra High Performance Liquid Chromatography ...... 12 1.2.1 General Mass Spectrometry ...... 14 1.2.1.1 Electrospray Ionization ...... 17 1.2.1.2 Quadrupole-Time-of-Flight (Q-TOF) Mass Analyzer ...... 19 1.3 Identification and Quantification of Stilbenoids in Wine Using Mass Spectrometry ...... 22 1.4 Research Objectives ...... 24 1.5 Hypotheses & Rationale ...... 24 1.6 Significance of study ...... 26 CHAPTER 2 – DEVELOPMENT OF A VALIDATED METHOD FOR QUANTIFICATION OF MONOMERIC STILBENES IN WINE ...... 27 2.1 Synopsis ...... 27 2.2 Experimental ...... 28 2.2.1 Test Samples ...... 28 2.2.2 Reagents and Materials ...... 28 2.2.3 Calibration Standards ...... 28 2.2.4 UHPLC-ESI-Q-TOF Analysis ...... 30 2.2.5 Validation ...... 32

v

2.2.5.1 Linearity, Limits of Detection and Quantification ...... 32 2.2.5.2 Intra-, Inter-day Repeatability, and Stability ...... 33 2.2.5.3 Recovery ...... 33 2.3 Results & Discussion ...... 35 2.3.1 Optimization of UHPLC-ESI-Q-TOF and Analysis ...... 35 2.3.1.1 Optimization of UHPLC Parameters ...... 35 2.3.1.2 Optimization of ESI-Q-TOF Parameters ...... 37 2.3.2 Validation ...... 43 2.3.2.1 Specificity ...... 43 2.3.2.2 Calibration Curves ...... 45 2.3.2.3 Linearity, Limits of Detection, and Quantification ...... 48 2.3.2.4 Precision, Stability, and Recovery ...... 50 2.3.4 Applications to Wine Samples ...... 52 2.3.4.1 Wine samples ...... 52 2.3.4.2 Quantification in Wine Samples ...... 53 CHAPTER 3 – INDENTIFICATION AND SEMI-QUANTIFICATION OF OLIGOMERIC STILBENOIDS IN RED WINE ...... 61 3.1 Synopsis ...... 61 3.2 Experimental ...... 63 3.2.1 Reagents, Samples & Solutions ...... 63 3.2.2 Sample Preparation ...... 63 3.2.2.1 Compound Elucidation Sample Preparation ...... 63 3.2.2.2 Quantitative Analysis Sample Preparation...... 63 3.2.3 Ultra-High Performance Liquid Chromatography Parameters ...... 64 3.2.4 Quadrupole-Time-of-Flight Parameters ...... 65 3.2.5 Quantitative Method Validation ...... 65 3.2.5.1 Calibration Curves ...... 65 3.2.5.2 Solid Phase Extraction Validation ...... 66 3.2.5.3 Linearity, Limits of Detection & Quantification ...... 68 3.2.5.4 Recovery, Inter & Intra-day Precision ...... 69 3.3 Results & Discussion ...... 70 3.3.1 Optimization of UHPLC and MS/MS parameters for qualitative analysis...... 70

vi

3.3.2 Strategy for Compound Identification without Standards...... 71 3.3.3 Identification of stilbene compounds in wine (by accurate MS/MS)...... 74 3.3.3.1 Monomers ...... 78 3.3.3.2 Dimers ...... 80 3.3.3.3 (R+P) Dimers ...... 82 3.3.3.4 Dimers (P+P) ...... 84 3.3.3.5 Glycosylated Dimers ...... 84 3.3.3.6 Methoxylated Dimers ...... 85 3.3.3.7 Oxidized Dimers ...... 86 3.3.3.8 Trimers ...... 87 3.3.3.9 Tetramers ...... 88 3.3.4 Quantitative Method Validation ...... 90 3.3.4.1 SPE Validation ...... 90 3.3.4.2 Calibration Curve, Linearity, Limits of Detection and Quantification ...... 92 3.3.4.3 Recovery, Intra- & Inter-day Precision ...... 93 3.3.5 Quantification Results ...... 95 3.3.5.1 Wine Samples ...... 95 3.3.5.2 Total Stilbenoid Levels in Okanagan and Quebec Wines ...... 96 CHAPTER 4 – CONCLUSIONS ...... 109 4.1 – Summary of Research ...... 109 4.2 – Research Novelty ...... 111 4.3 – Assumptions and Limitations...... 111 4.4 – Future Directions ...... 112 REFERENCES ...... 114 APPENDIX A: Additional Data ...... 128

vii

List of Tables

Table 2.1: Dilution series used to create calibration curves for trans-/cis-resveratrol and trans-/cis-piceid...... 29 Table 2.2: Outline of the serial dilution scheme used, defining the procedural details of the dilution...... 30 Table 2.3:UHPLC 6.75 minute gradient elution program; Solvent A = 0.1% formic acid in H2O, Solvent B = 0.1% formic acid in acetonitrile...... 31 Table 2.4: Validation result including the limits of detection and quantification, as well as the precision (intra- and inter-day) results...... 49 Table 2.5: Recovery experimental results (N=3 for each concentration) using four concentration levels in both red and white wine...... 52 Table 2.6: Quantification of all 6 monomer stilbenoids in 44 Okanagan wines. Error is displayed as standard deviation, indicating the error of the method including sample preparation and instrumental error...... 55 Table 2.7: Quantification results displaying the minimum and maximum total and individual stilbenoid concentrations, separated by variety and wine type. Error expressed as standard deviation, indicative of the variability of the sample pool...... 57 Table 3.1: Gradient elution method, A: H2O w/ 0.1 % Formic acid, B: Acetonitrile w/0.1 % Formic acid...... 64 Table 3.2: Serial dilution protocol for the preparation of the calibration curve standards...... 66 Table 3.3: Theoretical [M-H]- of many potential stilbenoid derivatives that were scanned by negative mode UHPLC-ESI-QTOF analysis in red wine ...... 72 Table 3.4: Fragmentation patterns and tentative assignments of stilbenoid compounds in red wine extract...... 75 Table 3.5: SPE extraction efficiency results for low, medium, and high concentrations ...... 91 Table 3.6: Calibration curve validation results, including Limits of detection and quantification...... 93 Table 3.7: Recovery results and for multiple days used to calculate both intra- and inter-day precision...... 94 Table 3.8: Condensed quantification results for each stilbenoid compound in each red wine variety ...... 97 Table 3.9: Wine quantification results from literature studies on known stilbenoids...... 100

Table A 1: Quantification data for all 41 compounds in each individual wine. Concentrations shown in ng/mL trans-piceid equivalents...... 128 Table A 2: Detailed MS/MS information for identified compounds 1-41...... 133

viii

List of Figures

Figure 1.1: Basic Flavonoid structure showing three subfamilies, the anthocyanins, the flavonols, and the flavan-3-ols...... 2 Figure 1.2: Biosynthetic pathway leading to the production of resveratrol and all of the stilbenoid derivatives. PAL, phenylalanine ammonia lyase; C4H, cinnamate-4- hydroxylase; 4CL, hydroxycinnamoyl CoA ligases; STS, stilbene synthase; CHS, chalcone synthase...... 3 Figure 1.3: Stilbene monomer structure, showing the primary stilbene derivatives that are present in a variety of plant species...... 5 Figure 1.4: Resveratrol may be oxidatively coupled to form higher oligomers such as the dimers trans-ε-viniferin and trans-δ-viniferin, the trimer α-viniferin, and the tetramer ...... 9 Figure 1.5: Box diagram of the major components of a mass spectrometer, and how they fit together...... 15 Figure 1.6: Diagram showing the Agilent AJS Electrospray source, highlighting the various gas flows, and applied voltages that can be altered to optimize ionization...... 18 Figure 1.7: Agilent 6530 QTOF diagram showing the inside of the mass spectrometer including the multiple vacuum stages, ion guides, mass analyzers, and reflectron...... 21

Figure 2.1: Chromatogram displaying the effect on analyte separation of increasing concentration of acetonitrile mobile phase over the 4 minute gradient. 1- trans- piceid; 2- cis piceid; 3-trans-resveratrol; 4-cis-resveratrol...... 36 Figure 2.2: Chromatogram displaying the effect of increasing column temperature on the separation, and retention time of the analytes. Darker colors indicate higher column temperature...... 36 Figure 2.3: Effect of increasing fragmentor voltage on the ionization of: A- trans-resveratrol; B-cis-resveratrol; C-trans-piceid; D-cis-piceid...... 38 Figure 2.4: The effect of increasing the capillary voltage on the ionization of each analyte...... 40 Figure 2.5: A- The effect on the peak shape when increasing the number of scans per second. B- A decrease in detection of ions as the number of scans per second is increased leading to decreased sensitivity. C-The negligible effect on the precision of repeated measurements resulting from increasing the scan rate of the instrument...... 42 Figure 2.6: A - MS/MS spectrum of [M-H]- 227.0714 m/z; resveratrol. B - MS/MS spectrum of [M-H]- 389.1242 m/z; piceid. C-MS/MS spectrum of [M-H]- 243.0663 m/z; piceatannol...... 44

ix

Figure 2.7: A-Combined EIC (m/z 227.0714 + 389.1242) of fresh resveratrol and piceid standards. B-Combined EIC (m/z 227.0714 + 389.1242) of resveratrol and piceid standards after exposure to sunlight for 6 hours, displaying the conversion from trans (A) to cis (B)...... 45 Figure 2.8:Combined EIC displaying the relative retention times between EGC3OG (A), trans-piceid (B), cis-piceid (C), trans-resveratrol (D), cis-resveratrol (E)...... 47 Figure 2.9: Calibration curves for (A) trans-resveratrol, (B) cis-resveratrol, (C) trans-piceid, (D) cis-piceid displaying the linear regression...... 48 Figure 2.10: Distribution of wine based on vintage and varietal quantified in this study (A). Total number of each variety analyzed (B)...... 53 Figure 2.11: Quantification results represented in graph form separated by red and white wine varieties...... 59 Figure 3.1: Top: MS/MS spectrum of authentic standard trans-ε-viniferin [M-H]- 453.1344 m/z . Bottom: previously unidentified MS/MS spectrum of [M-H]- 615.1872 m/z, tentatively identified as the 3-O-glucoside of trans-ε-viniferin...... 73 Figure 3.2: Tentative structures of the major compounds found in red wine extracts studied. 1,2 trans- and cis-resveratrol, 3 trans-piceatannol, 4,5 trans- and cis-piceid, 6 trans-, 7 , 8 parthenocissin A, 9 , 10 ampelopsin D, 11, 13 cis- and trans-ε-viniferin, 12, 14 cis- and trans-ω- viniferin, 16,17 trans- and cis-δ-viniferin, 19,20 trans- and cis-scirpusin A, 21 restrisol A or B, 25 scirpusin B, 26, 27 parthenostilbenin A and B, 28 ε- viniferin glucoside, 30 ampelopsin C, 31,32 E- and Z-, 39 hopeaphenol, 41 isohopeaphenol...... 77 Figure 3.3: Combined base peak chromatogram of m/z 227.0714, 243.0663, 389.1242, and 405.1191 showing compounds 1-6...... 78 Figure 3.4: Fragmentation pathway of the stilbene monomers (1, 2, and 3) and their glycosylated analogues (4, 5, and 6)...... 79 Figure 3.5: A: BPC scan of m/z 453.1344 showing compounds 7-17. B: MS/MS BPC showing the transition m/z 453.1344 -> 359.0925 displaying the indane type resveratrol dimers 7-10. C: MS/MS BPC of the transition m/z 453.1344 -> 411.1238 highlighting the "viniferin" type dimers 11-17...... 80 Figure 3.6: Combined BPC Chromatogram of the stilbenoid oligomers 18-41 found in Pinot Noir wine extract...... 83 Figure 3.7: trans-piceid calibration curve ranging from 0.01 - 30 mg/L ...... 92 Figure 3.8: Mean stilbenoid profile displaying mean concentration levels for N=41 red wines analyzed. All compounds above the dotted line have been reported in red wine, while all compounds below the line are first described in wine in this study...... 99 Figure 3.9: Average breakdown of monomer stilbenoids in each red wine variety analyzed (±SD)...... 101

x

Figure 3.10: Average breakdown of resveratrol dimer stilbenoids in each red wine variety analyzed (±SD)...... 103 Figure 3.11: Average breakdown of “modified’ dimer stilbenoids in each red wine variety analyzed (±SD)...... 104 Figure 3.12: Average breakdown of trimer stilbenoids in each red wine variety analyzed (±SD)...... 105 Figure 3.13: Average breakdown of tetramer stilbenoids in each red wine variety analyzed (±SD)...... 106 Figure 3.14: Average stilbenoid concentrations in red wines, separated by Vitis vinifera or hybrid varieties...... 107

xi

List of Abbreviations

AJS Agilent jet spray ANOVA analysis of variance APCI atmospheric pressure chemical ionization APPI atmospheric pressure photoionization BPC base peak chromatogram C4L 4-coenzyme A ligase CE collision energy CHS chalcone synthase CID collision induced dissociation DAD diode array detector ECD electron capture dissociation EIC extracted ion chromatogram ESI electrospray ionization FDA food and drug administration FL fluorescence detector GC gas chromatography HPLC high performance liquid chromatography ICH international conference on harmonization IS internal standard LC liquid chromatography NMR nuclear magnetic resonance LLE liquid/liquid extraction LOD limit of detection LOL limit of linearity LOQ limit of quantification m/z mass/charge MCP micro channel plate MS mass spectrometry MSn tandem mass spectrometers NMR nuclear magnetic resonance PAL phenylalanine ammonia lyase PTFE polytetrafluoroethylene QqQ triple quadrupole QTOF quadrupole time of flight RF radio frequency RP reversed phase RRHD rapid resolution high definition RSD relative standard deviation

xii

SAR structure activity relationship SD standard deviation SID surface induced dissociation SIL stable isotopically labelled SIM selective ion monitoring SPE solid phase extraction STOx stilbene oxidase STS stilbene synthase TOF time of flight UHPLC ultra high pressure liquid chromatography UV ultraviolet spectrophotometry UV/VIS ultraviolet/visible spectrophotometry VQA vintners quality alliance

xiii

Acknowledgements

The first person I would like to thank is my supervisor, Dr. Cédric Saucier, for sharing his support and experience and for offering me the opportunity to conduct research in his wine chemistry lab. I would also like to thank my committee members, Drs. Kevin Smith and James

Bailey. I would also like to thank the senior members of the wine chemistry lab Drs. Adeline

Delcambre, Eric Denis, and Yann Andre for their mentorship throughout my project. I would never have been able to achieve any of this, nor would I have had as much fun without you guys.

A thank you goes to all of the other lab mates from abroad who I’ve had the pleasure of getting to know for making the lab an enjoyable place to work. Finally, I would like to thank all of my fellow graduate students in both the chemistry, and biology faculties, for making graduate school a truly unforgettable experience.

This project was funded through grants from the Natural Sciences and Engineering Research

Council of Canada (NSERC), Canada Foundation for Innovation, and University of British

Columbia, awarded to Dr. Cédric Saucier.

xiv

CHAPTER 1 – INTRODUCTION

1.1 General Polyphenols

Polyphenols are a family of secondary metabolites produced in many plant species. Their name derives from their characteristic (poly)phenolic groups (an aromatic ring substituted by single or multiple hydroxyl -OH groups). The phenolic group(s) is acidic while the aromatic ring has nucleophilic character, which characterises some of their reactivity. Polyphenolic compounds are scientifically significant because they have a number of therapeutic uses and are known to possess potent antioxidant and anti-inflammatory activities.1,2 Many of these plant-derived compounds are found in fruits and vegetables, and their subsequent consumable products, endowing them as highly valuable to the medicinal and food chemical industries and research bodies.

Vitis vinifera (a species of grape) is most notable for its use in the production of wine, which is considered to be one of the biggest sources of polyphenolic compounds in the human diet.3 Based on their structures in plants, polyphenols are split into two large sub-families of compounds known as flavonoids and non-flavonoids.

1

Figure 1.1: Basic Flavonoid structure showing three subfamilies, the anthocyanins, the flavonols, and the flavan-3-ols.

Flavonoids are separated from other polyphenolic compounds based on their characteristic

C6-C3-C6 skeleton (Figure 1.1). The most important of these compounds are flavonols, anthocyanins and tannins. Flavonols act as potent UV-protectants for the plant, and also early stepping stone in the flavonoid pathway on the way to anthocyanins and condensed tannins.4

Anthocyanins are the principle color agents in plants and possess an array of different colors for flowers, fruits and vegetables. It has been shown that anthocyanins possess antioxidant effects in vitro but this has yet to be established in vivo.5 Condensed tannins or proanthocyanidins, are carbon-carbon linked polymers of flavan-3-ol monomers. They are very important to the taste and mouth feel of wine as they contribute the bitter and astringent characteristics. More recently, they have been getting some recognition for their potential health effects as well.6

2

trans-ε-viniferin

trans-piceid

cis-resveratrol

Figure 1.2: Biosynthetic pathway leading to the production of resveratrol and all of the stilbenoid derivatives. PAL, phenylalanine ammonia lyase; C4H, cinnamate-4-hydroxylase; 4CL, hydroxycinnamoyl CoA ligases; STS, stilbene synthase; CHS, chalcone synthase.

3

Non-flavonoid compounds are polyphenolic compounds created in the phenylpropanoid pathway which do not possess a C6-C3-C6 skeleton. Such compounds that fall into this category include: hydroxybenzoic acids, hydroxycinnamic acids, and stilbenes. The most abundant hydroxybenzoic acid is gallic acid, which is a by-product of the shikimate pathway. 3-

Dehydroshikimate can be reduced to gallic acid instead of continuing on to shikimate to complete flavonoid synthesis.7 From here, gallic acid is usually esterified to form complexes with flavonoid compounds such as epicatechin-3-O-gallate which eventually can form tannins. This is a reversible process though, as tannins can also be hydrolyzed to release gallic acid which can be abundant in wine. The hydroxycinnamic acids are important intermediate compounds in the phenylpropanoid pathway leading to flavonoids such as flavonols or anthocyanins as well as stilbenes (Figure 1.2). These C6-C3 molecules are biosynthesized through the enzymatic action of phenylalanine ammonia lyase (PAL), on phenylalanine, and to a lesser extent, tyrosine.8 The three main hydroxycinnamates, coumaric, ferulic and caffeic acids, are most commonly found in as esters with tartaric acid or as simple glucosides.9 These acids also have the potential to be acylated to the glucose moiety of pigmented anthocyanin molecules.10

4

1.1.1 Stilbenes

Stilbenes, or stilbenoids, are a class of polyphenolic compounds which are hydroxylated derivatives of stilbene. They are characterized by their general monomeric C6-C2-C6 structure

(1,2-diphenylethylene) (Figure 1.3). This group of specialized metabolites are known to possess numerous potential health benefits and because of this, are very well established in the pharmaceutical and scientific communities.11,12 Stilbenoids are biosynthesized in many plant families including pine (Pinaceae), peanut (Fabaceae), sorghum (Poaceae), and grape

(Vitaceae).13 Most stilbenoids are considered to be derivatives from trans-resveratrol, however, some other monomeric stilbenoid units do exist across plant species (Figure 1.3). 14

Stilbene Monomer R3 R5 R3’ R4’ resveratrol OH OH H OH piceid OGlu OH H OH piceatannol OH OH OH OH astringin OGlu OH OH OH pinosylvan OH OH H H

OCH3 OCH3 H OH

rhapontin OGlu OH OH OCH3 Figure 1.3: Stilbene monomer structure, showing the primary stilbene derivatives that are present in a variety of plant species.

5

The principal interest in resveratrol comes from the nutraceutical industry, because of its purported health properties; additionally, the cosmetic industry also has uses for resveratrol and its derivatives for products such as face creams.15

Japanese knotweed roots (Poligonum cuspidatum) are known to be one of the most resveratrol-rich sources in the plant kingdom; the roots have been used as an anti-inflammatory herbal medicine in Asia for hundreds of years.15-17 Nevertheless, when considering human consumption of resveratrol and related stilbenoids, the most relevant sources are pine, peanuts, grapes, bilberry, mulberry and hops.18-20 Consequently such crops, including grapes and wine, are considered to be a key dietary source of health promoting specialized metabolites, including polyphenols such as resveratrol.21 Contrary to popular belief, the abundance of resveratrol in wine is quite low, usually only reaching low mg/L quantities.22

1.1.2 Metabolism and Biosynthesis of Stilbenes in Vitis vinifera.

In grapes, stilbenes are mainly synthesized and located in the skins of the berry, although these compounds are also found within the stems and leaves of the plant. Resveratrol production in the grape was found to decline during berry ripening and in general, there is a negative correlation between the age of the grape berry, and the rate of resveratrol biosynthesis.23 This diminishes the ability of the grape berry to fight off pathogens as it reaches maturity.24

It appears as though stilbenoids are a constitutive part of grapevine wood, but only an inducible component of grape leaves and berries. Production can be elicited through certain types of injury such as UV light irradiation, or infection from bacteria or fungus.25 It has also been observed that physical trauma to the plant does not induce the production of stilbenes.25 Previous

6

research has found that upon infection with the Botrytis cinerea fungus, the concentration of stilbenes found in grape berries increases rapidly as the fungus develops.26 Another study showed that resveratrol concentration in grape berries peaked 24 h after being exposed to a stressor like fungus or UV exposure, however, 48-72h later, the action of a laccase-like stilbene oxidase triggers the detoxification of resveratrol.27-29

The biosynthesis of resveratrol begins through the shikimate pathway which results in the formation of both phenylalanine and tyrosine. The enzyme, PAL, begins the phenylpropanoid pathway by converting phenylalanine to hydroxycinnamic acid (Figure 1.2), which through the action of 4-coenzyme A-ligase (C4L) is converted to 4-coumaroyl-CoA. Depending on the cinnamic acid acted upon by C4L, other stilbene monomer compounds, such as piceatannol or , may be produced.

Flavonoids are produced through the action of a type III polyketide synthase, called chalcone synthase (CHS), while stilbenoids are produced by a similar enzyme in the same family, called stilbene synthase (STS). STS shares the same mechanism and substrates as CHS for the condensation of one molecule of 4-coumaroyl-CoA and three molecules of malonyl-CoA in succession, producing the enzyme bound tetraketide intermediate (Figure 1.2).30 STS then converts this intermediate to the stilbene trans-resveratrol through an enzyme mediated aldol condensation.31 From this point, many different modifications to resveratrol can be made, such as isomerization, oligomerization or glycosylation. These modifications will be discussed in the following paragraphs.

7

In grape berries resveratrol is predominantly produced as the trans isomer, but it can also exist in its cis form (Figure 1.2) through photochemical isomerization. When grape berries and leaves are protected from light, cis-resveratrol is not detected, indicating that the plant biosynthesizes trans-resveratrol. This suggests the only source of the cis isomer is through UV- light induced isomerization.32 This increases the number of analogues of resveratrol as most modifications and derivatives can also exist as both isomers. This is an important consideration as future research into stilbenes must consider that both isoforms may exist in different plant products and must be quantified separately. 33

Aside from the basic trans and cis forms (Figure 1.2), resveratrol can be β-glucosylated at the 3-O or 4’-O positions. These glucosides are known as piceid and respectively.

This is an enzymatic transformation performed in most organisms by a ubiquitous glucosyltranferase.34 Much of the reported total stilbene content in grapes and wine has been found to be in the form of trans- and cis-piceid.22 The rationale for glycosylation of stilbenoids in plants is not well understood, but could be involved in transport, storage or even protection from peroxidative degradation.14 There is also a small amount of evidence for the O-methylation of resveratrol in grapes through the presence of pterostilbene, which is an O-methylated analogue of resveratrol in the 3-O and 5-O (pterostilbene) positions. This compound has been observed in grape leaves and berries, but not in wine.35 It has been suggested that pterostilbene is not stable in air and is therefore possibly degraded during the winemaking process.

8

Figure 1.4: Resveratrol may be oxidatively coupled to form higher oligomers such as the dimers trans-ε-viniferin and trans-δ-viniferin, the trimer α-viniferin, and the tetramer hopeaphenol.

Resveratrol units may also be oxidatively coupled together to form many different types of oligomers (Figure 1.4). The first such compounds isolated in grapevines were termed viniferins, ε- viniferin (a stilbene dimer) and α-viniferin (a resveratrol trimer) were found in grape leaves after

UV irradiation.25,36 Further studies, aiming to determine the mechanism for this oxidative coupling of resveratrol used the enzyme horseradish peroxidase to control oxidative coupling.

This produced an oligomer analogous to ε-viniferin but with a different coupling pattern which was named δ-viniferin.37 More recently, numerous stilbenoid oligomers are now known to be produced in all tissues of the grapevine in response to biotic stresses such as Plasmopara viticola

(powdery mildew), and Botrytis cinerea (gray mold), and abiotic stresses like physical trauma,

UV irradiation, and certain chemicals, for example, silver acetate.38,39 These stilbenoid oligomers are produced constitutively within the stalks of grape vines. The resveratrol dimer, ε-viniferin is considered to be the major oligomer intermediate within the vinifera species, however, a wide range of dimers, trimers and tetramers have also been described and reviewed.13,40

9

Botrytis cinerea and Plasmopora viticola are known to contain multiple laccase-like stilbene oxidase (STOx) enzymes, possibly evolved to specifically breakdown the grapevine defences. These fungi are at least in part, responsible for the production of several stilbenoid oligomers.41,42 It is still unclear what role this type of stilbenoid metabolism plays in fungi, but it has been suggested that the oxidative coupling produces oligomers with low water solubility, protecting the fungus from continued exposure to the grapevine phytoalexins.38,43

Aside from the stilbene oxidase-mediated coupling of stilbenoid monomers, the grapevine has its own enzymes which can catalyze oxidative couplings. Several isoperoxidase enzymes isolated from within vacuoles of the cell wall, and apoplasts of the grapevine cells have been shown to oligomerize resveratrol and other stilbenoid monomers.44,45 Resveratrol production seems to induce the production of these peroxidases which have been proposed to be responsible for the elicitor-induced and constitutive accumulation of resveratrol oligomers.42,46

In vitro, the production of both ε-viniferin, and δ-viniferin from resveratrol can be elicited using silver acetate. This predominantly produces δ-viniferin which is quite stable under the reaction conditions, while ε-viniferin will undergo further oxidations.39 Many different single electron donating metal catalysts can carry out the dimerization of resveratrol, each with different products.47 These proposed mechanisms could help explain the diversity of stilbenoid class compounds found in nature.

10

1.1.3 Medical and Health Benefits of Stilbenes

Resveratrol and stilbenes have been implicated as contributing factors to the “French

Paradox”. It is known that a diet high in saturated fats is correlated with an increased incidence of, and mortality from, coronary heart disease. However, people in France are generally found to have a relatively low incidence of coronary heart disease but a high intake of saturated fats.3 Wine and general alcohol consumption is very high in France and has therefore been recognized as an important factor in explaining this phenomenon.12

Studies of the potential medicinal benefits of resveratrol are numerous.12 In vitro, resveratrol has exhibited anti-cancer, anti-oxidant, and anti-inflammatory activities, as well as improving cardiovascular health and inhibiting platelet aggregation.48-52 Experimental models have demonstrated the ability of resveratrol to aid in the prevention or slowing of a range of cardiovascular diseases such as hypertension, hypertrophy, and myocardial infarction.12,53

Extensive evidence both in animal models, and in vitro have also shown that resveratrol can delay the onset of age-associated diseases such as cancer, neurodegenerative disorders, or Alzheimer’s disease.12,54-56

These purported health properties are not limited to just resveratrol, piceatannol shares many of the same anti-tumor, anti-oxidant, and anti-inflammatory activities as resveratrol.57

Pterostilbene, a dimethylated derivative of resveratrol, has also been shown to possess enhanced antifungal activity and induces apoptosis in human melanoma cells.58 It has been suggested the structures of these compounds render them as potent anti-oxidants, which provide the basis of their health-promoting properties.

11

1.2 Instrumentation Background

Grapes and wines contain many different stilbenoid compounds at varying concentrations.

As such, one challenge in identification of these compounds lies in their analysis. One of the predominant techniques used for this purpose is high-performance liquid chromatography mass spectrometry (HPLC-MS). This technique is sensitive enough to enable identification and quantification in wine and plant extracts. When coupled to specialized detectors, it can also be used as gain structural information about separated compounds. The following sections discuss the HPLC and MS instruments in general, and also the equipment specifically used in this thesis.

1.2.1 Ultra High Performance Liquid Chromatography

Liquid chromatography is a separation method whereby solutes in a mixture are dispersed between a liquid mobile phase and a solid stationary phase. Stationary phases are often small particles which are packed tightly into steel columns. Pumping the liquid mobile phase through these tightly packed columns requires a relatively high pressure.59,60 The vast majority of HPLC systems in use for polyphenol applications use reversed-phase chromatography. This uses a non- polar stationary phase and a moderately polar mobile phase. This is in contrast to normal-phase

HPLC where the stationary phase is polar and the mobile phase is a non-polar, non-aqueous solvent. 60

The high pressures required are achieved using binary or quaternary piston pumps which precisely control both the flow and solvent mixture. Mobile phase gradients are achieved by using a single pump for each mobile phase component, mixing the solvents post-pump. Traditional

12

HPLC pumps have had an upper pressure limit of near 400 bar. To obtain a greater degree of separation the UHPLC (ultra-high performance liquid chromatography) systems, able to reach

1200 bar, were developed.59 These newer systems not only require specialized pumps, but most tubing, fittings, and columns must be specifically designed to withstand the greater pressure. As such, the greater resolution power of UHPLC that can be achieved over HPLC comes with the increased cost of these specialized components.

A temperature controlled column compartment in an HPLC system allows for stable and repeatable retention times. Increasing the temperature of the column can improve separation in two ways: it lowers the solvent viscosity, which in turn decreases the backpressure, allowing for higher flow rates, as well as improves the mass transfer between mobile and stationary phases.61

In addition to increased temperature, some systems can reliably keep the column compartment below room temperature for sensitive analysis.

HPLC columns are predominantly manufactured from stainless steel and can range in length from 50 – 250 mm. The internal diameter (ID) of the column can also range in size from

0.3 – 5 mm, and it has been shown that the smaller ID columns generally use much less solvent and can provide lower limits of detection.60 There is a great variety in the number of materials that can be used the stationary phase, but the most common is bonded silica. Many different substituents can be bonded to the silica through an ester or silicone polymeric linkage. Common bonded phases for RP-UHPLC are C4, C8, or C18 functional groups, which offer varying degrees of retention.60 The particles themselves are usually <5 µm, but can range in diameter between 1.0

– 200 µm. There is a trade-off with particles sizes, smaller particles offer higher separation efficiency, but generate a higher resistance to solvent flow. When using particles that are 13

especially small (<2 µm), the aforementioned UHPLC systems are a requirement due to the amount of back-pressure the particles create.59

A wide variety of detectors are used in conjunction with HPLC systems, these can be general or more specialized to a certain task. Most of the common detectors function via a post- column flow cell through which the dissolved analyte travels to create an electrical signal.60

Detectors can be used individually, but are also often linked in series to extract the most information possible from each analyte. Detectors are chosen based on the required specificity, detection limits, and properties of the analytes to be detected. Common HPLC detectors include, but are not limited to: fixed and variable wavelength spectrophotometers (UV-vis), diode array detectors (DAD), fluorometers (FL), and mass spectrometers (MS).62

1.2.1 General Mass Spectrometry

Mass spectrometry is one of the most versatile and prevalent analytical techniques used in modern science. The modern mass spectrometer generally consists of an ion source for creating ions, a mass analyzer for separating the ions based on mass/charge, and a detector which essentially detects and sums the number of ions of each mass/charge. These components exist in a vast number of combinations and/or configurations which can be highly specialized for different purposes (Figure 1.5).

14

Figure 1.5: Box diagram of the major components of a mass spectrometer, and how they fit together.

Analytes are ionized in the source, whether fed by HPLC or directly injected into the mass spectrometer. Depending on the ionization technique applied, the types of ion e.g. positive or negative, atomic, molecular or adduct produced will vary greatly. While many different types of ion sources and techniques for mass spectrometers exist, the two that are ordinarily coupled to

HPLC systems are atmospheric pressure chemical ionization (APCI) or electrospray ionization

(ESI). These are the most common techniques because the HPLC flow can be directly connected to the spray needle where the ionization occurs at atmospheric pressure.

Once ionized, the gas phase ions enter the mass spectrometer through a series of openings which connect the atmospheric pressure source to the high vacuum of the mass spectrometer. A number of ion lenses and guides focus the ion beam towards the mass analyzer. There are essentially three types of mass analyzers commonly used today. The first is a scanning type analyzer, such as a magnetic sector, quadrupole, or ion trap, which can modify either the magnetic or electric fields to allow only ions of a certain mass/charge to continue onto the detector. The second type is known as a time-of-flight (TOF) mass analyzer, where ions are “bunched” and then pushed through a flight tube at a constant kinetic energy. This causes the small molecules to reach

15

the detector before the larger molecules, and thus separates analytes according to mass. The final kind of mass analyzer differs from the first two as it functions as both the mass analyzer and the detector. The Fourier transform ion cyclotron resonance and the orbitrap are examples of such instruments. They measure the image current produced by ions that are cyclotroning inside of a magnetic field.

In addition to single stage mass spectrometry, more specialized mass spectrometers can have multiple mass filtering components combined in series. These instruments are collectively known as “tandem mass spectrometers”, or MSn for short. Two distinct categories of tandem spectrometers exist, tandem-in-space, and tandem-in-time. Tandem-in-space indicates the mass analyzers are physically positioned one after another, with a collision cell in between. This is different from tandem-in-time, where ions are held in a single mass analyzer by a magnetic or electric field, and mass filtering is carried out sequentially over time. The advantage of tandem mass spectrometry is that in between each mass filtering event, it is possible to fragment the parent ions that were allowed to pass the first mass analyzer, after which, the second mass analyzer can determine the product ion fragments. Essentially this enables the production of fragments that are unique to a single m/z parent ion. There are many different fragmentation methods in tandem mass spectrometry, including: collision induced dissociation (CID), electron- capture dissociation (ECD), and surface-induced dissocation (SID).

The final portion of a mass spectrometer, the detector, measures either the charge or induced current of an ion as it impacts the detector’s surface. In a scanning instrument, all ions detected during a scan of a particular mass can be attributed to that mass/charge. While in a TOF instrument, the time it takes for the ion to hit the detector, is what can be used to calculate the mass/charge. The most common detectors in modern mass spectrometers are the ion-to-photon 16

detector, faraday-cup, or for more sensitive instruments, an electron multiplier such as a micro- channel plate detector (MCP).63

1.2.1.1 Electrospray Ionization

Electrospray ionization is the preferred method of ionization for a mass spectrometer when coupled inline to an HPLC. Depending on the manufacturer and model of the instrument, there can be important differences in the operation or mechanisms of each source.64 While the basic principles behind electrospray ion formation are conserved across most hardware types, the

Agilent 6530 ESI-AJS (the primary instrument for this research) source will be the model described to avoid any ambiguities. Positive mode electrospray ionization is the most commonly used polarity for most research purposes. However, as polyphenols are more easily and selectively ionized in negative mode, the following discussion of the ESI mechanism will be focused on negative ion formation.65

Electrospray ionization begins with the UHPLC flow entering the nebulizer housing and exiting the electrospray needle inside the source (Figure 1.6). This infusion into the electrospray needle can occur at a flow rate anywhere from 100 µL/min to over 1 mL/min depending on the

HPLC or UHPLC flow rate. In the Agilent AJS-ESI system, the nebulizer is held at ground, while the curtain plate has a high positive voltage potential applied to it. This is referred to as the

“grounded emitter” system, where reductions of water inside the nebulizer needle form negative ions in solution.3,64,66 In this manner, the ESI source is functioning as an electrochemical cell.67 At the tip of the needle, the solution forms a Taylor cone, which emits a fine mist of droplets.68This process is aided by a nebulizing nitrogen gas flow that is concentrically streamed around the needle’s capillary.69

17

Figure 1.6: Diagram showing the Agilent AJS Electrospray source, highlighting the various gas flows, and applied voltages that can be altered to optimize ionization.

Assisted by heating gas, the droplets formed by the Taylor cone are evaporated as they travel towards the entrance of the mass spectrometer. The organic solvent is usually the first to evaporate, increasing the overall percentage of water in each droplet.70 The density of charge in the droplet increasing until the surface tension is balanced by coulombic repulsion. This is known as the Rayleigh limit, where coulomb fission of these droplets produces progeny droplets that are smaller, and highly charged.71,72 From this point, the mechanism of gas phase ion formation is somewhat uncertain; however, low molecular weight ions are thought to enter the gas phase as these progeny droplets again reach their Rayleigh limit, and the ion is ejected; initially connected by a string of solvent molecules.73,74 The product of this ejection is thought to be a gas phase ion

18

accompanied by a few solvent molecules, which are removed during the transition into the mass spectrometer by the flow of nitrogen curtain gas.75

1.2.1.2 Quadrupole-Time-of-Flight (Q-TOF) Mass Analyzer

The quadrupole-time-of-flight mass analyzer (Q-TOF) is a hybrid instrument that consists of a quadrupole mass filter, followed by a second quadrupole collision cell, which is used in conjunction with a time-of-flight mass analyzer (Figure 1.7). This hybrid setup is used because it can offer certain advantages that each component on its own cannot achieve.

A quadrupole is comprised of four parallel rods arranged such that each opposing rod is coupled together electrically. Mass filtering occurs as a result of a radio-frequency (RF) voltage potential being applied to all four rods, while a DC voltage is applied to a pair of opposing rods.

These radio and electric fields can be adjusted to only allow an ion of a certain m/z ratio through the length of the quadrupole. Ions that do not have the selected m/z will collide with the poles and be lost. Using this method, each m/z over a desired range can be “scanned” in order to determine a mass spectrum at any given time. It is important to note, however, that although the scan time is fast, many desired ions are lost during the sweep, decreasing the sensitivity.76

The time-of-flight mass (TOF) analyzer, as described above, is so called because rather than using an electric field to isolate masses in a scanning fashion, the TOF measures only the amount of time an ion takes to traverse a specific distance. A spectrum obtained from a single pulse of the TOF analyzer is known as a transient; and many modern instruments can perform several thousand transients per second to produce a single mass spectrum.66

The Agilent 6530 accurate mass Q-TOF begins with an octopole ion guide that traverses the first and second vacuum stages. This octopole is operated in wide band-pass mode and

19

functions only to collimate the ion beam. Following the first octopole is the quadrupole mass analyzer which when the mass spectrometer is operated in single stage mode (MS1), also acts as an ion guide. When functioning in tandem MS mode (MS2), the quadrupole is actively scanning the desired mass range and only allowing chosen m/z ions through. In the next stage of the mass spectrometer, the ions that have been selected to pass through the quadrupole enter a hexapole that has nitrogen gas fed into it. This high pressure region allows analyte ions to collide with neutral nitrogen gas molecules, increasing the internal energy of the ion, and causing fragmentation of the ion in a process known as collision induced dissociation (CID).77 The hexapole has a voltage potential applied across the length of it. This is known as the collision energy (CE) and can be adjusted depending on the size of the molecule and the desired degree of fragmentation. Once the ions have been fragmented, they pass through an area containing another octopole and multiple lenses that facilitate shaping of the ion beam as it enters the slicer and TOF pusher assembly.66

20

Figure 1.7: Agilent 6530 QTOF diagram showing the inside of the mass spectrometer including the multiple vacuum stages, ion guides, mass analyzers, and reflectron.

The QTOF setup has multiple advantages over a single stage mass spectrometer, starting with the detection of both the precursor mass and its fragment ions. A triple quadrupole can also perform this action, however, by placing a TOF analyzer in the place of the third quadrupole (Q3), mass accuracy of <10 ppm can be obtained with careful calibration.78 This allows either the full or partial structure elucidation of target analytes. Another advantage of using the TOF analyzer as the second MS analyzer is that it does not need to scan through m/z ratios, instead it can collect all fragment ions, increasing sensitivity over product ion scan with a triple quadrupole (QqQ) mass spectrometer.78

21

1.3 Identification and Quantification of Stilbenoids in Wine Using Mass Spectrometry

The identification and quantification of stilbenes in wine has been accomplished using a variety of analytical methods, including HPLC coupled to spectrophotometric detectors (UV-vis,

DAD, FL), mass spectrometric means (MS or MS/MS) or even using gas chromatography-mass spectrometry (GC-MS). GC-MS analysis of trans/cis-resveratrol has been performed with and without analyte derivatization with acceptable results.79,80 Another method of detection used for stilbenes specifically is fluorometric detection. This method is much more sensitive and selective than UV absorbance, however, it requires that the analyte fluoresces under UV light. These spectrophotometric methods are by far the most prevalent detectors used for HPLC detection of stilbenoid compounds. As ambient ionization techniques such as ESI, APCI, and APPI have continued to improve, the number of systems utilizing MS or MS/MS detection of stilbenoids has increased greatly.81

When working with a complex matrix such as wine, UV-vis and DAD chromatograms are often very crowded, causing difficulties in detection and quantification due to matrix and analyte interferences. Despite this, many studies have been carried out under these conditions.82-86

Fluorescence detection can also be used for stilbenoid quantification. It has been shown to be both more selective and an order of magnitude more sensitive than DAD methods.81,87,88

Better results have been obtained when wine is first extracted and concentrated by solid phase extraction (SPE) or liquid/liquid (L/L) extraction.89 SPE offers some advantages over L/L extraction, such as quantitative recoveries, minimal solvent usage, and it is much less labor intensive, greatly reducing the overall analysis time. 81,88

22

Mass spectrometry is becoming the preferred method of HPLC detection and quantification for stilbenoids in wine. This is especially true when considering that extremely sensitive hybrid instruments such as QqQ and QTOF can gain multiple pieces of information about an analyte from a single chromatogram alone. In terms of quantification of stilbenoids, MS detection has been shown to be more selective, sensitive, and accurate than its spectrophotometric counterparts.84,85,90,91 The one shortcoming of MS detection, when compared with DAD, is that intra- and inter-day repeatability is less precise.89 This is due to the day to day variability associated with atmospheric ionization.

When performing MS or MS/MS quantification on samples with an abundance of extraneous compounds, it is very likely that these compounds can negatively affect the ionization of the analyte(s). Even if these superfluous molecules are not detected by MS, they may co-elute and share retention times with the target compounds. These molecules can lead to significant ion suppression or enhancement, both of which severely affect the method’s accuracy.92 Attempting to control for these matrix effects is possible. Sample pre- extraction, or dilution results in less matrix loaded onto the column and longer gradients can ensure that interfering compounds are separated from the analytes.93,94 However, both of these methods come at the expense of increasing analysis time. Another commonly employed method to correct for the matrix effect is the use of an internal standard (IS) that is similar in mass, structure, and retention-time to the analyte. Instead of calibrating based on extracted ion chromatogram (EIC) peak area vs. concentration, a ratio of EIC peak areas (analyte/IS) vs. concentration is used. The best type of IS that can be used in MS or MS/MS analysis is a stable isotopically labelled (SIL) analogue of the analyte; which has the same retention time, properties, and ionization mechanism, but does not share the same mass, and can therefore be differentiated.93 Using an SIL is very convenient when

23

a commercial source of the labelled analogue is available. Synthesis of such labelled compounds is also possible but can be time-consuming, expensive and challenging, and therefore choosing an internal standard of similar structure to the analyte is often a better compromise than an SIL.

1.4 Research Objectives

1.3.1 To develop and optimize a fast, reliable and accurate method for quantification of

monomeric stilbenoids in red and white wine using UHPLC-MS.

1.3.2 To qualitatively screen and partially identify all the known and unknown stilbenoid

compounds detectable in red wine using UHPLC-ESI-QTOF.

1.3.3 To estimate the concentrations of the stilbenoid compounds in wine and determine

the influence of variety and type.

1.5 Hypotheses & Rationale

Hypothesis 1: Different varieties of red wine produced in the Okanagan valley possess significantly different quantities of the major stilbene monomers trans- and cis-resveratrol, trans- and cis-piceid, and trans- and cis-piceatannol.

Rationale 1: Previous literature has suggested that there is a great variance in levels of resveratrol, its isomers, and the mono-glycosides of each.22 It is of interest to determine whether red wine produced in the Okanagan Valley shows the same trend as in earlier studies.

Hypothesis 2: Monomeric glycosidic forms of resveratrol and piceatannol, other than piceid and astringin exist in red wine, including 4’-O-glycosides, and di- and tri-glycosides of various conformations.

24

Rationale 2: Stilbene mono-glycosides are the most abundant stilbenes in red wine.22 Studies using cell suspension cultures of Cabernet Sauvignon and other plants have elicited the production of diglycosides and triglycosides of resveratrol.95-97 If glucosyl transferases are capable of glycosylating the 5-O and 4’-O positions of resveratrol in cell suspension cultures, then it is quite likely that these compounds are also produced in vivo and thus can be extracted into wines.

Hypothesis 3: There exist multiple resveratrol derivatives, including dimers, trimers, and tetramers that have not yet been observed in red wine.

Rationale 3: While only few stilbene derivatives have been detected in wine, there have been hundreds of increasingly complex stilbene oligomers identified across many plant species.13,40

Perhaps these compounds are below the detection limit of previous methods of identification.

Resveratrol alone cannot account for the French paradox and the purported health benefits associated with red wine, perhaps it is the result of synergy between many stilbenoid oligomers rather than just a few.

Hypothesis 4: Non Vitis vinifera (hybrids) Quebec wines have higher stilbenoid contents compared to Okanagan Vitis vinifera based wines.

Rationale 4: varieties were bred to be very hardy in cold weather and more resistant to fungus, pests and disease.98-100 The increased resistance to fungus and bacteria could in part be explained by higher stilbenoid production. It has already been shown that tannin, anthocyanin and stilbene composition is significantly different between hybrids and the more common Vitis vinifera derived wines.101,102

25

1.6 Significance of study

Winegrowing regions around the world have been investigating the beneficial properties of wine for many decades. Research highlighting the differences in resveratrol and other polyphenols between the diverse growing areas is abundant.22,80,103 This project will be the first to determine the stilbenoid content of wine grown and made in the Okanagan valley.

In addition, this study also contributes to the knowledge of the collection of stilbenoid compounds that can be found in wine. Studies have observed that the concentration of previously known stilbenes present in wine does not fully account for the observed health benefits.12

Although the stilbenoid derivatives highlighted in this study are present in low concentrations, it may be that the cumulative concentration, or synergistic effects, of all of these antioxidant compounds can justify the health promoting claims. The semi-quantification data produced in this project can lead to research targeted at those stilbenoid compounds which are the most abundant in Okanagan red wines.

Scant data is available regarding the differences between the common Vitis vinifera species of grapes, and the hardier hybrid varieties of wine grapes that are grown in North America and

Europe. This study provides stilbenoid quantification data exploring the differences between these two groups of wine varieties.

26

CHAPTER 2 – DEVELOPMENT OF A VALIDATED METHOD FOR

QUANTIFICATION OF MONOMERIC STILBENES IN WINE

2.1 Synopsis

The aim of our study was to develop and validate a quick, robust, and selective method for quantification of the major stilbenoids in red and white wine using ultra-high-performance liquid chromatography coupled with electrospray ionization/quadrupole time-of-flight mass spectrometry (UHPLC-ESI-Q-TOF). Both isomers of resveratrol and piceid were quantified externally using authentic standards, while piceatannol was quantified in trans-piceid equivalents.

To improve the precision of the MS quantification, epigallocatechin-3-O-gallate was used as an internal standard. Sample preparation was limited to filtration and addition of the internal standard. Due to the minimal amount of sample preparation and the short method runtime, results were obtained rapidly and with low expenditure of energy, chemicals, and labor.

The first step of this work was the optimization of UHPLC-ESI-Q-TOF parameters. The elution gradient and column temperature was optimized to achieve a good separation of the analytes. Next, the mass spectrometry parameters were optimized to improve analyte sensitivity.

Finally, the method was validated with respect to linearity, limit of detection (LOD), limit of quantification (LOQ), accuracy, intra- and inter-day precision and stability. All six stilbene monomers were quantified in 44 British Columbian wines; the highest total stilbene concentration in an individual wine was 28.81 mg/L in Pinot Noir (average 13.81 ± 6.54 mg/L) which was found to be significantly higher than the total stilbene content of Merlot (average 6.96 ± 1.86 mg/L), while the average across all wines was 8.49 ± 6.25 mg/L.

27

2.2 Experimental

2.2.1 Test Samples

Red, white and Rosé wine samples (VQA, Okanagan) were both purchased from local stores as well as donated by Okanagan wineries. Multiple aliquots were taken from each bottle after opening, filtered through a 0.4 μm Millipore PTFE syringe filter into separate 1.7 mL microcentrifuge tubes, and then immediately frozen at -80 oC until further analysis.

2.2.2 Reagents and Materials

HPLC grade methanol and acetonitrile were purchased from Fischer Scientific (Waltham,

MA, USA). Deionized water was purified by a MilliQ water system (Millipore, Bedford, MA).

Formic acid (>99% HPLC grade) was purchased from Sigma Aldrich (St. Louis, MO, USA). ESI-

L undiluted tuning mix was purchased from Agilent Technologies (Santa Clara, CA). Reference standards for trans-resveratrol and trans-piceid were obtained as powder from Chromadex

(>99%, Irvine, CA, USA). The internal standard, epigallocatechin-3-O-gallate (>95%, EGC3OG), was purchased from Sigma Aldrich (St. Louis, MO, USA). A model wine solution was also prepared as follows: 88:12 water:ethanol, 5 g/L tartaric acid, and adjusted to pH 3.60 with 1.0 M

NaOH. This solution is commonly used to simulate wine in the laboratory setting.104

2.2.3 Calibration Standards

Stock solutions of resveratrol and piceid were prepared in triplicate by weighing out 10 mg of trans-resveratrol (transA1-3) and trans-piceid (transB1-3) into 100 mL volumetric flasks and diluted to the mark with model wine solution (12 % ethanol, 5 g/L tartaric acid, adjusted to pH 3.60 with 1.0 M NaOH) to afford 6 stock solutions of 100 mg/L. In addition to trans- resveratrol/piceid, cis-resveratrol/piceid stock solutions were obtained by exposing 20 mL of

28

solutions transA(1-3) and transB(1-3) to sunlight for 12 hours to produce cisA(1-3) and cisB(1-

3). The stock solution for the internal standard was prepared by weighing 5 mg of ECG3OG into a

50 mL volumetric flask and diluted to the mark with model wine.

Table 2.1: Dilution series used to create calibration curves for trans-/cis-resveratrol and trans-/cis-piceid.

Concentration of Concentration after Solution # analyte prepared internal standard (mg/L) addition (mg/L) SX 50 - S1 40 20 S2 30 15 S3 25 12.5 S4 20 10 S5 15 7.5 S6 10 5 S7 5 2.5 S8 2 1 S9 1 0.5 S10 0.2 0.1 S11 0.1 0.05 S12 0.05 0.025 S13 0.02 0.01 S14 0.01 0.005 S15 0.005 0.025

A mixed standard solution, “SX” was created by combining 5 mL of A and 5 mL of B to make 6, 50 mg/L mixed standards: trans(1-3) and cis(1-3). From each stock solution, a serial dilution was performed using concentrations as indicated in table 2.1. Solutions S1-S7 were prepared by pipetting 0.8, 0.6, 0.5, 0.4, 0.3, 0.2 and 0.1 mL respectively of SX and diluting to 1 mL using model wine solution. In order to reach the lower concentration ranges, serial dilution of

S6, S7 and S9 were used to make S8, S9 and S10, S11, S12 respectively. Finally, 0.2, 0.1 and 0.05 mL of S11 was used to create S13, S14, and S15 respectively (Table 2.2).

29

Table 2.2: Outline of the serial dilution scheme used, defining the procedural details of the dilution.

Concentration of [Concentrated Concentrated Solvent Solution # analyte prepared Solution] (mg/L); Solution Volume Volume (mg/L) From Solution # (mL) (mL) SX 50 100 5.000 - S1 40 50; SX 0.800 0.200 S2 30 50; SX 0.600 0.400 S3 25 50; SX 0.500 0.500 S4 20 50; SX 0.400 0.600 S5 15 50; SX 0.300 0.700 S6 10 50; SX 0.200 0.800 S7 5.0 50; SX 0.100 0.900 S8 2.0 10; S6 0.200 0.800 S9 1.0 5; S7 0.200 0.800 S10 0.2 1; S9 0.200 0.800 S11 0.1 1; S9 0.100 0.900 S12 0.05 1; S9 0.050 0.950 S13 0.02 0.1; S11 0.200 0.800 S14 0.01 0.1; S11 0.100 0.900 S15 0.005 0.1; S11 0.050 0.950

2.2.4 UHPLC-ESI-Q-TOF Analysis

Analyses of prepared standards and wine samples were performed on an Agilent 1290 infinity binary UHPLC system coupled to an Agilent accurate-mass 6530 UHPLC-ESI-Q-TOF

(Agilent Technologies, Santa Clara, CA). The column chosen for these analyses was a Zorbax

RRHD SB C18 (2.1 x 50 mm, 1.8 µm) (Agilent Technologies, Santa Clara, CA). Prior to injection, samples were held in a temperature controlled auto sampler at 4°C, and the column compartment was thermostatted to a constant 40°C during analyses.

The HPLC parameters were as follows. The mobile phases consisted of a binary mixture of water (A) and acetonitrile (B), both acidified with 0.1 % formic acid. A 6.75 minute gradient elution method with a 2 minute post-run for re-equilibration of the column was developed for optimal separation as is shown in table 2.3. The flow rate was fixed at 1.00 mL/min, and the injection volume was 2.5 µL.

30

Table 2.3:UHPLC 6.75 minute gradient elution program; Solvent A = 0.1% formic acid in H2O, Solvent B = 0.1% formic acid in acetonitrile.

Time (min) Mobile Phase A (%) Mobile Phase B (%) 0.00 100 0 0.25 95 5 4.00 68 32 4.50 0 100 6.00 0 100 6.50 100 0 6.75 100 0

The source parameters were optimized for each analyte as follows: drying gas and sheath gas were both N2 gas supplied by a membrane nitrogen generator fed with compressed air; drying gas flow rate 10 L/min; sheath gas flow rate 12 L/min; drying gas temperature 325 °C, sheath gas temperature 400 °C; nozzle voltage 0 V; capillary voltage (Vcap) 3000 V; skimmer voltage 75 V; finally, the fragmentor voltage was set to 175 V.

Q-TOF mass axis calibration was performed in the 2 GHz extended dynamic range instrument state. Analyses were all performed in negative ion mode. The calibration tuning mix was prepared by diluting 2.5 mL of ESI-L, with 1.9 mL of H2O and then diluting to the 100 mL mark with acetonitrile. Following the calibration of the mass axis using the auto calibration function, the mass errors from all ions were less than 0.08 ppm.

The QTOF detection parameters were optimized to achieve the greatest repeatability and stability between analyses. The mass range was fixed between 100-500 m/z, ion threshold was set to 30 counts or 0.005% of maximum, the acquisition rate was optimized to be 1.1 spectra/s which resulted in an acquisition time of 909.1 ms/spectrum and 9089 transients/ spectrum.

31

2.2.5 Validation

2.2.5.1 Linearity, Limits of Detection and Quantification

The calibration solutions ranged from 0.0025 to 20 mg/L with an added internal standard concentration of 0.5 mg/L. The plot of analyte(EIC)/Internal-Standard(EIC) versus concentration was constructed for each target stilbene compound. The linearity for each compound’s plot was evaluated based on the visual inspection of the plot, the residual plot, r2, and where required, a lack-of-fit test was performed.

The limits of detection (LOD) and quantification (LOQ) were determined by analyzing a series of 20 blank samples where no analyte was present. The instrument response at the retention time of each analyte was determined, and the following equations were used:

푦퐿표퐷 = 푦푏푙푎푛푘 + 3 ∗ 푆퐷푏푙푎푛푘

푦퐿표푄 = 푦푏푙푎푛푘 + 10 ∗ 푆퐷푏푙푎푛푘

In terms of concentration however, the blank possesses 0 mg/L of analyte and therefore the equations become:

3 ∗ 푆퐷 퐶 = (0) + 푏푙푎푛푘 퐿표퐷 푚

10 ∗ 푆퐷 퐶 = (0) + 푏푙푎푛푘 퐿표푄 푚

Where m represents the slope of the respective analyte’s calibration curve.

32

2.2.5.2 Intra-, Inter-day Repeatability, and Stability

The stability of wine samples whilst undergoing multiple freeze/thaw cycles was determined. Once removed from their respective bottles, internal standard was added to 9 red and

9 white wine samples (1 mL) to give a concentration of 0.5 mg/L and then immediately frozen at -

80°C. Prior to analysis, each wine sample was removed from the freezer and allowed to thaw to

4°C in the autosampler compartment for 1 h. After each injection, the samples were removed from the autosampler and refrozen at -80°C. This cycle was repeated 3 times.

The intra-day precision was determined with by preparing nine 1 mL red and white wine samples, with 0.5 mg/L of internal standard added. These 18 samples were analyzed 3 times in one day, with a 3 h period in between each set of analyses. During the analyses and wait times, the samples were kept in a thermostated autosampler held at a constant 4°C. The inter-day precision was measured by performing the same procedure as for the intraday test, with a replicate of the 18 wine samples, 14 days later. During this 14 day wait, the samples were frozen at -80°C.

2.2.5.3 Recovery

To test for the validity of the method when applied to the matrix of wine samples, the recovery was evaluated. Red and white wine samples were spiked with increasing concentrations at low, mid, and high ends of the method’s range. A 25 mg/L “mixed standard” was prepared from 100 mg/L stock solutions of each of the four stilbene analytes. To make the low concentration, 10 µL of the “mixed stock solution” and 5 µL of the 100 mg/L internal standard were added to 985 µL of wine to yield a final concentration of 0.25 mg/L. The midrange concentration of 1 mg/L was made by adding 40 µL of stock, 5 µL of internal standard and 955

µL of wine. Finally, the 5 mg/L spike was prepared by adding 200 µL of stock and 5 µL of internal standard to 795 µL of wine. In addition to these spike concentrations, a zero concentration 33

spike was made by adding just 5 µL of internal standard to 995 µL of wine. These samples were prepared in triplicate, and each was analysed three times.

34

2.3 Results & Discussion

2.3.1 Optimization of UHPLC-ESI-Q-TOF and Analysis

2.3.1.1 Optimization of UHPLC Parameters

The choice of mobile phase was predetermined to be water (solvent A) and acetonitrile

(solvent B) as these solvents are the most predominantly used mobile phases in RP-HPLC.

Although negative mode was used as the ionization polarity, and the target analytes themselves were weak acids; formic acid was chosen as the mobile phase modifier as it has been shown to increase ionization in negative mode.105

Optimization of the gradient was achieved by selecting a red wine that contained all analytes in relatively high quantities in order to optimize the separation in a wine sample. This ensured that the analytes with the same mass were well separated from each other, as well as from interfering compounds within the matrix that have similar masses. An initial linear increase from

0 % to 100 % B over 10 minutes was used as a “scanning” gradient to determine both the order and approximate organic phase content at which each analyte eluted. The order of elution was trans-piceid (1), cis-piceid (2), trans-resveratrol (3), cis-resveratrol (4). Increasing the percentage of solvent B caused the compounds to elute earlier, the general effect of changing the slope of the gradient is presented in Figure 2.1. By decreasing the final %B composition of the mobile phase, the compound will elute later, but will also increase peak resolution by becoming more separated from each other.

35

Figure 2.1: Chromatogram displaying the effect on analyte separation of increasing concentration of acetonitrile mobile phase over the 4 minute gradient. 1- trans-piceid; 2- cis piceid; 3-trans-resveratrol; 4-cis-resveratrol.

In a similar fashion, the temperature of the column compartment was optimized by performing multiple analyses, varying the temperature in each, from 15 - 50°C. As can be seen in

Figure 2.2, the general trend of increasing column temperature is decreasing retention time of the compounds, while simultaneously decreasing peak resolution. Ideally, the lowest retention time is desired, but not at the expense of causing the peaks to overlap, as is the case beginning at 35-

40°C. Therefore, the optimal column temperature was selected to be 30°C.

Figure 2.2: Chromatogram displaying the effect of increasing column temperature on the separation, and retention time of the analytes. Darker colors indicate higher column temperature.

36

2.3.1.2 Optimization of ESI-Q-TOF Parameters

Once the UHPLC conditions were optimized and fixed at their final values, the ESI-Q-

TOF parameters were optimized to increase the sensitivity and selectivity for the analytes. The complete optimization was split into three distinct sets. The first is the flow dependant parameters, sheath gas temperature/flow and the nebulizer pressure. These settings are contingent to an extent on the flow rate of the UHPLC. The second set of parameters is the compound specific settings, which can be highly reliant on the molecular weight, structure, and polarity of the target molecule(s).66 The final parameter to be optimized was the scan time of the TOF detector.

The first choice that must be made when analyzing compounds with a mass spectrometer is the ionization polarity. Traditionally, most MS analyses are done in positive mode, however, if the compound, or in this case, family of compounds (polyphenols), are more likely to exist as stable negative ions, then it can be much preferred over the former because of increased sensitivity, and less susceptibility to formation of adduct ions. Previous research on resveratrol has indicated that negative mode offers a 6-fold increase in sensitivity when compared to positive mode (REFID196).106 Because of this, all analyses were performed using negative ion mode.

The parameter that has the greatest effect on ionization within the Agilent Jetstream ESI

(AJS) source is the fragmentor voltage. This voltage potential is applied to the exit of the capillary at the interface of the ion source and the mass spectrometer. The optimization of the fragmentor voltage was performed by analyzing a red wine sample varying at 10 V increments with a range of 100-260 V in triplicate and then averaged. The EIC peak area was then integrated for each compound as shown in Figure 2.3.

37

Figure 2.3: Effect of increasing fragmentor voltage on the ionization of: A- trans-resveratrol; B-cis-resveratrol; C-trans- piceid; D-cis-piceid.

The optimal fragmentor voltage was determined to be 160 V for trans- and cis-resveratrol, and 190 V for trans- and cis-piceid. In each voltage curve it can be seen that the sensitivity is greatly increased by as much as 250 % from the default 100 V setting, which highlights the necessity in determining the optimal fragmentation voltage for each compound. A compromise was made by fixing the fragmentation voltage at 175 V, which is a value that is near the maximum on each curve. The alternative is to create a situation where the instrument will switch from a voltage A to voltage B over the course of the analysis. However, the optimal voltage for resveratrol and piceid are similar enough that the sensitivity was not greatly affected.

The AJS parameters can have a substantial effect on the ionization of compounds; they were the next to be optimized. These two parameters are highly dependent on the UHPLC flow

38

rate, which in the present method is relatively high at 1.0 mL/min. Sheath gas temperature and flow cannot be optimized independently as a high sheath gas temperature and a low sheath gas flow can cause overheating and subsequent instrument fault. The optimization was therefore carried out by varying the sheath gas flow from 5-12 L/min in triplicate. The sheath gas temperature was fixed at each flow rate as follows: 250°C at 5 L/min; 300°C at 6 L/min; 350°C at

7 L/min; and 400°C from 8-12 L/min. The ionization abundance of resveratrol is negatively affected by an increasing sheath gas temperature and flow, while piceid is positively influenced by this same increase. The effect in regards to resveratrol is less pronounced than on piceid, because of this and the high flow rate of the UHPLC, 400°C and 12 L/min were chosen as the optimal AJS parameters.

The nebulizer pressure is the pressure created by flowing nitrogen and the UHPLC flow through the nebulizer needle. The nebulizer pressure was varied between 25 – 55 psig and the effect was a very small increase in sensitivity for all compounds. Compared to the default setting of 45 psig, 55 psig produced an increase in ionization of trans- and cis-resveratrol by 6% and 5% respectively, and increased in trans- and cis-piceid by 0.6% in both.

The voltage that is applied to the front end of the capillary, where ions enter the mass spectrometer is known as the capillary voltage. This voltage was varied from 2000 – 4250 V in

250 V increments. The effect of changing capillary voltage is somewhat molecular weight dependant, and is certainly differentiable between compounds. Figure 2.4 demonstrates the opposite trends on ionization between trans-/cis-resveratrol and trans-/cis-piceid. The optimal capillary voltages are on completely opposite ends of each curve for resveratrol and piceid. The optimal capillary voltage was then compromised by choosing the intersection of the 4 curves, which was at 3000 V. When compared to the default of 3500 V, this led to an increase in 39

ionization of 9.3 % and 7.6 % for trans and cis-resveratrol, and a decrease in ionization of 7.5 % and 14.3 % for trans and cis-piceid respectively.

Figure 2.4: The effect of increasing the capillary voltage on the ionization of each analyte.

The skimmer is a cone shaped orifice situated after the capillary exit which allows only the center of the ion beam to pass. The voltage applied to this orifice ensures that solvent, gas, and neutral molecules are unable to continue into the mass spectrometer.107 This voltage was optimized by a step-wise increase from 50 – 90 V in 10 V increments. The optimal skimmer voltage was determined to be 75 V for all compounds.

The final source parameter to be optimized was the nozzle voltage. The nozzle is situated below the electrospray needle and this potential is used to direct ions from the spray into the front of the capillary. The effect of increasing the nozzle voltage was negligible; decreasing it from 500

40

V to 0 V gave an increase in ion abundance of 0.76 % and 0.94 % for trans-/cis-resveratrol and

1.63 % for both trans- and cis-piceid. The optimal nozzle voltage was therefore fixed at 0 V.

The final parameter that was explored was the scan time of the TOF detector. The goal for this optimization was to achieve the highest sensitivity without affecting the precision too drastically. Intuitively, fewer scans per second increases the intensity of the peak (and the sensitivity), but consequently will reduce the number of scans over the timeframe of the peak.

This effect is demonstrated in Figure 2.5, as the number of scans per second is increased, the intensity is reduced while the shape of the peak becomes “smoother” as there are effectively more points across the peak width. The optimal scan time was determined to be 1.0 scan/s because it delivers the greatest signal intensity (Figure 2.5B) without affecting the precision (Figure 2.5C).

41

Figure 2.5: A- The effect on the peak shape when increasing the number of scans per second. B- A decrease in detection of ions as the number of scans per second is increased leading to decreased sensitivity. C-The negligible effect on the precision of repeated measurements resulting from increasing the scan rate of the instrument.

42

2.3.2 Validation

The present method to quantify stilbenes was intended to provide an accurate, selective and precise determination of the major stilbenes present in wine. To do so, it is absolutely essential that the analytical method used for quantification be properly validated by establishing that the method is accurate, precise, selective, repeatable, and that the test samples and calibration standards are stable before and during analysis.108,109

2.3.2.1 Specificity

The four analytes, trans/cis-resveratrol, and trans/cis-piceid were unambiguously identified through retention time comparison with high purity reference standards, as well as accurate MS/MS characterization. Both trans and cis-resveratrol were well separated and displayed the same fragmentation pathway with an [M-H]- of 227.0714 m/z (0.00 Δppm,

- C14H11O3 )(figure 2.6A). Two characteristic fragments of resveratrol were present, the first at

- - 185.0590 m/z (9.76 Δppm, C12H9O2 ) and the second at 143.0501 m/z (0.70 Δppm, C10H7O1 )

110 which corresponds to the one and two losses of C2H2O2 respectively.

Similarly to resveratrol, trans- and cis-piceid were well separated, and both the retention times and MS/MS fragmentation were compared with authentic reference standards. Figure 2.6B displays the MS/MS mass spectrum of trans/cis-piceid, with an [M-H]- of 389.1253 (2.82 Δppm,

- C20H21O8 ). The main fragment produced comes from the loss of glucose, to give “resveratrol” at

- 227.0703 (4.84 Δppm, C14H11O3 ). Apart from this loss, the other two fragments are identical to those of resveratrol.

43

Figure 2.6: A - MS/MS spectrum of [M-H]- 227.0714 m/z; resveratrol. B - MS/MS spectrum of [M-H]- 389.1242 m/z; piceid. C-MS/MS spectrum of [M-H]- 243.0663 m/z; piceatannol.

Finally, trans/cis-piceatannol, for which reference standards were unavailable, were compared with literature MS/MS fragmentation information. Piceatannol differs in mass from resveratrol by +16 Da because of the hydroxyl group in the 3’ position. Consequently, its fragmentation is identical to resveratrol except that all fragments are shifted higher by +16 Da

- - (Figure 2.6C). The [M-H] of piceatannol was 243.0651 m/z (4.93 Δppm, C14H11O4 ) with two

- - main fragments at 201.0543 (5.76 Δppm, C12H9O3 ) and 159.0442 (5.65 Δppm, C10H7O2 )

110 corresponding to the consecutive losses of C2H2O, characteristic of stilbenoids. The peaks that produced these MS/MS fragments were well separated and easily discernable within all wine samples. 44

2.3.2.2 Calibration Curves

Calibration curves were prepared for both resveratrol, piceid, and their isomers individually because their ionization efficiencies are quite different, resulting in their respective response factors being vastly different. This is possible because trans-resveratrol and trans-piceid are commercially available and can be isomerized to their cis isomers by exposing them to sunlight in clear vials. While some sources state that the sunlight induced isomerization from trans to cis is problematic to reproduce, it was found in our experience that placing a solution of trans-resveratrol in the sunlight for 6 hours was suitable to yield at least 99% conversion to the cis form (Figure 2.7).111,112

Figure 2.7: A-Combined EIC (m/z 227.0714 + 389.1242) of fresh resveratrol and piceid standards. B-Combined EIC (m/z 227.0714 + 389.1242) of resveratrol and piceid standards after exposure to sunlight for 6 hours, displaying the conversion from trans (A) to cis (B).

45

The initial range for the four analytes was chosen to be 0-20 mg/L, using 15 concentration increments. These ranges were chosen because the average concentrations of resveratrol and piceid found in wine were reported within these ranges.22 Although a small percentage of studies have found stilbenoid levels greater than the upper limit of this chosen range, any wines that exceed this limit can be diluted and re-evaluated to bring the instrument response within the acceptable range.86,113,114

The internal standard epigallocatechin-3-O-gallate (EGC3OG) was chosen because it is a polyphenol that doesn’t exist in wine. Due to the multiple polyphenolic groups which share some structural similarities to resveratrol, the ionization mechanisms and retention time is also quite similar to the four analytes (Figure 2.8). EGC3OG also displayed a similar retention time as the analytes. Equally as important, EGC3OG has an ionization response factor that is similar with the stilbenoid analytes, allowing for a negligible amount to be added to each sample, giving a response that does not differ too greatly from that of the analytes.

46

Figure 2.8:Combined EIC displaying the relative retention times between EGC3OG (A), trans-piceid (B), cis-piceid (C), trans-resveratrol (D), cis-resveratrol (E). Calibration curves were constructed by plotting the EIC peak areas of analyte divided by internal standard against the concentration of each analyte. The four calibration curves were shown to be very linear until 14.1 mg/L for trans/cis-resveratrol, trans-piceid, and 12 mg/L for cis-piceid, where the calibration curve began to level off, indicating the limit of linearity (LOL) for the instrument/analyte. The linearity was assessed by the coefficient of determination (r2) which was found to be no less than 0.9977 for the four curves. This suggests that the linear regression represents the fitted data well. There is a great difference in the slopes of each calibration curve. This essentially equates to the sensitivity of the method for each analyte, a larger slope results in a greater difference between two very similar instrument responses, allowing the resolution of concentrations very near to the limit of detection. The order of sensitivity of this method is as follows: cis-piceid, trans-piceid, cis-resveratrol, trans-resveratrol.

47

Figure 2.9: Calibration curves for (A) trans-resveratrol, (B) cis-resveratrol, (C) trans-piceid, (D) cis-piceid displaying the linear regression 2.3.2.3 Linearity, Limits of Detection, and Quantification

The limit of detection (LOD) is defined as the lowest concentration of analyte that can elicit a noticeable instrument response. The LOD was determined by multiplying the standard deviation of the blank (N=25) response (at the retention time of the analyte) by 3, and then dividing by the slope of the calibration curve.108,109 Using this definition, the LOD was determined to be 0.032 and 0.0035 mg/L for trans-resveratrol and cis-resveratrol, respectively.

While the LOD for trans and cis piceid was determined to be 0.0039 and 0.0007 mg/L, respectively. These values are comparable to other studies using QqQ and QTOF quantification of stilbenoids in wine.84,115 Similar to the LOD, the limit of quantification (LOQ) is the lowest concentration of analyte that can be detected accurately and precisely with a high measure of repeatability. The LOQ was calculated by multiplying the standard deviation of the blank by 10

48

and dividing by the slope of the calibration curve (Table 2.4).108,109 The limits of quantification were also in line with previous MS quantification studies.84,115

It is important to note that this practice of using the variability of a blank matrix is only one of three methods that are commonly used to determine the LOD and LOQ of an analytical method. Calculation of the LOD and the LOQ in this fashion has the serious drawback of not taking into consideration: analyte specific errors, error derived from sample preparation, and to a certain extent, matrix effects. Despite these disadvantages, the AOAC and IUPAC both endorse the application of this method. 116-118 Alternatively, the FDA recommends using a sample with an analyte concentration that is near the instrument limit of detection with replicate injections to determine the LOD. This has the benefit of determining the LOD specifically for the analyte and sample preparation errors.116 Finally, the LOD can also be determined from the calibration curves themselves. Instead of using the standard deviation of the noised produced by multiple blank injections, one can use the standard deviation of the y intercept of the regression line multiplied by 3 (or 10 for the LOQ) and then divided by the slope of the curve.

Table 2.4: Validation result including the limits of detection and quantification, as well as the precision (intra- and inter- day) results.

Test LOQ LOD Red Wine White Wine Regression a b Analyte R2 Range (mg/L) (mg/L) Equation Intra-day Inter-day Stability Intra-day Inter-day Stability (mg/L) RSD (%)c RSD (%)d RSD (%)e RSD (%)c RSD (%)d RSD (%)e trans- Y = 0.0092x 0.999 0 – 14.1 0.107 0.032 5.23 6.04 4.78 10.04 16.81 9.84 resveratrol + 0.0007 cis- Y = 0.0897x 0.999 0 – 14.1 0.0118 0.0035 5.68 5.302 4.43 3.00 4.018 3.18 resveratrol - 0.0038 trans- Y = 0.1532x 0.999 0 – 14.1 0.0129 0.0039 3.81 3.99 2.98 1.80 8.31 5.70 piceid + 0.0068 cis- Y = 0.7774x 0.998 0 – 12 0.0023 0.0007 3.16 8.74 2.26 2.56 5.702 1.89 piceid + 0.0713

49

2.3.2.4 Precision, Stability, and Recovery

The precision of the method was determined by observing the relative standard deviation of each analyte in both a white and red wine, using nine replicate measurements, three hours apart, three times during a single day. This gave an intra-day RSD of 1.8 – 10.04 % for white wine and 3.16-5.68 % for red wine. The reason for the greater range of RSDs in white wine can be explained by the significantly lower concentration of stilbenoids. This causes the amount of resveratrol to be closer to the limit of quantification where the method is the least precise.

Conversely, when the concentration of the stilbenoid falls closer to the middle of the concentration range of the method, as with the other three stilbenes in this experiment, there is less variability throughout repeated measurements. This can be attributed to white wine having a far less complicated matrix, increasing the precision of the ionization. The same procedure was then repeated 14 days later, in order to determine the inter-day repeatability which produced an

RSD of 4.02-16.81 % for white and 3.99-8.74 % for red wine. These values fall well within the bioanalytical method guidelines showing that the method was precise within-run and between-run over several time intervals.

A common analytical practice in wine chemistry is the freezing of wine samples once they have been taken from a fresh bottle of wine.119 The stability of multiple freeze-thaw cycles was therefore of interest at least in relation to the possible effects it has on stilbenoid stability. To test this, the RSD was determined for nine replicates of red and white wine samples which were frozen three times and analyzed after each thawing (Table 2.4). The RSD values of the stability experiments in both wines (2.26-9.84%) were no greater than the intraday precision values (3.26-

10.04%). The constant freezing and thawing cycles that were used throughout this experiment

50

appear to have no significant effect on the stability of the analytes in wine. This validates the practice of freezing wine samples once the bottle is opened.

The recovery (or accuracy) of the method was determined by spiking both red wine and white wine with analyte at three concentration levels spanning the linear range of the calibration curve. Each sample was measured in triplicate, including the non-spiked wine. This validation procedure is very important as it acts as an evaluation of the calibration curve’s ability to accurately measure the analyte. The mean recovery for each analyte is displayed in Table 2.5. All four analytes show acceptable recovery efficiencies, the lowest belonging to trans-piceid at middle (1.00 mg/L) concentration at 95.93% in white wine. The highest recovery established was the mid-range (1.00 mg/L) cis-piceid at 102.2%. These values are well within the typical recovery experiment results published for LC-MS stilbene quantification.106

51

Table 2.5: Recovery experimental results (N=3 for each concentration) using four concentration levels in both red and white wine.

Red Wine White Wine Spike Measured Standard Subtracted RSD Recover Measured Standard Subtracted RSD Recovery [Conc] [Conc] Deviation [Conc]a (%) y Mean [Conc] Deviatio [Conc]a (%) Mean (mg/L) (mg/L) (mg/L) (mg/L) (%) (mg/L) n (mg/L) (mg/L) (%) trans-resveratrol 0 5.6 0.11 0.000 1.9 - 0.73 0.028 0.000 3.9 - 0.25 5.86 0.063 0.241 1.1 96.57 0.98 0.0269 0.251 2.7 100.57 1.00 6.61 0.074 0.989 1.1 98.91 1.71 0.043 0.978 2.5 97.78 5.00 10.6 0.18 4.946 1.7 98.92 5.6 0.19 4.878 3.4 97.55 cis-resveratrol 0 3.89 0.06 0.000 1.6 - 0.206 0.0054 0.000 2.6 - 0.25 4.13 0.036 0.242 0.9 96.65 0.459 0.0039 0.252 0.84 101.00 1.00 4.9 0.13 1.015 2.4 101.48 1.22 0.042 1.016 3.4 101.62 5.00 8.85 0.076 4.964 0.9 99.28 5 0.19 4.831 3.7 96.62 trans-piceid 0 3.87 0.042 0.000 1.1 - 0.53 0.0093 0.000 1.8 - 0.25 4.11 0.023 0.241 0.56 96.30 0.770 0.007 0.240 0.91 95.93 1.00 4.86 0.024 0.989 0.49 98.90 1.49 0.016 0.963 1.1 96.29 5.00 8.8 0.13 4.952 1.45 99.04 5.5 0.11 4.982 2.1 99.65

cis-piceid 0 7.24 0.06 0.000 0.83 - 1.11 0.0038 0.000 0.34 - 0.25 7.48 0.07 0.243 0.93 97.36 1.35 0.014 0.244 1 97.55 1.00 8.23 0.077 0.985 0.94 98.53 2.13 0.046 1.022 2.2 102.24 5.00 12.1 0.16 4.889 1.4 97.78 5.9 0.18 4.839 2.9 96.79 aSubtracted concentration is calculated by subtracting the 0 mg/L spike concentration from each measured concentration at each nonzero spike level

2.3.4 Applications to Wine Samples

2.3.4.1 Wine samples

In order to demonstrate the applicability of the validated quantification method, Canadian wines were analyzed. The UHPLC-QTOF MS method was applied to the 44 Canadian wines listed in Table 2.6, of which, 37 were red wines, 5 were white wines and 2 were rosé. Each wine was (filtered through a 0.45 µm PTFE syringe filter and 50 µL of internal standard was added) then analyzed in triplicate. The vintages and varieties of the wine samples are reported in Figure

2.10. Primary targets of this study were Pinot Noir, Cabernet Sauvignon and Merlot, which are very popular red wine varieties produced in the Okanagan valley.

52

Figure 2.10: Distribution of wine based on vintage and varietal quantified in this study (A). Total number of each variety analyzed (B).

2.3.4.2 Quantification in Wine Samples

The quantities of the six major stilbenes, trans-/cis-resveratrol, trans-/cis-piceid, and trans-/cis-piceatannol were determined and are displayed in Table 2.6. Mono-varietal red wine from around the world has been shown to contain levels of trans-resveratrol ranging from 0 – 14.3 mg/L and an average of 1.9 ± 1.7 mg/L using a sample size of 511 wines.22 The amount of trans- resveratrol in red wine determined with the present method ranged from trace - 4.70 mg/L with a mean of 1.67 ± 1.04 mg/L. This falls within the range, but lower than the mean value of previous studies, where Canadian wines were shown to have the highest average trans-resveratrol content of 3.16 ± 1.34 mg/L, and ranged from 1.2 to 5.8 mg/L for 36 red wine samples.80 Our results do not corroborate the claim that Canadian wines contain more trans-resveratrol on average;

53

however, since the previous study analyzed Ontario wines, this disparity could be the result of

Ontario vs. British Columbia winegrowing conditions. A total of 7 white wines were analyzed and in all samples the amount of trans-resveratrol was either not detected, or at least below the LOD of the method. This is somewhat expected as several studies have noted that white wines contain substantially lower amounts of stilbenes due to their limited skin maceration time.120,121

54

Table 2.6: Quantification of all 6 monomer stilbenoids in 44 Okanagan wines. Error is displayed as standard deviation, indicating the error of the method including sample preparation and instrumental error.

Concentration (mg/L) # Vintage Variety Winery t-resv c-resv t-piceid c-piceid t-ptannol c-ptannol 1 2012 Pinot Noir Ancient Hill 2.5 ± 0.1 6.2 ± 0.1 0.93 ± 0.03 4.2 ± 0.1 3.02 ± 0.08 1.59 ± 0.03 2 2010 Pinot Noir Arrowleaf 1.51 ± 0.09 8.5 ± 0.3 1.02 ± 0.01 2.1 ± 0.1 4.8 ± 0.2 1.04 ± 0.03 3 2011 Pinot Noir Calona Winery trace 3.32 ± 0.03 2.427 ± 0.005 3.88 ± 0.02 3.06 ± 0.03

55

The concentration of cis-resveratrol was found to be markedly lower than the trans form in many studies involving red and white wine.22,113 However, the present results indicate that cis- resveratrol had both a larger range (0.077 – 10.22 mg/L), and average (1.95 ± 2.5 mg/L) concentration than the trans form in red and white wine. These values agree with recent studies of trans- and cis-resveratrol using stable isotope dilution analysis (possibly the most accurate MS quantification technique), where the concentration ranged from trace – 11.97 mg/L.115

56

Table 2.7: Quantification results displaying the minimum and maximum total and individual stilbenoid concentrations, separated by variety and wine type. Error expressed as standard deviation, indicative of the variability of the sample pool.

Lowest Concentration Highest Concentration Mean Concentration Variety (mg/L) (mg/L) (mg/L) ± SD Merlot (N=11) trans-resveratrol trace 2.16 1.13 ± 0.54 cis-resveratrol 0.31 1.42 0.67 ± 0.39 trans-piceid 0.181 2.74 1.50 ± 0.87 cis-piceid 1.60 3.48 2.57 ± 0.76 trans-piceatannol 0.28 2.00 1.01 ± 0.65 cis-piceatannol

57

The mono-glucosides of resveratrol, trans- and cis-piceid, have been shown in numerous studies to be the predominant stilbene compounds in red and white wine with concentrations ranging from trace – 38.47 mg/L with an average of 5.4 ± 4.8 mg/L in red and 1.4 ± 2.4 mg/L in white.22,103,113 In the present study, we report average trans- and cis-piceid concentrations of 1.75

± 1.27 mg/L and 2.98 ± 1.65 mg/L, respectively. These values do in fact fall within the ranges in the literature; however, the average trans-piceid concentration is lower and cis-piceid’s average concentration is twofold higher. The highest concentration of trans-/cis- piceid was found in rosé at 9.57 mg/L, substantially lower than the maximum reported concentration of 38.47 mg/L found in Italian red wines.84

Piceatannol is a tetrahydroxylated stilbene that differs from resveratrol by the additional hydroxyl group in the 3’ position. The trans- and cis- isomers of this compound were quantified as equivalents of trans-piceid in this method because of the lack of availability of piceatannol.

While this quantification using the trans-piceid calibration curve will not give analytical accuracy for this compound, it can serve as an estimate of the concentration, and it can allow for the comparison of abundances within this method, if not between other sources. Both trans- and cis- isomers of piceatannol were found in the ranges of trace -6.42 mg/L, with an average of 1.55 ±

1.38 mg/L for trans and 0.45 ± 0.44 mg/L for cis in red wine. Levels in white wine were substantially lower with averages of 0.13 ± 0.06 mg/L and 0.04 ± 0.04 mg/L for trans and cis, respectively. Few studies have quantified trans-piceatannol, however, one such experiment found it ranging from trace – 5.22 mg/L in red wine which is quite similar to the concentrations that were determined here.84 It has been noted that cis-piceatannol has not been found in red wines, however, in the present study we did identify this isomer of piceatannol, although in few wines, and at very low concentrations.84

58

Figure 2.11: Quantification results represented in graph form separated by red and white wine varieties.

The average individual levels of stilbenes, in Merlot, Pinot Noir, and Cabernet Sauvignon wines are reported in Table 2.7 and displayed in Figure 2.11. A one-factor ANOVA was performed for each analyte, on each of the three varieties of red wine. A p-value of ≤ 0.0167

(Bonferroni, 3 comparisons) was reported for both trans- and cis-resveratrol, indicating a significant difference between the levels of these stilbenes between Merlot and Pinot Noir at 95 % confidence. Unsurprisingly, total stilbene content also showed this same disparity at a significant level. These values indicate that Pinot Noir is the most stilbene-rich mono-varietal wine analyzed via this experimental method. There is also substantial support for this finding within the literature.22,119

59

In Conclusion, an experimental method for the rapid quantification of stilbene monomers in wine samples by UHPLC-ESI-QTOF was developed. The method was fully validated as defined by the ICH and FDA bio-analytical guidelines.108,109 Additionally, the validity of the method was further established by applying it towards Okanagan wine samples. The average levels of stilbenes found in Pinot Noir were found to be significantly higher than in Merlot, validating the hypothesis that some varieties at least, are discernable by stilbene content. The quantification data generated showed that Okanagan wines have a stilbene content that falls within the known concentrations from wines around the world.

60

CHAPTER 3 – INDENTIFICATION AND SEMI-QUANTIFICATION OF

OLIGOMERIC STILBENOIDS IN RED WINE

3.1 Synopsis

Three red wine extracts were analyzed to obtain their stilbenoid profiles. In the initial targeted product ion scan experiment, 41 potential compounds (shown in Table 3.4), were detected. To our knowledge, 26 of these possible oligomeric stilbenes have not yet been identified in wine.

The dimeric forms of resveratrol that have been previously reported in wine include cis-ε- viniferin, trans-ε-viniferin, cis-δ-viniferin, trans-δ-viniferin, pallidol, parthenocissen A, and quadrangularin A.86,122,123 In the present study, compounds 7, 11, 13, 16 and 17 have been unambiguously identified as pallidol, cis-ε-viniferin, trans-ε-viniferin, trans-δ-viniferin, and cis-

δ-viniferin, respectively by comparison with authentic standard compounds. Parthenocissen A and quadrangularin A have been tentatively identified by comparison with accurate mass fragmentation patterns from the literature.

In addition to the homodimers of resveratrol, several derivatives have been identified in various plants. These include O-glycosylated dimers, methoxylated dimers, piceatannol- resveratrol heterodimers, piceatannol homodimers, and even additional oxidation products of these dimers.13 Although some of these compounds such as pallidol-3-O-glycoside (29), and ε- viniferin-glycoside (28) have been isolated in white wine, we report 10 dimers that have not yet been reported in wine but have been described in cell cultures of Vitis vinifera.124,125 Many of

61

these stilbenoid dimer derivatives have also been identified in other plant species as well as in the leaves and stems of Vitis vinifera.126

The stilbenoid trimers and tetramers are also well represented in many different species of plants, including Vitis vinifera.13 Although trimeric and tetrameric stilbenoids have been detected in the leaves and stems of grapevine, only the tetramer hopeaphenol (39), has been identified in wine.127 In this study, eight potential trimers (30-37) and four potential tetramers (38-41) have been found, displaying similar fragmentation patterns to stilbenoid oligomers in the literature. All precursor ion and fragment ion masses are supported by accurate mass measurements with mass error below 10 ppm.

A solid-phase extraction (SPE) method was developed for quantification of the stilbenoid oligomers in red wines. The 41 monomers and oligomers in red wine from the Okanagan Valley

(Cabernet Sauvignon, Merlot and Pinot Noir) and Québec (Maréchal Foch, Marquette, Sabrevois,

St. Croix, and Frontenac) wines were semi-quantified as equivalents of piceid using this method.

Okanagan Valley Vitis vinifera wines were found to have a significantly higher (5.25 ± 2.73 mg/L) average stilbenoid content than the Quebec-grown hybrid varieties (2.42 ± 2.49 mg/L). The highest concentration of total stilbenoids was 10.67 mg/L in Pinot noir with an overall average of

3.32 ± 2.86 mg/L in all wines.

62

3.2 Experimental

3.2.1 Reagents, Samples & Solutions

Deionised water was purified with a MilliQ water system (Millipore, Bedford, MA).

HPLC grade acetonitrile and diethyl ether were purchased from Fisher Scientific (Waltham,

MA,USA). Formic acid was purchased from Sigma Aldrich (St. Louis, MO, USA). The red wines used for quantification were Okanagan Valley (VQA, Okanagan Valley) Vitis vinifera varieties:

Pinot Noir, Merlot and Cabernet Sauvignon; and Quebec Vitis rupestris hybrid red wines

(Québec) Maréchal Foch, Marquette, Sabrevois, St. Croix, and Frontenac sampled in August

2011, and September 2013 respectively, and stored at -80oC before extraction.

3.2.2 Sample Preparation

3.2.2.1 Compound Elucidation Sample Preparation

One hundred milliliters of each red wine was extracted with diethyl ether (3 x 100 mL) and/or ethyl acetate (3 x 100 mL). The solvent was evaporated under vacuum at room temperature and re-suspended in 1 mL of MeOH to afford a concentrated red wine extract. The sample was maintained at 4oC until analysis by UHPLC-ESI-QTOF.

3.2.2.2 Quantitative Analysis Sample Preparation

The validated extraction procedure was carried out as follows: red wine samples (50 mL) were filtered through a 0.45 µm PTFE syringe filter. A 6 mL BondElut ENV SPE cartridge

(Agilent Technologies, Santa Clara, CA) was preconditioned with 3 mL of methanol followed by equilibration with 3 mL of 0.2 % aqueous acetic acid. Without allowing the column to dry, 15 mL of the prepared wine sample was loaded onto the column in 5 mL increments. The column was then dried using the vacuum manifold for 10 minutes prior to washing with 3 mL of H2O,

63

followed by another drying for 10 minutes. The analytes were eluted with 3 mL of methanol which was evaporated to dryness and then reconstituted in 1 mL of methanol to afford a 15x concentrated sample of wine.

3.2.3 Ultra-High Performance Liquid Chromatography Parameters

For separation, an Agilent 1290 series UHPLC (Agilent Technologies, Santa Clara, CA) equipped with Binary pump, solvent degasser, and a thermostatted column compartment. The columns used for separation were two reversed-phase Zorbax SB AQ (2.1 x 150 mm + 2.1x 100 mm in series, 1.8μm, Agilent Technologies, Santa Clara, CA). The mobile phase consisted of A: water (0.1 % formic acid) and B: acetonitrile (0.1 % formic acid). The 50 minute elution method started with 10-18 %B over 5 minutes. A linear gradient was used from 18-40 %B over the next

40 minutes, followed by an increase from 40-100 %B in 1 minute. An isocatric plate at 100 %B for 2 minutes followed by a decrease from 100-10 %B over 1 minute was used to clean the column. A 3 minute post-run isocratic step at 10 %B was used to re-equilibrate the column for the next analysis. The flow rate was held constant at 0.35 mL/min at 40oC throughout the analysis.

These UHPLC settings were used for both the qualitative structural elucidation studies as well as the quantification studies.

Table 3.1: Gradient elution method, A: H2O w/ 0.1 % Formic acid, B: Acetonitrile w/0.1 % Formic acid.

Time (min) Mobile Phase A (%) Mobile Phase B (%) 0.00 90 10 5.00 88 18 45.00 60 40 46.00 0 100 48.00 0 100 49.00 90 10 52.00 90 10

64

3.2.4 Quadrupole-Time-of-Flight Parameters

MS/MS experiments were performed using the Agilent UHPLC system above coupled to an Agilent 6530 Series accurate mass Q-TOF MS/MS spectrometer (Agilent Technologies, Santa

Clara, CA) operated in high mass resolution (4GHz) mode. An electrospray source in negative ion mode equipped with Agilent Jet Stream technology allowed for ionization. Agilent

MassHunter 5.0 software was utilized for both data acquisition and analysis and nitrogen was used both as the drying and collision gas.

The nebulizer was held at a pressure of 25 psig with a drying gas flow and temperature of

10 L.min-1 and 325oC, respectively. The nitrogen sheath gas had a temperature of 400oC and a flow rate of 12 L.min-1. The fragmentor and capillary voltages were optimized at 200 V and 2.25 kV respectively. Finally, for qualitative analysis, the collision energy (CE) was optimized for each compound separately using a range of 10-50 eV and is listed in Table 3.4.

3.2.5 Quantitative Method Validation

3.2.5.1 Calibration Curves

The stock solutions trans-piceid were prepared in triplicate by weighing out 10 mg piceid into a 100 mL volumetric flask and diluted to the mark with model wine solution to afford 3 stock solutions of 100 mg/L. The stock solution for the internal standard was prepared by weighing 4.9 mg of ECG3OG into a 50 mL volumetric flask and diluted to the mark with model wine.

65

Table 3.2: Serial dilution protocol for the preparation of the calibration curve standards.

Concentration of Concentration after Solution # analyte prepared Internal Standard (mg/L) Addition (mg/L) SX 100 - S1 60 30 S2 50 25 S3 40 20 S4 30 15 S5 20 10 S6 10 5 S7 5 2.5 S8 2 1 S9 1 0.5 S10 0.5 0.25 S11 0.2 0.1 S12 0.1 0.05 S13 0.05 0.025 S14 0.02 0.01

From the stock solution, a serial dilution was performed in triplicate using concentrations as indicated in table 3.2. Solutions S1-S6 were prepared by pipetting 0.6, 0.5, 0.4, 0.3, 0.2, and 0.1 mL respectively of SX and diluting to 1 mL using model wine solution. In order to reach the lower concentration ranges, serial dilution using 0.1 mL of S2, S5 and S6 were used to make S7,

S8 and S9 respectively. To create the lower concentrations, 0.1 mL of S7, S8, and S9 were used to make S10, S11, and S12. Finally, S13 and S14 were made by pipetting 0.1 mL of S10 and S11 and diluting to 1 mL.

3.2.5.2 Solid Phase Extraction Validation

The cartridge capacity was optimized by loading 1, 5, 10, 15, 20, 25, and 30 mL of wine in triplicate. The run-through was collected and put aside for further analysis. Once the analytes were eluted with methanol, they were evaporated under a stream of N2 and reconstituted, before being analyzed by UHPLC-ESI-QTOF. Every extracted sample was compared with its run- through for the monomers, dimers, trimers and tetramers.

66

The extraction efficiency was determined by using model wine solution (12 % ethanol, 5 g/L tartaric acid and brought to pH 3.60 with NaOH), spiked with 0.001, 0.050 and 1.000 mg/L of trans-piceid in replicate (n=6). A second set of replicates containing no spike was also prepared and extracted. After performing the SPE protocol, the internal standard was added to each sample; for the un-spiked samples, 0.015, 0.75, and 15.0 mg/L of trans-piceid was also added, these samples were treated as 100% recovery and EIC peak areas were compared to give the extraction efficiency at low, medium, and high concentration.

67

3.2.5.3 Linearity, Limits of Detection & Quantification

The calibration curve of EIC(peak area, piceid)/EIC(peak area ECG3OG) ranging from

0.01 – 30 mg/L had its linearity evaluated by visual inspection, r2, and inspection of the residual plots. The limits of detection and quantification were calculated by running 10 blank samples composed of model wine solution with 0.5 mg/L internal standard added, but without any analyte present. The EIC peak area of the blank sample at the retention time of trans-piceid was integrated and determined to be the signal of the noise. The LOD and LOQ were thus calculated as follows:

푦퐿표퐷 = 푦퐵푙푎푛푘 + 3 ∗ 푆퐷푏푙푎푛푘

푦퐿표푄 = 푦퐵푙푎푛푘 + 10 ∗ 푆퐷푏푙푎푛푘

In terms of concentration however, the blank theoretically possesses 0 mg/L of analyte and therefore the equations become:

3 ∗ 푆퐷 퐶 = (0) + 푏푙푎푛푘 퐿표퐷 푚

10 ∗ 푆퐷 퐶 = (0) + 푏푙푎푛푘 퐿표푄 푚

Where m represents the slope of the calibration curve.

68

3.2.5.4 Recovery, Inter & Intra-day Precision

The recovery or accuracy of the method was established by spiking 15 mL of red wine with three levels of piceid concentration (0.001, 0.05, and 1.00 mg/L) in triplicate. The red wine was then extracted using the above protocol, and when the wine is reconstituted, the internal standard is added at 0.5 mg/L. Using the calibration curve, the concentrations of trans-piceid are then calculated, and then compared with the known added spike.

The intra-day precision was measured by performing the recovery experiment twice on the same day, 12 hours apart. The %RSD was calculated between the two sets of data to give a measure of the intra-day precision. The inter-day precision was then calculated by performing the recovery experiment a third time, two weeks later, where the results are compared with the first two, in order to determine the inter-day precision.

69

3.3 Results & Discussion

3.3.1 Optimization of UHPLC and MS/MS parameters for qualitative analysis.

For the qualitative studies, the fragmentor, capillary, and skimmer voltages, drying and sheath gas flows, nebulizer pressure and Collision energy were optimized separately. However, the majority of the parameters were very similar across all of the compounds tested. The fragmentation voltage was the parameter that had the greatest individual effect for each compound, and the data is displayed in Table 3.4. A compromise fragmentor voltage of 200 V was chosen in order to ionize everything satisfactorily, even if it was not the optimal for some compounds.

While this parameter was not involved in the quantification because it was carried out using MS only and CE is not used in single stage scanning mode, it was very important for obtaining the characteristic fragments.

To optimize the fragmentor and capillary voltages for the best possible ionization of these compounds, a preliminary acquisition in simple MS mode was used. The fragmentor and capillary voltages were fixed at 200 V and 2.25 kV respectively. A targeted MS/MS acquisition was used (50-1200 m/z) with a fixed isolation peak width (1.3 m/z) and collision energy (CE) across all compounds initially. The CE was then varied between 10 - 50 eV, and optimal CE were chosen for each class of compounds to obtain the most characteristic fragments (Table 3.4).

70

3.3.2 Strategy for Compound Identification without Standards

Quadrupole-Time-of-Flight (QTOF) has an advantage over the triple quadrupole setup in that it can garner accurate mass measurements of both precursor and product ions. This can enable sample screening, leading to semi-structural elucidation of unknown compounds.128,129 However, true identification and characterization requires NMR data or comparison with authentic reference standards. This can be problematic as many secondary plant products are not commercially available and are troublesome to isolate in large enough quantities for NMR analysis. Yet there are a few studies wherein the authors have isolated some stilbenoid oligomers, and have reported the MS/MS fragmentation patterns, allowing for multiple ion comparisons with accurate mass, leading to a somewhat clear identification of these compounds in wine samples.

Utilizing the QTOF instrument in MS/MS mode implies the pre-selection of the parent

[M-H]- masses. Fundamentally this requires that the pseudo-molecular masses be known before- hand, so that they can be filtered in the quadrupole. Consequently the QTOF’s greatest advantage is in its ability to perform pre-target screening.130 High accuracy mass measurements can enable an increased confidence in the identity of precursor ion masses, fragmentations and neutral losses, since low parts-per-million accuracy (four decimal places) can only result from a low number of molecular compositions, especially when considering most biological molecules contain only carbon, hydrogen, nitrogen, and oxygen.131-133

71

Table 3.3: Theoretical [M-H]- of many potential stilbenoid derivatives that were scanned by negative mode UHPLC-ESI- QTOF analysis in red wine

# Exact Precursor Mass [M-H]- Prospective Compounds P1 227.0714 Resveratrol P2 389.1242 Resveratrol-Glucosides P3 551.1770 Resveratrol-diglucosides P4 713.2298 Resveratrol-triglucosides P5 243.0663 Piceatannol P6 405.1191 Piceatannol-glucosides P7 567.1719 Piceatannol-diglucosides P8 731.2247 Piceatannol-triglucosides P9 255.1027 Pterostilbene P10 387.1085 2,4,6-Trihydroxyphena nthrene 2-O-β-glucoside P11 451.1187 Oxidized Dimers P12 453.1344 Resveratrol Dimer P13 615.1872 Resveratrol Dimer-glucosides P14 777.2400 Resveratrol Dimer-diglucosides P15 939.2928 Resveratrol Dimer-triglucosides P16 469.1293 Resveratrol-Piceatannol Dimers P17 631.1821 Resveratrol-Piceatannol Dimer-glucosides P18 793.2349 Resveratrol-Piceatannol Dimers-diglucosides P19 485.1606 Resveratrol dimer (methylated) P20 485.1236 Piceatannol-Piceatannol Dimers P21 467.1136 Dehydro Resveratrol-piceatannol Dimers P22 471.1449 Oxidized Dimers P23 679.1974 Resveratrol Trimer P24 841.2502 Resveratrol Trimer-glucosides P25 1003.303 Resveratrol Trimer-diglucoside P26 677.1817 Resveratrol dehydro-Trimer P27 695.1923 Resv-Resv-Picea Trimer P28 711.1872 Resv-Picea-Picea Trimer P29 727.1821 Picea Trimer P30 905.2604 Resveratrol Tetramer P31 1067.313 Resveratrol Tetramer-glucosides P32 1229.366 Resveratrol Tetramer-diglucosides P33 921.2553 Oxidized Resveratrol Tetramer P34 1131.3233 Resveratrol Pentamer P35 1370.3942 Resveratrol Hexamer

A pool of known and theoretically possible stilbenoids, was constructed (Table 3.3). The precursor ion masses were scanned in the red wine extracts by inspecting the individual EICs for each mass to determine if any of the peaks produced warranted further MS/MS investigation. This required multiple sample injections, the first for MS scanning, and all subsequent injections based on the number of precursor masses that could be targeted without substantial loss of sensitivity, which were usually eight at a time. One would prefer to perform a single injection only, especially if the sample is precious and little. This highlights one of the few limitations of UHPLC-QTOF, the necessity to preselect ions in order to perform targeted MS/MS analysis.

72

-94 Da -106 Da

-42 Da -42 Da

-94 Da -162 Da (Glucose)

-106 Da

-42 Da -42 Da

[M- [M- - - 359.0925

347.0925

411.1238 369.1132

615.1872 453.1344

Figure 3.1: Top: MS/MS spectrum of authentic standard trans-ε-viniferin [M-H]- 453.1344 m/z . Bottom: previously unidentified MS/MS spectrum of [M-H]- 615.1872 m/z, tentatively identified as the 3-O-glucoside of trans-ε-viniferin.

Figure 3.1 is a representation of the strategy of compound identification without a standard. The top spectrum shown is the MS/MS fragmentation pattern of the authentic standard of trans-ε-viniferin (13). The characteristic neutral losses are 106 Da, 94 Da, and two consecutive losses of 42 Da. These losses are commonly found in nearly all spectra of stilbenoid oligomers.

The bottom spectrum in Figure 3.1 is the MS/MS fragmentation of an ion with an [M-H]- of

615.1872 m/z. This is the theoretical mass of a 3-O-glucoside of a stilbenoid dimer, which was identified as such by the identical fragmentation pattern after the loss of 162 Da (glucose). This is a good example of how standards can be used to aid in the identification of structurally similar compounds. Compounds 1-41 were identified in this manner with both more and less complex fragmentation pathways.

73

3.3.3 Identification of stilbene compounds in wine (by accurate MS/MS).

Once all wine extracts were screened in single stage MS mode, EICs that displayed well defined peaks displaying accurate mass ions with no more difference than 10 ppm from target ions were selected for further inquiry using targeted MS/MS mode. The results and proposed identification of compounds by MS/MS investigation are summarized in Table 3.4 and Figure 3.2, and are discussed in the following sections.

74

Table 3.4: Fragmentation patterns and tentative assignments of stilbenoid compounds in red wine extract.

RT Formula [M-H]- [M-H]- Mass Error Fragmentor CE MS/MS Peak MS/MS Fragments a Assignment*,b (min) [M-H-] Calculated Experimental (ppm) Voltage (V) (eV) References c 84,110,134 - 1 20.01 C14H11O3 227.0714 227.0725 4.84 160 22.5 185; 143 trans-resveratrol 84,110,134 - 2 21.96 C14H11O3 227.0714 227.0717 1.32 160 22.5 185; 143 cis-resveratrol 84,110,135,136 - 3 14.24 C14H11O4 243.0663 243.0659 -1.65 170 25 201; 159 Piceatannol 84,125,134,136 - 4 11.19 C20H21O8 389.1242 389.1235 -1.8 190 20 227 trans-Piceid 84,125,136 - 5 13.56 C20H21O8 389.1242 389.1263 5.4 190 20 227 cis-Piceid 84,125,136 - 6 8.21 C20H21O9 405.1191 405.1206 3.7 160 15 243; 201; 159 Astringin 136,137 - 7 21.46 C28H21O6 453.1344 453.1347 0.66 130 20 359; 265 Pallidol 137 - 8 23.96 C28H21O6 453.1344 453.1325 -4.19 200 20 359; 289 Parthenocissin A* 137 - 9 25.23 C28H21O6 453.1344 453.1354 2.21 210 20 359; 289 Quadrangularin A* 126 - 10 26.77 C28H21O6 453.1344 453.1363 4.19 200 20 359; 289 Ampelopsin D* 125,136,138 - 435; 411; 369; 359; 347; 11 31.45 C28H21O6 453.1344 453.1345 0.22 210 20 cis-ε-Viniferin 333; 225 126 - 435; 411; 369; 359; 347; 12 32.20 C28H21O6 453.1344 453.1332 -2.65 220 20 cis-ω-Viniferin* 333; 225 125,136,138 - 435; 411; 369; 359; 347; 13 33.83 C28H21O6 453.1344 453.1376 7.06 220 20 trans-ε-viniferin 333; 225 126 - 435; 411; 369; 359; 347; 14 34.34 C28H21O6 453.1344 453.1357 2.87 220 20 trans-ω-Viniferin* 333; 225

- 435; 411; 369; 359; 347; 15 36.13 C28H21O6 453.1344 453.1343 0.22 220 20 Dimer 1* - 333; 225 125,136 - 16 38.79 C28H21O6 453.1344 453.1356 1.77 230 25 435; 411; 369; 359; 333 trans-δ-viniferin 125,136 - 17 39.71 C28H21O6 453.1344 453.1329 -3.31 230 25 435; 411; 369; 359; 333 cis-δ-viniferin 135 - 451; 427; 385; 375; 359; 18 16.23 C28H21O7 469.1293 469.1305 2.56 150 25 R + P Dimer* 347; 265 135,138 - 451; 427; 385; 375; 359; 19 26.48 C28H21O7 469.1293 469.1301 1.71 210 25 cis-Scirpusin A* 347; 333; 241 135,138 - 451; 427; 385; 375; 359; 20 27.93 C28H21O7 469.1293 469.1307 2.98 230 25 trans-Scirpusin A* 347; 333; 241 125,134,139 - 21 11.30 C28H23O7 471.1449 471.1447 -0.42 200 20 377; 349; 255; 121 Restrisol A* 125 - 22 17.63 C28H23O7 471.1449 471.1456 1.49 230 20 349, 255, 121 Oxidized Dimer 1* 125 - 23 18.25 C28H23O7 471.1449 471.1477 5.94 120 20 387, 377, 349, 255, 121 Oxidized Dimer 2* 125 - 24 19.52 C28H23O7 471.1449 471.1463 2.97 160 20 349, 255, 241, 121 Oxidized Dimer 3*

- 467; 443; 401; 375; 363; 25 23.28 C28H21O8 485.1236 485.1248 2.47 140 25 P + P Dimer 1* 357; 333; 265; 241 137 - 26 19.64 C29H25O7 485.1606 485.1635 5.98 210 15 453; 391; 359; 289; 255; 187 Parthenostilbenin A*

137 - 27 21.09 C29H25O7 485.1606 485.1594 -2.47 220 15 453; 391; 359; 289; 255; 187 Parthenostilbenin B*

125 - 28 25.20 C34H31O11 615.1872 615.1878 0.98 230 20 453; 411; 359; 347 ε-viniferin glycoside*

- 29 18.39 C34H31O11 615.1883 615.1872 1.79 250 15 453; 359; 289 Dimer Glycoside 1* 136 30 30.04 C42H32O9 679.1974 679.1981 1.03 230 30 585; 573; 491; 479; 385 Trimer 1* (ampelopsin C) 136,137 - 661; 637; 585; 573; 555; Trimer 2* 31 35.37 C42H32O9 679.1974 679.1978 0.59 260 30 451; 479; 357; 345 (E-miyabenol C) 136,137 - 661; 637; 585; 573; 555; Trimer 3* 32 36.32 C42H32O9 679.1974 679.1984 1.47 240 30 479; 451; 357; 345 (Z-Miyabenol C) 137 - 661; 637; 585; 573; 555; 33 37.66 C42H32O9 679.1974 679.1985 1.62 260 30 Trimer 4* 479; 451; 357; 345

75

137 - 661; 637; 585; 573; 555; 34 36.87 C42H32O9 679.1974 679.2000 3.83 260 30 Trimer 5* 479; 451; 357; 345 137 - 661; 637; 585; 573; 555; 35 39.15 C42H32O9 679.1974 679.1916 -8.54 260 30 Trimer 6* 479; 451; 357; 345 137 - 661; 637; 585; 573; 555; 36 39.84 C42H32O9 679.1974 679.1998 3.53 260 30 Trimer 7* 479; 451; 357; 345 137 - 661; 637; 585; 573; 555; 37 40.35 C42H32O9 679.1974 679.2001 3.98 260 30 Trimer 8* 479; 451; 357; 345 135 - 887; 811; 799; 717; 705; 38 29.73 C56H41O12 905.2604 905.2577 -2.98 260 35 Tetramer 1 * 699; 611 135,136 - Tetramer 2 39 31.70 C56H41O12 905.2604 905.2573 -3.42 240 35 811; 717; 611; 451; 359; 265 (Hopeaphenol)* 135,136 - 40 34.72 C56H41O12 905.2604 905.2563 -4.53 260 35 811; 793; 717; 705; 611 Tetramer 3* 135 - 41 37.49 C56H41O12 905.2604 905.262 1.77 260 35 811; 717; 611; 451; 359; 265 Tetramer 4* aMajor fragments are shown in bold. *Tentative assignment was based on accurate mass MS/MS fragmentation pattern and comparison with literature. bBolded compounds are identified in wine for the first time. cReferences contain MS/MS fragmentation data for the tentatively assigned compounds.

76

Figure 3.2: Tentative structures of the major compounds found in red wine extracts studied. 1,2 trans- and cis-resveratrol, 3 trans-piceatannol, 4,5 trans- and cis-piceid, 6 trans-astringin, 7 pallidol, 8 parthenocissin A, 9 quadrangularin A, 10 ampelopsin D, 11, 13 cis- and trans-ε-viniferin, 12, 14 cis- and trans-ω-viniferin, 16,17 trans- and cis-δ-viniferin, 19,20 trans- and cis-scirpusin A, 21 restrisol A or B, 25 scirpusin B, 26, 27 parthenostilbenin A and B, 28 ε-viniferin glucoside, 30 ampelopsin C, 31,32 E- and Z-miyabenol C, 39 hopeaphenol, 41 isohopeaphenol.

77

3.3.3.1 Monomers

Monomeric stilbenes are the most abundant stilbenoids found in wine. These were characterized by both MS/MS fragmentation patterns and retention time as compared to authentic standards. Compounds 1 and 2 were a pair of cis/trans isomers with an [M-H]- ion at m/z 227, which corresponded to deprotonated resveratrol (Figure 3.3). The MS/MS spectra of these two compounds were identical, producing two fragment ions at m/z 187 and 143, which were derived from the loss of one and two C2H2O (42 Da) moieties respectively. The loss of C2H2O from the resorcinol ring is characteristic of resveratrol and most other stilbenoids (Figure 3.4).110 The presence of cis- and trans-resveratrol was confirmed by comparison of retention times of pure standards. A similar fragmentation pattern was observed for trans-piceatannol (3) with an [M-H]- at m/z 243, which yielded fragments at m/z 201 and 159 from the subsequent losses of C2H2O (42

Da) as above. Piceatannol was also confirmed through comparison with a pure standard's retention time and fragmentation pattern.

Figure 3.3: Combined base peak chromatogram of m/z 227.0714, 243.0663, 389.1242, and 405.1191 showing compounds 1-

6.

78

Piceid and astringin are the O-glycosides of resveratrol and piceatannol respectively. The identities of compounds 4 and 5 were confirmed as trans- and cis-piceid by the detection of an

[M-H]- ion at m/z 389. In the MS/MS spectra of 4 and 5, only one strong fragment at m/z 227 corresponding to the loss of glucose (162 Da) was observed. Compound 6 displayed a low intensity signal that corresponded with both retention time and fragmentation of trans-astringin, the glycoside of piceatannol. The loss of glucose (162 Da) resulting in an ion at m/z 243 was the only fragment visible in the MS/MS spectrum of compound 6 with a molecular ion at m/z 405.

Reference standards for cis/trans-piceid and trans-astringin corresponded to compounds 4, 5, and

6 respectively.

Figure 3.4: Fragmentation pathway of the stilbene monomers (1, 2, and 3) and their glycosylated analogues (4, 5, and 6).

79

3.3.3.2 Dimers

Two structurally separate groups of resveratrol dimers exist in nature. The “viniferins” possess a 2,3-dihydrobenzofuran ring system where a major fragment of m/z 347 is observed,.

The second group that are typified by an indane ring system which produces a highly abundant ion at m/z 359, such as pallidol.13 A product ion scan of m/z 453.1344 was acquired in order to detect 11 homodimers of resveratrol numbered 7-17 (Figure 3.5). The structural identification of stilbenoid oligomers were based on the presence of characteristic fragment ions that originate from the neutral losses of 42 Da (C2H2O), 94 Da (C6H6O), 106 Da (C7H6O) and 110 Da (C6H6O2).

Figure 3.5: A: BPC scan of m/z 453.1344 showing compounds 7-17. B: MS/MS BPC showing the transition m/z 453.1344 -> 359.0925 displaying the indane type resveratrol dimers 7-10. C: MS/MS BPC of the transition m/z 453.1344 -> 411.1238 highlighting the "viniferin" type dimers 11-17.

In the MS/MS spectrum of compound 7, the precursor ion at m/z 453 produces two main

- fragments. The most abundant ion at m/z 359 of composition C22H15O5 (measured m/z 359.0933, error 2.23 ppm) is generated from the loss of a single phenol group (94 Da). The loss of a second

- phenol group (94 Da) produces C16H9O4 with m/z 265 (measured 265.0515, error 3.40 ppm).

80

Compound 7 was identified as pallidol, a well-known symmetrical resveratrol dimer. A pure reference standard`s retention time and fragmentation pattern matched that observed for compound 7.

Three compounds 8, 9 and 10 displayed identical MS/MS spectra, which can be indicative of a set of structural isomers. Two high abundance fragments originated from the [M-H]- ion at

- m/z 453. An ion formed from the loss of phenol (94 Da) at m/z of 359 (C22H15O5 , measured

359.0907, error -4.19 ppm). Further sequential losses of CO (28 Da) and C2H2O (42 Da)

- produced an ion at m/z 289 (C19H13O3 , measured 289.0867, error 1.04 ppm). Minor ions at m/z

247, 221, and 205 of abundances less than 1% were observed that could correspond to the losses of C2H2O (42 Da), C3O2 (68 Da), and two C2H2O (84 Da) respectively from the ion m/z 289.

These low abundance ions have been isolated in the MS3 spectrum of the ion at m/z 289 in two compounds which are cis/trans isomers that have been isolated in Parthenocissus laetevirens.137 It is therefore proposed that two of these three compounds are the cis/trans isomers, quadrangularin

A and parthenocissin A, and the other is the regioisomer, ampelopsin D that was isolated from grapevine leaves.123,126

Compounds 11 and 13 are cis/trans isomers of cis/trans-ε-viniferin confirmed by their indistinguishable MS/MS spectra. A molecular ion of m/z 453 formed multiple product ions at

435, 411, 369, 359, 347 and 333. This was in contrast to the indane type dimers discussed above, which produced few highly abundant fragments. Losses of H2O (18 Da), C2H2O (42 Da), two

C2H2O (84 Da), phenol (94 Da), 4-methylenecyclohexa-2,5-dienone (106 Da), and C8H8O (120

- Da) were supported by accurate mass fragments of C28H19O5 (measured m/z 435.1237, error 0.23

- - ppm), C26H19O5 (measured m/z 411.1257, error 4.62 ppm), C24H17O4 (measured m/z 369.1126,

- - error 1.63 ppm), C22H15O5 (measured m/z 359.0927, error 0.56 ppm), C21H15O5 (measured m/z 81

- 347.0936, error 3.17 ppm), and C20H13O5 (measured m/z 333.0774, error 1.5 ppm), respectively.

In addition to cis/trans-ε-viniferin, the compounds 12, 14 and 15 also share this exact fragmentation pattern. Two compounds, cis- and trans-ω-viniferin, isolated in grapevine leaves infected with Plasmopara viticola have a near identical structure to ε-viniferin differing only in the stereochemistry of the H7a and H8a chiral centers.123,126 For this reason, the fragmentation pattern of these compounds would be indistinguishable from ε-viniferin and it is therefore proposed that the identity of two of 12, 14, and 15 correspond to cis/trans-ω-viniferin.

Compounds 16 and 17 have an [M-H]- ion at m/z 453 indicating a resveratrol dimer. The

MS/MS spectra for these compounds is similar to that of ε-viniferin with losses of H2O (18 Da),

C2H2O (42 Da), C3O2 (68 Da), two C2H2O (84 Da), and phenol (94 Da) which produce ions at m/z

435, 411, 385, 369, and 359, respectively. Accurate mass measurements support the identity of

- - C28H19O5 (measured m/z 435.1220, error 4.14 ppm), C26H19O5 (measured m/z 411.1222, error

- - 3.89 ppm), C24H17O4 (measured m/z 369.1136, error 1.08 ppm), C22H15O5 (measured m/z

- 359.0905, error 5.57 ppm), and C20H13O5 (measured m/z 333.0741, error 8.41 ppm). Although these fragments are similar to that of ε-viniferin, the ions at m/z 411, and 369 have a much higher relative abundance than in the MS/MS spectrum of ε-viniferin. Through comparison with authentic reference standards, 16 and 17 were identified as trans- and cis-δ-viniferin, respectively.

3.3.3.3 (R+P) Dimers

The MS spectra of compound 18 had a molecular ion at m/z 469 which could correspond to a resveratrol-piceatannol heterodimer (Figure 3.6). The ion at m/z 469 produced four characteristic fragments in the MS/MS spectrum at m/z 451, 385, 375, and 359 which were generated from the losses of H2O (18 Da), two C2H2O (84 Da), phenol (94 Da), and resorcinol

(110 Da) respectively. The identity of these fragments were again supported by accurate mass 82

- - measurements of C28H19O6 (measured m/z 451.1187, error 0.67 ppm), C24H17O5 (measured m/z

- - 385.1082, error 0.36 ppm), C22H15O6 (measured m/z 375.0899, error 6.67 ppm), C22H15O5

(measured 359.0918, error 1.95 ppm). Compounds 19 and 20 provide identical MS/MS spectra including the ions at m/z 451, 385, 375, and 359 ([M-H]- m/z 469). Compounds 19 and 20 differ from 18 as they display a fragment at m/z 241 rather than m/z 265. This could indicate that these compounds are structural isomers depending on which section of the dimer originates from

- piceatannol or resveratrol as the ion at m/z 241 (C14H9O4 , measured m/z 241.0514, error 3.32

- ppm) could correspond to a loss of resveratrol (228 Da), while m/z 265 (C16H9O4 , measured m/z

265.0492, error 5.28 ppm) is related to consecutive losses of phenol and catechol or resorcinol (94

+ 110 Da). Upon comparison with the literature, compounds 19 and 20 are tentatively identified as cis- and trans- scirpusin A, which have been identified in grapevine stems, while compound 18 may be an unreported structural isomer of scirpusin A or perhaps , which also shares this exact mass.138,140

Figure 3.6: Combined BPC Chromatogram of the stilbenoid oligomers 18-41 found in Pinot Noir wine extract.

83

3.3.3.4 Dimers (P+P)

Compound 25 showed an [M-H]- at m/z 485, which indicated a piceatannol dimer (Figure

6). In the MS/MS spectrum of m/z 485, the losses H2O (18 Da), two C2H2O (42 Da), resorcinol

(110 Da), 2-hydroxy-4-methylenecyclohexa-2,5-dienone (122 Da) yielded fragments at m/z 467,

- 401, 375, 363. Supporting accurate mass measurements confirmed these fragments as C28H19O7

- (measured m/z 467.1153, error 4.92 ppm), C24H17O6 (measured m/z 401.1012, error 2.99 ppm),

- - C22H15O6 (measured m/z 375.0895, error 7.19 ppm), and C21H17O6 (measured m/z 363.0878, error 2.75 ppm) respectively. While the consecutive losses of resorcinol and H2O (18 +110 Da),

C2H2O (42 + 110 Da), and another resorcinol (110 + 110 Da) produced the fragments m/z 357,

333, and 241. The fragmentation pattern here is very similar to that of ε-viniferin and except that instead of losses of phenol and 4-methylenecyclohexa-2,5-dienone groups, compound 25 loses resorcinol and 2-hydroxy-4-methylenecyclohexa-2,5-dienone. Because of this, 25 is tentatively identified as a cis- or trans- piceatannol homodimer similar to ε-viniferin and scirpusin A. The identity of 25 is possibly scirpusin B, which has been detected in various plants including passionfruit.141

3.3.3.5 Glycosylated Dimers

The MS/MS spectrum of compound 28 contained a [M-H]- ion at m/z 615 which respresented a resveratrol dimer with O-glycosylation. The loss of glucose (162 Da) formed the

- most abundant ion C28H21O6 (measured m/z 453.1371, error 5.96 ppm), which confirmed the presence of a glycosylated resveratrol dimer. The fragmentatation pattern below m/z 453 appeared identical to that of ε-viniferin with characteristic benzofuran type dimer fragments at m/z 359 and

347 due to loss of phenol (94 Da) and 4-methylenecyclohexa-2,5-dienone (106 Da) from the aglycone ion at m/z 453. Additional fragments were found below 1% relative abundance at m/z

84

435, 411, and 333 from losses of H2O (18 Da), C2H2O (42 Da), and C8H8O (120 Da) respectively.

Because ε-viniferin-O-glycoside has been reported in wine in previous works and the nearly identical fragmentation patterns, compound 28 is tentatively identified as cis- or trans-ε-viniferin-

O-glycoside.124

Compound 29 has an MS/MS spectra again containing the ion corresponding to a glycosylated resveratrol dimer at m/z 615. Similar to the fragmentation of compound 28, the most abundant fragment is generated from the loss of glucose (162 Da) resulting in an ion

- representative of a dimer of resveratrol C28H21O6 (measured m/z 453.1366, error 4.86 ppm). From here, fragment ions at m/z 359 and 289 produced from the loss of phenol (94 Da) and consecutive loss of two phenols (94 + 94 Da) and CO (28 Da) from m/z 453 respectively. Without taking into account the loss of glucose, the fragmentation pattern of compound 29 is comparable to that of compounds 8, 9 and 10 which are indane type dimers. Compound 29 could therefore be a glycosylated derivative of either parthenocissin A, quadrangularin A, or ampelopsin D.

3.3.3.6 Methoxylated Dimers

The [M-H]- ions of compounds 26 and 27 were observed at m/z 485, representative of a pair of potential methoxylated dimers. These compounds produced identical MS/MS spectra, indicating a pair of isomers. The precursor ion generated four main fragment ions at m/z 453, 391,

- 359, and 255. The ion C28H21O6 (measured m/z 453.1338, error 1.32 ppm) corresponding to a resveratrol dimer, had a mass difference from the molecular ion of 32 Da, which could be methanol, confirming the presence of a methoxylated dimer. The characteristic stilbenoid loss of a phenol moiety (94 Da) along with the consecutive loss of methanol and phenol (32 + 94 Da)

- - produced the ions C23H19O6 (measured m/z 391.1182, error 1.28 ppm) and C22H16O5 (measured m/z 359.0914, error 3.06 ppm) respectively. Loss of 4-(methoxymethylene)cyclohexa-2,5-dienone 85

- and phenol (136 + 94 Da) formed the most abundant ion C15H11O4 (measured m/z 255.0655, error

3.14 ppm). This fragmentation pattern indicated that these compounds were a pair of methoxylated indane type resveratrol dimers. In addition to the main fragments, ions at m/z 237,

227, 213, 211, and 187 were found below 1% abundance. The fragmentation pattern of compounds 26 and 27 are identical to a methoxylated dimer found in Parthenocissus laetevirens that was tentatively identified as parthenostilbenin A or B.137 Because of this, it is possible that compounds 26 and 27 are the cis/trans methoyxlated dimers known as parthenostilbenin A and B.

3.3.3.7 Oxidized Dimers

Five compounds numbered 21-24 were detected when a product ion scan of m/z 471.1449 was performed. Compound 21 produced four main fragments other than the [M-H]- ion at m/z

- 471. The ion at m/z 377 (C22H18O6 , measured m/z 377.1024, error 1.59 ppm) was related to the characteristic loss of phenol (94 Da), and consecutive loss of a phenol and a carbon monoxide

- (122 Da) formed the ion at m/z 349 (C21H18O5 , measured m/z 349.1089, error 2.29 ppm). Neutral loss of two phenols (188 Da) and a carbon monoxide (28 Da) resulted in the most abundant ion at

- m/z 255 (C15H12O4 , measured m/z 255.0667, error 1.84 ppm). Further loss of C8H6O2 (134 Da)

- from m/z 255 produced another highly abundant ion at m/z 121 (C7H6O2 , measured m/z

121.0290, error 3.30 ppm). The detection of a oxidized resvertrol dimer with this fragmentation pattern has been reported in a number of studies to date.47,139,142 These previous findings have tentatively identified compound 21 as one of the isomers, restrisol A or B.

The three compounds 22, 23, and 24 showed a comparable fragmentation pattern with an

[M-H]- ion of m/z 471. The difference between these three compounds and compound 19 is the absence of the [M-phenol]- ion at m/z 377 and a greater abundance of the ion at m/z 349 resultant from the loss of a phenol-CO (122 Da) group. As with compound 19, additional loss of C8H8O2 86

(134 Da) produces a highly abundant ion at m/z 121. Although compounds that fit this description have been found in the literature, they have not been characterized.125 Additional studies need to be performed about the oxidation products of resveratrol and its dimers including NMR studies.

3.3.3.8 Trimers

A total of eight peaks were obtained when m/z 679 was selected, which corresponded to trimers of resveratrol. Limited information is available in the literature about these compounds, and even less about their fragmentation behaviour is known. These potential trimers exhibit two noticeably different fragmentation patterns. Compound 30 had a ion at m/z 679 which produced three main fragments at m/z 585, 491, and 385. These fragments were generated from the characteristic stilbene losses of phenol (94 Da), two phenols (94 + 94 Da), and consecutive loss of two phenol (94 + 94 Da) and one 4-methylenecyclohexan-2,5-dienone (106 Da) respectively.

Minor fragments at m/z 573 and 479 which originates from the initial loss of 4- methylenecyclohexan-2,5-dienone (106 Da) follow by a subsequent loss of phenol (94 Da).

Meanwhile, compounds 31-37 show very similar fragmentation patterns that cannot be distinguished, indicative of a collection of trimeric isomers. Compound 31 will serve as a model to explain the fragmentation behaviour common to these seven peaks (Figure 3.6). The [M-H]- ion at m/z 679 indicated a resveratrol trimer. The MS/MS spectra of 31 generated many fragments at m/z 661, 637, 585, 573, 555, and 479, which were a result of the neutral losses of H2O (18 Da),

C2H2O (42 Da), phenol (94 Da), and consecutive loss of 4-methylenecyclohexan-2,5-dienone

(106 Da) with H2O (18 +106 Da) and phenol (106 + 94 Da) respectively. Supported by accurate

- mass, these ions were identified as C42H30O8 (measured m/z 661.1864, error 0.60 ppm),

- - C40H30O8 (measured m/z 637.1926, error 9.10 ppm), C36H26O8 (measured m/z 585.1572, error

- - 2.91 ppm), C35H26O8 (measured m/z 573.1561, error 1.05 ppm), C35H24O7 (measured m/z 87

- 555.1455, error 1.08 ppm), C29H20O7 (measured m/z 479.1169, error 6.89 ppm) respectively.

Further fragmentation is produced when CO (28 Da), CO-phenol (122 Da) and CO- 4- methylenecyclohexan-2,5-dienone (106 + 28 Da) is lost from the ion at m/z 479 to form m/z 451,

357, and 345. The proposed fragmentation pathway is depicted in scheme 3. This fragmentation pattern is characteristic of miyabenol C, which has been observed in parthenocissus laetevirens with authentic reference standard and fragmentation information available.137 Other studies have identified multiple isomers in Vitis vinifera.125 However, it was only very recently that a trimeric stilbene, ampelopsin C, was identified in wine by LC-NMR. 123

3.3.3.9 Tetramers

The two compounds 39 and 41 were determined to be resveratrol tetramers by the [M-H]- mass at m/z 905. Based on very similar fragmentation patterns, it is proposed that these compounds are isomers of each other. The MS/MS spectra of the ion at m/z 905 produced six characteristic stilbenoid fragments of m/z 811, 717, 611, 451, 359, and 265, supported by accurate

- - mass ions at C50H36O11 (measured m/z 811.2193, error 0.99 ppm), C44H30O10 (measured m/z

- - 717.1720, error 6.41 ppm), C37H24O9 (measured m/z 611.1322, error 4.09 ppm), C28H20O6

- (measured m/z 451.1171, error 3.77 ppm), C22H16O5 (measured m/z 359.0915, error 3.62),

- C16H10O4 (measured m/z 265.0494, error 6.79 ppm) respectively. The first three fragments originate from the consecutive losses of one (94 Da) and two (94+94 Da) phenol groups followed by the loss of 4-methylenecyclohexan-2,5-dienone (106 Da). Alternatively, minor fragments at m/z 799 and 705 indicate that the loss of one 4-methylenecyclohexan-2,5-dienone (106 Da) by itself as well as with a phenol (94 Da) is also occurring. The ion at m/z 451 and its minor counterpart at m/z 453 are the result of the loss of a dimer (454 Da) and dehydrodimer (452 Da) respectively. This suggests a symmetrical tetramer splitting into two dimers. Neutral loss from the

88

divided tetramer occurs as additional losses of one (94 Da) and two (94 + 94 Da) phenols produce the observed fragmentation at m/z 359 and 265. Although information about tetrameric stilbenes is scarce, the most well-known tetramer of resveratrol is hopeaphenol, which has been found in multiple plants and wine.127 A known isomer of hopeaphenol, isohopeaphenol has also been detected in many plants.126 It is therefore proposed that compounds 39 and 41 are isomers of hopeaphenol.

Compounds 38 and 40 produced similar MS/MS spectra to the other tetramers. An [M-H]- at m/z 905 followed by the major fragments at m/z 811, 717 and 611 resulting from the consecutive loss of one (94 Da) and two (94+94 Da) phenols and 4-methylenephenol (106 Da).

Minor fragments at m/z 887, 799, 793, and 705 corresponded to the loss of H2O (18 Da), 4- methylenecyclohexan-2,5-dienone (106 Da), phenol and subsequent loss of H2O (94 + 18 Da) and

4-methylenecyclohexan-2,5-dienone (94 + 106 Da). Although this compound shares much of the same fragmentation behaviour as the other three tetramers, no ions above 1% abundance were observed below m/z 611. These compounds may be non-symmetrical tetramers of resveratrol.

There were a few compounds that were surprisingly not detected in the present study. The di- and tri-glucosides of resveratrol were reported for the first time in Cabernet Sauvignon cell cultures.97 It had also been reported in preliminary studies by Dr. Dennis Taylor’s research team at the University of Adelaide that these compounds had been unofficially detected using authentic standards coupled with HPLC-DAD detection without extraction or pre-concentration. When we pursued this assertion with standards, m/z 551.1700 and 713.2298 were scanned and peaks containing the expected neutral loss of 162 Da were not found in any of the wines. It was suspected that these compounds were present, but in lower concentrations than the LOD. This led to the pursuit of extraction methods which resulted in the identification of these 41 stilbenoids, 89

none of which were the original target multi-glycosylated analogues of resveratrol. In addition to the monomer multi-glycosides, it was surprising to not find the resveratrol dimer diglucosides that had previously been reported in a white wine extraction study.124 Nevertheless, many compounds were identified for the first time in wine, which is a very exciting revelation that was brought about by the ever increasing sensitivity of analytical instruments.

3.3.4 Quantitative Method Validation

3.3.4.1 SPE Validation

Generally, sample preparations of wines are required prior to HPLC and MS analysis, such as filtration, extraction or concentration. The advantages and disadvantages of different extraction methods for stilbenoids in wine have been reviewed.81,88 The motivation for using SPE over liquid/liquid extraction (LLE) for the quantitative analysis was multi-faceted. The biggest appeal of SPE was the ability to perform over 20 extractions at a time using a vacuum manifold whereas using the LLE method, only 6 could feasibly be carried out at any given time. Although the recovery of the LLE extractions was not determined in our experiments, the literature supports the notion that SPE allows for superior recovery (usually above 90%) and precision of repeated extractions.81 Overall, SPE is more conducive to performing large number of analyses that require high repeatability, and recovery.

Most SPE studies involving the extraction of stilbenoids have been performed with C8 or

C18 sorbent.88 In the present study, we chose the Agilent BondElut ENV which uses a solid phase blend of polystyrene and divinyl-benzene with 125 µm packed particles. While this sorbent is intended for polar organic compounds, it showed a very good affinity for all of the stilbenoids of interest in this study.88

90

Experiments on the capacity of the sorbent bed for wine were carried out. It was determined that

25 mL of wine could be passed through the SPE cartridge without any analyte detected in the run- through. When 30 mL of wine was passed through, small amounts of resveratrol and piceid were detected in the run-through. Consequently, the volume of wine determined to be the ideal quantity for extraction was 15 mL. This ensured that in the chance that a specific wine had a larger amount of retained compounds than the wines tested, it would still have excellent extraction efficiency.

The extraction efficiency is a measure of the SPE protocol’s ability to retain desired compounds from the sample in the solid phase. The efficiency of our protocol was determined by preparing six trans-piceid solutions; 3 with no analyte, and 3 at low, medium and high concentrations in triplicate. These solutions were then extracted, evaporated, and reconstituted in internal standard solution. The second set of solutions containing no analyte was then spiked with the theoretical 100% yield concentrations of trans-piceid. The solutions were then analyzed by

UHPLC-ESI-QTOF and the two sets were compared to give the results shown in Table 3.5.

Table 3.5: SPE extraction efficiency results for low, medium, and high concentrations

Spiked [Conc] (mg/L)a Theoretical Mean Extraction Standard [Conc] (mg/L)b Efficiency (%)c Deviationd 0.001 0.015 97.07 5.86 0.05 0.75 96.45 1.43 1.00 15.00 98.23 0.94 a – Pre extraction spike concentrations b –Post extraction spike concentrations (theoretical 100%) c – calculated by (measured EIC / 100% EIC * 100%) d – Standard deviation of the efficiency (%)

91

The mean extraction efficiency was >90% for low, medium and high with values of

97.07%, 96.45%, and 98.23% respectively. These values are quite comparable with other studies of SPE extraction yield of stilbenes, where yields of 91-94% and 99-102% were observed respectively.111,143

3.3.4.2 Calibration Curve, Linearity, Limits of Detection and Quantification

A calibration curve was constructed using 14 concentrations ranging from 0.01 – 30 mg/L of piceid. This larger range was chosen as the prospective wine samples were pre-concentrated by

15x using SPE and it was important for all oligomers found to fall within the calibration range.

Figure 3.7 displays the calibration curve of trans-piceid that was constructed and analyzed using the long separation method described.

Figure 3.7: trans-piceid calibration curve ranging from 0.01 - 30 mg/L

92

The linearity of curve was evaluated by visual inspection as well as examination of the residual plot. The coefficient R2 was 0.999, while not a direct indicator of linearity, it can be a helpful to show how well the instrument response’s variance can be explained by the concentration of standards. Inspection of the residual plot showed that the response was adequately distributed around zero which is can be good measure of linearity and that the p value of the F statistic was <0.05 indicating that the regression is successful at approximating the real data.

Table 3.6: Calibration curve validation results, including Limits of detection and quantification.

Compound Regression R2 Range (mg/L) LOD (mg/L) LOQ (mg/L) trans-piceid y = 0.6751x + 0.0022 0.999 0.01-30 0.003 0.009

The limit of detection was defined as the lowest amount of analyte that could be detected but not accurately known, it was determined by replicate blank injections and the noise at the retention time to trans-piceid was integrated and the standard deviation of the noise was determined. The standard deviation of the noise was then multiplied by three and dividing by the slope of the calibration curve, giving a value of 0.003 mg/L (Table 3.6). The limit of quantification is the lowest amount of analyte that can accurately be detected and identified.108,109

The LOQ was calculated similarly to the LOD by multiplying the standard deviation of the noise by 10, and dividing by the slope of the calibration curve, giving a value of 0.009 mg/L. These are the limits of the calibration curve, but as each wine is concentrated by 15x, the lowest detectable amount of stilbenoids within a wine sample would be 0.2 ng/mL.

3.3.4.3 Recovery, Intra- & Inter-day Precision

The recovery is a measure of how accurately the entire method can predict the concentration of the analytes. The method’s recovery was performed by spiking red wine samples 93

with low, medium and high concentrations of trans-piceid in triplicate. The wines were then extracted, evaporated and reconstituted in internal standard solution. The samples were analyzed by UHPLC-ESI-QTOF and the concentrations were then calculated by using the calibration curve.

Subtraction of the un-spiked red wine from the spiked wine samples and then dividing by the spiked concentration (taking into account the 15x pre-concentration through SPE) gave a recovery in percent which is displayed in Table 3.7. The recovery was evaluated on each of the two days that were also used for the precision measurements. The recovery was above 90% for all of the concentration levels, which demonstrates that the method can be used to give accurate results within the extracted wine samples. These results are considered acceptable when compared with similar SPE-HPLC-MS stilbene quantification studies where the results range from 75-

99%.89,111,144

Table 3.7: Recovery results and for multiple days used to calculate both intra- and inter-day precision.

Intra-day Precision Inter-day Precision Spiked [Conc] Measured [Conc] Recovery Measured [Conc] Recovery %RSD %RSD (mg/L) Mean (mg/L) (%) Mean (mg/L) (%) 0.015 0.0142 10.25 94.82 0.0141 12.37 93.97 0.75 0.741 5.45 98.82 0.735 5.54 98.02 15.0 14.54 4.30 96.91 14.73 8.72 98.19

The precision of the method is defined as how similar reported values are from repeated measurements, without altering any other method parameters. Because ionization within the mass spectrometer source can change drastically over time, due to instrument cleanliness, constant recalibration, and unexpected faults within the system, the use of an internal standard is quite necessary when using mass spectrometry to quantify. The precision was evaluated both by performing the recovery analysis in triplicate, twice, on the same day and again two weeks later to give both the intra- and inter-day precision (Table 3.7). The intra-day precision was found to range from 4.30-10.25 %RSD which was below the 15 %RSD required for bio-analytical 94

methods.108,109 The inter-day precision values were expectedly higher than the intra-day precision values but still remained under 15 %RSD. These values show that the method was quite precise, largely due to the use of an internal standard.

3.3.5 Quantification Results

The fully validated semi quantification method, reporting concentrations of stilbenoid oligomers as trans-piceid equivalents was developed. In order to fully illustrate the usefulness of such a method, red wine samples were pre-concentrated and analyzed by UHPLC-ESI-QTOF. All compounds were quantified using peak area of the respective [M-H]- ions in the single stage EIC-

MS chromatograms. To confirm that the identities of the peaks were the same as detected in section 2.3.4, two MS/MS fragment ions for each compound were used as qualifying ions. If they were not present, the ion was assumed to derive from something other than the target analyte.

3.3.5.1 Wine Samples

A total of 39 red wines were sampled to demonstrate the utility of the developed method.

These included 3 of the most popular Okanagan red wine varietals: 4 Pinot Noir, 4 Merlot, and 4

Cabernet Sauvignon for a total of 12 wines. The remaining 27 red wines were hybrid varietals from the Quebec winegrowing regions; this included 6 Maréchal Foch, 6 Marquette, 3 Sabrevois,

6 St. Croix, and 6 Frontenac. Using multiple varieties of red wine for quantification studies allowed for the method to be tested with a wide range of concentrations and the ability to explore differences in stilbenoid concentrations between Vitis vinifera and disease-resistant hybrid varieties.

95

3.3.5.2 Total Stilbenoid Levels in Okanagan and Quebec Wines

The quantification results for each individual wine are displayed in the Appendix (Table A

1), while a condensed table presenting the mean concentrations of each stilbenoid within each red wine varietal is displayed in Table 3.8. All 41 compounds that were identified in the qualitative

MS/MS analysis were also detected in the majority of the red wines tested. This was not entirely unexpected as the goal of the SPE pre-concentration step was to sufficiently concentrate samples so that compounds could be detected in the low ng/mL range.

96

Table 3.8: Condensed quantification results for each stilbenoid compound in each red wine variety

Mean Concentration (ng/mL) Okanagan Wines Quebec Wines Retention Cabernet Pinot Maréchal Compound Time Merlot Marquette Sabrevois St. Croix Frontenac Sauvignon Noir Foch (min) trans-resveratrol 22.77 257.27 ± 234.69 600.49 ± 311.5 173.2 ± 146.49 98.12 ± 32.74 159.51 ± 86.3 39.16 ± 25.38 57.09 ± 38.03 6.64 ± 5.64 cis-resveratrol 25.20 372.97 ± 369.8 893.97 ± 570.41 200.92 ± 155.57 367.49 ± 218.79 934.84 ± 449.31 179.29 ± 100.55 188.26 ± 204.33 24.26 ± 20.37 trans-piceid 13.05 661.17 ± 146.08 403.95 ± 457.17 413.73 ± 193.37 49.53 ± 35.64 107.76 ± 156.43 5.59 ± 3.36 10.71 ± 11.25 11.49 ± 16.47 cis-piceid 15.90 1100.93 ± 658.72 633.16 ± 759.66 676.66 ± 366.58 703.21 ± 333.67 255.47 ± 243.59 62.32 ± 35.53 198.58 ± 189.68 67.86 ± 56.89 trans-piceatannol 16.43 739.87 ± 684.23 532.92 ± 496.41 330.35 ± 286.09 592.75 ± 292.34 494.47 ± 246.36 67.62 ± 39.25 145.95 ± 107.92 111.04 ± 60.2 cis-piceatannol 21.70 21.85 ± 18.72 62.52 ± 57.31 12.38 ± 9.85 39.69 ± 16.04 157.19 ± 83.57 10.88 ± 6.66 25.32 ± 39.01 13.67 ± 7.92 trans-astringin 9.72 344.02 ± 245.07 206.12 ± 248.12 181.11 ± 111.81 109.07 ± 85.82 64.54 ± 6.99 3.14 ± 3.36 4.67 ± 2.99 35 ± 10.22 cis-astringin 12.48 163.16 ± 114.97 60.89 ± 73.67 83.75 ± 48.68 138.08 ± 90.01 76.48 ± 20.27 5.1 ± 5.43 23.14 ± 22.29 73.05 ± 26.57 Total Monomers 3661.25 ± 2275.2 3394.02 ± 1502.53 2072.09 ± 1054.37 2097.93 ± 745.06 2250.26 ± 1025.03 373.09 ± 209.57 653.74 ± 408.42 343.02 ± 169.57 Pallidol 23.04 327.41 ± 83.26 793.08 ± 153.26 133.94 ± 94.8 748.95 ± 331.32 179.05 ± 72.4 41.44 ± 26.43 42.06 ± 29.08 43.72 ± 6.04 parthenocissin 25.63 93.96 ± 23.76 306.13 ± 153.34 42.36 ± 31.54 132.9 ± 79.84 50.33 ± 21.04 5.41 ± 3.64 9.61 ± 6.62 5.25 ± 4.08 QuadrangularinA* 27.42 9.38 ± 2.86 23.63 ± 10.23 2.95 ± 3.69 20.8 ± 7.11 11.09 ± 9.31 0.35 ± 0.4 1.03 ± 2.31 trace AmpelopsinA* D* 28.48 183.45 ± 256.49 312.63 ± 205.32 49.18 ± 37.43 76.35 ± 26.31 50.28 ± 21.67 9.54 ± 6.56 3.62 ± 3.6 1.57 ± 1.59 cis-ε-viniferin 33.45 84.12 ± 49.14 165.31 ± 83.25 37.84 ± 30.52 132.91 ± 56.62 158.25 ± 79.53 15.63 ± 11.04 20.56 ± 11.65 5.14 ± 4.59 cis-ω-viniferin 34.39 17.3 ± 12.76 14.67 ± 7.2 8.78 ± 6.35 104.92 ± 63.41 35 ± 18.68 6.5 ± 4.2 11.29 ± 8.96 4.3 ± 4.79 trans-ε-viniferin 35.67 156.62 ± 104.37 241.23 ± 104.36 341.99 ± 450.08 282.19 ± 167.59 87.8 ± 44.85 17.34 ± 14.4 42.63 ± 42.28 8.56 ± 9 trans-ω-viniferin 36.62 12.79 ± 8.97 21.88 ± 10.56 7.6 ± 3.95 13.03 ± 5.33 10.03 ± 5.55 0.04 ± 0.05 2.56 ± 2.72 trace Dimer 1* 38.31 44.42 ± 30.35 81.86 ± 63.52 29.92 ± 22.79 75.4 ± 39.8 16.83 ± 8.79 3.53 ± 3 5.85 ± 5.48 1.39 ± 1.99 trans-δ-viniferin 40.74 9.83 ± 7.15 18.79 ± 13.41 11.44 ± 7 9.22 ± 5.64 2.02 ± 2.96

97

While most of these compounds are quantified here for the first time, there is one key limitation of the present method. The results cannot be viewed as accurate concentrations of each analyte, as many compounds (even very similarly structured) can have vast differences in ionization efficiency in an atmospheric setting (as with ESI). This is quite apparent in Chapter 2.0, where even compounds as similar as trans- and cis-resveratrol have very different calibration curve slopes. The difference in ionization between a monomer such as piceid, and a dimer, trimer or tetramer is most likely even greater than that of trans- and cis-resveratrol. Despite all of this, expressing the concentration of all analytes as trans-piceid equivalents was a necessary compromise. The vast majority of these compounds are not commercially available, and it is a laborious process to isolate or synthesize even a few of these compounds in sufficient quantities to make accurate standards for quantification. The results should therefore be viewed as approximations of the real concentrations; and comparisons of the values within the present method are undoubtedly scientifically valid, since they are relative contrasts between results calculated with the same equivalence.

98

Figure 3.8: Mean stilbenoid profile displaying mean concentration levels for N=41 red wines analyzed. All compounds above the dotted line have been reported in red wine, while all compounds below the line are first described in wine in this study.

The average stilbenoid profile of the red wines tested is displayed in Figure 3.8. It is quite interesting to note that all previous studies that have quantified stilbenes have detected only those with concentrations greater than ~0.075 mg/L (above the indicated line). This highlights the effectiveness of both the SPE pre-concentration protocol coupled with the high sensitivity that the

UHPLC-ESI-QTOF offers.

99

Table 3.9: Wine quantification results from literature studies on known stilbenoids.

Concentration Compound References Range (mg/L) 84,86,111,113,115,119,127,145-148 trans-resveratrol 0-36.1 22,84,86,113,115,119,146,148,149 cis-resveratrol 0-23.2 22,84,86,113,114,119,145,150 trans-piceid 0-45 22,84,86,113,114,119,145 cis-piceid 0-38.5 84 trans-piceatannol 0-5.2 84,86,113,146 trans-astringin 0-38.1 84 cis-astringin 0-1.6 146,151-153 pallidol 0-7.0 152-154 trans-ε-viniferin 0-4.3 155 cis-ε-viniferin 0-1.1 154 trans-δ-viniferin 0 – 22.4 145,152 Hopeaphenol 0 – 2.7

The wines that contained the highest relative mean concentration of stilbenoid compounds were in order: Pinot noir (7.36 mg/L), Maréchal Foch (5.75 mg/L), Cabernet Sauvignon (5.56 mg/L) and Merlot (3.21 mg/L). While the three varieties containing the lowest overall stilbenoid content were Sabrevois (0.624 mg/L), Frontenac (0.696 mg/L) and St. Croix (0.973 mg/L).

Significant differences in the total stilbenoid concentration were first confirmed with a single- factor ANOVA resulting in a p-value less than 0.05. Multiple comparison tests were performed with a Bonferroni correction for 28 comparisons (α=0.0018). Frontenac was found to have significantly lower total stilbenoid concentration than Cabernet Sauvignon, Pinot Noir, Maréchal

100

Foch, and Marquette while Pinot Noir was found to have a significantly higher total stilbenoid concentration than St. Croix.

The breakdown of the monomers detected is presented in Figure 3.8. The highest mean total monomer concentration was found in Cabernet Sauvignon wines at 3.66 ± 2.28 mg/L, cis- piceid was the main contributor to the monomer concentration at 1.10 ± 0.64 mg/L. Previous studies reporting monomer concentration in wines support this finding as cis-piceid is often the predominant stilbene present in red wine.84,146 Unsurprisingly, the mean monomer concentrations followed the same trend as total stilbenoid content, with Sabrevois, Frontenac, and St. Croix wines containing very low average amounts (<0.7 mg/L) of all monomers. To determine the significance of the differences in monomer concentrations between varieties, a single-factor

ANOVA was performed. Frontenac (0.34 ± 0.17 mg/L) was determined to have a significantly lower overall monomer concentration than Merlot, Pinot Noir, Maréchal Foch, and Marquette at

95% confidence (p < 0.0018).

Figure 3.9: Average breakdown of monomer stilbenoids in each red wine variety analyzed (±SD).

101

Eleven dimers of resveratrol were quantified, and they were observed in the majority of red wines analyzed. The total average concentration of dimers was 0.726 mg/L across all wines, and the variety that contained the greatest amount of total dimers was Pinot Noir at 1.99 mg/L and

Maréchal Foch at 1.60 mg/L (Figure 3.10). Similar to the monomer concentrations, Sabrevois, St.

Croix and Frontenac all contained less than 0.10 mg/L total dimers. Individually, the highest concentration dimer was pallidol, the symmetrical resveratrol dimer, with a range of 0.011 –

1.158 mg/L and an average concentration of 0.279 ± 0.322 mg/L across all wines. This is quite a bit lower than previous quantification studies which determined that pallidol had been quantified as high as 7.0 mg/L (Table 3.9). The next most abundant dimer was trans-ε-viniferin, which has also been reported with relatively high concentration in the literature, rivaling the monomers in some cases (Table 3.9). However, in the present study, it was found in levels between 0 – 1.01 mg/L with an average concentration of 0.137 ± 0.187 mg/L. The rest of the resveratrol dimers, trans-/cis- ω- and δ-viniferin, parthenocissen A, quadrangularin A, and ampelopsin D had relatively low average concentrations ranging from 0 – 0.076 mg/L and none exceeded 0.563 mg/L in any one wine.

Similarly to the monomer concentrations, Frontenac (0.217± 0.105 mg/L) was found to have a significantly lower concentration of dimers than Cabernet Sauvignon, Merlot, Maréchal

Foch, and Marquette. Contrarily, Pinot Noir was found to have a significantly higher concentration of dimers than Frontenac, Marquette, and St. Croix.

102

Figure 3.10: Average breakdown of resveratrol dimer stilbenoids in each red wine variety analyzed (±SD).

The so called “modified dimers” were an assortment of oxidized, hydroxylated, glycosylated, and hetero- dimers. It is therefore nearly meaningless to report the average overall modified dimer concentration, as they are essentially categorized arbitrarily into a miscellaneous group. The highest average of these modified dimers was the compound identified as restrisol A, an oxidized dimer with an average concentration in all wines of 0.122 ± 0.160 mg/L. The rest of these compounds were found to be very minor contributors to the overall stilbenoid count with none averaged above 0.05 mg/L (Figure 3.11).

103

Figure 3.11: Average breakdown of “modified’ dimer stilbenoids in each red wine variety analyzed (±SD).

Some resveratrol trimers have been identified previously in red wines; however, this is the first study to attempt to semi-quantify them in wine (Figure 3.12). The average amount of total trimers found in the red wines tested ranged between trace-0.485 mg/L with a median value of

0.040 mg/L. Two trimers that have been identified in red wine samples previously, were ampelopsin C, and Z-miyabenol C.123 These two trimers were found to be the most abundant with concentration ranges of 0 – 0.075 mg/L and 0 – 0.256 mg/L respectively. On average, Z- miyabenol C and ampelopsin C were found to have concentrations of 0.015 ± 0.015 mg/L and

0.032 ± 0.060 mg/L respectively. Marquette was found to have a significantly higher total trimer concentration than Sabervois, St. Croix, and Frontenac varieties.

Considering that these compounds are present in such low concentrations, it isn’t surprising that they have until now, been detected within wine samples. To put this into perspective, analysis of these two trimers by HPLC-NMR took 3.0 L of red wine to positively

104

identify the structures.123 In contrast, if these trimers were available, the present study has the ability to quantify with 15.0 mL of wine.

Figure 3.12: Average breakdown of trimer stilbenoids in each red wine variety analyzed (±SD).

Resveratrol tetramers were the highest order oligomers that have been detected in wine samples.127 Surprisingly, the concentration of the tetramer hopeaphenol is one of the most abundant stilbenoids that were quantified in this study (Figure 3.13). However, this could simply be due to a higher ionization efficiency. The average level of hopeaphenol found across all wines was 0.291 ± 0.454 mg/L; though this value is decreased significantly by the low concentrations found in Marquette, Sabrevois, St. Croix, and Frontenac. The range of concentrations that hopeaphenol displayed in the tested wines was 0.150 – 1.912 mg/L with the highest average levels found within the Maréchal Foch hybrid variety (0.987 ± 0.772 mg/L). The remaining three tetramers were found in only Cabernet Sauvignon, Pinot Noir, Merlot, and Maréchal Foch wines; with concentrations that did not exceed 0.143 mg/L at their highest. Due to the conservative nature of the Bonferroni correction, no significant differences were found between tetramers.

105

Figure 3.13: Average breakdown of tetramer stilbenoids in each red wine variety analyzed (±SD).

One of the goals of this experiment was to determine if there was a substantial difference between the stilbenoid concentrations of traditional Vitis vinifera varieties grown in the Okanagan

Valley and the highly resistant hybrid varieties grown in Quebec (Figure 3.14). The stilbenoid monomers found in wines grown in the Okanagan valley had an average concentration of 3.01 ±

1.62 mg/L while the Quebec varieties showed a significantly lower average concentration at 1.20

± 1.05 mg/L (p<0.0083, 6 comparisons, 95% confidence). The dimers, trimers and tetramers did not show a significant difference in concentration between Okanagan and Quebec grown varieties.

Due to the high proportion of monomers compared to other stilbenoids, the total stilbenoid content was found to be significantly higher in Okanagan-grown Vitis vinifera varieties compared to the Quebec-grown hybrid varieties.

106

Figure 3.14 – Average stilbenoid concentrations in red wines, separated by Vitis vinifera or hybrid varieties.

This was an interesting surprise because it was hypothesized that the increased resistance to pests, fungus, and bacteria was thought to be derived, in part, from increased stilbenoid production in hybrid varieties.102 Much of the observed differences could be a result of growing conditions such as soil or climate, or even the winemaking process. The results are odd as

Marechal Foch, unlike the other hybrids, does contain in some cases the highest concentrations of stilbenoid compounds, while the other four hybrids have nearly non-existent levels. For

Marquette, St. Croix, Frontenac, and Sabrevois wines, perhaps the resistance to these diseases is the result of another defence mechanism, rendering the use of stilbenoid phytoalexins by the plant, unnecessary and therefore under-expressed. It is also possible that because these varieties of grapes are bred for their extreme cold hardiness, they may not be grown under conditions that are suitable for fungal growth. Because fungal growth is a main elicitor of stilbenoid production, these growing conditions may be responsible for the observed differences in these wine varieties.

107

The present study used MS/MS analysis to detect and partially identify multiple monomer and oligomer stilbenoids in extracted red wine samples. This was facilitated by accurate mass

ESI-QTOF which provided sufficient resolution to identify both precursor and product ions that were characteristic of stilbenoids known in the literature and some potentially new compounds. A total of 41 stilbenoid compounds were detected, 6 of which were monomers, up to as many as 23 dimers including their many derivatives (glycosylated, methyoxylated and oxidized), as well as resveratrol and piceatannol homo- and heterodimers, 8 trimers and 4 tetramers. Some of these compounds have previously been identified in wine; however, in this study 23 were detected for the first time in red wine.

Additionally, all 41 of the detected stilbenoid compounds were quantified in 8 red wine varieties using trans-piceid equivalents. The goal in this semi-quantification was to better understand the quantities that are expressed in different varietals of wine. Pinot Noir and

Maréchal Foch showed a substantially higher content of oligomeric derivatives of resveratrol. The data demonstrates a significant stilbene concentration difference between the Vitis vinifera

Okanagan-grown and the hybrid varietal Quebec-grown wines.

108

CHAPTER 4 – CONCLUSIONS

4.1 – Summary of Research

In this study, two UHPLC-QTOF methods were developed, validated, and applied to red and white wine samples from the Okanagan valley. The first method was a rapid quantification method for the accurate and precise determination of the health promoting compound resveratrol, along with its monomeric derivatives in red and white wine. The second was a long gradient run used to screen and semi-quantify possible oligomeric derivatives of resveratrol. After the optimization of each of these methods, they were applied to red and white wines from around the

Okanagan valley and hybrid varieties from Quebec.

The short method was fully optimized for separation, sensitivity, accuracy, and repeatability, with respect to the UHPLC and ESI-QTOF parameters. This rapid method had a total run time of 6.75 minutes with a 2 minute post-run equilibration time, and was successful at separating all six stilbene monomers. While “Canadian” wines were reported to have significantly higher trans-resveratrol content than other growing regions with an average of 3.16 ± 1.34 mg/L; the present study found a significantly lower average trans-resveratrol quantity of 1.67 ± 1.04 mg/L.22 The monoglycosides of resveratrol, trans- and cis- piceid were also determined in red wine with average concentrations of 1.75 ± 1.27 mg/L and 2.98 ± 1.65 mg/L respectively. This was also lower than the average quantities in previous studies but did fall within the known ranges.22 This contributes to the idea that wines in this region have stilbenoid production due to the higher number of degree days, and therefore less chance of infection leading to stilbenoid synthesis. Lastly, in Chapter 2, Hypothesis 1 was partially accepted as levels of stilbene monomers were found to be significantly different at 95% confidence between Merlot and Pinot

109

Noir red wines. However, Cabernet Sauvignon was not differentiable using stilbene concentration.

In Chapter 3, thorough screening of multiple red wine samples from both the Okanagan

Valley and Quebec resulted in the rejection of Hypothesis 2, since surprisingly, none of the potential multi-glycosylated derivatives that have been produced in vitro were detected in the present study.97

More recent studies of stilbenes in wine have detected some dimers of resveratrol present in significant quantities.84,86,155 This directed our research at identifying all stilbenoid-class compounds that could be present in red wine. To do this, scanning experiments using MS/MS spectrometry with accurate mass fragmentation analyses were performed on concentrated wine samples, in search of stilbenoid compounds potentially present. In total, 41 stilbenoid compounds were tentatively identified through accurate mass measurements, retention time with standards, and comparison with literature spectra. Of the 41 compounds, 6 were monomers, 23 were dimers or derivatives thereof, 8 were trimers and 4 were tetramers, many of which had never been detected in wine. To our knowledge, 23 of these oligomeric stilbenoids in this study have been identified for the first time in red wine. In light of these discoveries, Hypothesis 3 was accepted as multiple new derivatives of resveratrol were identified in Canadian wines.

Three Okanagan Valley-produced red wine varieties were compared with five Quebec- produced hybrid varieties in stilbenoid concentration. This served to test our newly developed oligomer quantification method, and compare the two differently derived red wine categories

(Vitis vinifera vs. hybrid). Hypothesis 4 was rejected because the expected trend that hybrid varieties would contain a higher content of phytoalexin stilbenoids was not observed. In fact, the

110

opposite trend was seen. The Okanagan Vitis vinifera-derived wines contained a significantly higher content of monomer and total stilbenoids.

4.2 – Research Novelty

This work has provided accurate stilbene quantification data for the red and white wines produced in the Okanagan Valley. These data can be used to compare with other winegrowing regions of the world, adding to the already extensive collection of studies investigating the stilbene content of wine.22,88,111In addition, this study also provides a model with settings, gradient information and relative retention times for other high-throughput analyses of monomeric stilbenes.

The 23 oligomeric stilbenoids that were newly identified in wine offer an increased pool of compounds that can be potentially contributing to the health properties purported as a result of drinking red wine. The chromatographic and mass spectrometric data supplied in this study can aid in the identification and classification of new stilbenoids as well as provide for comparison studies involving the separation of these compounds in wine. The fully functional and validated

SPE quantification method was, to our knowledge, the first to concentrate red wine adequately enough to quantify these minor oligomeric compounds. The SPE method also compared two categories of red wine varieties Vitis vinifera derived wines and the hardier lesser studied hybrid varieties of North America.

4.3 – Assumptions and Limitations

The main limitation of the developed oligomer quantification method is that all compounds were expressed as equivalents of trans-piceid. A major assumption in quantifying in this manner is that all compounds have the same instrument response at comparable

111

concentrations, not taking into account the substantially different ionization efficiencies of individual compounds. This however, was a necessary concession to make because of the unavailability of standards for most of these compounds, and the difficulty in obtaining them through extraction or synthesis in purified form. Consequently, the quantification results presented in Chapter 3 cannot be considered true concentrations of these oligomeric compounds, but rather estimations of their relative abundances.

The comparison data between Okanagan and Quebec grown wines should be viewed as preliminary in nature. The samples (especially the Okanagan wines) should be more numerous in order to get a further representative sample of the region’s wines.

4.4 – Future Directions

The high-throughput monomeric stilbene quantification method is designed to enable the rapid analysis of many samples. Joint experiments with grape growers to monitor the evolution of stilbene compounds within grapes, while varying the growing conditions, could lead to a better understanding of what factors influence the concentration of these compounds in wine. Similarly, the method could be used in experiments observing the differences in stilbenoid abundance created by altering the winemaking process.

To confirm the definite structures of many of the oligomeric stilbenoids, efforts could be taken to extract many of these compounds from other plant species and tissues where the compounds exist in higher quantities. These extracted and purified compounds could positively identify the compounds tentatively characterized in the present studies. In addition to compound confirmation, it would be advantageous to isolate many of these compounds in purified form to analyze for their bioactivities in relation to resveratrol.156

112

The SPE quantification method is intended for pre-concentration and maximum separation between stilbenoids, however, it is certainly possible to be used to for additional polyphenolic compounds that are low in concentration but have yet to be recognized in wine and plant extracts.

A study that evaluates the differences between hybrid varieties and the traditional Vitis vinifera derived wines would be much improved with more samples from each category to make a more balanced study.

113

REFERENCES

1. Frankel, E.; Kanner, J.; German, J.; Parks, E.; Kinsella, J. Inhibition of Oxidation of Human Low-Density-Lipoprotein by Phenolic Substances in Red Wine. Lancet 1993, 341, 454-457.

2. Monagas, M.; Bartolomé, B.; Gómez-Cordovés, C. Updated Knowledge About the Presence of Phenolic Compounds in Wine. Critical Reviews in Food Science & Nutrition 2005, 45, 85- 118.

3. Renaud, S.; Delorgeril, M. Wine, Alcohol, Platelets, and the French Paradox for Coronary Heart-Disease. Lancet 1992, 339, 1523-1526.

4. Ribereau-Gayon, R.; Glories, Y.; Maujean, A.; Dubourdieu, D., Eds.; In Handbook of Enology: The Chemistry of Wine Stabilization and Treatments; 2006; Vol. 2, pp 141-151.

5. Lotito, S. B.; Frei, B. Consumption of flavonoid-rich foods and increased plasma antioxidant capacity in humans: Cause, consequence, or epiphenomenon? Free Radic. Biol. Med. 2006, 41, 1727-1746.

6. Dixon, R.; Xie, D.; Sharma, S. Proanthocyanidins - a final frontier in flavonoid research? New Phytol. 2005, 165, 9-28.

7. Dey, P. M.; Harborne, J. B. Plant Biochemistry; Academic Press: Waltham, Massachusetts, 1997; , pp 390-412.

8. KOUKOL, J.; CONN, E. Metabolism of Aromatic Compounds in Higher Plants .4. Purification and Properties of Phenylalanine Deaminase of Hordeum Vulgare. J. Biol. Chem. 1961, 236, 2692-&.

9. Lu, Y.; Foo, L. The polyphenol constituents of grape pomace. Food Chem. 1999, 65, 1-8.

10. Mazza G; Miniati E Anthocyanins in Fruits, Vegetables,and Grains. CRC Press: Boca Raton, 1993; .

11. Athar, M.; Back, J. H.; Tang, X.; Kim, K. H.; Kopelovich, L.; Bickers, D. R.; Kim, A. L. Resveratrol: A review of preclinical studies for human cancer prevention. Toxicol. Appl. Pharmacol. 2007, 224, 274-283.

12. Baur, J.; Sinclair, D. Therapeutic potential of resveratrol: the in vivo evidence. Nat. Rev. Drug Discov. 2006, 5, 493-506.

13. Riviere, C.; Pawlus, A. D.; Merillon, J. Natural stilbenoids: distribution in the plant kingdom and chemotaxonomic interest in Vitaceae. Nat. Prod. Rep. 2012, 29, 1317-1333. 114

14. Chong, J.; Poutaraud, A.; Hugueney, P. Metabolism and roles of stilbenes in plants. Plant Science 2009, 177, 143-155.

15. Donnez, D.; Jeandet, P.; Clement, C.; Courot, E. Bioproduction of resveratrol and stilbene derivatives by plant cells and microorganisms. Trends Biotechnol. 2009, 27, 706-713.

16. Vastano, B.; Chen, Y.; Zhu, N.; Ho, C.; Zhou, Z.; Rosen, R. Isolation and identification of stilbenes in two varieties of Polygonum cuspidatum. J. Agric. Food Chem. 2000, 48, 253- 256.

17. Chen, H.; Tuck, T.; Ji, X.; Zhou, X.; Kelly, G.; Cuerrier, A.; Zhang, J. Quality Assessment of Japanese Knotweed (Fallopia japonica) Grown on Prince Edward Island as a Source of Resveratrol. J. Agric. Food Chem. 2013, 61, 6383-6392.

18. Manach, C.; Scalbert, A.; Morand, C.; Remesy, C.; Jimenez, L. Polyphenols: food sources and bioavailability. Am. J. Clin. Nutr. 2004, 79, 727-747.

19. Counet, C.; Callemien, D.; Collin, S. Chocolate and cocoa: New sources of trans-resveratrol and trans-piceid. Food Chem. 2006, 98, 649-657.

20. Roupe, K. A.; Remsberg, C. M.; Yanez, J. A.; Davies, N. M. Pharmacometrics of stilbenes: seguing towards the clinic. Current clinical pharmacology 2006, 1, 81-101.

21. Paredes-Lopez, O.; Cervantes-Ceja, M. L.; Vigna-Perez, M.; Hernandez-Perez, T. Berries: Improving Human Health and Healthy Aging, and Promoting Quality Life-A Review. Plant Food Hum. Nutr. 2010, 65, 299-308.

22. Stervbo, U.; Vang, O.; Bonnesen, C. A review of the content of the putative chemopreventive phytoalexin resveratrol in red wine. Food Chem. 2007, 101, 449-457.

23. Jeandet, P.; Bessis, R.; Gautheron, B. The Production of Resveratrol (3,5,4'- Trihydroxystilbene) by Grape Berries in Different Developmental Stages. Am. J. Enol. Vitic. 1991, 42, 41-46.

24. Creasy, L.; Coffee, M. Phytoalexin Production Potential of Grape Berries. J. Am. Soc. Hort. Sci. 1988, 113, 230-234.

25. Langcake, P.; Pryce, R. J. The production of resveratrol and the viniferins by grapevines in response to ultraviolet irradiation. Phytochemistry 1977, 16, 1193-1196.

26. Perrone, G.; Nicoletti, I.; Pascale, M.; De Rossi, A.; De Girolamo, A.; Visconti, A. Positive correlation between high levels of ochratoxin A and resveratrol-related compounds in red wines. J. Agric. Food Chem. 2007, 55, 6807-6812.

27. Soleas, G.; Diamandis, E.; Goldberg, D. Resveratrol: A molecule whose time has come? And gone? Clin. Biochem. 1997, 30, 91-113.

115

28. Pezet, R.; Pont, V.; Hoangvan, K. Evidence for Oxidative Detoxication of Pterostilbene and Resveratrol by a Laccase-Like Stilbene Oxidase Produced by Botrytis-Cinerea. Physiol. Mol. Plant Pathol. 1991, 39, 441-450.

29. Sbaghi, M.; Jeandet, P.; Bessis, R.; Leroux, P. Degradation of stilbene-type phytoalexins in relation to the pathogenicity of Botrytis cinerea to grapevines. Plant Pathol. 1996, 45, 139- 144.

30. Jeandet, P.; Delaunois, B.; Conreux, A.; Donnez, D.; Nuzzo, V.; Cordelier, S.; Clement, C.; Courot, E. Biosynthesis, metabolism, molecular engineering and biological functions of stilbene phytoalexins in plants. Biofactors 2010, 36, 331-341.

31. Austin, M.; Bowman, M.; Ferrer, J.; Schroder, J.; Noel, J. An aldol switch discovered in stilbene synthases mediates cyclization specificity of type III polyketide synthases. Chem. Biol. 2004, 11, 1179-1194.

32. Langcake, P.; Pryce, R. J. The production of resveratrol by Vitis vinifera and other members of the Vitaceae as a response to infection or injury. Physiological Plant Pathology 1976, 9, 77-86.

33. Romero-Perez, A.; Ibern-Gomez, M.; Lamuela-Raventos, R.; de la Torre-Boronat, M. Piceid, the major resveratrol derivative in grape juices. J. Agric. Food Chem. 1999, 47, 1533-1536.

34. Gachon, C.; Langlois-Meurinne, M.; Saindrenan, P. Plant secondary metabolism glycosyltransferases: the emerging functional analysis. Trends Plant Sci. 2005, 10, 542-549.

35. Schmidlin, L.; Poutaraud, A.; Claudel, P.; Mestre, P.; Prado, E.; Santos-Rosa, M.; Wiedemann-Merdinoglu, S.; Karst, F.; Merdinoglu, D.; Hugueney, P. A Stress-Inducible Resveratrol O-Methyltransferase Involved in the Biosynthesis of Pterostilbene in Grapevine. Plant Physiol. 2008, 148, 1630-1639.

36. Pryce, R.; Langcake, P. Alpha-Viniferin - Antifungal Resveratrol Trimer from Grapevines. Phytochemistry 1977, 16, 1452-1454.

37. Langcake, P.; Pryce, R. Oxidative Dimerization of 4-Hydroxystilbenes Invitro - Production of a Grapevine Phytoalexin Mimic. J. Chem. Soc. -Chem. Commun. 1977, 208-210.

38. Jeandet, P.; Douillt-Breuil, A.; Bessis, R.; Debord, S.; Sbaghi, M.; Adrian, M. Phytoalexins from the vitaceae: biosynthesis, phytoalexin gene expression in transgenic plants, antifungal activity, and metabolism. J. Agric. Food Chem. 2002, 50, 2731-2741.

39. Sako, M.; Hosokawa, H.; Ito, T.; Iinuma, M. Regioselective oxidative coupling of 4- hydroxystilbenes: Synthesis of resveratrol and epsilon-viniferin (E)-dehydrodimers. J. Org. Chem. 2004, 69, 2598-2600.

116

40. Shen, T.; Wang, X.; Lou, H. Natural stilbenes: an overview. Nat. Prod. Rep. 2009, 26, 916- 935.

41. Pezet, R. Purification and characterization of a 32-kDa laccase-like stilbene oxidase produced by Botrytis cinerea Pers.: Fr. FEMS Microbiol. Lett. 1998, 167, 203-208.

42. Calderon, A.; Zapata, J.; Barcelo, A. Peroxidase-Mediated Formation of Resveratrol Oxidation-Products during the Hypersensitive-Like Reaction of Grapevine Cells to an Elicitor from Trichoderma-Viride. Physiol. Mol. Plant Pathol. 1994, 44, 289-299.

43. Breuil, A.; Jeandet, P.; Adrian, M.; Chopin, F.; Pirio, N.; Meunier, P.; Bessis, R. Characterization of a pterostilbene dehydrodimer produced by laccase of Botrytis cinerea. Phytopathology 1999, 89, 298-302.

44. Barcelo, A.; Pomar, F.; Lopez-Serrano, M.; Pedreno, M. Peroxidase: a multifunctional enzyme in grapevines. Funct. Plant Biol. 2003, 30, 577-591.

45. Calderon, A.; Zapata, J.; Pedreno, M.; Munoz, R.; Barcelo, A. Levels of 4-Hydroxystilbene- Oxidizing Isoperoxidases Related to Constitutive Disease Resistance in Invitro-Cultured Grapevine. Plant Cell Tissue Organ Cult. 1992, 29, 63-70.

46. Morales, M.; Barcelo, A. A basic peroxidase isoenzyme from vacuoles and cell walls of Vitis vinifera. Phytochemistry 1997, 45, 229-232.

47. Velu, S. S.; Buniyamin, I.; Ching, L. K.; Feroz, F.; Noorbatcha, I.; Gee, L. C.; Awang, K.; Wahab, I. A.; Weber, J. F. Regio- and Stereoselective Biomimetic Synthesis of Dimers from Resveratrol Analogues: Influence of the Solvent, Oxidant, and Substitution. Chem. -Eur. J. 2008, 14, 11376-11384.

48. Hung, L.; Chen, J.; Huang, S.; Lee, R.; Su, M. Cardioprotective effect of resveratrol, a natural antioxidant derived from grapes. Cardiovasc. Res. 2000, 47, 549-555.

49. Jang, M.; Cai, E.; Udeani, G.; Slowing, K.; Thomas, C.; Beecher, C.; Fong, H.; Farnsworth, N.; Kinghorn, A.; Mehta, R.; Moon, R.; Pezzuto, J. Cancer chemopreventive activity of resveratrol, a natural product derived from grapes. Science 1997, 275, 218-220.

50. Fremont, L.; Belguendouz, L.; Delpal, S. Antioxidant activity of resveratrol and alcohol-free wine polyphenols related to LDL oxidation and polyunsaturated fatty acids. Life Sci. 1999, 64, 2511-2521.

51. Bertelli, A.; Giovannini, L.; Giannessi, D.; Migliori, M.; Bernini, W.; Fregoni, M.; Bertelli, A. Antiplatelet Activity of Synthetic and Natural Resveratrol in Red Wine. Int. J. Tissue React. - Exp. Clin. Asp. 1995, 17, 1-3.

52. Paceasciak, C.; Hahn, S.; Diamandis, E.; Soleas, G.; Goldberg, D. The Red Wine Phenolics Trans-Resveratrol and Quercetin Block Human Platelet-Aggregation and Eicosanoid

117

Synthesis - Implications for Protection Against Coronary Heart-Disease. Clin. Chim. Acta 1995, 235, 207-219.

53. Opie, L. H.; Lecour, S. The red wine hypothesis: from concepts to protective signalling molecules. Eur. Heart J. 2007, 28, 1683-1693.

54. Aggarwal, B.; Bhardwaj, A.; Aggarwal, R.; Seeram, N.; Shishodia, S.; Takada, Y. Role of resveratrol in prevention and therapy of cancer: Preclinical and clinical studies. Anticancer Res. 2004, 24, 2783-2840.

55. Saiko, P.; Szakmary, A.; Jaeger, W.; Szekeres, T. Resveratrol and its analogs: Defense against cancer, coronary disease and neurodegenerative maladies or just a fad? Mutat. Res. - Rev. Mut. Res. 2008, 658, 68-94.

56. Baur, J. A.; Pearson, K. J.; Price, N. L.; Jamieson, H. A.; Lerin, C.; Kalra, A.; Prabhu, V. V.; Allard, J. S.; Lopez-Lluch, G.; Lewis, K.; Pistell, P. J.; Poosala, S.; Becker, K. G.; Boss, O.; Gwinn, D.; Wang, M.; Ramaswamy, S.; Fishbein, K. W.; Spencer, R. G.; Lakatta, E. G.; Le Couteur, D.; Shaw, R. J.; Navas, P.; Puigserver, P.; Ingram, D. K.; de Cabo, R.; Sinclair, D. A. Resveratrol improves health and survival of mice on a high-calorie diet. Nature 2006, 444, 337-342.

57. Piotrowska, H.; Kucinska, M.; Murias, M. Biological activity of piceatannol: Leaving the shadow of resveratrol. Mutat. Res. - Rev. Mut. Res. 2012, 750, 60-82.

58. Larrosa, M.; Tomas-Barberan, F.; Espin, J. The grape and wine polyphenol piceatannol is a potent inducer of apoptosis in human SK-Mel-28 melanoma cells. Eur. J. Nutr. 2004, 43, 275-284.

59. Wu, N.; Clausen, A. M. Fundamental and practical aspects of ultrahigh pressure liquid chromatography for fast separations. J. Sep. Sci. 2007, 30, 1167-1182.

60. Burtis, C.; Ashwood, E.; Bruns, D., Eds.; In Tietz Textbook of Clinical Chemistry and Molecular Diagnostics, 5th Edition; 2011; , pp 2256.

61. Vanhoenacker, G.; Sandra, P. Elevated temperature and temperature programming in conventional liquid chromatography - fundamentals and applications. J. Sep. Sci. 2006, 29, 1822-1835.

62. Lough, W. J.; Wainer, I. W., Eds.; In High Performance Liquid Chromatography - Fundamental Principles and Practice; 1996; , pp 276.

63. Dubois, F.; Knochenmuss, R.; Zenobi, R.; Brunelle, A.; Deprun, C.; Le Beyec, Y. A comparison between ion-to-photon and microchannel plate detectors. Rapid Commun. Mass Spectrom. 1999, 13, 786-791.

118

64. Van Berkel, G. J.; Kertesz, V. Using the electrochemistry of the electrospray ion source. Anal. Chem. 2007, 79, 5510-5520.

65. Watson, R., Ed.; In Polyphenols in Plants: Isolation, Purification, and Extraction preparation; 2014; , pp 360.

66. Agilent Technologies Agilent 6200 Series TOF and 6500 Series Q-TOF LC/MS System Concepts Guide. Agilent Technologies 2011, 116.

67. de la Mora, J.; Van Berkel, G.; Enke, C.; Cole, R.; Martinez-Sanchez, M.; Fenn, J. Electrochemical processes in electrospray ionization mass spectrometry - Discussion. J. Mass Spectrom. 2000, 35, 939-952.

68. Wu, X.; Oleschuk, R. D.; Cann, N. M. Characterization of microstructured fibre emitters: in pursuit of improved nano electrospray ionization performance. Analyst 2012, 137, 4150- 4161.

69. Covey, T. R.; Thomson, B. A.; Schneider, B. B. Atmospheric Pressure Ion Sources. Mass Spectrom. Rev. 2009, 28, 870-897.

70. Wang, R.; Zenobi, R. Evolution of the Solvent Polarity in an Electrospray Plume. J. Am. Soc. Mass Spectrom. 2010, 21, 378-385.

71. Cech, N.; Enke, C. Practical implications of some recent studies in electrospray ionization fundamentals. Mass Spectrom. Rev. 2001, 20, 362-387.

72. Nguyen, S.; Fenn, J. B. Gas-phase ions of solute species from charged droplets of solutions. Proc. Natl. Acad. Sci. U. S. A. 2007, 104, 1111-1117.

73. Iribarne, J.; Thomson, B. Evaporation of Small Ions from Charged Droplets. J. Chem. Phys. 1976, 64, 2287-2294.

74. Ahadi, E.; Konermann, L. Ejection of Solvated Ions from Electrosprayed Methanol/Water Nanodroplets Studied by Molecular Dynamics Simulations. J. Am. Chem. Soc. 2011, 133, 9354-9363.

75. Daub, C. D.; Cann, N. M. How Are Completely Desolvated Ions Produced in Electrospray Ionization: Insights from Molecular Dynamics Simulations. Anal. Chem. 2011, 83, 8372- 8376.

76. , C. G.; Johnstone, R. A. W., Eds.; In Mass Spectrometry Basics; 2003; , pp 474.

77. Sleno, L.; Volmer, D. Ion activation methods for tandem mass spectrometry. J. Mass Spectrom. 2004, 39, 1091-1112.

119

78. Ens, W.; Standing, K. Hybrid quadrupole/time-of-flight mass spectrometers for analysis of biomolecules. Biol. Mass Spectrom. 2005, 402, 49-78.

79. Jeandet, P.; Bessis, R.; Maume, B. F.; Sbaghi, M. Analysis of Resveratrol in Burgundy Wines. Journal of Wine Research 1993, 4, 79.

80. Goldberg, D.; Yan, J.; Ng, E.; Diamandis, E.; Karumanchiri, A.; Soleas, G.; Waterhouse, A. A Global Survey of Trans-Resveratrol Concentrations in Commercial Wines. Am. J. Enol. Vitic. 1995, 46, 159-165.

81. Fan, E.; Lin, S.; Du, D.; Jia, Y.; Kang, L.; Zhang, K. Current separative strategies used for resveratrol determination from natural sources. Anal. Methods 2011, 3, 2454-2462.

82. Kolouchova-Hanzlikova, I.; Melzoch, K.; Filip, V.; Smidrkal, J. Rapid method for resveratrol determination by HPLC with electrochemical and UV detections in wines. Food Chem. 2004, 87, 151-158.

83. Vian, M.; Tomao, V.; Gallet, S.; Coulomb, P.; Lacombe, J. Simple and rapid method for cis- and trans-resveratrol and piceid isomers determination in wine by high-performance liquid chromatography using Chromolith columns. J. Chromatogr. A 2005, 1085, 224-229.

84. Buiarelli, F.; Coccioli, F.; Jasionowska, R.; Merolle, M.; Terracciano, A. Analysis of some stilbenes in Italian wines by liquid chromatography/tandem mass spectrometry. Rapid Commun. Mass Spectrom. 2007, 21, 2955-2964.

85. Mark, L.; Nikfardjam, M.; Avar, P.; Ohmacht, R. A validated HPLC method for the quantitative analysis of trans-resveratrol and trans-piceid in Hungarian wines. J. Chromatogr. Sci. 2005, 43, 445-449.

86. Vitrac, X.; Bornet, A.; Vanderlinde, R.; Valls, J.; Richard, T.; Delaunay, J.; Merillon, J.; Teissedre, P. Determination of stilbenes (delta-viniferin, trans-astringin, trans-piceid, cis- and trans-resveratrol, epsilon-viniferin) in Brazilian wines. J. Agric. Food Chem. 2005, 53, 5664- 5669.

87. Rodriguez-Delgado, M.; Gonzalez, G.; Perez-Trujillo, J.; Garcia-Montelongo, F. Trans- resveratrol in wines from the Canary Islands (Spain). Analysis by high performance liquid chromatography. Food Chem. 2002, 76, 371-375.

88. Gocan, S. Analysis of Stilbenes in Wine by HPLC: Recent Approaches. J. Liq. Chromatogr. Rel. Technol. 2009, 32, 1598-1643.

89. Dominguez, C.; Guillen, D.; Barroso, C. Automated solid-phase extraction for sample preparation followed by high-performance liquid chromatography with diode array and mass spectrometric detection for the analysis of resveratrol derivatives in wine. J. Chromatogr. A 2001, 918, 303-310.

120

90. Careri, M.; Corradini, C.; Elviri, L.; Nicoletti, I.; Zagnoni, I. Liquid chromatography- electrospray tandem mass spectrometry of cis-resveratrol and trans-resveratrol: Development, validation, and application of the method to red wine, grape, and winemaking byproducts. J. Agric. Food Chem. 2004, 52, 6868-6874.

91. Buiarelli, F.; Coccioli, F.; Jasionowska, R.; Merolle, M.; Terracciano, A. Chromatographic analysis of trans resveratrol in Italian wines: Comparisons between FL, UV and MS detection. Chromatographia 2006, 64, 475-481.

92. Matuszewski, B.; Constanzer, M.; Chavez-Eng, C. Strategies for the assessment of matrix effect in quantitative bioanalytical methods based on HPLC-MS/MS. Anal. Chem. 2003, 75, 3019-3030.

93. Chin, C.; Zhang, Z.; Karnes, H. A study of matrix effects on an LC/MS/MS assay for olanzapine and desmethyl olanzapine. J. Pharm. Biomed. Anal. 2004, 35, 1149-1167.

94. Wang, S.; Cyronak, M.; Yang, E. Does a stable isotopically labeled internal standard always correct analyte response? A matrix effect study on a LC/MS/MS method for the determination of carvedilol enantiomers in human plasma. J. Pharm. Biomed. Anal. 2007, 43, 701-707.

95. Lo, C.; Le Blanc, J. C. Y.; Yu, C. K. Y.; Sze, K. H.; Ng, D. C. M.; Chu, I. K. Detection, characterization, and quantification of resveratrol glycosides in transgenic arabidopsis a over- expressing a sorghum stilbene synthase gene by liquid chromatography/tandem mass spectrometry. Rapid Commun. Mass Spectrom. 2007, 21, 4101-4108.

96. Decendit, A.; Waffo-Teguo, P.; Richard, T.; Krisa, S.; Vercauteren, J.; Monti, J.; Deffieux, G.; Merillon, J. Galloylated catechins and stilbene diglucosides in Vitis vinifera cell suspension cultures. Phytochemistry 2002, 60, 795-798.

97. Larronde, F.; Richard, T.; Delaunay, J.; Decendit, A.; Monti, E.; Krisa, S.; Merillon, J. New stilbenoid glucosides isolated from Vitis vinifera cell suspension cultures (cv. cabernet sauvignon). Planta Med. 2005, 71, 888-890.

98. Bautista, J.; Dangl, G. S.; Yang, J.; Reisch, B.; Stover, E. Use of genetic markers to assess pedigrees of grape cultivars and breeding program selections. Am. J. Enol. Vitic. 2008, 59, 248-254.

99. Vivier, M. A.; Pretorius, I. S. Genetic Improvement of Grapevine: Tailoring Grape Varieties for The
Third Millennium - A Review. S. Afr. J. Enol. Vitie. 2000, 21, 5-26.

100. Daniela, P.; Federica, G.; Mirella, G.; Diego, T. Performance of interspecific grapevine varieties in north-east Italy. Agricultural Sciences 2013, 4, 91-101.

121

101. Manns, D. C.; Lenerz, C. T. M. C.; Mansfield, A. K. Impact of Processing Parameters on the Phenolic Profile of Wines Produced from Hybrid Red Grapes , , and Marquette. J. Food Sci. 2013, 78, C696-C702.

102. Babikova, P.; Vrchotova, N.; Triska, J.; Kyselakova, M. Content of Trans-resveratrol in Leaves and Berries of Interspecific Grapevine (Vitis sp.) Varieties. Czech. J. Food Sci. 2008, 26, S13-S17.

103. Goldberg, D.; Ng, E.; Karumanchiri, A.; Diamandis, E.; Soleas, G. Resveratrol glucosides are important components of commercial wines. Am. J. Enol. Vitic. 1996, 47, 415-420.

104. Jose Jara-Palacios, M.; Gordillo, B.; Lourdes Gonzalez-Miret, M.; Hernanz, D.; Luisa Escudero-Gilete, M.; Heredia, F. J. Comparative Study of the Enological Potential of Different Winemaking Byproducts: Implications in the Antioxidant Activity and Color Expression of Red Wine Anthocyanins in a Model Solution. J. Agric. Food Chem. 2014, 62, 6975-6983.

105. Wu, Z.; Gao, W.; Phelps, M.; Wu, D.; Miller, D.; Dalton, J. Favorable effects of weak acids on negative-ion electrospray ionization mass spectrometry. Anal. Chem. 2004, 76, 839-847.

106. Urpi-Sarda, M.; Jauregui, O.; Lamuela-Raventos, R.; Jaeger, W.; Miksits, M.; Covas, M.; Andres-Lacueva, C. Uptake of diet resveratrol into the human low-density lipoprotein. identification and quantification of resveratrol metabolites by liquid chromatography coupled with tandem mass spectrometry. Anal. Chem. 2005, 77, 3149-3155.

107. Kruve, A.; Herodes, K.; Leito, I. Optimization of electrospray interface and quadrupole ion trap mass spectrometer parameters in pesticide liquid chromatography/electrospray ionization mass spectrometry analysis. Rapid Commun. Mass Spectrom. 2010, 24, 919-926.

108. International Conference on Harmonization (ICH) Validation of Analytical Procedures: Methodology ICH Q2, http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Quality/Q2_R1/St ep4/Q2_R1__Guideline.pdf. July 10th 2014).

109. US Department of Health and Human Services: Food and Drug Administration FDA: Guidance for Industry, Bioanalytical Method Validation. http://www.fda.gov/downloads/Drugs/Guidances/ucm070107.pdf. July 10th 2014).

110. Stella, L.; De Rosso, M.; Panighel, A.; Vedova, A. D.; Flamini, R.; Traldi, P. Collisionally induced fragmentation of [M-H](-) species of resveratrol and piceatannol investigated by deuterium labelling and accurate mass measurements. Rapid Commun. Mass Spectrom. 2008, 22, 3867-3872.

111. Paulo, L.; Domingues, F.; Queiroz, J. A.; Gallardo, E. Development and Validation of an Analytical Method for the Determination of trans- and cis-Resveratrol in Wine: Analysis of Its Contents in 186 Portuguese Red Wines. J. Agric. Food Chem. 2011, 59, 2157-2168.

122

112. Trela, B.; Waterhouse, A. Resveratrol: Isomeric molar absorptivities and stability. J. Agric. Food Chem. 1996, 44, 1253-1257.

113. Vitrac, X.; Monti, J.; Vercauteren, J.; Deffieux, G.; Merillon, J. Direct liquid chromatographic analysis of resveratrol derivatives and flavanonols in wines with absorbance and fluorescence detection. Anal. Chim. Acta 2002, 458, 103-110.

114. Ribeiro de Lima, M. T.; Waffo-Teguo, P.; Teissedre, P. L.; Pujolas, A.; Vercauteren, J.; Cabanis, J. C.; Merillon, J. M. Determination of stilbenes (trans-astringin, cis- and trans- piceid, and cis- and trans-resveratrol) in Portuguese wines. J. Agric. Food Chem. 1999, 47, 2666-2670.

115. Stark, T.; Wollmann, N.; Loesch, S.; Hofmann, T. Quantitation of Resveratrol in Red Wines by Means of Stable Isotope Dilution Analysis-Ultra-Performance Liquid Chromatography- Quan-Time-of-Flight Mass Spectrometry and Cross Validation. Anal. Chem. 2011, 83, 3398- 3405.

116. Betz, J. M.; Brown, P. N.; Roman, M. C. Accuracy, precision, and reliability of chemical measurements in natural products research. Fitoterapia 2011, 82, 44-52.

117. AOAC International AOAC Guidelines for Single Laboratory Validation of Chemical Methods for Dietary Supplements and Botanicals. http://www.aoac.org/imis15_prod/AOAC_Docs/StandardsDevelopment/SLV_Guidelines_Di etary_Supplements.pdf. 2014).

118. Thompson, M.; Ellison, S.; Wood, R. Harmonized guidelines for single-laboratory validation of methods of analysis - (IUPAC technical report). Pure Appl. Chem. 2002, 74, 835-855.

119. Lee, J.; Rennaker, C. Antioxidant capacity and stilbene contents of wines produced in the Snake River Valley of Idaho. Food Chem. 2007, 105, 195-203.

120. Jeandet, P.; Bessis, R.; Maume, B.; Meunier, P.; Peyron, D.; Trollat, P. Effect of Enological Practices on the Resveratrol Isomer Content of Wine. J. Agric. Food Chem. 1995, 43, 316- 319.

121. Pezet, R.; Cuenat, P. Resveratrol in wine: Extraction from skin during fermentation and post- fermentation standing of must from Gamay grapes. Am. J. Enol. Vitic. 1996, 47, 287-290.

122. Vitrac, X.; Castagnino, C.; Waffo-Teguo, P.; Delaunay, J.; Vercauteren, J.; Monti, J.; Deffieux, G.; Merillon, J. Polyphenols newly extracted in red wine from southwestern France by centrifugal partition chromatography. J. Agric. Food Chem. 2001, 49, 5934-5938.

123. Pawlus, A. D.; Cantos-Villar, E.; Richard, T.; Bisson, J.; Poupard, P.; Papastamoulis, Y.; Monti, J.; Teissedre, P.; Waffo-Teguo, P.; Merillon, J. Chemical dereplication of wine stilbenoids using high performance liquid chromatography-nuclear magnetic resonance spectroscopy. Journal of chromatography.A 2013, 1289, 19-26.

123

124. Baderschneider, B.; Winterhalter, P. Isolation and characterization of novel stilbene derivatives from Riesling wine. J. Agric. Food Chem. 2000, 48, 2681-2686.

125. Mulinacci, N.; Innocenti, M.; Santamaria, A. R.; la Marca, G.; Pasqua, G. High-performance liquid chromatography/electrospray ionization tandem mass spectrometric investigation of stilbenoids in cell cultures of Vitis vinifera L., cv. Malvasia. Rapid Commun. Mass Spectrom. 2010, 24, 2065-2073.

126. Mattivi, F.; Vrhovsek, U.; Malacarne, G.; Masuero, D.; Zulini, L.; Stefanini, M.; Moser, C.; Velasco, R.; Guella, G. Profiling of Resveratrol Oligomers, Important Stress Metabolites, Accumulating in the Leaves of Hybrid Vitis vinifera ( x Teroldego) Genotypes Infected with Plasmopara viticola. J. Agric. Food Chem. 2011, 59, 5364-5375.

127. Guebailia, H. A.; Chira, K.; Richard, T.; Mabrouk, T.; Furiga, A.; Vitrac, X.; Monti, J.; Delaunay, J.; Merillon, J. Hopeaphenol: The first resveratrol tetramer in wines from North Africa. J. Agric. Food Chem. 2006, 54, 9559-9564.

128. Ibanez, M.; Sancho, J.; Pozo, O.; Hernandez, F. Use of quadrupole time-of-flight mass spectrometry in environmental analysis: Elucidation of transformation products of triazine herbicides in water after UV exposure. Anal. Chem. 2004, 76, 1328-1335.

129. Stolker, A.; Niesing, W.; Fuchs, R.; Vreeken, R.; Niessen, W.; Brinkman, U. Liquid chromatography with triple-quadrupole and quadrupole-time-of-flight mass spectrometry for the determination of micro-constituents - a comparison. Anal. Bioanal. Chem. 2004, 378, 1754-1761.

130. Sancho, J. V.; Pozo, O. J.; Ibanez, M.; Hernandez, F. Potential of liquid chromatography/time-of-flight mass spectrometry for the determination of pesticides and transformation products in water. Anal. Bioanal. Chem. 2006, 386, 987-997.

131. Han, L.; Boakye-Yiadom, M.; Liu, E.; Zhang, Y.; Li, W.; Song, X.; Fu, F.; Gao, X. Structural Characterisation and Identification of Phenylethanoid Glycosides from Cistanches deserticola YC Ma by UHPLC/ESI-QTOF-MS/MS. Phytochem. Anal. 2012, 23, 668-676.

132. Hu, Y.; Qu, J.; Liu, Y.; Yu, S.; Li, J.; Zhang, J.; Du, D. Structural characterization of trace stilbene glycosides in Lysidice brevicalyx Wei using liquid chromatography/diode-array detection/electrospray ionization tandem mass spectrometry. J. Chromatogr. B 2010, 878, 1- 7.

133. Xu, G.; Li, L.; Fang, D.; Li, G.; Zhang, G.; Wu, Z. Electrospray ionization tandem mass spectrometry of chaetoglobosins. Rapid Commun. Mass Spectrom. 2012, 26, 2115-2122.

134. Perestrelo, R.; Lu, Y.; Santos, S. A. O.; Silvestre, A. J. D.; Neto, C. P.; Camara, J. S.; Rocha, S. M. Phenolic profile of Sercial and Tinta Negra Vitis vinifera L. grape skins by HPLC- DAD-ESI-MSn Novel phenolic compounds in Vitis vinifera L. grape. Food Chem. 2012, 135, 94-104.

124

135. Pussa, T.; Floren, J.; Kuldepp, P.; Raal, A. Survey of grapevine Vitis vinifera stem polyphenols by liquid chromatography-diode array detection-tandem mass spectrometry. J. Agric. Food Chem. 2006, 54, 7488-7494.

136. Flamini, R.; Rosso, M.; Marchi, F.; Dalla Vedova, A.; Panighel, A.; Gardiman, M.; Maoz, I.; Bavaresco, L. An innovative approach to grape metabolomics: stilbene profiling by suspect screening analysis. Metabolomics 2013, 1-11.

137. Chen, J.; He, S.; Mao, H.; Sun, C.; Pan, Y. Characterization of polyphenol compounds from the roots and stems of Parthenocissus laetevirens by high-performance liquid chromatography/tandem mass spectrometry. Rapid Commun. Mass Spectrom. 2009, 23, 737- 744.

138. Kong, Q. J.; Ren, X. Y.; Hu, N.; Sun, C. R.; Ran, Y. J. Identification of isomers of resveratrol dimer and their analogues from wine grapes by HPLC/MSn and HPLC/DAD-UV. Food Chem. 2011, 127, 727-734.

139. Jean-Denis, J.; Pezet, R.; Tabacchi, R. Rapid analysis of stilbenes and derivatives from downy mildew-infected grapevine leaves by liquid chromatography-atmospheric pressure photoionisation mass spectrometry. J. Chromatogr. A 2006, 1112, 263-268.

140. Nakajima, K.; Taguchi, H.; Endo, T.; Yosioka, I. Constituents of Scirpus-Fluviatilis (Torr) Gray,a. .1. Structures of 2 New Hydroxystilbene Dimers, Scirpusin-a and Scirpusin-B. Chem. Pharm. Bull. 1978, 26, 3050-3057.

141. Sano, S.; Sugiyama, K.; Ito, T.; Katano, Y.; Ishihata, A. Identification of the Strong Vasorelaxing Substance Scirpusin B, a Dimer of Piceatannol, from Passion Fruit (Passiflora edulis) Seeds. J. Agric. Food Chem. 2011, 59, 6209-6213.

142. Tamboli, V. F.; Re, N.; Coletti, C.; Defant, A.; Mancini, I.; Tosi, P. A joint experimental and theoretical investigation on the oxidative coupling of resveratrol induced by copper and iron ions. Int. J. Mass Spectrom. 2012, 319, 55-63.

143. Mercolini, L.; Saracino, M. A.; Bugamelli, F.; Ferranti, A.; Malaguti, M.; Hrelia, S.; Raggi, M. A. HPLC-F analysis of melatonin and resveratroll isomers in wine using an SPE procedure. J. Sep. Sci. 2008, 31, 1007-1014.

144. Gamoh, K.; Nakashima, K. Liquid chromatography mass spectrometric determination of trans-resveratrol in wine using a tandem solid-phase extraction method. Rapid Commun. Mass Spectrom. 1999, 13, 1112-1115.

145. Boutegrabet, L.; Fekete, A.; Hertkorn, N.; Papastamoulis, Y.; Waffo-Teguo, P.; Merillon, J. M.; Jeandet, P.; Gougeon, R. D.; Schmitt-Kopplin, P. Determination of stilbene derivatives in Burgundy red wines by ultra-high-pressure liquid chromatography. Anal. Bioanal. Chem. 2011, 401, 1513-1521.

125

146. Naugler, C.; McCallum, J. L.; Klassen, G.; Strommer, J. Concentrations of trans-resveratrol and related stilbenes in Nova Scotia wines. Am. J. Enol. Vitic. 2007, 58, 117-119.

147. Gerogiannaki-Christopoulou, M.; Athanasopoulos, P.; Kyriakidis, N.; Gerogiannaki, I.; Spanos, M. trans-resveratrol in wines from the major Greek red and white grape varieties. Food Control 2006, 17, 700-706.

148. Feijóo, O.; Moreno, A.; Falqué, E. Content of trans- and cis-resveratrol in Galician white and red wines. Journal of Food Composition and Analysis 2008, 21, 608-613.

149. Moreno, A.; Castro, M.; Falque, E. Evolution of trans- and cis-resveratrol content in red grapes (Vitis vinifera L. cv Mencia, Albarello and Merenzao) during ripening. Eur. Food Res. Technol. 2008, 227, 667-674.

150. Moreno-Labanda, J.; Mallavia, R.; Perez-Fons, L.; Lizama, V.; Saura, D.; Micol, V. Determination of piceid and resveratrol in Spanish wines deriving from Monastrell (Vitis vinifera L.) grape variety. J. Agric. Food Chem. 2004, 52, 5396-5403.

151. Vitrac, X.; Monti, J.; Vercauteren, J.; Deffieux, G.; Merillon, J. Direct liquid chromatographic analysis of resveratrol derivatives and flavanonols in wines with absorbance and fluorescence detection. Anal. Chim. Acta 2002, 458, 103-110.

152. Guebailia, H. A.; Chira, K.; Richard, T.; Mabrouk, T.; Furiga, A.; Vitrac, X.; Monti, J.; Delaunay, J.; Merillon, J. Hopeaphenol: The first resveratrol tetramer in wines from North Africa. J. Agric. Food Chem. 2006, 54, 9559-9564.

153. Landrault, N.; Larronde, F.; Delaunay, J.; Castagnino, C.; Vercauteren, J.; Merillon, J.; Gasc, F.; Cros, G.; Teissedre, P. Levels of stilbene oligomers and astilbin in French varietal wines and in grapes during noble rot development. J. Agric. Food Chem. 2002, 50, 2046-2052.

154. Vitrac, X.; Bornet, A.; Vanderlinde, R.; Valls, J.; Richard, T.; Delaunay, J.; Merillon, J.; Teissedre, P. Determination of stilbenes (delta-viniferin, trans-astringin, trans-piceid, cis- and trans-resveratrol, epsilon-viniferin) in Brazilian wines. J. Agric. Food Chem. 2005, 53, 5664- 5669.

155. Amira-Guebailia, H.; Valls, J.; Richard, T.; Vitrac, X.; Monti, J.; Delaunay, J.; Merillon, J. Centrifugal partition chromatography followed by HPLC for the isolation of cis-epsilon- viniferin, a resveratrol dimer newly extracted from a red Algerian wine. Food Chem. 2009, 113, 320-324.

156. Caruso, F.; Tanski, J.; Villegas-Estrada, A.; Rossi, M. Structural basis for antioxidant activity of trans-resveratrol: Ab initio calculations and crystal and molecular structure. J. Agric. Food Chem. 2004, 52, 7279-7285.

126

127

APPENDIX A: ADDITIONAL DATA

Table A 1: Quantification data for all 41 compounds in each individual wine. Concentrations shown in ng/mL trans-piceid equivalents.

[M-H]- 227.0714 227.0714 389.1242 389.1242 243.0663 243.0663 405.1191 RT 22.773 25.204 13.05 15.903 16.432 21.698 9.72 Sample # Winery; Variety trans-resveratrol cis-resveratrol trans-piceid cis-piceid trans-piceatannol cis-piceatannol trans-astringin 1 2009 Jackson Triggs Cab Sauv 0.054 ± 0.047 0.064 ± 0.059 665.717 ± 8.23 336.691 ± 7.251 0.082 ± 0.01 0.101 ± 0.008 0.092 ± 0.006 2 2010 Copper Moon Cab Sauv 284.054 ± 5.435 214.659 ± 15.646 469.386 ± 6.309 802.393 ± 5.038 655.996 ± 4.466 23.381 ± 0.224 372.284 ± 7.276 3 2010 Inniskillin Cab Sauv 561.98 ± 21.291 867.751 ± 41.147 824.393 ± 11.27 1810.949 ± 47.981 1657.154 ± 24.509 45.595 ± 1.769 426.799 ± 12.212 4 2010 Jackson Triggs Cab Sauv 183.004 ± 1.771 409.426 ± 13.533 685.175 ± 26.995 1453.693 ± 12.309 646.266 ± 8.631 18.328 ± 0.843 576.904 ± 10.442 5 Coteau St. Paul Frontenac 0.211 ± 0.072 0.226 ± 0.076 0.17 ± 0.028 3.07 ± 0 93.657 ± 0 8.073 ± 0 29.559 ± 0 6 La Halte des Pèlerins Frontenac 15.735 ± 0.511 62.354 ± 1.451 45.998 ± 1.171 174.085 ± 1.368 210.809 ± 2.011 9.754 ± 0.906 41.671 ± 0.601 7 Les Petits Cailloux Frontenac 11.028 ± 0.033 6.648 ± 0.646 5.013 ± 0.22 6.08 ± 0.453 25.553 ± 0.505 6.958 ± 0.285 45.347 ± 0.857 8 Ste-Pétronille Frontenac 3.848 ± 0.026 16.459 ± 0.529 0.12 ± 0.04 69.366 ± 0.353 66.414 ± 0.127 17.438 ± 1.077 38.621 ± 0.979 9 Isle de Bacchus Frontenac 0.516 ± 0.12 33.596 ± 0.633 0.528 ± 0.072 74.238 ± 0.419 157.731 ± 3.099 29.858 ± 1.581 14.598 ± 0.093 10 Ste-Pétronille Frontenac 7.972 ± 0.14 24.856 ± 0.592 15.038 ± 0.35 74.08 ± 0.051 109.217 ± 0.133 10.879 ± 0.03 38.785 ± 0.626 11 Côte de Vaudreuil Frontenac 7.178 ± 0.122 25.704 ± 0.591 13.558 ± 0.324 74.095 ± 0.081 113.925 ± 0.394 12.721 ± 0.137 36.437 ± 0.569 12 Domaine de lavoie Maréchal Foch 98.832 ± 2.093 205.124 ± 5.576 70.346 ± 1.626 834.749 ± 12.989 761.753 ± 12.471 10.464 ± 0.296 173.766 ± 3.448 13 Domaine de lavoie Maréchal Foch 100.929 ± 2.417 211.395 ± 1.06 105.295 ± 5.37 1248.069 ± 12.106 1050.479 ± 12.853 50.891 ± 1.044 233.993 ± 2.317 14 Isle de Bacchus Maréchal Foch 41.748 ± 1.855 231.396 ± 7.69 7.07 ± 0.11 268.727 ± 1.949 293.836 ± 7.431 38.569 ± 1.724 28.315 ± 0.911 15 Orpailleur Maréchal Foch 144.516 ± 4.52 778.774 ± 5.039 17.886 ± 0.448 471.563 ± 10.581 276.814 ± 8.142 57.152 ± 0.657 6.622 ± 0.565 16 l'ange-Gardien Maréchal Foch 101.344 ± 0.736 389.117 ± 7.006 48.303 ± 0.503 698.069 ± 5.115 586.823 ± 6.57 40.526 ± 1.304 105.849 ± 0.504 17 Les Bromes Maréchal Foch 101.342 ± 0.736 389.111 ± 7.006 48.302 ± 0.503 698.057 ± 5.115 586.813 ± 6.569 40.525 ± 1.304 105.847 ± 0.504 18 ste. Marguerite Marquette 210.777 ± 1.602 1049.455 ± 45.766 25.927 ± 2.054 710.054 ± 9.396 415.625 ± 20.112 135.899 ± 1.179 67.881 ± 2.858 19 Alain Breault Marquette 0.424 ± 0.049 130.526 ± 3.136 8.389 ± 1.152 23.388 ± 1.178 114.906 ± 1.874 29.609 ± 0.041 69.632 ± 3.165 20 La Bauge Marquette 193.448 ± 2.385 1240.436 ± 269.609 15.199 ± 1.064 119.577 ± 23.791 716.864 ± 181.556 209.227 ± 84.879 56.068 ± 6.352 21 Le Domaine Bergeville Marquette 247.251 ± 1.28 1423.468 ± 166.111 418.266 ± 337.041 129.752 ± 3.016 806.964 ± 72.762 279.064 ± 126.443 71.818 ± 4.253 22 Isle De Bacchus Marquette 165.843 ± 0.01 959.164 ± 42.447 97.145 ± 66.319 298.913 ± 4.403 495.827 ± 34.253 157.243 ± 4.683 66.223 ± 0.934 23 Artisans Marquette 139.33 ± 5.08 805.991 ± 45.699 81.605 ± 54.652 251.154 ± 9.854 416.63 ± 32.055 132.102 ± 6.157 55.615 ± 2.176 24 2009 Camelot Merlot 24.796 ± 0.102 68.336 ± 1.122 162.166 ± 1.487 441.642 ± 6.129 22.151 ± 1.036 0.104 ± 0.007 56.301 ± 1.122 25 2009 Inniskillin Merlot 210.284 ± 8.446 314.22 ± 17.894 484.76 ± 12.953 1195.521 ± 16.656 714.615 ± 29.992 15.15 ± 0.16 318.586 ± 8.988 26 2009 Sumac Ridge Merlot 96.981 ± 0.514 65.641 ± 0.508 386.605 ± 8.078 673.339 ± 6.703 290.072 ± 2.706 10.511 ± 1.047 213.173 ± 2.097 27 2009 Jackson Triggs Merlot (Box) 360.746 ± 2.459 355.488 ± 9.196 621.37 ± 1.47 396.144 ± 5.062 294.56 ± 5.601 23.754 ± 0.173 136.363 ± 7.338 28 2008 Inniskillin Pinot Noir 731.403 ± 0.393 660.597 ± 7.345 915.093 ± 13.466 1621.055 ± 41.588 1250.766 ± 39.13 44.139 ± 1.383 310.693 ± 1.614 29 2011 DB Pinot Noir 134.608 ± 37.834 215.233 ± 60.676 664.491 ± 186.716 838.545 ± 235.387 254.488 ± 72.434 17.835 ± 5.147 508.005 ± 143.721 30 2009 Jackson Triggs Pinot Noir (RF) 747.886 ± 14.921 1499.989 ± 8.106 25.528 ± 0.1 60.759 ± 2.507 471.134 ± 20.808 146.64 ± 1.856 0.308 ± 0.022 31 2010 Jackson Triggs Pinot Noir (TT) 788.051 ± 13.98 1200.079 ± 27.17 10.681 ± 1.35 12.292 ± 1.646 155.272 ± 7.581 41.46 ± 1.626 5.475 ± 0.778 32 Limousin Sabrevois 9.911 ± 0.056 78.593 ± 1.934 2.68 ± 0.033 29.363 ± 0.419 23.182 ± 0.462 5.7 ± 0.194 2.627 ± 0.318 33 Côte de Vaudreuil Sabrevois 55.351 ± 0.872 179.564 ± 1.863 4.805 ± 0.128 57.622 ± 0.81 82.115 ± 4.131 8.554 ± 0.534 0.061 ± 0.048 34 La Mission Sabrevois 52.23 ± 0.824 279.699 ± 1.436 9.27 ± 0.101 99.96 ± 0.501 97.57 ± 2.202 18.399 ± 0.524 6.729 ± 0.665 35 2010 Cedar Creek Shiraz-cabernet 73.167 ± 1.59 103.551 ± 2.673 456.487 ± 8.344 1199.585 ± 19.642 253.246 ± 4.217 6.515 ± 0.63 297.188 ± 14.53 36 Domaine de lavoie St. Croix 37.709 ± 0.279 77.014 ± 1.682 3.575 ± 0.105 159.743 ± 4.446 101.489 ± 1.162 4.178 ± 0.038 8.998 ± 0.07 37 Isle de Bacchus St. Croix 85.79 ± 3.683 127.449 ± 4.879 3.828 ± 0.251 162.442 ± 3.315 91.894 ± 0.639 4.96 ± 0.652 5.83 ± 0.138 38 Les Petits Cailloux St. Croix 26.637 ± 0.387 56.243 ± 0.479 10.693 ± 0.518 131.121 ± 3.219 61.619 ± 3.158 8.785 ± 0.284 0.032 ± 0.012 39 La Halte des Pèlerins St. Croix 115.572 ± 3.478 152.496 ± 2.453 6.579 ± 0.224 576.469 ± 11.539 319.843 ± 13.963 18.062 ± 0.957 3.982 ± 0.061 40 Vignoble du Mitan St. Croix 60.648 ± 0.485 599.249 ± 17.795 6.514 ± 0.584 49.079 ± 1.72 239.741 ± 3.613 104.278 ± 0.765 5.749 ± 0.219 41 Ste-Pétronille St.Croix 16.199 ± 1 117.129 ± 3.694 33.065 ± 1.565 112.649 ± 0.993 61.143 ± 3.215 11.643 ± 0.16 3.455 ± 0.628

128

[M-H]- [M-H]- 405.1191 453.1344 453.1344 453.1344 453.1344 453.1344 453.1344 RT 12.479 23.043 25.627 27.424 28.481 33.448 34.399 Sample # Winery; Variety cis-astringin Pallidol parthenocissin A* Quadrangularin A* Ampelopsin D* cis-ε-viniferin cis-ω-viniferin 1 2009 Jackson Triggs Cab Sauv 0.043 ± 0.029 400.025 ± 12.093 82.263 ± 3.619 10.409 ± 0.089 5.69 ± 1.066 12.58 ± 1.184 0.036 ± 0.007 2 2010 Copper Moon Cab Sauv 164.855 ± 8.089 378.12 ± 11.034 126.665 ± 3.384 6.451 ± 0.383 108.439 ± 3.105 99.953 ± 3.615 17.161 ± 0.084 3 2010 Inniskillin Cab Sauv 240.795 ± 11.304 317.455 ± 12.647 95 ± 5.371 12.894 ± 0.138 563.01 ± 447.989 124.694 ± 4.463 21.576 ± 0.324 4 2010 Jackson Triggs Cab Sauv 246.942 ± 15.187 214.034 ± 0.991 71.906 ± 1.007 7.783 ± 0.388 56.653 ± 1.023 99.239 ± 1.716 30.423 ± 1.268 5 Coteau St. Paul Frontenac 31.924 ± 0 49.874 ± 0 0.2 ± 0.069 0.204 ± 0.072 0.252 ± 0.046 0.215 ± 0.035 0.224 ± 0.06 6 La Halte des Pèlerins Frontenac 118.683 ± 4.015 48.836 ± 0.207 12.277 ± 1.369 0.046 ± 0.01 4.727 ± 0.29 13.88 ± 0.147 13.675 ± 0.597 7 Les Petits Cailloux Frontenac 83.886 ± 8.199 39.631 ± 2.119 5.468 ± 1.234 2.099 ± 0.116 2.047 ± 0.154 3.907 ± 0.332 0.303 ± 0.063 8 Ste-Pétronille Frontenac 66.234 ± 0.187 32.911 ± 0.153 6.078 ± 0.771 0.146 ± 0.006 0.145 ± 0.068 5.337 ± 0.373 5.034 ± 0.149 9 Isle de Bacchus Frontenac 56.775 ± 2.108 48.257 ± 0.572 0.505 ± 0.136 0.547 ± 0.128 0.507 ± 0.051 0.459 ± 0.043 0.466 ± 0.138 10 Ste-Pétronille Frontenac 77.953 ± 2.834 43.02 ± 0.503 6.425 ± 0.008 0.412 ± 0.025 1.746 ± 0.06 6.414 ± 0.186 5.467 ± 0.107 11 Côte de Vaudreuil Frontenac 75.899 ± 2.366 43.528 ± 0.401 5.78 ± 0.01 0.351 ± 0.023 1.555 ± 0.055 5.77 ± 0.172 4.916 ± 0.099 12 Domaine de lavoie Maréchal Foch 154.561 ± 4.279 1064.328 ± 7.416 203.233 ± 3.56 24.997 ± 4.847 88.76 ± 0.787 169.32 ± 0.516 158.301 ± 2.553 13 Domaine de lavoie Maréchal Foch 292.581 ± 3.744 1158.2 ± 3.641 239.983 ± 5.976 31.137 ± 3.724 114.283 ± 2.503 203.848 ± 3.053 191.554 ± 2.318 14 Isle de Bacchus Maréchal Foch 95.542 ± 1.208 291.649 ± 8.236 34.677 ± 0.784 10.328 ± 1.052 34.089 ± 2.076 35.216 ± 2.124 24.268 ± 1.521 15 Orpailleur Maréchal Foch 17.541 ± 2.492 479.145 ± 34.46 56.708 ± 2.352 16.48 ± 0.385 67.283 ± 0.921 121.669 ± 1.272 47.999 ± 2.826 16 l'ange-Gardien Maréchal Foch 134.122 ± 0.502 750.196 ± 15.527 131.412 ± 2.65 20.93 ± 0.098 76.855 ± 0.271 133.692 ± 1.926 103.712 ± 2.4 17 Les Bromes Maréchal Foch 134.12 ± 0.502 750.183 ± 15.526 131.41 ± 2.65 20.93 ± 0.098 76.853 ± 0.271 133.69 ± 1.926 103.71 ± 2.4 18 ste. Marguerite Marquette 108.118 ± 2.649 184.553 ± 12.347 58.284 ± 4.138 5.441 ± 1.363 51.58 ± 7.548 158.428 ± 8.776 52.878 ± 1.787 19 Alain Breault Marquette 45.57 ± 2.765 60.016 ± 1.288 12.276 ± 0.13 0.392 ± 0.03 13.621 ± 0.231 25.167 ± 3.197 0.439 ± 0.091 20 La Bauge Marquette 78.709 ± 2.797 215.4 ± 20.483 72.705 ± 18.718 13.74 ± 1.488 78.207 ± 32.826 208.009 ± 50.424 40.198 ± 0.308 21 Le Domaine Bergeville Marquette 77.639 ± 9.862 279.075 ± 8.874 62.791 ± 9.183 27.653 ± 2.847 63.368 ± 6.833 262.564 ± 88.502 48.569 ± 1.533 22 Isle De Bacchus Marquette 80.89 ± 3.829 182.179 ± 0.916 52.134 ± 1.473 10.508 ± 0.671 51.59 ± 4.018 160.506 ± 7.253 36.929 ± 0.746 23 Artisans Marquette 67.94 ± 4.017 153.057 ± 5.633 43.775 ± 2.005 8.796 ± 0.643 43.318 ± 3.661 134.845 ± 7.745 30.998 ± 1.293 24 2009 Camelot Merlot 92.088 ± 0.057 49.525 ± 2.222 9.462 ± 0.435 0.125 ± 0.034 9.63 ± 0.648 8.941 ± 0.475 2.698 ± 0.286 25 2009 Inniskillin Merlot 149.079 ± 4.586 231.288 ± 9.148 65.838 ± 2.401 7.812 ± 0.625 74.919 ± 2.939 78.311 ± 1.769 17.113 ± 0.158 26 2009 Sumac Ridge Merlot 41.518 ± 2.442 55.786 ± 2.988 21.501 ± 0.431 0.046 ± 0.048 25.41 ± 0.832 20.768 ± 1.906 5.189 ± 0.063 27 2009 Jackson Triggs Merlot (Box) 52.297 ± 1.184 199.155 ± 0.483 72.628 ± 3.377 3.828 ± 0.148 86.755 ± 2.785 43.329 ± 0.479 10.113 ± 0.211 28 2008 Inniskillin Pinot Noir 65.154 ± 1.742 932.395 ± 19.52 299.425 ± 0.106 27.839 ± 1.189 462.461 ± 3.323 138.987 ± 2.593 12.397 ± 0.023 29 2011 DB Pinot Noir 163.244 ± 45.906 575.511 ± 161.758 114.744 ± 32.253 8.452 ± 2.842 84.724 ± 23.839 60.827 ± 17.162 10.576 ± 3.068 30 2009 Jackson Triggs Pinot Noir (RF) 0.39 ± 0.016 850.419 ± 5.302 320.75 ± 1.598 27.405 ± 1.171 508.241 ± 12.142 212.943 ± 5.181 25.379 ± 0.102 31 2010 Jackson Triggs Pinot Noir (TT) 14.782 ± 2.599 813.995 ± 3.473 489.583 ± 4.618 30.832 ± 0.442 195.082 ± 0.366 248.464 ± 2.83 10.338 ± 0.667 32 Limousin Sabrevois 4.346 ± 0.272 10.932 ± 0.017 1.216 ± 0.059 0.15 ± 0.077 2.025 ± 0.486 3.096 ± 0.153 1.647 ± 0.071 33 Côte de Vaudreuil Sabrevois 0.089 ± 0.037 57.571 ± 1.436 7.765 ± 0.592 0.09 ± 0.014 14.157 ± 0.384 23.874 ± 0.846 8.947 ± 0.403 34 La Mission Sabrevois 10.863 ± 0.492 55.803 ± 1.249 7.263 ± 0.369 0.817 ± 0.178 12.433 ± 1.23 19.935 ± 0.602 8.896 ± 0.13 35 2010 Cedar Creek Shiraz-cabernet 212.845 ± 3.507 192.807 ± 4.799 51.371 ± 1.002 5.987 ± 0.104 50.459 ± 2.3 31.372 ± 1.614 13.325 ± 0.541 36 Domaine de lavoie St. Croix 11.301 ± 1.901 21.541 ± 1.264 6.881 ± 1.094 0.097 ± 0.029 0.095 ± 0.023 9.043 ± 0.581 7.25 ± 1.159 37 Isle de Bacchus St. Croix 17.352 ± 0.562 36.013 ± 0.298 14.235 ± 2.775

129

[M-H]- 469.1293 471.1449 471.1449 471.1449 471.1449 485.1606 485.1606 RT 29.749 12.837 19.601 20.024 23.594 21.716 23.196 Sample # Winery; Variety trans-Scirpusin A* Restrisol A* Oxidized Dimer 1* Oxidized Dimer 2* Oxidized Dimer 3*Parthenostilbenin A*Parthenostilbenin B* 1 2009 Jackson Triggs Cab Sauv 5.403 ± 0.835 102.279 ± 0.482 41.049 ± 0.863 9.084 ± 0.316 19.781 ± 0.084 26.78 ± 0.732 20.916 ± 0.089 2 2010 Copper Moon Cab Sauv 40.08 ± 0.904 59.557 ± 0.412 33.806 ± 0.317 8.389 ± 0.29 7.485 ± 0.173 10.321 ± 0.391 27.267 ± 0.7 3 2010 Inniskillin Cab Sauv 45.779 ± 0.429 129.484 ± 1.233 17.538 ± 0.835 7.655 ± 0.034 9.136 ± 0.39 4.67 ± 0.343 29.752 ± 0.947 4 2010 Jackson Triggs Cab Sauv 55.562 ± 0.249 73.383 ± 0.691 10.529 ± 0.123 2.823 ± 0.013 5.838 ± 0.263 4.8 ± 0.092 24.352 ± 0.637 5 Coteau St. Paul Frontenac 4.758 ± 0 23.078 ± 0 4.562 ± 0 8.764 ± 0 1.629 ± 0 11.715 ± 0 15.093 ± 0 6 La Halte des Pèlerins Frontenac 24.363 ± 1.659 23.076 ± 0.088 6.855 ± 0.118 3.887 ± 0.001 61.276 ± 0.284 0.125 ± 0.034 4.741 ± 0.516 7 Les Petits Cailloux Frontenac 19.082 ± 2.238 28.036 ± 0.466 6.6 ± 0.066 7.028 ± 0.109 3.844 ± 0.271 0.232 ± 0.046 10.746 ± 0.106 8 Ste-Pétronille Frontenac 0.381 ± 0.012 28.857 ± 0.589 6.132 ± 0.509 2.369 ± 0.2 10.569 ± 0.396 0.151 ± 0.04 6.51 ± 0.037 9 Isle de Bacchus Frontenac 0.625 ± 0.025 55.953 ± 2.67 10.537 ± 0.052 4.925 ± 0.508 0.601 ± 0.12 0.616 ± 0.041 41.133 ± 1.435 10 Ste-Pétronille Frontenac 12.406 ± 0.107 25.587 ± 0.22 6.084 ± 0.185 5.213 ± 0.095 22.829 ± 0.358 2.73 ± 0.009 8.763 ± 0.164 11 Côte de Vaudreuil Frontenac 11.262 ± 0.101 28.534 ± 0.044 6.516 ± 0.169 5.333 ± 0.384 20.671 ± 0.3 2.525 ± 0.013 11.904 ± 0.026 12 Domaine de lavoie Maréchal Foch 97.108 ± 0.694 338.028 ± 4.331 18.91 ± 0.018 12.28 ± 0.103 4.923 ± 0.411 26.901 ± 1.863 54.873 ± 2.243 13 Domaine de lavoie Maréchal Foch 97.737 ± 0.425 485.237 ± 0.877 22.173 ± 0.293 10.047 ± 0.275 8.582 ± 0.242 26.351 ± 0.576 73.54 ± 0.903 14 Isle de Bacchus Maréchal Foch 13.995 ± 0.707 170.298 ± 0.237 4.651 ± 0.206 5.521 ± 0.045 19.638 ± 0.531 8.969 ± 0.003 20.008 ± 0.677 15 Orpailleur Maréchal Foch 17.691 ± 1.182 201.568 ± 1.306 8.622 ± 0.999 2.367 ± 0.258 5.135 ± 0.328 17.357 ± 0.33 37.505 ± 0.25 16 l'ange-Gardien Maréchal Foch 55.813 ± 0.914 298.055 ± 1.709 13.558 ± 0.167 7.625 ± 0.102 9.912 ± 0.169 20.309 ± 0.747 46.638 ± 1.118 17 Les Bromes Maréchal Foch 55.812 ± 0.914 298.05 ± 1.709 13.557 ± 0.167 7.625 ± 0.102 9.912 ± 0.169 20.309 ± 0.747 46.638 ± 1.118 18 ste. Marguerite Marquette 16.672 ± 0.352 130.973 ± 1.367 13.779 ± 0.534 6.729 ± 0.186 36.081 ± 0.694 8.962 ± 0.594 43.336 ± 1.262 19 Alain Breault Marquette 13.011 ± 0.198 46.028 ± 1.329 9.548 ± 0.167 8.904 ± 0.102 8.203 ± 0.57 7.113 ± 0.939 19.954 ± 0.479 20 La Bauge Marquette 27.589 ± 0.24 176.162 ± 63.413 21.283 ± 1.086 9.974 ± 0.117 6.298 ± 0.269 16.621 ± 0.935 36.14 ± 0.033 21 Le Domaine Bergeville Marquette 26.455 ± 0.084 196.411 ± 35.174 28.372 ± 1.108 17.706 ± 1.969 10.654 ± 0.625 35.066 ± 0.774 37.217 ± 1.339 22 Isle De Bacchus Marquette 20.363 ± 0.14 135.347 ± 8.274 17.477 ± 0.336 10.163 ± 0.471 17.637 ± 0.099 15.508 ± 0.362 35.094 ± 0.481 23 Artisans Marquette 17.077 ± 0.64 113.702 ± 7.958 14.652 ± 0.607 8.506 ± 0.502 14.786 ± 0.552 12.997 ± 0.565 29.456 ± 1.152 24 2009 Camelot Merlot 4.312 ± 0.328 11.511 ± 0.057 6.642 ± 0.647 2.656 ± 0.212 1.964 ± 0.436 0.144 ± 0.012 2.328 ± 0.091 25 2009 Inniskillin Merlot 38.85 ± 9.336 45.713 ± 0.447 8.387 ± 0.257 5.07 ± 0.063 6.197 ± 0.263 2.992 ± 0.325 14.685 ± 0.025 26 2009 Sumac Ridge Merlot 18.193 ± 2.632 8.932 ± 0.766 6.937 ± 0.126 2.438 ± 0.01 7.186 ± 0.064 3.859 ± 0.281 4.59 ± 0.056 27 2009 Jackson Triggs Merlot (Box) 43.313 ± 0.217 28.671 ± 0.283 17.532 ± 0.136 7.955 ± 0.141 0.077 ± 0.056 5.552 ± 0.021 11.035 ± 0.067 28 2008 Inniskillin Pinot Noir 134.563 ± 2.876 171.332 ± 2.509 111.254 ± 3.305 14.511 ± 0.192 11.793 ± 0.226 10.52 ± 0.313 36.957 ± 1.213 29 2011 DB Pinot Noir 113.317 ± 31.861 98.217 ± 27.636 17.472 ± 4.964 5.695 ± 1.673 12.341 ± 3.534 10.051 ± 3.09 22.779 ± 6.461 30 2009 Jackson Triggs Pinot Noir (RF) 146.563 ± 0.492 786.405 ± 9.582 269.326 ± 9.884 79.136 ± 5.675 0.307 ± 0.074 58.322 ± 1.339 41.495 ± 0.862 31 2010 Jackson Triggs Pinot Noir (TT) 98.829 ± 0.865 453.755 ± 5.74 212.181 ± 1.062 31.274 ± 0.548 21.525 ± 1.059 82.304 ± 0.968 25.672 ± 0.175 32 Limousin Sabrevois 1.349 ± 0.015 6.465 ± 0.232 2.67 ± 0.039 0.263 ± 0.033 0.246 ± 0.065 1.803 ± 0.098 4.196 ± 0.116 33 Côte de Vaudreuil Sabrevois 6.888 ± 0.221 20.109 ± 0.344 3.708 ± 0.232 0.063 ± 0.032 3.619 ± 0.1 4.239 ± 0.28 6.184 ± 0.082 34 La Mission Sabrevois 7.136 ± 0.077 26.103 ± 1.005 8.689 ± 0.09 1.076 ± 0.102 2.837 ± 0.176 6.883 ± 0.175 13.596 ± 0.165 35 2010 Cedar Creek Shiraz-cabernet 37.003 ± 1.062 41.173 ± 0.766 16.76 ± 0.401 5.773 ± 0.101 6.403 ± 0.12 2.981 ± 0.311 9.963 ± 0.598 36 Domaine de lavoie St. Croix 10.294 ± 0.291 12.945 ± 0.104 2.566 ± 0.116 1.722 ± 0.162 38.348 ± 0.315 1.348 ± 0.171 6.446 ± 0.159 37 Isle de Bacchus St. Croix 13.079 ± 0.596 11.412 ± 0.132 4.974 ± 0.373 4.462 ± 0.141 24.715 ± 0.176 2.85 ± 0.077 9.066 ± 0.217 38 Les Petits Cailloux St. Croix 11.695 ± 0.475 9.734 ± 0.852 0.082 ± 0.037 0.095 ± 0.035 2.981 ± 0.17 8.024 ± 0.33 2.797 ± 0.056 39 La Halte des Pèlerins St. Croix 18.159 ± 0.131 6.682 ± 0.003 6.109 ± 0.145 2.306 ± 0.023 3.054 ± 0.11 0.106 ± 0.048 0.157 ± 0.005 40 Vignoble du Mitan St. Croix 6.264 ± 0.363 96.243 ± 0.923 9.255 ± 0.553 2.01 ± 0.25 4.109 ± 0.26 11.418 ± 0.591 15.749 ± 0.701 41 Ste-Pétronille St.Croix 2.746 ± 0.179 16.446 ± 0.818 6.998 ± 0.101 1.811 ± 0.256 29.523 ± 1.387 1.356 ± 0.169 3.87 ± 0.44

[M-H]- 485.1236 615.1872 615.1872 679.1974 679.1974 679.1974 679.1974 RT 25.396 19.814 26.789 31.44 36.936 37.781 39.155 Sample # Winery; Variety P+P Dimer* ε-viniferin glycoside*Dimer glycoside 1* Ampelopsin C* E-Miyabenol C Z-Miyabenol C Trimer 4* 1 2009 Jackson Triggs Cab Sauv 20.24 ± 0.242 0.044 ± 0.022 0.057 ± 0.051 6.114 ± 0.08 4.557 ± 0.386 0.082 ± 0.019 0.095 ± 0.013 2 2010 Copper Moon Cab Sauv 20.045 ± 0.069 31.564 ± 0.164 26.096 ± 0.343 25.405 ± 0.86 34.348 ± 1.123 58.921 ± 0.672 19.249 ± 1.192 3 2010 Inniskillin Cab Sauv 135.24 ± 2.406 45.404 ± 0.619 29.502 ± 0.215 12.124 ± 0.181 20.159 ± 0.52 29.06 ± 0.205 7.667 ± 0.099 4 2010 Jackson Triggs Cab Sauv 87.415 ± 0.262 15.631 ± 0.039 18.878 ± 0.059 11.76 ± 0.301 40.953 ± 0.287 61.29 ± 0.914 20.44 ± 0.453 5 Coteau St. Paul Frontenac 0.226 ± 0.044 0.181 ± 0.064 0.186 ± 0.035 10.941 ± 0 0.237 ± 0.072 0.234 ± 0.047 0.193 ± 0.057 6 La Halte des Pèlerins Frontenac 57.002 ± 0.572 9.063 ± 0.046 4.221 ± 0.098 0.105 ± 0.013 0.1 ± 0.029 5.638 ± 0.556 2.82 ± 0.046 7 Les Petits Cailloux Frontenac 29.485 ± 0.569 0.332 ± 0.033 0.22 ± 0.022 8.201 ± 0.846 0.335 ± 0.063 0.352 ± 0.009 0.186 ± 0.016 8 Ste-Pétronille Frontenac 17.588 ± 0.133 0.174 ± 0.036 0.148 ± 0.069 9.066 ± 0.109 0.134 ± 0.049 0.167 ± 0.089 0.187 ± 0.023 9 Isle de Bacchus Frontenac 62.538 ± 1.727 145.433 ± 1.164 0.509 ± 0.085 24.088 ± 0.64 0.596 ± 0.024 0.515 ± 0.047 0.507 ± 0.077 10 Ste-Pétronille Frontenac 27.95 ± 0.285 2.642 ± 0.001 1.15 ± 0.037 6.467 ± 0.181 0.16 ± 0.04 1.586 ± 0.162 0.719 ± 0.019 11 Côte de Vaudreuil Frontenac 31.306 ± 0.109 16.501 ± 0.036 1.017 ± 0.033 8.177 ± 0.11 0.202 ± 0.032 1.411 ± 0.147 0.628 ± 0.017 12 Domaine de lavoie Maréchal Foch 47.666 ± 1.831 18.278 ± 0.099 27.724 ± 0.029 24.475 ± 0.235 65.128 ± 0.774 250.654 ± 6.447 25.618 ± 1.155 13 Domaine de lavoie Maréchal Foch 53.872 ± 0.916 24.157 ± 0.051 30.834 ± 0.35 62.08 ± 0.864 73.171 ± 0.889 255.677 ± 1.606 29.475 ± 0.607 14 Isle de Bacchus Maréchal Foch 13.272 ± 0.105 0.108 ± 0.029 0.114 ± 0.006 8.199 ± 0.382 5.834 ± 0.035 14.756 ± 0.078 7.162 ± 0.391 15 Orpailleur Maréchal Foch 27.243 ± 0.417 17.884 ± 0.643 23.308 ± 0.205 4.736 ± 0.249 16.281 ± 0.643 20.954 ± 0.372 11.467 ± 0.076 16 l'ange-Gardien Maréchal Foch 35.758 ± 0.496 15.193 ± 0.122 20.745 ± 0.067 23.251 ± 0.071 39.362 ± 0.21 131.812 ± 0.939 18.412 ± 0.099 17 Les Bromes Maréchal Foch 35.757 ± 0.496 15.193 ± 0.122 20.744 ± 0.067 23.251 ± 0.071 39.361 ± 0.21 131.81 ± 0.939 18.412 ± 0.099 18 ste. Marguerite Marquette 0.235 ± 0.015 4.906 ± 0.339 64.096 ± 0.414 11.727 ± 0.09 8.196 ± 0.369 10.892 ± 0.616 9.782 ± 0.33 19 Alain Breault Marquette 21.356 ± 0.234 0.392 ± 0.041 8.988 ± 0.197 16.769 ± 0.702 0.498 ± 0.072 0.466 ± 0.034 0.458 ± 0.101 20 La Bauge Marquette 0.4 ± 0.044 12.429 ± 2.234 45.484 ± 10.274 22.666 ± 0.416 11.832 ± 0.777 14.231 ± 0.077 7.992 ± 0.429 21 Le Domaine Bergeville Marquette 0.527 ± 0.156 27.753 ± 5.812 60.487 ± 20.677 28.923 ± 0.638 11.725 ± 0.204 13.586 ± 0.237 0.451 ± 0.036 22 Isle De Bacchus Marquette 4.692 ± 0.148 10.045 ± 0.508 46.443 ± 1.495 18.865 ± 0.175 7.843 ± 0.014 9.687 ± 0.096 5.107 ± 0.156 23 Artisans Marquette 3.908 ± 0.194 8.406 ± 0.524 38.993 ± 1.888 15.818 ± 0.601 6.556 ± 0.247 8.106 ± 0.313 4.257 ± 0.208 24 2009 Camelot Merlot 26.173 ± 0.692 0.121 ± 0.036 0.106 ± 0.047 3.261 ± 0.144 0.118 ± 0.034 0.135 ± 0.025 0.122 ± 0.029 25 2009 Inniskillin Merlot 61.816 ± 0.594 19.722 ± 0.005 15.461 ± 0.351 16.435 ± 0.034 18.779 ± 0.087 28.974 ± 0.44 9.774 ± 0.071 26 2009 Sumac Ridge Merlot 73.471 ± 0.7 19.543 ± 0.416 8.246 ± 0.064 6.904 ± 0.444 0.084 ± 0.032 13.618 ± 0.258 4.715 ± 0.712 27 2009 Jackson Triggs Merlot (Box) 63.678 ± 0.839 39.754 ± 0.054 21.823 ± 0 18.479 ± 0.094 8.405 ± 0.148 23.364 ± 0.346 8.971 ± 0.046 28 2008 Inniskillin Pinot Noir 56.967 ± 0.922 179.804 ± 1.09 51.968 ± 0.37 74.872 ± 0.266 31.745 ± 0.596 104.256 ± 1.626 36.776 ± 1.47 29 2011 DB Pinot Noir 14.324 ± 4.111 29.66 ± 8.413 43.964 ± 12.395 18.878 ± 5.386 5.249 ± 2.067 32.748 ± 9.253 14.617 ± 4.242 30 2009 Jackson Triggs Pinot Noir (RF) 38.007 ± 0.104 181.827 ± 2.59 32.142 ± 0.534 20.423 ± 5.591 28.597 ± 0.449 33.165 ± 0.183 12.506 ± 0.775 31 2010 Jackson Triggs Pinot Noir (TT) 28.745 ± 0.269 112.039 ± 0.676 17.436 ± 0.899 8.594 ± 0.083 34.832 ± 0.524 15.194 ± 0.25 6.939 ± 0.685 32 Limousin Sabrevois 2.346 ± 0.106 18.194 ± 0.246 3.647 ± 0.042 1.233 ± 0.029

130

[M-H]- 679.1974 679.1974 679.1974 905.2604 905.2604 905.2604 905.2604 RT 40.331 41.058 42.009 30.383 30.911 31.546 32.814 Sample # Winery; Variety Trimer 5* Trimer 6* Trimer 7* Tetramer* Tetramer1* Tetramer2* Hopeaphenol* 1 2009 Jackson Triggs Cab Sauv 0.083 ± 0.036 0.039 ± 0.014 0.046 ± 0.03 0.043 ± 0.007 9.375 ± 0.193 12.916 ± 0.51 256.266 ± 1.221 2 2010 Copper Moon Cab Sauv 0.085 ± 0.043 13.763 ± 0.089 11.92 ± 0.218 13.254 ± 0.116 24.575 ± 0.272 25.426 ± 0.286 781.292 ± 7.719 3 2010 Inniskillin Cab Sauv 0.144 ± 0.042 7.229 ± 0.381 10.247 ± 0.323 11.462 ± 0.268 8.657 ± 0.209 9.393 ± 0.023 230.639 ± 3.897 4 2010 Jackson Triggs Cab Sauv 4.967 ± 0.758 14.559 ± 0.342 18.84 ± 0.092 12.568 ± 0.032 7.03 ± 0.103 16.715 ± 0.089 215.341 ± 1.009 5 Coteau St. Paul Frontenac 0.155 ± 0.048 0.201 ± 0.048 0.25 ± 0.009 0.197 ± 0.043 0.131 ± 0.023 0.149 ± 0.025 55.175 ± 0 6 La Halte des Pèlerins Frontenac 0.101 ± 0.034 0.063 ± 0.026 0.072 ± 0.064 0.117 ± 0.054 0.099 ± 0.057 0.117 ± 0.043 57.117 ± 1.058 7 Les Petits Cailloux Frontenac 0.214 ± 0.033 0.326 ± 0.037 0.304 ± 0.121 0.235 ± 0.059 0.283 ± 0.035 0.216 ± 0.03 75.897 ± 1.269 8 Ste-Pétronille Frontenac 0.171 ± 0.058 0.171 ± 0.066 0.144 ± 0.067 0.132 ± 0.048 0.144 ± 0.022 0.16 ± 0.029 22.748 ± 0.205 9 Isle de Bacchus Frontenac 0.413 ± 0.11 0.532 ± 0.072 0.481 ± 0.049 0.412 ± 0.051 0.448 ± 0.075 0.484 ± 0.061 38.482 ± 0.571 10 Ste-Pétronille Frontenac 0.212 ± 0.035 0.237 ± 0.011 0.14 ± 0.033 0.115 ± 0.033 0.201 ± 0.058 0.15 ± 0.054 51.124 ± 0.302 11 Côte de Vaudreuil Frontenac 0.214 ± 0.034 0.187 ± 0.034 0.16 ± 0.043 0.25 ± 0.037 0.229 ± 0.016 0.166 ± 0.051 49.897 ± 0.335 12 Domaine de lavoie Maréchal Foch 7.102 ± 0.254 23.288 ± 0.772 35.202 ± 0.609 24.398 ± 1.043 24.769 ± 0.311 20.369 ± 0.716 1805.069 ± 25.362 13 Domaine de lavoie Maréchal Foch 0.157 ± 0.071 28.449 ± 0.096 35.919 ± 0.354 28.733 ± 0.07 28.408 ± 0.198 22.329 ± 0.674 1912.557 ± 4.46 14 Isle de Bacchus Maréchal Foch 0.089 ± 0.012 3.755 ± 0.093 5.193 ± 0.302 0.15 ± 0.005 0.055 ± 0.032 6.263 ± 0.133 81.402 ± 0.417 15 Orpailleur Maréchal Foch 10.867 ± 0.426 7.87 ± 0.394 5.826 ± 0.234 4.713 ± 0.256 6.755 ± 0.557 2.899 ± 0.145 187.307 ± 1.354 16 l'ange-Gardien Maréchal Foch 5.004 ± 0.181 15.602 ± 0.104 20.201 ± 0.216 13.979 ± 0.322 14.606 ± 0.135 12.751 ± 0.047 968.128 ± 3.065 17 Les Bromes Maréchal Foch 5.003 ± 0.181 15.602 ± 0.104 20.201 ± 0.216 13.978 ± 0.322 14.606 ± 0.135 12.751 ± 0.047 968.112 ± 3.065 18 ste. Marguerite Marquette 0.176 ± 0.034 14.261 ± 0.43 0.264 ± 0.069 0.239 ± 0.073 7.696 ± 0.254 7.287 ± 0.294 145.959 ± 7.008 19 Alain Breault Marquette 0.445 ± 0.098 0.41 ± 0.098 0.452 ± 0.105 0.442 ± 0.137 0.508 ± 0.061 0.397 ± 0.051 19.911 ± 0.176 20 La Bauge Marquette 0.393 ± 0.089 0.445 ± 0.119 0.523 ± 0.007 0.367 ± 0.037 9.609 ± 0.98 0.339 ± 0.03 131.659 ± 44.029 21 Le Domaine Bergeville Marquette 0.557 ± 0.138 0.6 ± 0.104 0.555 ± 0.121 0.584 ± 0.084 0.647 ± 0.057 0.469 ± 0.097 171.496 ± 24.078 22 Isle De Bacchus Marquette 0.502 ± 0.028 4.67 ± 0.222 0.418 ± 0.069 0.441 ± 0.087 4.79 ± 0.1 2.314 ± 0.059 118.859 ± 7.61 23 Artisans Marquette 0.532 ± 0.142 3.89 ± 0.236 0.502 ± 0.102 0.554 ± 0.074 3.99 ± 0.174 1.91 ± 0.092 99.848 ± 7.255 24 2009 Camelot Merlot 0.059 ± 0.006 0.073 ± 0.018 0.109 ± 0.061 0.107 ± 0.059 0.094 ± 0.029 0.112 ± 0.041 20.023 ± 0.806 25 2009 Inniskillin Merlot 0.071 ± 0.043 6.68 ± 0.282 8.798 ± 0.171 13.355 ± 0.352 8.31 ± 0.341 18.3 ± 0.138 260.232 ± 1.208 26 2009 Sumac Ridge Merlot 0.065 ± 0.039 0.039 ± 0.03 4.119 ± 0.177 0.072 ± 0.029 0.049 ± 0.059 19.106 ± 0.101 91.621 ± 0.704 27 2009 Jackson Triggs Merlot (Box) 0.095 ± 0.032 8.154 ± 0.005 6.844 ± 0.126 5.611 ± 0.266 12.143 ± 0.315 6.44 ± 0.033 271.549 ± 3.08 28 2008 Inniskillin Pinot Noir 0.231 ± 0.075 11.53 ± 1.486 31.093 ± 0.876 30.571 ± 0.295 37.306 ± 1.74 11.617 ± 0.383 1151.292 ± 13.346 29 2011 DB Pinot Noir 14.004 ± 3.99 6.977 ± 2.026 10.694 ± 3.087 4.254 ± 1.255 12.867 ± 3.679 19.757 ± 5.616 418.837 ± 117.75 30 2009 Jackson Triggs Pinot Noir (RF) 0.301 ± 0.085 12.234 ± 0.469 9.371 ± 0.189 20.48 ± 0.756 21.294 ± 0.437 16.265 ± 0.635 377.498 ± 1.326 31 2010 Jackson Triggs Pinot Noir (TT) 0.226 ± 0.062 5.338 ± 0.085 3.894 ± 0.203 0.217 ± 0.058 13.442 ± 0.259 0.249 ± 0.061 346.905 ± 0.705 32 Limousin Sabrevois

131

[M-H]- 905.2604 905.2604 Total RT 35.879 38.732 Stilbenes Sample # Winery; Variety Tetramer 3* Tetramer 4* 1 2009 Jackson Triggs Cab Sauv 26.503 ± 0.489 6.911 ± 0.066 2116.09 ± 42.417 2 2010 Copper Moon Cab Sauv 94.366 ± 0.589 27.873 ± 0.095 5485.557 ± 100.96 3 2010 Inniskillin Cab Sauv 27.912 ± 0.397 12.173 ± 0.1978794.258 ± 668.235 4 2010 Jackson Triggs Cab Sauv 28.773 ± 0.406 10.679 ± 0.4455831.594 ± 113.833 5 Coteau St. Paul Frontenac 0.173 ± 0.049 0.226 ± 0.076 445.535 ± 1.358 6 La Halte des Pèlerins Frontenac 5.407 ± 0.517 0.115 ± 0.045 1122.064 ± 25.149 7 Les Petits Cailloux Frontenac 0.354 ± 0.041 0.302 ± 0.076 531.809 ± 26.961 8 Ste-Pétronille Frontenac 0.158 ± 0.034 0.158 ± 0.039 476.09 ± 9.378 9 Isle de Bacchus Frontenac 0.512 ± 0.065 0.439 ± 0.135 923.029 ± 20.907 10 Ste-Pétronille Frontenac 1.516 ± 0.168 0.115 ± 0.025 675.647 ± 11.378 11 Côte de Vaudreuil Frontenac 1.348 ± 0.151 0.221 ± 0.061 698.89 ± 10.322 12 Domaine de lavoie Maréchal Foch 140.843 ± 3.595 95.213 ± 1.597943.342 ± 141.333 13 Domaine de lavoie Maréchal Foch 143.437 ± 0.059 96.779 ± 0.979 9622.216 ± 84.291 14 Isle de Bacchus Maréchal Foch 8.332 ± 0.067 7.355 ± 0.054 1972.222 ± 49.059 15 Orpailleur Maréchal Foch 18.81 ± 0.195 11.722 ± 0.513 3502.382 ± 91.573 16 l'ange-Gardien Maréchal Foch 76.084 ± 0.743 51.57 ± 0.069 5726.593 ± 62.913 17 Les Bromes Maréchal Foch 76.082 ± 0.743 51.57 ± 0.069 5726.495 ± 62.912 18 ste. Marguerite Marquette 13.459 ± 0.135 0.244 ± 0.0863990.649 ± 147.733 19 Alain Breault Marquette 0.417 ± 0.016 0.484 ± 0.083 844.558 ± 27.252 20 La Bauge Marquette 11.138 ± 0.437 0.371 ± 0.0754115.651 ± 835.143 21 Le Domaine Bergeville Marquette 0.61 ± 0.044 0.518 ± 0.0475173.646 ± 949.479 22 Isle De Bacchus Marquette 7.096 ± 0.071 0.409 ± 0.13533.412 ± 200.608 23 Artisans Marquette 5.929 ± 0.231 0.456 ± 0.0572968.445 ± 215.078 24 2009 Camelot Merlot 5.217 ± 0.053 2.436 ± 0.034 1066.485 ± 21.541 25 2009 Inniskillin Merlot 43.429 ± 0.19 10.072 ± 0.1855719.227 ± 955.282 26 2009 Sumac Ridge Merlot 13.15 ± 0.216 8.558 ± 0.236 2438.782 ± 50.75 27 2009 Jackson Triggs Merlot (Box) 30.461 ± 0.309 9.092 ± 0.187 3603.331 ± 57.617 28 2008 Inniskillin Pinot Noir 135.335 ± 1.21 42.186 ± 1.99810672.215 ± 185.946 29 2011 DB Pinot Noir 55.848 ± 15.73 15.927 ± 4.535055.041 ± 1425.987 30 2009 Jackson Triggs Pinot Noir (RF) 36.825 ± 0.775 13.22 ± 0.3857688.993 ± 130.991 31 2010 Jackson Triggs Pinot Noir (TT) 37.34 ± 0.125 12.265 ± 0.522 6004.317 ± 90.594 32 Limousin Sabrevois

132

Table A 2: Detailed MS/MS information for identified compounds 1-41.

Peak RT Precursor ion Fragment mass Formula Theoretical fragment Δppm (min) (m/z) mass (m/z) - 1 20.005 227.0725 185.0620 C12H9O2 185.0608 6.48 - 143.0510 C10H7O1 143.0502 5.59

- 2 21.96 227.0717 185.0607 C12H9O2 185.0608 0.54 - 143.0499 C10H7O1 143.0502 2.10

- 3 14.243 243.0659 201.0551 C12H9O3 201.0557 2.98 - 159.0459 C10H7O2 159.0451 5.03

- 4 11.185 389.1235 227.0712 C14H11O3 227.0714 0.88

- 5 13.561 389.1263 227.0711 C14H11O3 227.0714 1.32

- 6 8.209 405.1206 243.0691 C14H11O4 243.0669 2.47 - 201.0574 C12H9O3 201.0564 3.48 - 159.0467 C10H7O2 159.0463 7.54

- 7 21.459 453.1347 359.0933 C22H15O5 359.0925 2.23 - 265.0515 C16H9O4 265.0506 3.40

- 8 23.959 453.1325 359.0907 C22H15O5 359.0925 5.01 - 289.0867 C19H13O3 289.0870 1.04

- 9 25.225 453.1354 359.0940 C22H15O5 359.0925 4.18 - 289.0889 C19H13O3 289.0870 6.57

- 10 26.765 453.1363 359.0933 C22H15O5 359.0944 3.06 - 289.0874 C19H13O3 289.0889 5.19

- 11 31.448 453.1345 435.1238 C28H19O5 435.1238 0.00 - 411.1257 C26H19O5 411.1238 4.62 - 369.1126 C24H17O4 369.1132 1.63 - 359.0927 C22H15O5 359.0925 0.56 - 347.0936 C21H15O5 347.0925 3.17 - 333.0774 C20H13O5 333.0763 1.80 225.0567 225.0557 4.44

- 12 32.202 453.1332 435.1205 C28H19O5 435.1238 7.58 - 411.1246 C26H19O5 411.1238 1.95 - 369.1138 C24H17O4 369.1132 1.63 - 359.0911 C22H15O5 359.0925 3.90 - 347.0935 C21H15O5 347.0925 2.88 - 333.0772 C20H13O5 333.0763 1.20 - 225.0546 C14H9O3 225.0557 4.89

- 13 33.833 453.1376 435.1263 C28H19O5 435.1238 5.75 - 411.1259 C26H19O5 411.1238 5.11 - 369.1142 C24H17O4 369.1132 2.71 - 359.0940 C22H15O5 359.0925 4.18 - 347.0945 C21H15O5 347.0925 5.76 - 333.0785 C20H13O5 333.0763 5.10 - 225.0571 C14H9O3 225.0557 6.22

- 14 34.34 453.1357 435.1233 C28H19O5 435.1238 1.15 - 411.1252 C26H19O5 411.1238 3.41 - 369.1103 C24H17O4 369.1132 7.86 - 359.0912 C22H15O5 359.0925 3.62 - 347.0902 C21H15O5 347.0925 6.63 - 333.0776 C20H13O5 333.0763 2.40 - 225.0568 C14H9O3 225.0557 4.89

- 15 36.127 453.1343 435.1244 C28H19O5 435.1238 1.38 - 411.1224 C26H19O5 411.1238 3.41 - 369.1129 C24H17O4 369.1132 0.81 - 359.0928 C22H15O5 359.0925 0.84 - 347.0917 C21H15O5 347.0925 2.30 225.0556 C14H9O3- 225.0557 0.44

133

- 16 38.789 453.1345 435.1220 C28H19O5 435.1238 4.14 - 411.1222 C26H19O5 411.1238 3.89 - 369.1136 C24H17O4 369.1132 1.08 - 359.0905 C22H15O5 359.0925 5.57 - 333.0741 C20H13O5 333.0763 8.41

- 17 39.705 453.1332 435.1218 C28H19O5 435.1238 4.60 - 411.1222 C26H19O5 411.1238 3.89 - 369.1118 C24H17O4 369.1132 3.79 - 359.0908 C22H15O5 359.0925 4.73 - 333.0762 C20H13O5 333.0763 2.10

- 18 16.228 469.1302 451.119 C28H19O6 451.1187 0.67 - 427.1147 C26H19O6 427.1182 9.37 - 385.1082 C24H17O5 385.1081 0.26 - 375.0899 C22H15O6 375.0874 6.67 - 359.0918 C22H15O5 359.0925 1.95 - 347.0921 C21H15O5 347.0925 1.15 - 265.0492 C16H9O4 265.0506 5.28

- 19 26.482 469.1301 451.1178 C28H19O6 451.1187 2.00 - 427.1188 C26H19O6 427.1182 0.23 - 385.1087 C24H17O5 385.1081 1.56 - 375.0893 C22H15O6 375.0874 5.07 - 359.0932 C22H15O5 359.0925 1.95 - 347.0913 C21H15O5 347.0925 3.46 - 333.0793 C20H13O5 333.0763 7.21 - 241.0495 C14H9O4 241.0506 4.56

- 20 27.93 469.1307 451.1188 C28H19O6 451.1187 0.22 - 427.1224 C26H19O6 427.1182 8.66 - 385.1086 C24H17O5 385.1081 1.30 - 375.0883 C22H15O6 375.0874 2.40 - 359.0942 C22H15O5 359.0925 4.73 - 347.0914 C21H15O5 347.0925 3.17 - 333.0782 C20H13O5 333.0763 3.90 - 241.0514 C14H9O4 241.0506 3.32

- 21 11.303 471.1447 377.1024 C22H18O6 377.1030 1.59 - 349.1089 C21H18O5 349.1081 2.29 - 255.0667 C15H12O4 255.0662 1.84 - 121.0290 C7H6O2 121.0294 3.30

- 22 17.626 471.1456 349.1089 C21H18O5 349.1081 2.29 - 255.0658 C15H12O4 255.0662 1.57 - 121.0294 C7H6O2 121.0294 0.00

- 23 18.251 471.1477 387.1262 C24H19O5 387.1237 6.46 - 377.1091 C22H18O6 377.1030 1.59 - 349.1065 C21H18O5 349.1081 4.58 - 255.0680 C15H12O4 255.0662 7.06 - 121.0284 C7H6O2 121.0294 8.26

- 24 19.524 471.1463 349.1121 C21H18O5 349.1081 0.57 - 255.0691 C15H12O4 255.0662 9.41 - 241.0496 C14H9O4 241.0506 4.15 - 121.0295 C7H6O2 121.0294 0.83

- 25 23.276 485.1248 467.1153 C28H19O7 467.1130 4.92 - 443.1095 C26H19O7 443.1130 7.90 - 401.1012 C24H17O6 401.1024 2.99 - 375.0895 C22H15O6 375.0868 7.20 - 363.0878 C21H17O6 363.0868 2.75 - 357.0795 C22H13O5 357.0763 8.96 - 333.0778 C20H13O5 333.0762 4.80 - 265.0506 C16H9O4 265.0506 2.26 - 241.0508 C14H9O4 241.0506 3.32

- 26 19.64 485.1635 453.1360 C28H21O6 453.1344 3.53 - 391.1211 C23H19O6 391.1187 6.14 - 359.0951 C22H15O5 359.0925 7.24 - 289.0895 C19H13O3 289.0870 8.65 - 255.0676 C15H12O4 255.0663 5.10 134

- 187.0777 C12H11O2 187.0765 6.41

- 27 21.088 485.1594 453.1338 C28H21O6 453.1344 1.32 - 391.1182 C23H19O6 391.1187 1.28 - 359.0914 C22H15O5 359.0925 3.06 - 289.0864 C19H13O3 289.0870 2.08 - 255.0655 C15H12O4 255.0663 3.14 - 187.0762 C12H11O2 187.0765 1.60

- 28 25.202 615.1878 453.1339 C28H21O6 453.1344 1.10 - 411.1262 C26H19O5 411.1238 5.84 - 359.093 C22H15O5 359.0925 1.39 - 347.0957 C21H15O5 347.0925 9.22

- 29 18.392 615.1883 453.1371 C28H21O6 453.1344 5.96 - 359.0958 C22H15O5 359.0925 9.19 - 289.0892 C19H13O3 289.0870 7.61

- 30 30.044 679.1981 585.1538 C36H26O8 585.1555 2.91 - 573.1569 C35H26O8 573.1555 2.44 - 491.1143 C30H19O7 491.1136 1.43 - 479.1178 C29H20O7 479.1136 8.77 - 385.0726 C23H13O6 385.0717 2.34

- 31 35.372 679.1978 661.1817 C42H30O8 661.1868 7.71 - 637.1903 C40H30O8 637.1868 5.49 - 585.1596 C36H26O8 585.1555 7.01 - 573.1592 C35H26O8 573.1555 6.46 - 555.1461 C35H24O7 555.1449 2.16 - 479.1151 C29H20O7 479.1136 3.13 - 451.1176 C28H19O6 451.1187 2.44 - 357.079 C22H13O5 357.0768 6.16 - 345.0792 C21H13O5 345.0768 6.95

- 32 36.322 679.1984 661.1864 C42H30O8 661.1868 0.60 - 637.1926 C40H30O8 637.1868 9.10 - 585.1572 C36H26O8 585.1555 2.91 - 573.1561 C35H26O8 573.1555 1.05 - 555.1455 C35H24O7 555.1449 1.08 - 479.1169 C29H20O7 479.1136 6.89 - 451.1168 C28H19O6 451.1187 4.21 - 357.0744 C22H13O5 357.0768 6.72 - 345.0792 C21H13O5 345.0768 6.95

- 33 37.661 679.1985 661.1834 C42H30O8 661.1868 5.14 - 637.1827 C40H30O8 637.1868 6.43 - 585.154 C36H26O8 585.1555 2.56 - 573.1606 C35H26O8 573.1555 8.90 - 555.1503 C35H24O7 555.1449 9.73 - 479.1156 C29H20O7 479.1136 4.17 - 451.1164 C28H19O6 451.1187 5.10 - 357.0743 C22H13O5 357.0768 7.00 - 345.0764 C21H13O5 345.0768 1.16

- 34 36.866 679.2000 661.1917 C42H30O8 661.1868 7.41 - 637.1869 C40H30O8 637.1868 0.16 - 585.1585 C36H26O8 585.1555 5.13 - 573.159 C35H26O8 573.1555 6.11 - 555.1418 C35H24O7 555.1449 5.58 - 479.1123 C29H20O7 479.1136 2.71 - 451.121 C28H19O6 451.1187 5.10 - 357.0769 C22H13O5 357.0768 0.28 - 345.0786 C21H13O5 345.0768 5.22

- 35 39.152 679.1916 661.1853 C42H30O8 661.1868 2.27 - 637.1805 C40H30O8 637.1868 9.89 - 585.1595 C36H26O8 585.1555 6.84 - 573.1532 C35H26O8 573.1555 4.01 - 555.1435 C35H24O7 555.1449 2.52 - 479.1116 C29H20O7 479.1136 4.17 - 451.1223 C28H19O6 451.1187 7.98 357.075 C22H13O5- 357.0768 5.04 135

- 345.0784 C21H13O5 345.0768 4.64

- 36 39.835 679.1998 661.1858 C42H30O8 661.1868 1.51 - 637.1925 C40H30O8 637.1868 8.95 - 585.1535 C36H26O8 585.1555 3.42 - 573.1502 C35H26O8 573.1555 9.25 - 555.1468 C35H24O7 555.1449 3.42 - 479.1114 C29H20O7 479.1136 4.59 - 451.1147 C28H19O6 451.1187 8.87 - 357.0792 C22H13O5 357.0768 6.72 - 345.0764 C21H13O5 345.0768 1.16

- 37 40.347 679.2001 661.1833 C42H30O8 661.1868 5.29 - 637.1916 C40H30O8 637.1868 7.53 - 585.1598 C36H26O8 585.1555 7.35 - 573.1551 C35H26O8 573.1555 0.70 - 555.1419 C35H24O7 555.1449 5.40

- 479.1154 C29H20O7 479.1136 3.76 - 451.1151 C28H19O6 451.1187 7.98 - 357.0781 C22H13O5 357.0768 3.64 - 345.0788 C21H13O5 345.0768 5.80

- 38 29.734 905.2577 887.2585 C56H39O11 887.2498 9.81 - 811.2193 C50H35O11 811.2185 0.99 - 799.2186 C49H35O11 799.2185 0.13

- 717.1738 C44H29O10 717.1766 3.90 - 705.1814 C43H29O10 705.1766 6.81 - 699.1607 C44H27O9 699.1660 7.58 - 611.1370 C37H24O9 611.1347 3.76

- 39 31.695 905.2573 811.2157 C50H35O11 811.2185 3.45 - 717.1720 C44H30O10 717.1766 6.41 - 611.1322 C37H24O9 611.1347 4.09

- 451.1171 C28H19O6 451.1188 3.77 - 359.0915 C22H15O5 359.0928 3.62 - 265.0494 C16H9O4 265.0512 6.79

- 40 34.718 905.2563 811.2170 C50H35O11 811.2185 1.85 - 793.2095 C50H33O10 793.2079 2.02 - 717.1741 C44H30O10 717.1766 3.49 - 705.1767 C43H29O10 705.1766 0.14

- 611.1332 C37H24O9 611.1347 2.45

- 41 37.494 905.2620 811.2108 C50H35O11 811.2185 9.49 - 717.1757 C44H30O10 717.1766 1.25 - 451.12 C28H19O6 451.1188 2.66 - 359.0941 C22H15O5 359.0928 3.62 - 265.0489 C16H9O4 265.0512 8.68

136