The Impacts of Ion Exchange on the Volatile Composition of Wine and the Characterization of Missouri Wine Aromas

Home , Catawba (grape)

THE IMPACTS OF ION EXCHANGE ON THE VOLATILE COMPOSITION OF WINE AND THE CHARACTERIZATION OF MISSOURI WINE AROMAS

______

A Thesis

presented to

the Faculty of the Graduate School

at the University of Missouri-Columbia

______

In Partial Fulfillment

of the Requirements for the Degree

Master of Science

______

BRIAN WAYNE

Dr. Misha Kwasniewski, Thesis Supervisor

DECEMBER 2015

The undersigned, appointed by the Dean of the Graduate School, have examined the thesis entitled

THE IMPACTS OF ION EXCHANGE ON THE VOLATILE COMPOSITION OF WINE AND THE CHARACTERIZATION OF MISSOURI WINE AROMAS presented by Brian Wayne, a candidate for the degree of Master of Science, and hereby certify that, in their opinion, it is worthy of acceptance.

______Professor Misha Kwasniewski

______Professor Ingolf Gruen

______Professor Karthik Panchanathan

ACKNOWLEDGEMENTS

I would like to thank Dr. Misha Kwasniewski for his endless guidance on this thesis. By working closely with him for the past few years, he has added new dimensions to my understanding of analytical chemistry, chromatography, and mass spectrometry. He tirelessly led me to answer questions and question answers through research and critical thinking, going above and beyond a typical advisor. Thanks to Dr. Ingolf Gruen and Dr. Karthik Panchanathan for proofreading this thesis and providing research ideas. Thanks to Connie Liu for passing on her knowledge of GC-MS maintenance and troubleshooting, but also stepping in to keep the instruments running and calibrated when problems became overwhelming.

I would like to thank my fellow GWI graduate students for providing pleasant lunchtime conversation, putting up with my scattered messy desk, and always maintain a positive work environment. Specifically, I would like to thank Megan Wasielewski for her collaborative efforts on the ion exchange project through the design and implementation of our treatment system and for providing vital information through her sensory analysis and organic acid experiments. I would like to thank Dr. Rob Walker, Keity Walker, Stephanie Grau, Paula Grossi, Courtney

Duncan, Connie Liu and Abhi Datta who all provided logistical assistance in the form of housing and transportation, facilitating my continued thesis work. I want to thank Kameron Oser for writing a Python script that enhanced my chromatogram data analysis. Mary Schneier also helped perform some of the initial experiments in the ion exchange research.

I need to thank Jackie Harris, Dr. Arianna Bozzolo, Ashley, Tim, Britta, Spencer,

Harrison and the rest of the vineyard crew for spending so much of their time in our vineyards, bringing us back excellent fruit while I spent most of my time in graduate school sitting in an air- conditioned lab. iii

Finally, thanks to my parents, Janice and Paul, who have over the past 26 years instilled in me all the attributes required for successfully completing this thesis.

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ...... ii

LIST OF TABLES ...... vii

LIST OF FIGURES ...... viii

CHAPTER 1: INTRODUCTION TO ION EXCHANGE IN THE WINE INDUSTRY

1. INTRODUCTION ...... 1

2. WINE AROMA ...... 2

3. CONTINENTAL VITICULTURE ...... 5

4. WINE ACIDITY ...... 6

5. POTASSIUM ...... 8

6. ION EXCHANGE ...... 9

6.1 Principles ...... 10

6.2 Operation ...... 10

6.3 Wine Quality Concerns ...... 12

REFERENCES ...... 16

CHAPTER 2: IMPACTS OF ION EXCHANGE ON CATIONS, ORGANIC ACIDS, AND ETHANOL IN MISSOURI WINES

ABSTRACT ...... 23

1. INTRODUCTION ...... 23

1.1 Wine Acidity ...... 23

1.2 Potassium ...... 25

1.3 Ion Exchange ...... 26

2. MATERIALS AND METHODS ...... 29

2.1 Winemaking...... 29 v

2.2 Ion Exchange Treatment ...... 31

2.3 Analysis ...... 32

3. RESULTS ...... 33

4. DISCUSSION ...... 34

5. CONCLUSION ...... 37

REFERENCES ...... 38

TABLE AND FIGURES ...... 42

CHAPTER 3: COMPARISON OF MUSCAT ODORANTS PRESENT IN VALVIN MUSCAT WITH OTHER AROMATIC WHITE WINES

ABSTRACT ...... 47

1. INTRODUCTION ...... 48

1.1 Varietal Aroma...... 48

1.2 Hybridization Impacts on Wine Aroma...... 51

1.3 Valvin Muscat ...... 52

2. MATERIALS AND METHODS ...... 54

2.1 Chemicals ...... 52

2.2 Wines ...... 52

2.3 Basic Analysis ...... 53

2.4 Volatile Analysis ...... 54

2.4 Data Compilation...... 55

3. RESULTS ...... 55

3.1 Basic Analysis ...... 55

3.2 Odorants in Valvin Muscat ...... 56

3.3 Odorants Among Cultivars ...... 57

4. DISCUSSION ...... 57

5. CONCLUSION ...... 60 vi

REFERENCES ...... 61

TABLES AND FIGURES ...... 66

CHAPTER 4: IMPACTS OF ION EXCHANGE ON VOLATILE AROMA COMPOUNDS IN MISSOURI WINES

ABSTRACT ...... 68

1. INTRODUCTION ...... 69

1.1 Wine Acidity ...... 69

1.2 Ion Exchange ...... 70

2. MATERIALS AND METHODS ...... 72

2.1 Winemaking...... 72

2.2 Ion Exchange Treatment...... 74

2.3GC-MS Analysis...... 75

2.4 Data Processing……………...... 77

3. RESULTS…………………… ...... 78

3.1 pH Dependancy…...... 78

3.2 XCMS Detection…...... 78

3.3 Manual Detection…...... 80

4. DISCUSSION ...... 81

4.1 pH Dependancy ...... 81

4.2 XCMS Detection...... 82

4.2 Manual Detection...... 82

5. CONCLUSION ...... 83

REFERENCES ...... 84

TABLES AND FIGURES ...... 88

APPENDIX ...... 96 vii

LIST OF TABLES

Table Page

2.1 Concentration (mg/L) of Potassium, Calcium, and Magnesium Cations During 2013 Chambourcin Ion Exchange Treatment ...... 47

2.2 Effects of Ion Exchange Treatment of Percent Alcohol (2013) ...... 47

3.1 Basic chemistry of wines selected for volatile analysis ...... 66

3.2 Concentration (ug/L) of important Muscat odorants in selected Valvin Muscat wines ...... 67

3.3 Mean concentrations (ug/L) of important Muscat odorants in selected white wines ...... 67

4.1 Peak areas of standards extracted at different pH values ...... 88

4.2 Ion features determined by XCMS analysis to be significantly different between control and ion exchange treatments ...... 92

4.3 Concentration (ug/L) of volatile aroma compounds in wines treated through ion exchange ...... 93

4.4 Relative peak areas of 2013 and 2014 Chambourcin, Norton, and Valvin Muscat wines treated using ion exchange ...... 93

4.5 Relative peak area percentages of compounds significantly impacted by ion Exchange ...... 94

4.6 Concentration (ug/L) of Muscat Odorants in Valvin Muscat Treated with Ion Exchange ...... 95

viii

LIST OF FIGURES

Figure Page

2.1 Wine pH for each bed volume (330ml) over the course of ion exchange treatments in 2013 (A) Valvin Muscat,(B) Chambourcin and (C) Norton wines. Error bars indicate 95% confidence intervals for three replicate treatments...... 43

2.2 Concentration (g/L) of tartaric, malic, and lactic acids in each collected fraction during ion exchange treatment in (A) 2013 Valvin Muscat, (B) 2013Norton, (C) 2013 Chambourcin, and (D) Syrah wines. Error bars indicate 95% confidence Intervals for three replicate trials. Malic acid was not present in Norton or Chambourcin wines due to malolactic fermentation ...... 45

4.1 PCA plots of 3 extraction by 3 injection (9 total) replicates of (A) 2013 (B) 2014 Chambourcin, (C) 2013 (D) 2014 Norton, (E) 2013, (F) 2014 Valvin Muscat chromatograms analyzed by XCMS software...... 89

A.1 Percent ethanol reductions observed in a 100ml model wine solution (12.5%ABV, 5g tartaric acid, pH 3.9 adjusted with 5M NaOH) when introduced to various Amounts of Amberlite IR-120 resin and agitation for ten minutes………………...96

CHAPTER 1: INTRODUCTION TO ION EXCHANGE IN THE WINE INDUSTRY

1. INTRODUCTION

Wine varies significantly in chemical composition depending on the location and

manner in which it was produced. These variations have significant impacts on a wine’s price and demand by consumers. Wine aroma plays a central role in quality determinations. While some aspects of wine aroma are manipulated, both intentionally and unintentionally, through decisions made by viticulturists and winemakers, other aspects of wine aroma cannot be easily changed. They are inherent to particular cultivars and constitute varietal aroma. Wine producers in the state of Missouri are significantly disadvantaged by climactic factors that largely restrict viticulture to the planting of hybrid cultivars which produce wines generally do not have as large a market as pure vinifera- based European cultivars due to their flavor profiles. Specifically, wines made from hybrid cultivars often lack sufficient tannin, contain undesirable flavors, and are more likely to possess a pH/TA imbalance. Research in the field of grape breeding seeks to produce new cultivars that do not exhibit these negative characteristics. Viticultural and enological research attempts to improve wine quality in current hybrid cultivars.

Resin-based ion exchange technology has been successfully applied to correcting pH/TA imbalance issues found in hybrid cultivars. However, its adoption by the Missouri wine industry and elsewhere has been hindered by concerns regarding its tendency to strip flavor and aroma, thus reducing wine quality. Research regarding the impacts of ion exchange on wine chemistry, specifically flavor and aroma, is conflicting 2

and insufficient. A thorough analysis into the impacts of ion exchange on aroma

compounds present in hybrid wines will lead Missouri winemakers to make better-

informed decisions about whether to adopt this technology.

2. WINE AROMA

Wine aroma is determined by interactions among grape cultivar, terroir, as well as

viticultural and winemaking practices. Aromas can be described as primary, secondary or

tertiary for those derived or released from the grape, fermentation or additives

respectively. For the most part tertiary flavors are process influenced whereas primary

and secondary aromas are more heavily influenced by the fruit. Furan and furanone are

examples of tertiary aromas which are extracted and formed in wines from interactions

with wood, with cultivar playing a minimal role in determining the type and

concentration of furans present in a wine (Robinson et al., 2013). Other compounds, like

esters, are formed through a complex of factors occurring during fermentation (yeast

strain, temperature, redox state) along with ester precursors determined by cultivar (Soles

et al., 1982; Farina et al., 2012). Varietal aromas are derived from specific compounds

whose origins can be traced to particular grape cultivars. They define the unique aroma

profile that some wines exhibit (Styger et al., 2011), some of which are present in the

fruit while others are bound and only released during fermentations and ageing (Bisotto

et al., 2015). Four classes of compounds are responsible for the varietal aroma observed

in many wines; C13 norisoprenoids, thiols, methoxypyrazines, and monoterpenes.

C13 norisoprenoids contribute to the varietal aroma of some white wines, particularly Rieslings and Gewurztraminers, but are also present in many other wines at lower, but above threshold concentrations (Martinez-Gil et al., 2013; Yuan & Qian, 3

2016). They are formed as degradation products of carotenoids during berry maturation,

bound to glucose molecules in the fruit and released during fermentation, and wine

ageing (Mendes-Pinto, 2009). Sunlight exposure at certain stages of grape maturation

can greatly increase carotenoid production, leading to increased concentrations of some

C13 norisoprenoids such as TDN (1,1,6 trimethyl-1,2 dihydronapthalene) and vitispirane in resulting wines (Kwasniewski et al., 2010). Currently, about six C13 norisoprenoids

are considered important in wine aroma; β-damascenone, β-ionone, vitispirane, actinidiol,

TDN and riesling acetal (Mendes-Pinto, 2009). The abundance of specific C13 norisoprenoids is heavily impacted by grape variety. Most notably, Riesling aroma is dominated by TDN, vitispirane, β-damascenone and riesling acetal, along with linalool, linalool oxides and 2-phenylethanol (Chisolm et al., 1994, 1995; Mendes-Pinto, 2009;

Meyers et al. 2013; Sacks et al., 2012), while Gewurztraminer aroma is attributed to β- damascenone, in addition to monoterpenes (cis-rose oxide and geraniol) and esters (Ong and Acree 1999).

Thiols are a class of compounds containing sulfhydryl (-SH) groups. Several thiols define the aroma profile of Sauvignon blanc and Chenin blanc wines; 4MMP (4- mercapto-4-methylpentan-2-one), 3MH (3-mercaptohexan-1-ol), and 3MHA (3- mercaptohexan-1-ol acetate) (Roland et al., 2011). Obtaining information about thiol concentrations in wine is difficult, since they are odor-active at very low concentrations

(ng/L). Determining viticultural influences on thiol concentration is also difficult since they are formed during fermentation from unknown precursors.

Methoxypyrazines contribute a ‘vegetative’ varietal aroma to Sauvignon blanc and

Cabernet Sauvignon wines (Allen et al., 1991; Sala et al., 2004). Several 4

methoxypyrazines, particularly IBMP (3-Isobutyl-2-methoxypyrazine), SBMP (3-sec- butyl-2-methoxypyrazine), and IPMP (3-isopropyl-2-methoxy-pyrazine) can add to a wine’s aroma profile at concentrations as low as 2ng/L (Sala et al., 2004). While some amount of vegetative character is considered desirable, especially in Sauvignon blanc wines, restricting excessive concentrations is important when growing these cultivars.

Methoxypyrazines concentrations are higher in grapes harvested prematurely and grapes grown in cool climates that have less exposure to sunlight (Sala et al., 2004).

Monoterpenes are a class of aroma compounds found to some degree in most grape

varieties. However, they are closely associated with Muscat wines. About 50 different

monoterpenes have been found in winegrapes in both free and glycosidically bound

forms, imparting a variety of floral and citrus flavors and aromas (Ebeler & Thorngate,

2009; Mateo &Jimenez, 2000). They are synthesized in the grape berry from acetyl co-

enzyme A (Mateo & Jimenez, 2000). During berry maturation, concentrations of both

free and bound monoterpenes vary dramatically. Most compounds achieve maximum

concentrations at or after harvest maturity, except for geraniol where concentrations peak

near berry set, then diminish steadily (Wilson et al., 1984). Berry sunlight exposure will

increase monoterpene concentrations (Reynolds & Wardle, 1989). Monoterpene

concentrations are not influenced by nitrogen fertilization (Webster et al., 1993).

In the winery, monoterpene extraction from fruit can be increased by carbonic

anaerobiosis or skin contact before pressing, and by using glucosidase enzymes to cleave

glycosylated precursors (Bitteur et al., 1992; Palomo et al., 2006). Monoterpene

degradation begins quickly in wine and significant losses can be seen within months

(Rapp & Mandery, 1986). Numerous studies have confirmed the findings of Ribereau- 5

Gayon et al. (1975) who determined that the typical aroma of Muscat grapes and wines

is attributed to roughly eight monoterpenes; linalool, nerol, geraniol, α-terpineol, and four oxides of linalool (Bordiga et al., 2013; Palomo et al., 2006; Skinkis, 2008) .

Monoterpenes also contribute to the varietal aroma of other aromatic white wines.

3. CONTINENTAL VITICULTURE

The state of Missouri has a rich history of wine production and research.

However, the climate in Missouri puts significant constraints on the types of grapevines that will produce an economically viable quantity of fruit. Viticulturists in Missouri must deal with fluctuating winter temperatures that leave vines susceptible to winter injury, high summer temperatures that can negatively impact fruit quality, and high humidity that promotes fruit and canopy rot (Gu et al., 2000). Most European (Vitis vinifera) grapevines cannot withstand these conditions (Sun et al., 2011). Consequently, alternative grape cultivars have been selected and bred that are better suited to Missouri’s climate. Their success in this continental climate comes from the incorporation of various amounts of Native American grapevine genetics, which have been bred into them.

Some of these cultivars are heavily derived from Native American species, making them well adapted to Missouri’s climate. Concord, Catawba and Niagara cultivars fit into this category since they are strictly native V. labrusca species However, the grapevines most easily cultivated in continental climates do not produce fruit that is ideal for some styles of winemaking. The wines produced by these varieties lack tannin, produce undesirable ‘foxy’ aromas and flavors, low alcohol content and contain high concentrations of potassium leading to pH/TA imbalance (Morris & Cawthon, 1982). 6

Attempts to create hybrid grapevine cultivars that possess the abilities to both

withstand the harsh conditions of continental climates and also produce wines with

characteristics more desirable to consumers have been occurring since Europeans first

began colonizing the Eastern United States (Alleweldt, 1997). Some progress has been

made in that some hybrid cultivars such as ‘Chambourcin’, and ‘Marechal Foch’ do not

possess the typical ‘foxy’ character used to describe V. labrusca cultivars because they

have been bred using V. riparia, V. aestivalis, and V. rupestris species, which were determined to have more agreeable flavors (Sun et al., 2011). However, these cultivars still exhibit some of the challenges to quality common in Native American species.

Hybrid breeding programs continue to release new cultivars with improved wine quality.

However, global acceptance of hybrid cultivars remains weak and V. vinifera cultivars are still in demand (Bisson et al., 2002).

4. WINE ACIDITY An important element of wine stability is pH. A sufficiently low pH is necessary in order for SO2 to function effectively as an antimocrobial (Zoecklin et al., 1995). It also

partially determines the perception of sourness, which is an important consideration in

consumer preference. Titratable acidity (TA) is also an important consideration in the

perception of wine sourness, with high TA values correlating with higher perceptions of

acidity (Plane et al., 1980).

Low pH is necessary for the microbial protection of wine because it inhibits the

growth of many types of spoilage organisms. Sulfur dioxide is the main antimicrobial

used in wine production (King et al., 1982), and is generally added as potassium

metabisulfite. In wine solutions, sulfur dioxide (SO2) exists in pH dependant equilibrium 7

- 2- with bisulfite (H2SO3 ) and sulfite (SO3 ), which do not posses the antimicrobial benefits of SO2 (King et al., 1982). The pKa of SO2 is 1.77. In other words, at pH 1.77, fifty

percent of the sulfur will be in the form SO2, while the other fifty percent will be in the

- - form HSO3 . At typical wine pH values, the equilibrium between SO2 and HSO3 is

- heavily in favor of the HSO3 ion, both of which do not exhibit antimicrobial or

antioxidant activity. At pH 3, about eight percent of the sulfur is in the SO2 form. While

at pH 4, about one percent is available. For this reason, acidifying high pH wines greatly

increases the antimicrobial effects of sulfur dioxide.

When potassium metabisulfite has not been added to wines, pH still significantly

affects microbial growth. Malolactic baterica such as Leuconostic oenos grow slower in

wines at pH 3.2 than wines at pH 3.8, retarding their ability to conduct malolactic

fermentations from 14 days to 164 days (Bousbouras & Kunkee, 1971). Other malolactic

bacteria such as Pediococcus and Lactobacillus are unable to grow below pH 3.5

(Wibowo et al., 1985).

Low pH values do not always correlate with high TA. Under conditions where

excessive cation concentrations are present, high pH/ TA wines will occur. Reducing the

pH of these wines to levels necessary for the antimicrobial effectiveness of SO2 is difficult. The commonly used method of adding tartaric acid to reduce pH is sometimes impossible since tartaric acid precipitates from the wine solution when excessive salt levels are present. In instances where the tartaric acid does not precipiate, it’s addition further increases TA to levels which are percieved as overly sour by consumers (Jackson,

2014; Walker et al., 2002).

5. POTASSIUM

High pH/TA wines are commonly encountered in wines made from hybrid grape

varietals, particularly Norton, which often contain excessive concentrations of potassium

cations (K+)(Morris et al., 1983). An inverse relationship exists between K+ and

hydrogen ions (H+) in grape juice, so that any increase in K+ concentration results in a

pH increase (Boulton, 1980). The K+ content of a wine is influenced by numerous factors beginning with soil. Morris et al., (1983) determined that potassium fertilization resulted in direct exchange of K+ for H+ in grape berries, increasing pH substantially.

However, potassium uptake by roots is influenced by the presence of other cations, particularly sodium (Na+) (Chow et al., 1990), whereby, high Na+ concentrations will compete with K+ for uptake into the plant. Rootstocks also vary considerably in their tendency to uptake K+ (Mpelasoka et al., 2003).

Once inside grapevines, K+ is translocated from older tissues to growing tissues, which serve as K+ sinks (Mpelasoka et al., 2003). Leaves, which begin growing at

budbreak, become the initial K+ sinks. During veraison, berries become the predominate

K+ sink as cells inside the grape expand and ripen. When grapes reach harvest maturity,

K+ is the dominant cation present in grape berries, while calcium, magnesium, sodium,

iron, copper, and Manganese are also present at lower concentrations (Hrazdina et al.,

1984). This effect is magnified in over-ripe grapes as K+ concentrations continue to

increase after harvest maturity (Iland & Coombe, 1988; Kliewer et al., 1967).

Sunlight exposure is a critical factor in determining K+ concentrations, whereby

shading either clusters or canopy leads to increased fruit K+(Rojas-Lara & Morrison,

1989). Irrigation also increases K+ in berries (Freeman & Kliewer, 1983). While 9

substantial research has been conducted on factors influencing K+ movement in

grapevines, more research is required to determine specific viticultural practices that

reduce K+ in grape berries (Mpelasoka et al., 2003).

While mitigation of high K+ issues should be attempted with in the vineyard

(Mpelasoka et al, 2003), high potassium grapes still frequently enter Missouri wineries

due to the limited effectiveness of vineyard best pratices on solving this problem.

Several methods of pH/TA imbalance correction are often employed to reduce wine pH

without further increasing the TA, such as double salt-deacidification, electrodialysis, cold-stability manipulation, and ion exchange. Each technique is capable of lowering wine pH with minimal influence on TA. With the exception of cold-stability manipulation, they all require specialized equipment, can be complex, and require significant time and space commitments (Garcia-Ruiz et al, 1991; Cole and Boulton,

1989; Nagel et al. 1988; Soares et al. 2009).

6. ION EXCHANGE

Ion exchange resins have been used in wine remediation since the 1950’s. In addition to adjusting pH, they have been used for tartrate stabilization, concentrating grape must, and to remove metals, proteins, volatile acidity, and color ( Bazinet &

Firdaous, 2011;Benitez et al, 2002a; Bonorden et al., 1986; Esau & Amerine, 1966;

Lasanta & Gomez, 2012; Mira et al., 2006). Ion exchange has successfully been applied to correcting high pH/TA wines by removing K+ and replacing it with H+, thereby reducing pH by the direct addition of H+ (Lasanta et al., 2013; Walker et al., 2002).

Resulting wines are perceived as less acidic than those remedied by tartaric acid addition because acid anions, which are perceived as acidic by humans, are not added in the 10

process (Plane et al., 1980). Resin-based ion exchange treatment systems are inexpensive and widely available due to their use in water treatment systems (Carreon-Alvarez et al.,

2011; Qiang & Yan, 2013; Cuvelier, 1995). A cost analysis performed by Walker et al.,

(2004) shows that resin based cation exchange systems are more cost effective than membrane-based cation exchange systems for pH/TA correction.

6.1 Principles

The principles of ion exchange were first discovered in 1850 by measuring the flow of cations through soil, which acts as a natural ion exchange medium (Nasef & Ujang,

2012). As ions in water enter soil, exchanges of different ions will take place in binding sites naturally present in the soil. Synthetic resins that were able to mimic the exchange properties of soil were synthesized in the early 20th century. At first they were used as a

method of water softening (Nasef & Ujang, 2012). When applied to the correction of

high pH/TA wines, an efficient exchange of hydrogen ions present on the resin surface

for potassium ions present in the wine matrix is desired.

6.2 Operation

Three components are necessary for a potassium-hydrogen exchange to occur; resin

beads, column, and an acid regenerant. Many commercial resins are available,

manufactured in various shapes and sizes ranging from 0.3 to 1.2mm (Nasef and Ujang,

2012). Within these sizes, there are also specifications for size distribution ranges

(Zaganiaris, 2011). Two common bead shapes include gel-type and macroreticular (or

macroporous). One important difference between the two bead shapes is that the gel-type

requires swelling by an appropriate solvent while macroreticular beads do not require 11

swelling to function. The backbone structure of resin beads generally consist of

polystyrene, although polyphenol, acrylic copolymer, or phenol formaldehyde condensate

are also used (Esau & Amerine, 1966). Within these synthetic structures, divinylbenzene

is added at different concentrations to promote crosslinking, with highly crosslinked

structures displaying fewer nonpolar interactions with nonpolar solute compounds,

slower exchange rates, lower flow-rates, and higher cation exchange capacity

(Zaganiaris, 2011). Some research has been conducted to determine the impacts of resin- type on wine quality when correcting high pH/TA wines, with small differences determined among the resins employed (Walker et al., 2002). Amberlite IR-120 has been the most widely used resin in high pH/TA wine correction research (Walker et al., 2002,

Lasanta et al, 2013).

The exchange of K+ and H+ on resin beads most often takes place inside a resin column. The dimensions of the resin column and layout design impact the overall functioning of the system. The following guidelines have been suggested for an efficient exchange to occur. Staying within these dimensions lowers the possibility of

“maldistribution” or uneven flow of liquid through the resin (Inglezakis, 2012). Briefly, if bed height is greater than five times the bed diameter and greater than fifty times the particle diameter, and if bed diameter is greater than twelve times the particle diameter,

then a thorough and efficient exchange of K+ and H+ will take place. In addition, Esau

& Amerine (1966) state that the cation exchange capacity of any resin can be optimized by using a column that has a diameter 20 times larger than the diameter of the resin. The resin bed should also occupy only 60% of the volume of the column to ensure successful backwashing. 12

After the exchange of H+ and K+ takes place, the resin is exhausted of H+ and must

be regenerated with acid before it can be used again. Some authors recommend dilute

hydrochloric acid (Esau & Amerine, 1966; Palacios, 2001). However, chloride ions in hydrochloric acid are extremely corrosive to stainless steel even at dilute concentrations.

If stainless steel is used in ion exchange operations, dilute sulfuric acid is recommended.

6.3 Wine Quality Concerns

Since the first studies on IER treatments of wine, it has been understood that resins, like most other methods of wine remediation, bind to a large variety of compounds within the wine matrix regardless of their intended use (Esau & Amerine,

1966). Esau & Amerine (1966) proposed a list of mechanisms to explain the unintended

binding of wine components to IER's which can partially predict the loss of specific wine

components. However, the authors acknowledged that the impacts of IER treatments on

specific wines cannot be practically predicted. Therefore, they suggested that experimental evidence is an important tool for assessing IER impacts on wine components. While Ion exchange is an effective candidate for correcting pH/TA imbalance, its effects on the concentration of volatile aromas has been cited as a major reason to reject its use in the winery (Zoecklin et al., 1995).

During the ion exchange process, cations other than potassium interact with resin.

Substantial decreases in Calcium, iron, magnesium and Copper have been reported

(Palacios et al., 2001; Lasanta et al., 2013). Reductions in ethanol between 0.1-0.5% have been reported (Lasanta et al., 2013; Ibeas et al., 2015). Walker et al., (2002) analyzed tartaric acid and determined no difference in concentrations between control and ion exchange-treated wines. Wasielewski (2015) determined decreases in lactic acid 13 in some wines, increases in tartaric acid, and no difference in citric acid concentrations after ion exchange treatment. Since the wines in that study were analyzed several months after bottling, the higher concentrations of tartaric acid were most likely due to potassium bitartrate precipitation in the control wines, since they contained substantially more K+.

The wines treated by Wasielewski showed inconsistent patterns in organic acid concentrations among treatments. A “depleted resin” treatment was also employed in this research that was sometimes much higher or lower than the control wine in concentrations of some organic acids in some wines.

When wine is passed through an ion exchange column, its pH drops sharply as K+ and other cations exchange with H+ that are saturated on the resin. For this reason, the wine that first exits the column will have a low pH (around 2.1). As more wine flows through the column, less H+ is available to exchange with K+ (Lasanta et al., 2013).

Eventually equilibrium is reached between H+ present in the wine and resin. At this point, significant exchange no longer takes place and wine pH is not altered as it passes through the resin. By collecting fractions of the eluent at regular intervals during treatment, it is possible to obtain a better understanding of what is happening during treatment. Tannins, pH, anthocyanins, hue, color intensity and polyphenols have been measured by collecting fractions throughout the ion exchange treatment (Palacios et al.,

2001; Lasanta et al., 2013). Fractional analysis of organic acid concentrations during ion exchange treatment has not been performed previously, and may be able to better explain changes in organic acid that occur during the course of ion exchange treatment.

Sensory analysis conducted on wines treated with ion exchange suggests that 14

flavor and aroma is impacted by the treatment and that small reductions of flavor components may occur in some wines (Bonorden et al, 1986; Walker et al., 2002; Ibeas et al. 2015, Lasanta et al. 2013). Other sensory studies have found no difference in flavor or aroma after ion exchange treatment (Mira et al., 2012). Gas chromatography has also been applied to determine differences in wines treated with ion exchange (Benitez et al,

2002a; Lasanta et al., 2013). However, these studies had conflicting findings potentially due to the differences in methodology. Interestingly, in opposition to some of the sensory work, both GC-based studies found that some compounds may have increased in concentration after ion exchange. A possible explanation for this occurrence may come from a pH dependent extraction step used in GC sample preparation. Lasanta et al.,

(2013) used a solid phase extraction procedure that has can favor the retention of acidic or basic compounds based on sample pH (Supelco, 1998). Neutral compounds are not thought to be impacted. The effect of pH on the selective extraction of volatile aroma compounds must be considered, since ion exchange treated wines are lower in pH than control wines.

Previous experiments have also taken a targeted approach to determine differences in the aroma profile of ion exchange-treated wines (Benitez et al, 2002; Lasanta et al.,

2013). This approach involves preselecting a number of wine aroma compounds and determining their concentration. Due to the complexity of wine chromatograms, thousands of aroma compounds that have potentially been altered by ion exchange treatment go unmeasured (Arbulu et al., 2015). Recent advances in the field of metabolomics have produced computer software capable of providing untargeted analysis of chromatograms. In untargeted analysis, all differences in peak areas among 15 chromatograms are determined. This technique has been applied to a several wine chromatogram analysis projects (Robinson et al., 2011; Springer et al., 2014; Vaclavik et al., 2011).

REFERENCES

Allen, M.S., Lacey, M.J., Harris, R.L.N, Brown, W.V. (1991). Contribution of Methoxypyrazines to sauvignon blanc wine aroma. Am. J. Enol. Vitic. 42(2), 1991.

Alleweldt, G. (1997). Genetics of grapevine breeding. In Progress in Botany. Springer Berlin Heidelberg. Pp.441-454.

Arbulu, M., Sampedro, M.C., Gomez-Caballero, A., Goicolea, M.A., Barrio, R.J. (2015) Untargeted metabolomic analysis using liquid chromatography quadrupole time- of-flight mass spectrometry for non-volatile profiling of wines. Analytica Chimica Acta. 858, 32-41.

Bazinet, L. & Firdaous, L. (2011). Recent patented applications of ion-exchange membranes in the agrifood sector. Recent patents Chem. Eng. 4 (2), 207-216.

Benitez, P., Castro, R., Barroso, C.G. (2002a). Removal of iron, copper, and manganese from white wines through ion exchange techniques: effects on their organoleptic characteristics and susceptibility to browning. Analytica Chimica Acta. 458, 197- 202.

Benitez, J.G., Palacios, V.M., Pazo, J.A., Rodriguez, P.L. (2002b). Industrial development of proton exchange for tartrate stabilization of sherry wines. Eur. Food Res. Tech. 21, 418-422.

Bisson, L.F., Waterhouse, A.L., Ebeler, S.E., Walker, M.A., Lapsley, J.T. (2002). The present and future of the international wine industry. Nature. 418(6898). 696-699.

Bisotto, A., Julien, A., Schneider, R., Salmon, J.M. (2015). Evaluation of the inherent capacity of commercial yeast strains to release glycosidic aroma precursors from Muscat grape must. Aus. J. Grape Wine Res. 21(2), 194-199.

Bitteur, S., Tesniere, C., Sarris, J., Baumes, R., Bayonove, C., Flanzy, C. (1992). Carbonic anaerobiosis of Muscat grapes. I. Changes in the profiles of free and bound volatiles. Am. J. Enol. Vitic. 43(1), 1992.

Bonorden, W.R., Nagel, C.W., Powers, J.R. (1986). The adjustment of high pH/high titratable acidity wines by ion exchange. Amer. J. Enol. Vitic. 37 (2), 143-148.

Boulton, R. (1980). The general relationship between potassium sodium and pH in grape juice and wine. Am. J. Enol. Vitic. 31(2), 1980.

Bousbouras, G.E. & Kunkee, R.E. (1971). Effect of pH on malo-lactic fermentation in 17

wine. Am. J. enol. Vitic. 22(3), 121-126.

Bordiga, M., Rinaldi, M., Locatelli, M., Piana, G., Travaglia, F., Coisson, J.D., Arlorio, M. (2013).Characterization of Muscat wines aroma evolution using comprehensive gas chromatography followed by a post-analytic approach to 2D contour plots comparison. Food Chem. 140 (1), 57-67.

Carreon-Alvarez, A., Herrera-Gonzalez, A., Casillas, N., Prado-Ramirez, R., Estarron- Espinosa, M., Soto, V., de la Cruz, W., Barcena-Soto, M., & Gomez-Salazar, S. (2011). Cu (II) removal from tequila using an ion-exchange resin. Food Chem. 127(4), 1503-9. Chisolm, M.G., Guiher, L.A., Vonah, T.M., Beaumont, J.L. (1994). Comparison of some French-American hybrid wines with White Riesling using gas chromatography- olfactometry. Am. J. Enol. Vitic. 45 (2), 201-212.

Chisolm, M.G., Guiher, L.A., Zaczkiewicz, S.M. (1995). Aroma characteristics of aged Vidal Blanc wine. Am. J. Enol. Vitic. 46 (1), 56-62.

Chow, W.S., Ball., M.C., Anderson, J.M. (1990). Growth and photosynthetic responses of spinach to salinity: implications of K+ nutrition for salt tolerance. Func. Plant Bio. 17(5), 563-578.

Cole, J. and Boulton, R. (1989). A study of calcium salt precipitation in solutions of malic acid and tartaric acid. Vtis, 28, 177-190. Cuvelier, J.M. (1995). Ion exchange resins used in the food industry and in the drinking water preparation in Europe. Reactive Polymers. 24, 109-116. Ebeler, S.E., Thorngate, J.H. (2009). Wine chemistry and flavor: looking into the crystal glass. J. Agric. Food Chem. 57, 8098-8108.

Esau, P. & Amerine, M.A. (1966). Synthetic ion-exchange resins in wine research. Amer. J.Enol. Vitic. 17 (4), 268-276.

Farina, L., Medina, K., Urruty, M., Boido, E., Dellacassa, E., Carrau, F. (2012). Redox effect on volatile compound formation in wine during fermentation by Saccharomyces cerevisiae. Food Chem. 134, 933-939.

Freeman, B.M. & Kliewer, W.M. (1983). Effect of irrigation, crop level and potassium fertilization on Carignane vines II. Grape and wine quality. Am. J. Enol. Vitic. 34, 197-207. Garcia-Ruiz, J.M., Alcantara, R., & Martin, J. (1991). Evaluation of wine stability to potassium hydrogen tartrate precipitation. Am. J. Enol. Vitic. 42, 336-40. 18

Gu., S., Wilker, K.L., Dharmadhikari, M.R., Moore, J.F., Howard, S., Edson, C.E. Evaluation of grapevine cultivars and breeding lines for adaptability to Missouri conditions. Acta Hort. 528(2), 569- 574.

Hrazdina, G., Parsons, G.F., Mattick, L.R. (1984). Physiological and biochemical events during development and maturation of grape berries. Am. J. Enol. Vitic. 35(4), 220-227. .Ibeas ,V., Correia, A., Jordao, A. (2015). Wine tartrate stabilization by different levels of cation exchange resin treatments: impact on chemical composition, phenolic profile, and organoleptic properties of red wine. Food Res. Int. 69, 364-372.

Kliewer, W.M., Howorth, L., Omori, M. (1967). Concentrations of tartaric acid and malic acid and their salts in Vitis vinifera grapes. Am. J. Enol. Vitic. 18, 42-54. Kwasniewski, M.T., Vanden Heuvel, J.E., Pan, B.S., Sacks, G.L. (2010). Timing of cluster light environment manipulation during grape development affects C13 norisoprenoids and carotenoid concentrations in Riesling. J. Agric. Food Chem. 58, 6841-6849.

Lasanta, C. & Gomez, J. (2012). Tartrate stabilization of wines. Trends in Food Sci. and Tech. 28, 52-59.

Lasanta, C., Caro, I., Perez, L. (2013). The influence of cation exchange treatment on the final characteristics of red wines. Food Chem. 138, 1072-1078.

Martinez-Gil, A.M., Angenieux, M., Pardo-Garcia, A.I., Alonso, G.L., Ojeda, H., Salinas, M.R. (2013). Glycosidic aroma precursors of Syrah and Chardonnay grapes after an oak extract application to the grapevines. Food Chem. 138(2), 956-965.

Mateo, J.J., Jimenez, M. (2000). Monoterpenes in grape juice and wines. J. Chromatogr. A. 881, 557-567

Mendes-Pinto, M.M. (2009). Carotenoid breakdown products-the norisoprenoids-in wine aroma. Archives Biochem. Biophys. 483(2). 236-245. 19

Meyers, J.M., Sacks, G.L., Heuvel, J.E.V. (2013). Glycosylated aroma compound responses in ‘Riesling’ wine grapes to cluster exposure and vine yield. Hortechnology. 23 (5), 581-588.

Mira, H., Leite, P., Ricardo-Da-Silva, J.M., Curvelo-Garcia, A.S. (2006). Use of ion exchange resins for tartrate wine stabilization. J. Int. Sci. Vigne Vin. 40 (4), 223- 246.

Morris, J.R. & Cawthon, D.L. (1982). Effect of irrigation, fruit load, and potassium fertilization on yield, quality, and petiole analysis of Concord (Vitis labrusca L.) grapes. Am. J. Enol. Vitic. 33(3), 145-148.

Morris, J.R. Sims, C.A., Cawthon D.L. (1983). Effects of excessive potassium levels on pH, acidity and color of resh and stored grape juice. Am. J. Enol. Vitic. 34(1), 35- 39.

Mpelasoka, B.S., Schachtman, D.P., Treeby, M.T., Thomas, M.R. (2003). A review of potassium nutrition in grapevines with special emphasis on berry accumulation. Austr. J. Grape Wine Res. 9:154-168.

Nagel, C.W., Weller, K/, & Filiatreau, D. (1988). Adjustment of high pH-high TA musts and wines. In: Proceedings of the 2nd International Symposium for Cool Climate Viticulture and Oenology (R.E. Smart et al, eds.), pp. 222-4. New Zealand Society for Viticulture and Oenology, Auckland, New Zealand.

Nasef, M.M. & Ujang, Z. (2012). Introduction to Ion Exchange Processes. Ion exchange technology I. Springer. 1-39.

Ong, P.K.C & Acree, T.E. (1999). Similarities in the aroma chemistry of Gewurztraminer variety wines and lychee (Litchi chinesis Sonn.) fruit. J. Agric. Food Chem. 47, 665-670.

Palacios, V.M., Caro, I., Perez, L. (2001). Application of Ion Exchange Techniques to industrial process of metal ions removed from wine.

Palomo, E. S., Perez-Coello, M.S., Diaz-Maroto, M.C., Vinas, M.G., Cabezudo, M.D. (2006). Contribution of free and glycosidically-bound volatile compounds to the aroma of Muscat “a petit grains” wines and effect of skin contact." Food Chem. 95 (2), 279-289.

Plane, R.A., Mattick, L.R., Weirs, L.D. (1980). An acidity index for the taste of wine. Am. J. Enol. Vitic. 31 (3), 265-268. 20

Qiang, Y. & Yan, Y. (2013). Study on the deacidifying condition of litchi wine by ion- exchange resin 314. Modern Food Sci. & Tech. 29, 380-2. Rapp, A., Mandery, H. (1986). New progress in vine and wine research. Experientia. 42(87), 3-84.

Reynolds, A.G., Wardle, D.A. (1989). Influence of fruit microclimate on monoterpene levels of Gewürztraminer. Am. J. Enol. Vitic. 40(3), 149-154.

Ribereau-Gayon, P., Boidron, J.N., Terrier, A. (1975). Aroma of Muscat grape varieties. J. Agric. Food Chem. 23 (6), 1042-1047.

Robinson, A.L., Boss, P.K., Heymann, H., Solomon, P.S., Trengrove, R.D. (2011). Development of a sensitive non-targeted method for characterizing the wine volatile profile using headspace solid-phase microextraction comprehensive two- dimensional gas chromatography time-of-flight mass spectrometry. J. Chromatography A. 1218 (3), 504-517.

Robinson, A.L., Boss, P.K., Solomon, P.S., Trengove, R.D., Heymann, H., Ebeler, S.E. (2013). Origins of grape and wine aroma. Part 1. Chemical components and viticultural impacts. Am. J. Enol. Vitic. 65(1), 1-24.

Rojas-Lara, B.A., Morrison, J.C. (1989). Differential effects of shading fruit or foliage on the development and composition of grape berries. Vitis. 28, 199-208.

Roland, A., Schneider, R., Razungles, A., Cavelier, F. (2011). Varietal thiols in wine: discovery, analysis and applications. Chem. Reviews. 111(11), 7355-7376.

Sacks, G.L., Gates, M.J., Ferry, F.X., Lavin, E.H., Kurtz, A.J., Acree, T.E. (2012). Sensory threshold of 1,1,6-Trimethyl-1,2-dihydronaphthalene (TDN) and concentrations in young Riesling and non-Riesling wines. J. Agric. Food Chem. 60, 2998-3004.

Sala, C., Busto, O., Guasch, J., Zamora, F. (2004). Influence of vine training and sunlight exposure on the 3-alkyl-2-methoxypyrazines content in musts and wines from the Vitis vinifera variety Cabernet Sauvignon. J. Agric. Food Chem. 52, 3492-3497.

Skinkis, P.A., Bordelon, B.P., Wood, K.V. (2008). Comparison of monoterpene constituents in Traminette, Gewurztraminer, and Riesling winegrapes. Am. J. Enol. Vitic. 59 (4), 440-445.

Soles, R.M., Ough, C.S., Kunkee, R.E. 1982. Ester concentration differences in wine fermented by various species and strains of yeasts. Am. J. Enol. Vitic. 33(2), 94- 98.

Springer, A.E., Riedl, J., Esslinger, S., Roth, T., Glomb, M.A., Fauhl-Hassek, C. (2014). Validated modeling for German white wine varietal authentication based on headspace solid-phase microextraction online coupled with gas chromatography mass spectrometry fingerprinting. J. Agric. Food Chem. 62 (28), 6844-6851.

Styger, G., Prior, B., Bauer, F.F. (2011). Wine flavor and aroma. J. Indust. Micro. & Biochem. 38(9), 1145-1159.

Sun, Q., Gates, M.J., Lavin, E.H., Acree, T.E., Sacks, G.L. (2011). Comparison of odor- active compounds in grapes and wines from Vitis vinifera and non-foxy american grape species. J. Agric. Food Chem. 59, 10657-10664.

Supelco. (1998). Guide to solid phase extraction. Bulletin 910. Supelco, Bellefonte, PA.

Vaclavik, L., Lacina, O., Hajslova, J., Zweigenbaum, J. (2011). The use of high performance liquid chromatography-quadrupole time-of-flight mass spectrometry coupled to advanced data mining and chemometric tools for discrimination and classification of red wines according to their variety. Analytica chimica acta. 685(1), 45-51.

Walker, T., Morris, J., Threlfall, R., Main, G. (2002). pH modification of Cynthiana wine using cationic exchange. J. Agric. Food Chem. 50 (22), 6346-6352.

Walker, T., Morris, J., Threlfall, R., Main, G. (2004). Quality, sensory and cost comparison for pH reduction of Syrah wine using ion exchange or tartaric acid. J. Food Qual. 27, 483-496

Wasielewski, M. (2015). Combating high pH/high TA Missouri wines and its impacts on non-volatile and sensory characteristics. MS Thesis. University of Missouri- Columbia.

Webster, D.R., Edwards, C.G., Spayd, S.E., Peterson, J.C., Seymour, B.S. (1993). Influence of ineyard nitrogen fertilization on the concentrations of monoterpenes, higher alcohols, and esters in aged Riesling wines.

Wibowo, D., Eschenbruch, R., Davis, C.R., Fleet, G.H., Lee, T.H. (1985). Occurrence and growth of lactic acid bacteria in wine: a review. Am J. Enol. Vitic. 36(4), 302- 313.

Wilson, B., Strauss, C.R., Williams, P.J. (1984). Changes in free and glycosidically bound monoterpenes in developing Muscat grapes. J. Agric. Food Chem. 32, 919- 924.

Yuan, F. & Qian, M.C. (2016). Development of C 13-norisporenoids, carotenoids, and other volatile compounds in Vitis vinifera L. Cv. Pinot noir grapes. Food Chem. 192, 633-641.

Zaganiaris, E.J. (2011). Ion exchange resins and synthetic adsorbents in food processing. Books on demand GmbH. Paris, France.

Zoecklin, B.W., Fugelsang, K.C., Gump, B.H., Nury, F.S. (1995). Wine Analysis and Production. New York: Chapman & Hall.

CHAPTER 2: IMPACTS OF ION EXCHANGE ON CATIONS, ORGANIC ACIDS, AND ETHANOL IN MISSOURI WINES

ABSTRACT Cation exchange is a promising tool for managing high pH/TA wines, including in Missouri where wines with this chemical profile are commonly encountered. High pH/TA wines occur due to cultivar choice, weather, soil and viticultural practices that result in grapes with high potassium concentrations at harvest. However, cation exchange has not been widely adopted due to concerns about its tendency to strip flavor and aroma, thus reducing overall wine quality. Four wines (Valvin Muscat,

Chambourcin, Norton, and Syrah) were treated using ion exchange. Organic acid, potassium, sodium, Calcium, and Magnesium concentrations were monitored during

treatment. Percent alcohol was determined after treatment. The wines displayed a 50%

reduction on average in tartaric, malic, and lactic acid, and a 90% reduction in potassium,

sodium, calcium and magnesium during the initial stages of treatment. Only sodium and

calcium concentrations remained lower than initial levels even at the end of treatment

indicating that these ions may be preferentially scavenged. Percent alcohol was reduced

by 0.2-0.7% overall.

1. INTRODUCTION

1.1 Wine Acidity An important element of wine stability is pH. A sufficiently low pH is necessary in

order for SO2 to function effectively as an antimocrobial (Zoecklin et al., 1995). It also

partially determines the perception of sourness, which is an important consideration in

consumer preference. Titratable acidity (TA) is also an important consideration in the 24

perception of wine sourness, with high TA values correlating with higher perceptions of acidity (Plane et al., 1980).

Low pH is necessary for the microbial protection of wine because it inhibits the growth of many types of spoilage organisms. Sulfur dioxide is the main antimicrobial used in wine production (King et al., 1982), and is generally added as potassium metabisulfite. In wine solutions, sulfur dioxide (SO2) exists in pH dependant equilibrium

- 2- with bisulfite (HSO3 ) and sulfite (SO3 ), which do not posses the antimicrobial benefits

of SO2 (King et al., 1982). The pKa of SO2 is 1.77. In other words, at pH 1.77, fifty

percent of the sulfur will be in the form SO2, while the other fifty percent will be in the

- - form HSO3 . At typical wine pH values, the equilibrium between SO2 and H2SO3 is

- heavily in favor of the HSO3 ion, both of which do not exhibit antimicrobial or

antioxidant activity. At pH 3, about eight percent of the sulfur is in the SO2 form. While

at pH 4, about one percent is available. For this reason, acidifying high pH wines greatly

increases the antimicrobial effects of sulfur dioxide.

When potassium metabisulfite has not been added to wines, pH still significantly

affects microbial growth. Malolactic baterica such as Leuconostic oenos grow slower in wines at pH 3.2 than wines at pH 3.8, retarding their ability to conduct malolactic fermentations from 14 days to 164 days (Bousbouras & Kunkee, 1971). Other malolactic bacteria such as Pediococcus and Lactobacillus are unable to grow below pH 3.5

(Wibowo et al., 1985).

Low pH values do not always correlate with high TA. Under conditions where

excessive cation concentrations are present, high pH/ TA wines will occur. The high pH is due to both the buffering capacity of potassium as well as the negative correlation 25 between hydrogen and potassium in grape berriers described in section 1.2 (Boulton,

1980;Bonorden et al., 1986) Reducing the pH of these wines to levels necessary for the antimicrobial effectiveness of SO2 is difficult. The commonly used method of adding tartaric acid to reduce pH is sometimes impossible since tartaric acid precipitates from the wine solution when excessive salt levels are present. In instances where the tartaric acid does not precipiate, it’s addition further increases TA to levels which are percieved as overly sour by consumers (Jackson, 2014; Walker et al., 2002).

1.2 Potassium

High pH/TA wines are commonly encountered in wines made from hybrid grape varietals, particularly Norton, which often contain excessive concentrations of potassium cations (K+)(Morris et al., 1983). An inverse relationship exists between K+ and hydrogen ions (H+) in grape juice, so that an increase in K+ concentration results in a pH increase (Boulton, 1980). The K+ content of a wine is influenced by numerous factors beginning with soil. Morris et al., (1983) determined that potassium fertilization resulted in direct exchange of K+ for H+ in grape berries, increasing pH substantially. However, potassium uptake by roots is influenced by the presence of other cations, particularly sodium (Na+) (Chow et al., 1990). Whereby, high Na+ concentrations will compete with

K+ for uptake into the plant. Rootstocks also vary considerably in their tendency to uptake K+ (Mpelasoka et al., 2003).

Once inside grapevines, K+ is translocated from older tissues to growing tissues, which serve as K+ sinks (Mpelasoka et al., 2003). Leaves, which begin growing at budbreak, become the initial K+ sinks. During veraison, berries become the predominate

K+ sink as cells inside the grape expand and ripen. When grapes reach harvest maturity, 26

K+ is the dominant cation present in grape berries, while calcium, magnesium, sodium,

iron, Copper, and Manganese are also present at lower concentrations (Hrazdina et al.,

1984). This effect is magnified in over-ripe grapes as K+ concentrations continue to

increase after harvest maturity (Iland & Coombe, 1988; Kliewer et al., 1967).

Sunlight exposure can impact K+ concentrations, whereby shading either clusters or

canopy leads to increased fruit K+(Rojas-Lara & Morrison, 1989). Irrigation also increases K+ in berries (Freeman & Kliewer, 1983). While substantial research has been conducted on factors influencing K+ movement in grapevines, more research is required

to determine specific viticultural practices that reduce K+ in grape berries (Mpelasoka et

al., 2003).

While excessive K+ issues are ideally dealt with in the vineyard (Mpelasoka et al,

2003), high potassium grapes still frequently enter Missouri wineries because best

management practices are not necessarily enough to mitigate the problem. Several

methods of pH/TA imbalance correction are often employed to reduce wine pH without

further increasing the TA; Double salt-deacidification, electrodialysis, cold-stability manipulation, and ion exchange. Each tehnique is capable of lowering wine pH with minimal influence on TA. With the exception of cold-stability manipulation, they all require specialized equipment, can be complex, and require significant time and space commitments (Garcia-Ruiz et al, 1991; Cole and Boulton, 1989; Nagel et al. 1988; Soares et al. 2009).

1.3 Ion Exchange

Ion exchange resins have been used in wine remediation since the 1950’s. In addition to adjusting pH, they have been used for tartrate stabilization, concentrating 27

grape must, and to remove metals, proteins, volatile acidity, and color ( Bazinet &

Firdaous, 2011; Bonorden et al., 1986; Esau & Amerine, 1966; Lasanta & Gomez, 2012;

Mira et al., 2006). Ion exchange has successfully been applied to correcting high pH/TA wines by removing K+ and replacing it with H+, thereby reducing pH by the direct addition of H+ (Lasanta et al., 2013; Walker et al., 2002). Resulting wines are perceived as less acidic than those remedied by tartaric acid addition because acid anions, which are perceived as acidic by humans, are not added in the process (Plane et al., 1980). Resin-

based ion exchange treatment systems are inexpensive and widely available due to their

use in water treatment systems (Carreon-Alvarez et al., 2011; Qiang & Yan, 2013;

Cuvelier, 1995). A cost analysis performed by Walker et al., (2004) shows that it resin based cation exchange systems are more cost effective than membrane-based cation exchange systems for pH/TA correction (Walker et al., 2004). However, resin-based cation exchange technology has not been adopted by some wine producers due to concerns over unintended color and flavor stripping (Zoecklin et al., 1995).

During the ion exchange process, cations other than potassium interact with resin.

Substantial decreases in Calcium, iron, magnesium and Copper have been reported

(Palacios et al., 2001; Lasanta et al., 2013). Reductions in ethanol by 0.1% have been reported (Lasanta et al., 2013) Walker et al., (2002) analyzed tartaric acid and determined no difference in concentrations between control and ion exchange-treated wines.

Wasielewski (2015) determined decreases in lactic acid in some wines, increases in tartaric acid, and no difference in citric acid concentrations after ion exchange treatment.

Since the wines in that study were analyzed several months after bottling, the higher concentrations of tartaric acid were most likely due to potassium bitartrate precipitation 28

in the control wines, since they contained substantially more K+. The wines treated by

Wasielewski showed inconsistent patterns in organic acid concentrations among

treatments. A “depleted resin” treatment was also employed in this research that was

sometimes much higher or lower than the control wine in concentrations of some organic

acids in some wines.

When wine is passed through an ion exchange column, its pH drops sharply as K+

and other cations exchange with H+ that are saturated on the resin. For this reason, the

wine that first exits the column will have a low pH (around 2.1). As more wine flows

through the column, less H+ is available to exchange with K+ (Lasanta et al., 2013).

Eventually equilibrium is reached between H+ present in the wine and resin. At this

point, significant exchange no longer takes place and wine pH is not altered as it passes

through the resin. By collecting fractions of the eluent at regular intervals during

treatment, it is possible to obtain a better understanding of what is happening during

treatment. pH, tannins, anthocyanins, hue, color intensity and polyphenols have been

measured by collecting fractions throughout the ion exchange treatment (Palacios et al.,

2001; Lasanta et al., 2013). Fractional analysis of organic acid concentrations during ion

exchange treatment has not been performed previously, and may be able to better explain

changes in organic acid that occur during the course of ion exchange treatment.

2. MATERIALS AND METHODS

2.1 Winemaking Three grape varieties commonly utilized in the Midwest wine industry were sourced from Missouri vineyards and vinified at the University of Missouri experimental winery.

In 2014, Syrah grapes were sourced from Lodi, CA and vinified. The resulting wines 29

were used for ion exchange treatments approximately two months after primary

fermentation. Control wines were compared with ion exchange treated wine and depleted

resin treated contol (described in section 2.2) .

Valvin Muscat grapes were machine harvested early on September 5th, 2013 at

Meyers Farm, Mt. Vernon, MO (37.0784°N, 93.7752°W). Approximately 450kg grapes

were received at the MU Experimental winery in a half-ton picking bin. The grapes were

pressed at 1.2 bar with a Pecaws 500L SK pneumatic press (Prospero, Geneva, NY). An

addition of 50mg/L SO2 was added as potassium metabisulfite to press pan. The pressed

volume was approximately 350 liters. The juice was cold settled (5oC) for 24 hours,

racked, and inoculated with Saccharomyces cerevisiae strain DV10 yeast (Lallemand,

Petaluma, CA) previously rehydrated in 260mg/L GoFerm (Lallemand) according to

manufacturer’s guidelines. Fermentation was conducted at 13oC until dry, which was determined using Clinitest (Bayer). Post-fermentation, the wine was at pH 3.5 with a TA of 8.0g/L. A final addition of 50mg/L SO2 was added as potassium metabisulfite.

Chambourcin grapes were hand-harvested on September 22nd, 2013 from

University of Missouri managed vineyards in Mount Vernon, Mo (37.0784°N,

93.7752°W). Approximately 450 kg was received at the MU experimental winery in picking lugs. The grapes lugs were stored at 7oC overnight. The next morning, they

were destemmed and lightly crushed into a 600 liter stainless steel fermenter. The wine

was then and inoculated with Saccharomyces cerevisiae strain GRE yeast (Lallemand,

Petaluma, CA) previously rehydrated in 260mg/L GoFerm (Lallemand) according to

manufacturer’s guidelines. Fermentation was conducted in a refrigerated room set at 30

19oC. After 8 days, the end of alcoholic fermentation was determined by a 0.25% residual sugar reading using Clinitest tablets. The must was then pressed in a membrane press to 1.2 bar. The wine was racked twice before ion exchange treatment. Malolactic fermentation took place before treatments.

Norton grapes were hand-harvested on October 11th, 2013 at Les Bougois Vineyards,

Rocheport, Missouri (38.9794°N, 92.5622°W) . Approximately 450kg. was received at the MU experimental winery in a ½ ton bin. On the day of harvest, the grapes were destemmed and lightly crushed into an open top fermenter. The wine was then inoculated with Saccharomyces cerevisiae strain GRE yeast (Lallemand, Petaluma, CA) previously rehydrated in 260mg/L GoFerm (Lallemand) according to manufacturers guidelines.

Fermentation was conducted in a refrigerated room set at 20oC. After 7 days, the end of alcoholic fermentation was determined by a 0.25% residual sugar reading using Clinitest tablets. The must was then pressed in a membrane press to 1.2 bar. The pH and TA post- fermentation were 3.74 and 9.64g/L respectively. The wine was racked twice before ion exchange treatment. Malolactic fermentation took place before treatments.

Approximately 500kg of Syrah grapes was purchased from S and L Vineyards in

Lodi, Calif. (38.2173°N, 121.4101°W). It was picked on September 7th 2014, flash frozen, and transported to the University of Missouri Winery. It arrived partially frozen on September 9th and was destemmed into a 600L stainless steel tank. The wine was then then inoculated with Saccharomyces cerevisiae strain GRE yeast (Lallemand, Petaluma,

CA) previously rehydrated in 260mg/L GoFerm (Lallemand) according to manufacturers guidelines. Fermentation was conducted in a refrigerated room set at 20oC. The fermentation showed signs of reduction, so a second addition of Di-ammonium 31

Phosphate was made and the fermentation was heavily aerated. After seven days, the end

of alcoholic fermentation was determined by a 0.25% residual sugar reading using

Clinitest tablets. The must was then pressed in a membrane press to 1.2 bar. After

receiving treatments, the wines were stabilized with 70mg/L SO2 and bottled under

nitrogen for further analysis.

2.2 Ion Exchange Treatment Amberlite IR-120(H+) ion exchange resin was purchased from Alfa Aesar, Ward

Hill, Mass. 500 grams dry resin was placed into a glass beaker containing deionized

water for 72 hours prior to treatment to facilitate swelling. Shortly before treatment, the

resin slurry was packed into a 0.33L stainless steel column fitted with tight mesh screens

on each end. Sanitary winery tubing was connected from dispatch tank though a variable

speed pump to the entrance of the column. Tubing at the treated end of the column

directed the wine into 1L plastic graduated cylinders which facilitated the calculation of

flow rate and volume treated. Treated wine was stored in 20L glass carboys. The

column was regenerated by pumping an excess volume of 5% sulfuric acid solution

through the resin at a rate of 250mL/min. This was followed by a rinse with purified

water at a rate of 200ml/min to remove all sulfuric acid in the system. The inlet hose was then attached to a 200L stainless steel tank containing 150L of homogenated wine. Wine

was pumped through the ion exchange system at a rate of approximately 500 mL per

minute (80 bed volumes/Hr). The wine was collected in 1 liter graduated cylinders in

960mL increments to represent 3 bed volumes. A 50 mL sample was collected every

3BV and later used for fractional analysis. After collecting the sample, the rest of the

treated wine was transferred to a glass carboy. The treatment process was stopped upon 32

resin depletion which was determined once the pH of the treated eluent matched the

initial pH (Hanna Instruments HI 2222 pH meter).

Each wine was treated in triplicate using the same resin. After each treatment, the

resin was rinsed and regenerated using the above procedure. New resin was used for

each of the three wines. Five additional gallons were taken from the stock tank to be

bottled and used as a control.

After completing the treatments, the wines were bottled under nitrogen in Stelvin

screwcap closures and stored at 13oC until analysis.

2.3 Analysis

A Hanna Instruments HI 2222 pH meter was used for pH measurments.

Titratable acidity was measured using official methods of analysis (AOAC 942.15). The

pH and TA of each collected fraction was determined, along with control, depleted resin

and treatment wines prior to bottling. K+, calcium, and magnesium concentrations were

determined in 2013 Chambourcin ion exchange treated and control wines by the

University of Missouri Soil and Plant Testing laboratory using atomic absorption

spectrophotometry. Ethanol was measured using ebulliometry (Dujardin-Salleron,

France).

Ion Exchange fractions collected during treatment were analyzed for Tartaric and

Malic acid by High Performance Liquid Chromatography (HPLC) using the following

method based on Agilent (2003). Briefly 3g PVPP was added to 5mL of each Norton,

Chambourcin, and Syrah and mixed for 20 seconds to achieve decolorization. The mixtures were centrifuged at 12,000RPM for 5 minutes (Eppendorf, Germany). Then the

supernatant was collected and filtered through a 0.22uM membrane. HPLC was 33

performed on a (VARIAN, Inc. ProStar, Palo Alto, Calif.) consisting of a 410

Autosampler, 210 pump with in-line degasser, 335 LC dual path diode array UV-Visible

detector, and operated via the Galaxie chromatography manager software (Version

1.9.302.530, Agilent Technologies, Santa Clara, Calif.). A Zorbax SB-Aq column

(4.6x250 mm with 5 mm pore size) was operated at 35°C (Agilent Technologies, Santa

Clara, Calif.). The aqueous mobile phase was isocratic with 99% 20mM sodium phosphate and 1% acetonitrile at a total flow rate of 1mL per minute (total run time of 10 minutes). The UV-Visible diode-array operated at 210nm to detect tartaric, malic, and lactic, acid. Each acid was identified by reference standard and quantified with a 5-point calibration curve ranging from 0.25 g/L - 10 g/L. Sample injection volume was 10 μL.

3. RESULTS

In all wines, pH was reduced dramatically at the intial stages of ion exchange treatment (figure 2.1). Initial eluted fractions were at pH 2.3-2.7. As treatment progressed, wine pH began to steadily rise. The increase in pH occurred after approximately 15 bed volumes (5L) in Norton and Chambourcin and 27 bed volumes

(9L) in Valvin Muscat. In all cases, the wines reached their initial feed pH (3.9 for

Norton and Chambourcin, 3.6 for Valvin Muscat). This occurred between 25-33 bed volumes (8.3-11L) in Norton, 39 bed volumes (13L) in Chambourcin, and 39-51 bed volumes (13-17L) in Valvin Muscat. Once the eluent pH reached the feed pH, no further changes in pH occurred.

The ion exchange treatments reduced K+, calcium, and magnesium concentrations in all Chambourcin replicates (table 2.1). A 90% reduction in all cation concentrations 34

was observed in the early fractions. K+ was initially reduced from 1242mg/L to between

140 and 230mg/L. Calcium was reduced from 122mg/L to between 20 and 25mg/L.

Magnesium was reduced from 86mg/L to between 18 and 24mg/L. As the treatment progressed, there was a decrease in the reduction of cations. Late in the ion exchange

cycle, K+ concentrations rose to near their initial concentrations (between 1150 and

1195mg/L). However, calcium and magnesium concentrations remained below feed

concentrations even as pH leveled off and the resin was determined to be depleted.

Calcium remained between 35 and 53mg/L and magnesium remained between 43 and

53mg/L, nearly 50% below feed concentrations.

All measured organic acids were reduced in the earliest fractions of all wines (Figure

2.2). Lactic acid was initially reduced by 25-50%. Tartaric acid was reduced by 35-55%.

Malic acid was reduced by 40-60%. As the treatment progressed, organic acids reached

a maximum concentration near the halfway point in Valvin Muscat and Norton wines. In

the Chambourcin and Syrah wines, organic acids increased throughout treatment, with

the highest concentrations found at the end of treatment. Malic acid was not detected in

Norton or Chambourcin wines due to malolactic fermentation.

Significant decreases in alcohol percent were observed in all measured wines. In

Chambourcin, a 0.2% reduction was determined. In Norton and Valvin Muscat, the reduction was nearly 0.6%.

4. DISCUSSION

The cation exchange treatment was able to dramatically reduce the pH of all

wines in all trials (see Wasiekewski, (2015)). For each wine, similar patterns were seen

in the rate of pH increase for each trial. Initial drastic pH reductions were followed by 35 steady increases in pH until thefeed pH was reached. This pattern shows similarities to the ion exchange elution curve presented by Lasanta et al. (2013). However, the wines they treated spent a longer time at lower pH values before increasing to feed pH, suggesting that their ion exchange system may have caused different effects on the treated wines. This type of difference could be explained by using an ion exchange comumn with a smaller diameter and larger bed (see chapter 1 section 6.2).

Potassium, Calcium, and Magnesium cations are removed during the process.

Calcium and Magnesium concentrations remained below feed concentrations even after the resin was exhausted of H+. This finding is in agreement with Palacios et al., (2001), who determined that Calcium and Magnesium reach feed concentrations much later than

K+ or pH. However, they used a resin specifically designed for purpose of metal removal. While only three cations were analyzed in this experiment, the characteristics of Calcium and Magnesium suggests that other ions not measured will behave in ways characteristic to their affinity to Amberlite IR-120 resin (Singare et al., 2009).

Calcium, magnesium, sodium and K+ are required by yeasts to conduct fermentation (Pohl, 2007). For this reason, performing ion exchange pre-fermentation may create complications during alcohol fermentation, due to insufficient nutrient availability for yeast growth. Manganese has been identified as essential for the growth of malolactic bacteria (Terrade & Orduna, 2009). Conversely, ion exchange post fermentation can help to restrict the growth of spoilage organisms by removing essential metals required for their growth. Heavy metals such as cadmium , arsenic, and lead are sometimes present in some wines, posing health concerns (Marias & Blackhurst, 2009).

The impact of ion exchange on heavy metals in wine is unknown for the tested resin type. 36

However, the metal-removing resin used by Palacios et al., (2001) was able to remove

copper and lead from wine.

Cations, with the exception of Na+, cannot be perceived by human taste at their concentrations found in wine (Maltman, 2013). Sodium is generally considered to be detrimental to wine quality, so its removal should enhance wine attributes. However, cations can indirectly affect wine flavor and aroma (Pohl, 2007). It has been shown that the removal of multivalent metals from wine can inhibit oxidative browning, strengthing the ageability of wines (Lasanta et al., 2013). Salt concentrations in solutions also affect the partitioning of volatile compounds, which tend to volatilize more easily in high salt solutions (Friant & Suffet, 1979). Ion exchange treated wines may have less aromatic qualities, even if the same concentrations of volatile aroma compounds exists, due to the fact that their liquid-vapor partitioning equilibrium has shifted towards the liquid.

Organic acids are initially lowered by ion exchange resin, with reductions from feed concentrations between 30-50% observed. However, their concentration in the eluent increases throughout treatment. The treatments had the strongest effect on malic acid. The reduction of these acids should lower the titratable acidity and overall perception of acidity. While only tartaric, lactic, and malic acids were analyzed, the ion exchange resin have similar effects on other minor acids, such as acetic acid, which substantially impact wine flavor (Du Toit & Lambrechts, 2002).

Reductions in ethanol concentration were more severe than reported by Lasanta et al., (2013), but consistent with the 0.5% reduction reported by Ibeas et al., (2015).

Ibease et al., (2015) stated “this (ethanol) loss could be due to evaporation during passage through the column, pump, and pipes rather than a consequence of the cation exchange 37

process itself”. Further benchtop experiments conducted on this issue are presented in the

appendix (Figure A1). The results of these benchtop experiments suggest that the resin is

directly responsible for the ethanol reductions. Ethanol is removed from wine

proportional to the amount of Amberlite IR-120 ion exchange resin introduced.

Reductions of 1.7% can be achieved when 22g resin is introduced to 100ml model wine.

This ethanol reduction is also seen in depleted resin in which no pH change takes place, making it an interesting method of de-alcoholization. De-alcoholization can diminish the perception of ‘body’ and sweetness, and alter the composition of volatile aromas in the headspace (Jordao et al., 2015). However, there is debate about whether or not a preferred alcohol concentration exists (King & Heymann, 2014). Reduced alcohol can be percieved as either a positive or negative effect depending on the wines original composition and the intended wine style.

5. CONCLUSION

Resin cation exchange is an effective method of pH reduction in Missouri cultivars, lowering pH with a less severe increase in titratable acidity than tartaric acid additions. In addition to the sensory benefits over other treatment methods, it has been shown to be predictable and easily implemented with minimal equipment. Reductions in potassium, calcium, magnesium, organic acids, and ethanol are accompanied with the treatment with impacts varying dependent on the time point in the process. These temporal differences leave open the potential for further optimization in treating a wine to induce desired chemical changes other than just pH reduction (e.g. organic acids, ethanol and cations). 38

REFERENCES

Agilent. (2003). Agilent Zorbax SB-Aq datasheet.

Bazinet, L. & Firdaous, L. (2011). Recent patented applications of ion-exchange membranes in the agrifood sector. Recent patents Chem. Eng. 4 (2), 207-216.

Bonorden, W.R., Nagel, C.W., Powers, J.R. (1986). The adjustment of high pH/high titratable acidity wines by ion exchange. Amer. J. Enol. Vitic. 37 (2), 143-148.

Boulton, R. (1980). The general relationship between potassium sodium and pH in grape juice and wine. Am. J. Enol. Vitic. 31(2), 182-186.

Bousbouras, G.E. & Kunkee, R.E. (1971). Effect of pH on malo-lactic fermentation in wine. Am. J. enol. Vitic. 22(3), 121-126.

Carreon-Alvarez, A., Herrera-Gonzalez, A., Casillas, N., Prado-Ramirez, R., Estarron- Espinosa, M., Soto, V., de la Cruz, W., Barcena-Soto, M., & Gomez-Salazar, S. (2011). Cu (II) removal from tequila using an ion-exchange resin. Food Chem. 127(4), 1503-9. Chow, W.S., Ball., M.C., Anderson, J.M. (1990). Growth and photosynthetic responses of spinach to salinity: implications of K+ nutrition for salt tolerance. Func. Plant Bio. 17(5), 563-578.

Cole, J. and Boulton, R. (1989). A study of calcium salt precipitation in solutions of malic acid and tartaric acid. Vtis, 28, 177-190. Cuvelier, J.M. (1995). Ion exchange resins used in the food industry and in the drinking water preparation in Europe. Reactive Polymers. 24, 109-116. Du Toit, W.J. & Lambrechts, M.G. (2002). The enumeration and identification of acetic acid bacterial from South African red wine fermentations. Int. J. Food Micr. 74(1), 57-64.

Esau, P. & Amerine, M.A. (1966). Synthetic ion-exchange resins in wine research. Am. J.Enol. Vitic. 17 (4), 268-276.

Hepner, Y. & Bravado, B., (1985). Effect of crop level and drip irrigation scheduling on the potassium status of Cabernet Sauvignon and Carignane vines and its must and wine composition and quality. Am. J. Enol. Vitic. 36, 140-147. Hrazdina, G., Parsons, G.F., Mattick, L.R. (1984). Physiological and biochemical events during development and maturation of grape berries. Am. J. Enol. Vitic. 35(4), 220-227. Ibeas ,V., Correia, A., Jordao, A. (2015). Wine tartrate stabilization by different levels of cation exchange resin treatments: impact on chemical composition, phenolic profile, and organoleptic properties of red wine. Food Res. Int. 69, 364-372.

Iland, P.G. & Coombe, B.G. (1988). Malate, tartrate, potassium , and sodium in flesh and skin of Shiraz grapes during ripening: concentration and compartmentation. Am. J. Enol. Vitic. 39, 71-76. Jackson R.S. (2014). Wine science: principles and applications (4th ed.). Amsterdam: Elsevier/Academic Press. Jordao, A.M., Vilela, A., Cosme, F. (2015). From sugar of grape to alcohol of wine: sensorial impact of alcohol in wine. Beverages. 1(4). 292-310. Kliewer, W.M., Howorth, L., Omori, M. (1967). Concentrations of tartaric acid and malic acid and their salts in Vitis vinifera grapes. Am. J. Enol. Vitic. 18, 42-54. King, A.D., Ponting, J.D., Sanshuck, D.W., Jackson, R., Mihara, K. (1981). Factors affecting death of yeast by sulfur dioxide. J. Food Prot. 44, 92-97. King, E.S., Heymann, H. (2014). The effect of reduced alcohol on the sensory profiles and consumer preferences of white wine. J. Sensory Studies. 29, 33-42. Lasanta, C. & Gomez, J. (2012). Tartrate stabilization of wines. Trends in Food Sci. and Tech. 28, 52-59.

Lasanta, C., Caro, I., Perez, L. (2013). The influence of cation exchange treatment on the final characteristics of red wines. Food Chem. 138, 1072-1078.

Maltman, A. (2013). Minerality in wine: a geological perspective. J. Wine Res.24(3), 169-181.

Marias, A.D. & Blackhurst, D.M. (2009). Do heavy metals counter the potential health benefits of wine?. J. Endocrinology, Metabolism, and Diabetes of South Africa. 14(2), 77-79.

Maathuis, F.J., & Sanders, D. (1996). Mechanisms of potassium absorption by higher plant roots. Physiologia Plantarum, 96(1), 158-168. 40

Mira, H., Leite, P., Ricardo-Da-Silva, J.M., Curvelo-Garcia, A.S. (2006). Use of ion exchange resins for tartrate wine stabilization. J. Int. Sci. Vigne Vin. 40 (4), 223- 246.

Morris, J.R. Sims, C.A., Cawthon D.L. (1983). Effects of excessive potassium levels on pH, acidity and color of resh and stored grape juice. Am. J. Enol. Vitic. 34(1), 35- 39.

Mpelasoka, B.S., Schachtman, D.P., Treeby, M.T., Thomas, M.R. 2003. A review of potassium nutrition in grapevines with special emphasis on berry accumulation. Australian J. Grape Wine Res. 9, 154-158.

Palacios, V.M., Caro, I., Perez, L. (2001). Application of ion exchange techniques to industrial process of metal ions removal from wine. Adsorption. 7(2), 131-138. Plane, R.A., Mattick, L.R., & Weirs, L.D. (1980). An acidity index for the taste of wines. Am. J. Enol. Vitic. 31(3), 265-8. Pohl, P. 2007. What do metals tell us about wine?. Trends Anal. Chem. 26(9), 941-949. Qiang, Y. & Yan, Y. (2013). Study on the deacidifying condition of litchi wine by ion- exchange resin 314. Modern Food Sci. & Tech. 29, 380-2. Rojas-Lara, B.A., Morrison, J.C. (1989). Differential effects of shading fruit or foliage on the development and composition of grape berries. Vitis. 28, 199-208. Singare, P., Lokhande, R., Samant, N. (2009). Studies of uni-univalent ion exchange reactions using strongly acidic cation exchange resin Amberlite IR-120. Nat. Sci. 1(2), 124-128. Soares, P.A.M.H., Geraldes, V., Fernandes, C., Cameira dos Santos, P., Norberta de Pinho, M. (2009). Wine tartaric stabilization by electrodialysis: Prediction of required deionization degree. Am. J. Enol. Vitic. 60, (2), 183-188. Terrade, N., Orduna, R.M. (2009). Determination of the essential nutrient requirements of wine-related bacteria from the genera Oenococcus and Lactobacillus. Int. J. Food Micr. 133(1-2), 8-13. Walker, T., Morris, J., Threlfall, R., Main, G. (2002). pH modification of Cynthiana wine using cationic exchange. J. Agric. Food Chem. 50 (22), 6346-6352.

Walker, T., Morris, J., Threlfall, R., Main, G. (2004). Quality, sensory and cost 41

comparison for pH reduction of Syrah wine using ion exchange or tartaric acid. J. Food Qual. 27(6), 483-496.

Wasielewski, M. (2015). Combating high pH/high TA Missouri wines and its impacts on non-volatile and sensory characteristics. MS Thesis. University of Missouri- Columbia.

Wibowo, D., Eschenbruch, R., Davis, C.R., Fleet, G.H., Lee, T.H. (1985). Occurrence and growth of lactic acid bacteria in wine: a review. Am J. Enol. Vitic. 36(4), 302- 313.

Zoecklin, B.W., Fugelsang, K.C., Gump, B.H., Nury, F.S. (1995). Wine Analysis and Production. New York: Chapman & Hall.

TABLES AND FIGURES

CHAPTER 3: COMPARISON OF ODORANTS PRESENT IN VALVIN MUSCAT WINES WITH OTHER AROMATIC WHITE WINES

ABSTRACT

Since its release in 2006 of cv. Valvin Muscat (interspecific hybrid) by Cornell, the acreage under cultivation has grown rapidly in the Midwest and Northeast U.S. While descriptors such as “floral” and “spicy” are used to both describe Valvin Muscat and V. vinifera Muscat varieties, little is known about Valvin Muscat’s actual aroma chemistry.

This information is needed as a first step in allowing aroma-driven optimization of viticultural and vinification parameters of the new cultivar. Seven monoterpenes known to be important to classic Muscat aroma and TDN were quantified by GC-MS, following concentration by solid phase extraction. Ten Valvin Muscat wines, from three different states (Missouri, New York and Indiana), were compared to wines made from six different examples of Muscat varieties, three Rieslings, three Gewurztraminers, four

Vidal blancs and a three Traminettes. The concentration of linalool in Valvin Muscat was far greater than those found in V. vinifera Muscat varieties (a mean of 1,859 μg/L versus 240 μg/L respectively). Total linalool oxides (cis and trans) concentrations were comparable between Valvin Muscat and V. vinifera Muscat’s (with mean concentrations of 982 μg/L and 595 μg/L respectively). While there was large differences in specific monoterpene concentrations within Valvin Muscat, up to 10x, it was found that generally

Valvin Muscat wines have concentrations equal or higher to those concentrations found in V. vinifera Muscat’s for all compounds quantified. TDN was found in highest quantities in Riesling wines (6.4ug/L on average).

1. INTRODUCTION

1.1 Varietal Aroma

Wine aroma is determined by interactions among grape cultivar, terroir, as well as

viticultural and winemaking practices. Aromas can be described as primary, secondary or

tertiary for those derived or released from the grape, fermentation or additives

respectively. For the most part tertiary flavors are process influenced whereas primary

and secondary aromas are more heavily influenced by the fruit. Furan and furanone are

examples of tertiary aromas which are extracted and formed in wines from interactions

with wood, with cultivar playing a minimal role in determining the type and concentration of furans present in a wine (Robinson et al., 2013). Other compounds, like esters, are formed through a complex of factors occurring during fermentation (yeast strain, temperature, redox state) along with ester precursors determined by cultivar (Soles et al., 1982; Farina et al., 2012). Varietal aromas are derived from specific compounds whose origins can be traced to particular grape cultivars. They define the unique aroma profile that some wines exhibit (Styger et al., 2011), some of which are present in the fruit while others are bound and only released during fermentations and ageing (Bisotto et al., 2015). Four classes of compounds are responsible for the varietal aroma observed in many wines; C13 norisoprenoids, thiols, methoxypyrazines, and monoterpenes.

C13 norisoprenoids contribute to the varietal aroma of some white wines,

particularly Rieslings and Gewurztraminers, but are also present in many other wines at

lower, but above threshold concentrations (Martinez-Gil et al., 2013; Yuan & Qian,

2016). They are formed as degradation products of carotenoids during berry maturation,

bound to glucose molecules in the fruit and released during fermentation, and wine 49

ageing (Mendes-Pinto, 2009). Sunlight exposure at certain stages of grape maturation

can greatly increase carotenoid production, leading to increased concentrations of some

C13 norisoprenoids such as TDN (1,1,6 trimethyl-1,2 dihydronapthalene) and vitispirane in resulting wines (Kwasniewski et al., 2010). Currently, about six C13 norisoprenoids

are considered important in wine aroma; β-damascenone, β-ionone, vitispirane, actinidiol,

Meyers et al. 2013; Sacks et al., 2012). While Gewurztraminer aroma is attributed to β- damascenone, in addition to monoterpenes (cis-rose oxide and geraniol) and esters (Ong and Acree 1999).

(ng/L). Specialized extraction steps are required to quantify thiols (Tominaga et al., 2000)

Determining viticultural influences on thiol concentration is also difficult since they are formed during fermentation from unknown precursors.

Methoxypyrazines contribute a ‘vegetative’ varietal aroma to Sauvignon blanc and Cabernet Sauvignon wines (Allen et al., 1991; Sala et al., 2004). Several methoxypyrazines, particularly IBMP (3-Isobutyl-2-methoxypyrazine), SBMP (3-sec- 50

butyl-2-methoxypyrazine), and IPMP (3-isopropyl-2-methoxy-pyrazine) can add to a wine’s aroma profile at concentrations as low as 2ng/L (Sala et al., 2004). While some amount of vegetative character is considered desirable, especially in Sauvignon blanc wines, restricting excessive concentrations is important when growing these cultivars.

Methoxypyrazines concentrations are higher in grapes harvested prematurely and grapes grown in cool climates that have less exposure to sunlight (Sala et al., 2004).

Monoterpenes are a class of aroma compounds found to some degree in most

grape varieties. However, they are closely associated with Muscat wines. About 50

different monoterpenes have been found in winegrapes in both free and glycosidically

bound forms, imparting a variety of floral and citrus flavors and aromas (Ebeler &

Thorngate, 2009; Mateo &Jimenez, 2000). They are synthesized in the grape berry from

acetyl co-enzyme A (Mateo & Jimenez, 2000). During berry maturation, concentrations

of both free and bound monoterpenes vary dramatically. Most compounds achieve

maximum concentrations at or after harvest maturity, except for geraniol where

concentrations peak near berry set, then diminish steadily (Wilson et al., 1984). Berry

sunlight exposure will increase monoterpene concentrations (Reynolds & Wardle, 1989).

Monoterpene concentrations are not influenced by nitrogen fertilization (Webster et al.,

1993).

In the winery, monoterpene extraction from fruit can be increased by carbonic

anaerobiosis or skin contact before pressing, and by using glucosidase enzymes to cleave

gylcosylated precursors (Bitteur et al., 1992; Palomo et al., 2006). Monoterpene

degradation begins quickly in wine and significant losses can be seen within months

(Rapp & Mandery, 1986). Numerous studies have confirmed the findings of Ribereau- 51

Gayon et al. (1975) who determined that the typical aroma of Muscat grapes and wines

is attributed to roughly eight monoterpenes; linalool, nerol, geraniol, α-terpineol, and four oxides of linalool (Bordiga et al., 2013; Palomo et al., 2006; Skinkis, 2008) .

Monoterpenes also contribute to the varietal aroma of other aromatic white wines.

1.2 Hybridization Impacts on Wine Aroma

Hybrid cultivars are created from crossing Vitis vinifera species with various species in the Vitis genera native to North America. These hybrid crosses are cultivated

in regions where V. vinifera lacks sufficient disease resistance or cold hardiness to grow

successfully. The aroma profile of wines produced from hybrid cultivars is influenced by

the characteristics of its inherited genetics. North American grape species impart a pronounced varietal aroma to hybrid wines. V. labrusca wines have a “foxy” aroma, due to three compounds; AAP (2-aminoacetophenone), methyl anthranilate, and furaneol.

Some varietal odorants in wines made from V. riparia, V. cinerea, and V. rupestris grapes are eugenol, cis-3-hexenol, 1,8-cineole, 3-isobutyl-2methoxypyrazine (IBMP) and 3- isopropyl-2-methoxypyrazine (IPMP) (Sun et al., 2011). Hybrid wines also contain aromas present in the parent V. vinifera species. Traminette is a cross between

Gewurztraminer and Joannes Seyve 23-416. It has an aroma profile similar to

Gewurztraminer (Ong and Acree 1999; Skinkis et al. 2008). Vidal Blanc is a hybrid cultivar that produces wine that is similar to Riesling wines. It tends to contain higher concentrations of monoterpenes, leading to a more intense aroma. Vidal Blanc also tends to produce undesirable aromas such as methyl anthranilate (Chisolm et al. 1994, 1995).

The aroma profile of many other hybrid cultivars has not been evaluated in depth.

1.3 Valvin Muscat 52

Valvin Muscat is an interspecific hybrid cultivar derived from a cross between

Muscat Ottonel and Muscat du Moulin (couderac 299-35), that was released by Cornell

University in 1995 and has quickly gained popularity and acreage in the United States

(Reisch et al. 2006). It was selected for its characteristic Muscat aroma, winter hardiness and disease resistance (Reisch et al. 2006). While the number of acres planted with Valvin Muscat continues to grow, there are still numerous questions left related to the optimum conditions for cultivation and vinification of this cultivar, given the limited time it has been produced commercially and the diversity of the regions in which it is being grown.

In the Four Corners Region of New Mexico, it has been shown to be an early budding cultivar. Consequently, primary buds are susceptible to damage by late spring frosts (Lombard et al., 2013). However, it can successfully set fruit on secondary buds after such events (Allen, 2007; Lombard et al., 2013). Moderate bud mortality has also been observed following severe winter freeze events. However, the extent of bud damage has been less severe than in other hybrid white cultivars such as Vignoles and

Chardonel (Allen et al., 2011). When compared with V. vinifera Muscat cultivar, Spring frost resistance is highly improved (Stafne, 2007). Studies concerning Valvin Muscat’s disease resistance are limited and inconclusive (Wilson et al., 2007).

The volatile aromas in the cultivars used to produce Valvin Muscat (Muscat du

Moulin and Muscat Ottonel) have been evaluated to some degree. Muscat du Moulin is a

French-American hybrid that contains negligible quantities of methyl anthranilate and subsequently no ‘foxy’ V. labrusca aromas (Fuleki, 1982). Muscat Ottonel contains typical Muscat aromas (Mateo & Jiminez, 2000). However, it lacks the cold-hardiness 53 and disease resistance required for Missouri’s climate. Volatile analysis has not yet been performed on wines made from Valvin Muscat. With a better understanding of the aroma compounds that contribute to Valvin Muscat wines, and how their concentrations compare to other varieties, it will be easier to understand the best practices needed in making wine from this variety. By understanding more about the underlying biochemistry, it is possible to draw on the decades of research on those varieties with similar aromas to jump ahead in bringing new varieties to its full potential quickly.

2. MATERIALS AND METHODS

2.1 Chemicals

Dichloromethane (≤ 99.8%, Sigma, St. Louis, MO), methanol (≤99.9%, Sigma,

St. Louis, MO) ,2-octanol (99%+, Acros, New Jersey), α-terpineol (≤ 97%, Fluka, St.

Louis, MO), rose oxide (Analytical Grade, Sigma), linalool oxide (≤ 97%, Sigma), citronellol (99%, Sigma), TDN ( 1,1,6 trimethyl-1,2 dihydronapthalene)( ≤ 99%, donated by Sacks Lab, Cornell University)(Kwasniewski et al., 2010).

2.2 Wines

Wines were either purchased from commercial producers or made in house. For in house produced wines, Vintage 2013 and 2014 ‘Valvin Muscat’ grapes were machine harvested on Sep. 5th both years and sourced from Meyers Farm, Mt. Vernon, MO

(37.0784°N, 93.7752°W). The grapes received approximately 12 hours of skin contact before pressing (bladder press, 1.2 bar). The juice was cold settled for 24 hours, racked, and inoculated with Saccharomyces cerevisiae strain DV10 yeast (Lallemand, Petaluma,

CA) previously rehydrated in 260mg/L GoFerm (Lallemand) according to manufacturer’s 54

guidelines. Fermentation lasted approximately 21 days at 13oC until dryness was

determined using Clinitest (Bayer). The wine was stabilized with 70mg/L SO2 as KMBS

and bottled after 2 months. In addition, 27 commercially available wines ranging in price

between $12 and $25 were purchased from local stores or directly from the wineries.

These wines consisted of 8 Valvin Muscats, 6 Muscats, 3 Gewurztraminers, 3 Rieslings,

4 Vidal Blancs, and 3 Traminettes. Region and vintage information is listed in table 3.1.

2.3 Basic Analysis

Residual sugar was determined using Clinitest (Bayer). Percent alcohol was

determined using ebulliometry (supplier) when not supplied by the producer. pH was

determined using a Hanna Instruments HI 2222 pH meter.

2.4 Volatile Analysis

Solid Phase Extraction was performed for each wine on an Alltech 12-port SPE

manifold (Fischer, St. Louis, MO) in triplicate using the methods based on Pineiro et al.

2004. Lichrolut EN SPE cartridges (Merck, Billerica, MA) (200mg solid phase) were

used. Prior to SPE extraction 2-octanol was added at a concentration of 20mg/L to be used as an internal standard. A 0.66 bar vacuum suction was used with 4 ml of water for washing added after loading. Dichloromethane (2ml) was employed as the extraction solvent. The solution was then dried under nitrogen to a final volume of 200ul before injection. Separation and quantification were achieved on a Varian 431-GC and 220-MS.

The chromatographic conditions were the following: injection volume, 1 ul; injector temp, 200oC; guard column, base deactivated FS (1m, 0.25m diameter) (Bruker,

Billerica, MA); main column, DB-WAXETR (30m, 0.25mm diameter, 0.25um film thickness) (Agilent, Santa Clara, CA); split ratio, splitless (0-3min), 100:1 (3min-51min); 55 carrier gas, helium; flow rate, 1ml/min; oven ramp, 40oC (hold 6min), 145oC (5oC/min),

175oC (2.5oC/min), 220oC (hold 10min). All injections were performed in triplicate. All extractions and analyses were performed between May and July 2015.

2.5 Data Compilation

Compound identification and quantification was achieved by generating calibration curves using pure standards in a model wine (12% abv and 0.8g tartaric acid/L with the pH adjusted to 3.3 using 5M NaOH). Calibration curves were assembled for α-terpineol, rose oxide, linalool oxide, citronellol, and TDN using a 1/x weighted regression analysis, using 2-octanol as an internal standard. For all compounds quantified, the limits of quantification achieved was below reported aroma threshold values, except Linalool

(LOQ 40ug/L)(threshold, 25ug/L), cis-rose oxide and trans-rose oxide (LOQ

1.5ug/L)(threshold 0.2ug/L). Odor activity values (OAV’s) were established by dividing a compounds concentration by published aroma threshold concentration.

3. RESULTS

3.1 Basic Analysis

The wines varied in vintage from 2011 to 2014 (table 3.1). However, vintages were not designated for two Vidal blancs and one Traminette. The average vintage of the

Valvin Muscats (2012.9) was close to the average vintage of the other Muscats (2013).

Gewurztraminer and Riesling were older on average (2012.3). Regional differences also confound potential findings since the hybrid cultivars (Valvin Muscat, Vidal Blanc, and

Traminette) were grown in Missouri, New York, and Indiana, while the other cultivars were grown in the West Coast of the United States and in Europe. The average pH of 56

Valvin Muscat (3.55) was significantly higher than that of the other Muscats (3.21) and the Rieslings (3.03). No differences were seen among the average residual sugar or percent alcohol among cultivars. However, residual sugar did vary among the wines tested, from 0.2% to >5%. While a number of aromatic white wines were included in this study it should be noted that the scope was limited to better understanding the compounds contributing to Valvin Muscat aroma in context of other wines, rather than allowing for extensive comparison between different wines, regions, vintages etc., all of which are factors known to influence wine aroma.

3.2 Odorants in Valvin Muscat

The Valvin Muscat wines contained important Muscat odorants at odor-active concentrations (table 3.2). Most notably, linalool was detected well above its threshold concentration (25ug/L)(Ferreira et al., 2000) in all but one measured wine, up to

4.6mg/L. It may have been above threshold in this wine, but below the 40ug/L LOQ.

All wines were also above threshold concentrations in α-Terpineol (250ug/L)(Ferreira et al., 2000). Rose oxide was only detected in some wines. When detected, rose oxide was above threshold concentrations (0.2ug/L)(Guth, 1997). Citronellol and TDN were present near published threshold levels (100ug/L)(Guth, 1997) and (2ug/L)(Sacks et al.,

2012) respectively. Linalool oxides were detected significantly below threshold concentration (4600ug/L)(Marias, 1983) in all cases. Vintage appears to influence odorant concentrations, with older vintages having overall lower concentrations of aroma compounds.

3.3 Odorants Among Cultivars

Many of the aroma compounds known to contribute to other aromatic white wines were found at concentrations above threshold in at least some of the wines (table 3.3).

However, there was large variability amongst the wines measured, as would be expected for wines of different cultivars, regions, and vintages. Significant differences were determined among varieties in five of the seven measured compounds. Linalool, linalool oxide, and α-terpineol were highest in the Valvin Muscat varieties. These compounds were present at somewhat lower concentrations in the other Muscats. In the other cultivars, they were substantially lower in concentration than any of the Muscat varieties analyzed. Linalool oxide was below threshold in all wines (OAV<0.15). Total monoterpenes (sum of measured monoterpenes) were also highest in Valvin Muscat

(3.6mg/L compared with >1.4mg/L in all other cultivars).

TDN was found in highest concentration in Riesling wines (6.4ug/L), where it was on average three times higher than the threshold concentration (2ug/L). The other Muscat category was also significant higher in TDN than Valvin Muscat (4.5ug/L compared with

1.8ug/L). However, TDN concentrations were still at the threshold value in Valvin

Muscat (OAV 0.9).

4. DISCUSSION

The Muscat wines selected were on average younger than the other wine cultivars

(2013 compared with 2012.3). Monoterpenes are known to degrade over time (Simpson,

1978). This may have led to larger exaggerations in the difference in monoterpene concentration between all Muscats and the other cultivars. However, vintage effects 58

cannot explain large discrepancy between monoterpene concentrations between Valvin

Muscats and other Muscats (3.6ug/L compared with 1.4ug/L), since they were of the

same average vintage. The pH of Valvin Muscat is in the upper range of what is

expected for a white wine. This is possibly due to high pH/TA issues caused by excess

potassium (Morris et al., 1983). The large variation in residual sugar determined among

all wines reflects differences in winemaking styles. The influence of residual sugar (RS) on monoterpene concentration is unknown. The possibility that high RS wines contained fewer monoterpenes, due to incomplete cleavage of glycoslyated precursors during fermentation was considered. However, no correlation was determined between residual

sugar and total monoterpene concentration (R2 = 0.0036).

Several Monoterpenes were significantly higher in concentration in Valvin Muscat

than other Muscats. The linalool concentration in Valvin Muscat (1.8mg/L on average, up to 4.6mg/L) is also higher than other reported values, which ranged from 0.1 to 0.6 mg/L (Bordiga et al., 2013; Dziadas & Jeleri, 2010; Lambri et al., 2012; Palomo et al.,

2006). The values obtained by these researchers were in the same range as the other

Muscats category (0.24mg/L on average), suggesting that an analytical error was not

responsible for the high values. Rose oxide was found in some of the wines. It probably

exists above threshold values for other wines. However, it was not detected due to the

difference between the LOQ (1.5ug/L) and aroma threshold (0.2ug/L). Since only one

2011 and one 2012 wine was analyzed, insufficient data was obtained in order to

determine whether vintage influenced the concentration of monoterpenes.

The most significant difference between Valvin Muscat and all other wines is the

high concentration of linalool (1.8mg/L on average, up to 4.6mg/L), with an average 59

OAV of 80. Linalool was not detected in non-Muscat wines. Valvin Muscat also contained the highest concentration of Linalool oxide. While still far below threshold concentrations, linalool oxide has been shown to have synergistic influences on overall monoterpene aroma (Styger et al., 2011). Therefore, it may still be contributing to the aroma profile of Valvin Muscat. α-Terpineol was also highest in Valvin Muscat, most likely contributing to the aroma profile (OAV 3.2). Similar concentrations were determined in other Muscats.

Concentrations of cis-rose oxide were lower in Traminette than Gewurztraminer, in agreement with Ong and Acree, 1999. However, Skinkis et al. 2008 found the opposite to be true. Low concentrations of monoterpenes were found in both Vidal Blanc and

Riesling. This is expected as the varietal aroma of wines made from these grapes consists of C13 norisoprenoids, not monoterpenes. Yet the Vidal Blanc also did not contain high

concentrations of TDN, while TDN concentrations in Riesling were at previously

reported values (Sacks et al., 2012). Hybrid wines are known to inherent aroma

characteristics from the cultivars used to make them. For this reason it is not surprising

that Valvin Muscat contains monoterpene concentrations similar to other Muscats.

However, the varietal aromas attributed to Native American grapes such as methyl

anthranilate, 2-AAP, and Furaneol were not measured. The extent to which Valvin

Muscat contains these aroma compounds, compared with V. vinifera cultivars, would be

of interest to grape breeders and grape growers.

Valvin Muscat can be characterized as a highly aromatic wine with monoterpene

concentrations exceeding most other Muscats. This makes Valvin Muscat potentially 60 important for use as a blending grape, to add aroma to neutral wines. Further work is required to determine the extent of hybrid character that exists in Valvin Muscat.

5. CONCLUSION

Valvin Muscat had very high values for several odor compounds important to

Muscat aroma. Concentrations are so high that it will still exhibit intense aroma even when grown or vinified in ways that may reduce these compounds. This makes Valvin

Muscat important for blending, to add Muscat aroma characteristics to a neutral wine.

REFERENCES

Allen, A. (2007). What we learned from the Easter freeze in Missouri. Understanding and Preventing Freeze Damage in Vineyards. pp 95-106. University of Missouri- Columbia.

Allen, A., Harris, J., Strieglier, K., Bergmeier, E. (2011). 2011 Missouri grapevine bud mortality survey report. University of Missouri, 1-8.

Allen, M.S., Lacey, M.J., Harris, R.L.N, Brown, W.V. (1991). Contribution of methoxypyrazines to sauvignon blanc wine aroma. Am. J. Enol. Vitic. 42(2), 109- 112.

Barbera, D., Avellone, G., Filizzola, F., Monte, L.G., Catanzaro, P., Agozzino, P. (2013). Determination of terpene alcohols in Sicilian Muscat wines by HS-SPME-GC- MS. Nat. Prod. Res. 27 (6), 541-547.

Belancic, A., Agosin, E., Ibacache, A., Bordeu, E., Baumes, R., Razungles, A., Bayonove, C. (1997). Influence of sun exposure on the aromatic composition of Chilean Muscat grape cultivars Moscatel de Alejandria and Moscatel Rosada. Am. J. Enol. Vitic. 48 (2), 181-186.

Bowen, A.J, Reynolds, A.G. (2012). Odor potency of aroma compounds in Riesling and Vidal Blanc table wines and icewines by gas chromatography-olfactometry-mass spectrometry. J.Agric. Food Chem. 60, 2874-2883.

Chisolm, M.G., Guiher, L.A., Vonah, T.M., Beaumont, J.L. (1994). Comparison of some French-American hybrid wines with White Riesling using gas chromatography- olfactometry. Am. J. Enol. Vitic. 45 (2), 201-212.

Chisolm, M.G., Guiher, L.A., Zaczkiewicz, S.M. (1995). Aroma characteristics of aged Vidal Blanc wine. Am. J. Enol. Vitic. 46 (1), 56-62.

Delcroix, A., Gunata, Z., Sapis, J.C., Salmon, J.M., Bayonove, C. (1994). Glycosidase activities of three enological yeast strains during winemaking: effect on the terpenol content of Muscat wine. Am. J. Enol. Vitic. 45 (3), 291-296.

Demir, K.O.K & Erdinc, B.A.L. (2014). The response of Monoterpene compounds of cv. Gewurztraminer grape (Vitis vinifera L.) to various doses of Prohexadione- Calcium applied at different periods. Turkish J. Agric. Nat. Sci. 1, 1231-1235.

Dziadas, M., Jeleri, H.H. (2010). Analysis of terpenes in white wine using SPE-SPME- GC/MS approach. Analytica chimica Acta. 667(1), 43-49.

Ebeler, S.E., Thorngate, J.H. (2009). Wine chemistry and flavor: looking into the crystal glass. J. Agric. Food Chem. 57, 8098-8108.

Ferreira, A.C.S & Pinho, P.G. (2004). Nor-isoprenoid profile during port wine ageing – influence of some technological parameters. Analytica Chimica Acta. 513(1), 169- 176.

Fuleki, T. (1982). The vineland grape flavor index – a new objective method for the accelerated screening of grape seedlings on the basis of flavor character. Vitis. 21, 111-120.

Gunata, Y.Z., Bayonove, C.L., Baumes, R.L., Cordonnier, R.E. (1986). Stability of free and bound fractions of some aroma components of grapes cv. Muscat during the wine processing: preliminary results. Am J. Vitic. Enol. 37(2), 112-114.

Guth, H. (1997). Quantitation and sensory studies of character impact odorants of different white wine varieties. J. Agric. Food Chem. 45(8), 3027-3032.

Kwasniewski, M.T., Vanden Heuvel, J.E., Pan, B.S., Sacks, G.L. (2010). Timing of cluster light environment manipulation during grape development affects C13 norisoprenoids and carotenoid concentrations in Riesling. J. Agric. Food Chem. 58, 6841-6849.

Lambri, M., Dordoni, R., Silva, A., De Faveri, D.M. (2012). Comparing the impact of bentonite addition for both must clarification and wine fining on the chemical profile of wine from Chambave Muscat grapes. Int. J. Food Sci. Tech. 47 (1), 1- 12.

Lombard, K. , Maier, B., Thomas, F.J., O'Neill, M., Allen, S., Heyduck, R. (2013). Wine grape cultivar performance in the four corners region of New Mexico in 2010-12. Hortechnology. 23 (5), 699-709. 63

Marias, J. (1983). Terpenes in the aroma of grapes and wines: a review. S. Afr. J. Enol. Vitic. 4(2), 49-60.

Mateo, J.J., Jimenez, M. (2000). Monoterpenes in grape juice and wines. J. Chromatogr. A. 881, 557-567

Mendes-Pinto, M.M. (2009). Carotenoid breakdown products-the norisoprenoids-in wine aroma. Archives Biochem. Biophys. 483(2). 236-245.

Meyers, J.M., Sacks, G.L., Heuvel, J.E.V. (2013). Glycosylated aroma compound responses in ‘Riesling’ wine grapes to cluster exposure and vine yield. Hortechnology. 23 (5), 581-588.

Morris, J.R. Sims, C.A., Cawthon D.L. (1983). Effects of excessive potassium levels on pH, acidity and color of resh and stored grape juice. Am. J. Enol. Vitic. 34(1), 35- 39.

Ong, P.K.C & Acree, T.E. (1999). Similarities in the aroma chemistry of Gewurztraminer variety wines and lychee (Litchi chinesis Sonn.) fruit. J. Agric. Food Chem. 47, 665-670.

Pineiro, Z., Palma, M., Barroso, C.G. (2004). Determination of terpenoids in wines by solid phase extraction and gas chromatography. Analytica Chimica Acta. 513, 209-214.

Rapp, A., Mandery, H. (1986). New progress in vine and wine research. Experientia. 42(87), 3-84.

Rapp, A., Versini, G., Ullemeyer, H. (1993). 2-Aminoacetophenone-casual component of untypical aging flavor (napthalene note, hybrid note) of wine. Vitis. 32(1), 61-62.

Reisch, B.I., Luce, R.S., Bordelon, B., Henick-Kling T. (2006). Valvin Muscat grape. New York’s Food and Life Sciences Bulletin, 161. New York. New York State Agricultural Experiment Station.

Reynolds, A.G., Wardle, D.A. (1989). Influence of fruit microclimate on monoterpene l evels of Gewürztraminer. Am. J. Enol. Vitic. 40(3), 149-154.

Ribereau-Gayon, P., Boidron, J.N., Terrier, A. (1975). Aroma of Muscat grape varieties. J. Agric. Food Chem. 23 (6), 1042-1047.

Roland, A., Schneider, R., Razungles, A., Cavelier, F. (2011). Varietal thiols in wine: discovery, analysis and applications. Chem. Reviews. 111(11), 7355-7376.

Ruiz-Garcia, L., Hellin, P., Flores, P., Fenoll, J. (2014).Prediction of Muscat aroma in table grape by analysis of rose oxide. Food Chem. 154, 151-157.

Selli, S., Canbas, A., Cabaroglu, T. (2006). Aroma components of cv. Muscat of Bornova wines and influence of skin contact treatment. Food chem. 94, 319-326.

Simpson, R.F. (1978). Aroma and compositional changes in wine with oxidation, storage and ageing. Vitis. 17, 274-287.

Skinkis, P.A., Bordelon, B.P., Wood, K.V. (2008). Comparison of monoterpene constituents in Traminette, Gewurztraminer, and Riesling winegrapes. Am. J. Enol. Vitic. 59 (4), 440-445.

Soles, R.M., Ough, C.S., Kunkee, R.E. 1982. Ester concentration differences in wine

fermented by various species and strains of yeasts. Am. J. Enol. Vitic. 33(2), 94- 98.

Stafne, E.T. (2007). Analysis of the April 2007 Freeze Event in Oklahoma . Oklahoma 65

State University, pp. 1-21.

Styger, G., Prior, B., Bauer, F.F. (2011). Wine flavor and aroma. J. Indust. Micro. & Biochem. 38(9), 1145-1159.

Tominaga, T., Baltenweck-Guyot, R., Peyrot Des Gachons, C., Dubourdieu, D. (2000). Contribution of volatile thiols to the aromas of white wines made from several Vitic vinifera grape varieties. Am. J. Enol. Vitic. 51(2), 178-181.

Versari, A., Laurie, V.F., Ricci, A., Laghi, L., Parpinello, G.P. (2014). Progress in authentication, typification and traceability of grapes and wines by chemometric approaches. Food Res. Int. 60, 2-18.

Webster, D.R., Edwards, C.G., Spayd, S.E., Peterson, J.C., Seymour, B.S. (1993). Influence of vineyard nitrogen fertilization on the concentrations of monoterpenes, higher alcohols, and esters in aged Riesling wines. Am. J. Enol. Vitic. 44(3), 275-284.

Wilson, B., Strauss, C.R., Williams, P.J. (1984). Changes in free and glycosidically bound monoterpenes in developing Muscat grapes. J. Agric. Food Chem. 32, 919- 924.

Wilson, P., Kunze, E., Kurtural, S.K. (2007). A comparison of ectoparasitic nematode populations in American and French-American hybrid grapevines. Ed. Coolong, T., Snyder, J., Smigell, C. pp. 32-34. University of Kentucky.

Yuan, F. & Qian, M.C. (2016). Development of C 13-norisporenoids, carotenoids, and other volatile compounds in Vitis vinifera L. Cv. Pinot noir grapes. Food Chem. 192, 633-641.

Zhang, H., Fan, P., Liu, C., Wu, B., Li, S., Liang, Z. (2014). Sunlight exclusion from Muscat grape alters volatile profiles during berry development. Food Chem. 164, 242-250.

TABLES AND FIGURES

CHAPTER 4: UNTARGETED ASSESSMENT OF THE IMPACTS OF ION EXCHANGE RESIN ON VOLATILE COMPOUNDS PRESENT IN MISSOURI WINES USING GC-MS

ABSTRACT

Cation exchange resins are promising tools for use in the Missouri wine industry for the correction of high pH/TA wines, which are commonly encountered in region due to viticultural conditions and practices that result in grapes with high potassium concentrations at harvest. However, cation exchange has not been widely adopted due to concerns about its tendency to strip flavor and aroma, thus reducing overall wine quality.

Three wines (Valvin Muscat, Chambourcin, Norton) were treated using ion exchange over two vintages (2013 and 2014), then analyzed by GC-MS. Differences in chromatograms were determined using untargeted analysis software. Treated wines were distinguished by principal component analysis. Unique features (85) were determined as being different between control and treatment wines. Phenolic monomers and monoterpenes were most significantly impacted by treatment.

1. INTRODUCTION

1.1 Wine Acidity

An important element of wine stability is pH. A sufficiently low pH is necessary in order for SO2 to function effectively as an antimocrobial (Zoecklin et al., 1995). It also

partially determines the perception of sourness, which is an important consideration in

consumer preference. Titratable acidity (TA) is also an important consideration in the

perception of wine sourness, with high TA values correlating with higher perceptions of

acidity (Plane et al., 1980). 69

Low pH is necessary for the microbial protection of wine because it inhibits the

growth of many types of spoilage organisms. Sulfur dioxide is the main antimicrobial

used in wine production (King et al., 1981), and is generally added as potassium

metabisulfite. In wine solutions, sulfur dioxide (SO2) exists in pH dependant equilibrium

- 2- with bisulfite (HSO3 ) and sulfite (SO3 ), which do not posses the antimicrobial benefits

of SO2 (King et al., 1981). The pKa of SO2 is 1.77. In other words, at pH 1.77, fifty

percent of the sulfur will be in the form SO2, while the other fifty percent will be in the

- - form HSO3 . At typical wine pH values, the equilibrium between SO2 and HSO3 is

- heavily in favor of the HSO3 ion. At pH 3, about eight percent of the sulfur is in the SO2

form. While at pH 4, about one percent is available. For this reason, acidifying high pH

wines greatly increases the antimicrobial effects of sulfur dioxide.

When potassium metabisulfite has not been added to wines, pH still significantly

(Wibowo et al., 1985).

Low pH values do not always correlate with high TA. Under conditions where excessive cation concentrations are present, high pH/ TA wines will occur. Reducing the pH of these wines to levels necessary for the antimicrobial effectiveness of SO2 is difficult. The commonly used method of adding tartaric acid to reduce pH is sometimes impossible since tartaric acid precipitates from the wine solution when excessive salt levels are present. In instances where the tartaric acid does not precipiate, it’s addition 70

further increases TA to levels which are percieved as overly sour by consumers (Jackson,

2014; Walker et al., 2002).

High pH/TA wines are commonly encountered in wines made from hybrid grape

varietals, often due to excessive concentrations of potassium cations. Several methods of

pH/TA imbalance correction are often employed to reduce wine pH without further

increasing the TA; Double salt-deacidification, electrodialysis, cold-stability

manipulation, and ion exchange. Each tehnique is capable of lowering wine pH with

minimal influence on TA. With the exception of cold-stability manipulation, they all

require specialized equipment, can be complex, and require significant time and space

commitments (Garcia-Ruiz et al, 1991; Cole and Boulton, 1989; Nagel et al. 1988; Soares

et al. 2009).

1.2 Ion Exchange

Ion exchange resins have been used in wine remediation since the 1950’s. In

addition to adjusting pH, they have been used for tartrate stabilization, concentrating

grape must, and to remove metals, proteins, volatile acidity, and color ( Bazinet &

Firdaous, 2011; Bonorden et al., 1986; Esau & Amerine, 1966; Lasanta & Gomez, 2012;

based ion exchange treatment systems are inexpensive and widely available due to their

use in water treatment systems (Carreon-Alvarez et al., 2011; Qiang & Yan, 2013; 71

Cuvelier, 1995). A cost analysis performed by Walker et al., (2004) shows that resin

based cation exchange systems are more cost effective than membrane-based cation exchange systems for pH/TA correction (Walker et al., 2004). However, resin-based cation exchange technology has not been adopted by some wine producers due to concerns over unintended color and flavor stripping (Zoecklin et al., 1995).

It was acknowledged in the 1960’s that a wide variety of compounds present in wine could potentially become attached to resins or undergo chemical transformations during treatment (Esau & Amerine, 1966). Esau & Amerine (1966) argued that given the complexity of the wine matrix, experimental evidence is necessary to determine actual impacts of ion exchange. Sensory analysis conducted on wines treated with ion exchange suggests that flavor and aroma is impacted by the treatment and that small amounts of flavor ‘stripping’ may occur in some wines (Bonorden et al, 1986; Walker et al., 2002;

Ibeas et al. 2015, Lasanta et al. 2013). Other sensory studies have found no difference in flavor or aroma after ion exchange treatment (Mira et al., 2006). Gas chromatography has also been applied to determine differences in wines treated with ion exchange (Benitez et al, 2002; Lasanta et al., 2013). However, these studies were designed so that only limited conclusions could be drawn. Both studies also produced results suggesting that some compounds may have increased in concentration after ion exchange. A possible explanation for this occurrence may come from a pH dependent extraction step used in

GC sample preparation. Lasanta et al., (2013) used a solid phase extraction procedure that can favor the retention of acidic or basic compounds based on sample pH (Supelco,

1998). Neutral compounds are not thought to be impacted. The effect of pH on the 72

selective extraction of volatile aroma compounds must be considered, since ion exchange

treated wines are lower in pH than control wines.

Previous experiments have also taken a targeted approach to determine differences in

the aroma profile of ion exchange-treated wines (Benitez et al, 2002; Lasanta et al.,

2013). This approach involves preselecting a number of wine aroma compounds and

determining their concentration. Due to the complexity of wine chromatograms,

thousands of aroma compounds that have potentially been altered by ion exchange

treatment go unmeasured (Arbulu et al., 2015). Recent advances in the field of

metabolomics have produced computer software capable of providing untargeted analysis

of chromatograms. In untargeted analysis, all differences in peak areas among

chromatograms are determined. This technique has been applied to a several wine

chromatogram analysis projects (Robinson et al., 2011; Springer et al., 2014; Vaclavik et

al., 2011).

2. MATERIALS AND METHODS

2.1 Winemaking

Three grape varieties commonly utilized in the Midwest wine industry were

sourced from Missouri vineyards and vinified at the University of Missouri experimental

winery. Valvin Muscat grapes were machine harvested early on September 5th, 2013 and

2014 at Meyers Farm, Mt. Vernon, MO (37.0784°N, 93.7752°W). Approximately 450kg

grapes were received at the MU Experimental winery in a half-ton picking bin. The grapes were loaded into a A Pecaws 500L SK pneumatic press (Prospero, Geneva, NY) and pressed to 1.2 bar. 50mg/L SO2 was added as potassium metabisulfite at the press pan. The pressed volume was approximately 350 L. The juice was pumped into two, 73

200L stainless steel tanks and cold settled for 48 hours at 7oC. The juice was cold settled

for 24 hours, racked, and inoculated with Saccharomyces cerevisiae strain DV10 yeast

(Lallemand, Petaluma, CA) previously rehydrated in 260mg/L GoFerm (Lallemand)

according to manufacturers guidelines. Fermentation lasted approximately 21 days at

13oC until dryness was determined using Clinitest (Bayer). The pH and TA post-

fermentation were 3.51 and 8.0g/L (2013) and 3.87 and 5.5g/L (2014). 50mg/L SO2 was added as potassium metabisulfite to prohibit malolactic fermentation.

Chambourcin grapes were hand-harvested on September 22nd, 2013 and September

28th, 2014 from University of Missouri managed vineyards in Mount Vernon, Mo

(37.0784°N, 93.7752°W). Approximately 450 kg was received at the MU experimental

winery in picking lugs. The grapes lugs were stored at 7oC overnight. The next morning,

they were destemmed and lightly crushed into a 600L stainless steel fermenter. The wine

was then and inoculated with Saccharomyces cerevisiae strain GRE yeast (Lallemand,

Petaluma, CA) previously rehydrated in 260mg/L GoFerm (Lallemand) according to

manufacturer’s guidelines. Fermentation was conducted in a refrigerated room set at

19oC. After 8 days, the end of alcoholic fermentation was determined by a 0.25%

residual sugar reading using Clinitest tablets (Bayer). The must was then pressed in a

membrane press to 1.2 bar. The wine was racked twice before ion exchange treatment.

Malolactic fermentation took place before treatments.

Norton grapes were hand-harvested on October 11th, 2013 and October 8th, 2014 at

Les Bougois Vineyards, Rocheport, Missouri (38.9794°N, 92.5622°W) . Approximately

450kg. was received at the MU experimental winery in an open top fermentor. On the

day of harvest, the grapes were destemmed and lightly crushed into a ½ ton bin. The wine 74

was then inoculated with Saccharomyces cerevisiae strain GRE yeast (Lallemand,

Petaluma, CA) previously rehydrated in 260mg/L GoFerm (Lallemand) according to

manufacturers guidelines. Fermentation was conducted in a refrigerated room set at

20oC. After 7 days, the end of alcoholic fermentation was determined by a 0.25%

residual sugar reading using Clinitest tablets (Bayer). The must was then pressed in a membrane press to 1.2 bar. The pH and TA post-fermentation were 3.74 and 9.64g/L respectively. The wine was racked twice before ion exchange treatment. Malolactic fermentation took place before treatments.

2.2 Ion Exchange Treatment

Amberlite IR-120(H+) ion exchange resin was purchased from Alfa Aesar, Ward

Hill, Mass. 500 grams dry resin was placed into a glass beaker containing deionized water for 72 hours prior to treatment to facilitate swelling. Shortly before treatment, the resin slurry was packed into a 330cm3 stainless steel column fitted with tight mesh screens on each end. Sanitary winery tubing was connected from dispatch tank though a

variable speed pump to the entrance of the column. Tubing at the treated end of the

column directed the wine into 1L plastic graduated cylinders which facilitated the calculation of flow rate and volume treated. Treated wine was stored in 5 gallon glass

carboys. The column was regenerated by pumping an excess volume of 5% sulfuric acid

solution through the resin at a rate of 250mL/min. This was followed by a rinse with

purified water at a rate of 200ml/min to remove all sulfuric acid in the system. The inlet

hose was then attached to a 200L stainless steel tank containing 150L of homogenated

wine. Wine was pumped through the ion exchange system at a rate of approximately 500

mL per minute (80 bed volumes/Hr). The wine was collected in 1 liter graduated 75 cylinders in 960mL increments to represent 3 bed volumes (BV). A 50 mL sample was collected every 3BV and later used for fractional analysis. After collecting the sample, the rest of the treated wine was transferred to a glass carboy. The treatment process was stopped upon resin depletion which was determined once the pH of the treated eluent matched the initial pH (Hanna Instruments HI 2222 pH meter).

After the resin had been determined to be depleted, five gallons of the wine was pumped from the stock tank through the resin and collected as a depleted resin treatment for Norton and Chambourcin wines. This treatment was included to provide insight into how the wine reacts with resin that is not charged with hydrogen. Depleted resin-treated wineundergoes the influences that occur during the treatment process, such as pumping and aeration, that occuring during ion exchange treatment, but without a pH reduction.

Each wine was treated in triplicate using the same resin. After each treatment, the resin was rinsed and regenerated using the above procedure. New resin was used for each of the three wines. Five additional gallons were taken from the stock tank and used as a control. After completing the treatments, the wines were bottled under nitrogen in

Stelvin screwcap closures and stored at 13oC until analysis.

2.3 GC-MS Analysis

Chemicals. Dichloromethane (≤ 99.8%, Sigma, St. Louis, MO), methanol (≤99.9%,

Sigma, St. Louis, MO) ,2-octanol (99%+, Acros, New Jersey), furfural (≤99% Sigma), β- damascenone (≤90% Sigma), 2 –phenylethanol (≤99% Sigma), para-cymene (≤99.5%

Sigma), nerol (≤99.5% Sigma), geraniol (≤99% Sigma), α-terpineol (≤ 97%, Fluka, St.

Louis, MO), rose oxide (Analytical Grade, Sigma), linalool oxide (≤ 97%, Sigma), 76

citronellol (99%, Sigma), TDN (1,1,6 trimethyl-1,2 dihydronapthalene)( ≤ 99%, donated by Sacks Lab, Cornell University) (Kwasniewski et al., 2010).

Volatile analysis. Solid Phase Extraction was performed for each wine on an Alltech 12- port SPE manifold (Fischer, St. Louis, MO) in triplicate using the methods based on

Pineiro et al. 2004. Lichrolut EN SPE cartridges (Merck, Billerica, MA) (200mg solid phase) were used. Prior to SPE extraction, 2-octanol was added at a concentration of

20mg/L to be used as an internal standard. A 0.66 bar vacuum suction was used with 4 ml of water for washing added after loading. Dichloromethane (2ml) was employed as the extraction solvent. The solution was then dried under nitrogen to a final volume of

200ul before injection. Separation and quantification were achieved on a Varian 431-GC

and 220-MS. The chromatographic conditions were the following: injection volume, 1 ul;

injector temp, 200oC; guard column, base deactivated FS (1m, 0.25m diameter) (Bruker,

Billerica, MA); main column, DB-WAXETR (30m, 0.25mm diameter, 0.25um film thickness) (Agilent, Santa Clara, CA); split ratio, splitless (0-3min), 100:1 (3min-51min);

carrier gas, helium; flow rate, 1ml/min; oven ramp, 40oC (hold 6min), 145oC (5oC/min),

175oC (2.5oC/min), 220oC (hold 10min). All injections were performed in triplicate. All

extractions and analyses were performed between July and September 2015.

pH dependency determination. Furfural (10 mg/L), β -Damascenone(3 mg/L), Nerol

(12 mg/L), Geraniol (1 mg/L), 2-Phenylethanol (3 mg/L), and Para-Cymene (4 mg/L) were dissolved in a model wine solution at three different pH values (2.5, 3.3, and 4.0).

Each model wine solution was then extracted in triplicate using SPE cartridges and injected using the procedures listed in section 2.3. Peak areas corresponding to each 77

standard were compared across the three pH values to determine pH influence on

extraction.

Quantification. Compound identification and quantification was achieved using pure

standards in a model wine (12% abv and 0.8g tartaric acid/L with the pH adjusted to 3.3

using 5M NaOH). Calibration curves were assembled for each compound using a 1/x

weighted regression analysis, using 2-octanol as an internal standard. Odor activity

values (OAV’s) were established by dividing a compounds concentration by its published

aroma threshold concentration.

2.4 Data Processing

The files were converted from MS Workstation file types (.SMS) to .CHROM types

using Openchrom (Community Edition) software (Wenig & Odermatt, 2010). A script

capable of dividing all values in a chromatogram was developed in-house using the

Python programming language. Peak areas of the internal standard peaks in each chromatogram were manually obtained and used in the script for chromatogram division.

The files were then converted from .CHROM to a .CDF file extension using Openchrom and loaded onto the XCMS metabolomics database (Tautenhahn et al., 2012). A

pairwise comparison was performed using the GC/SingleQuad (centwave) parameters.

This job produced a list of features that were determined to be different between control,

ion exchange-treated, and depleted resin-treated chromatograms using an unpaired parametric t-test (Welch t-test).

Chromatograms were also assessed visually to determine differences in peak area among treatments. Peaks that appeared to be different in size among treatments were 78

tentatively identified and quantified. Significant differences in quantified compounds

were determined by ANOVA (2-way).

3. RESULTS

3.1 pH Dependency

Differences were observed in the peak areas of some compounds depending on their extraction efficiency (table 4.1). Both β-damascenone and Nerol displayed peak areas

15-20% higher in samples extracted at pH 4 than the samples extracted at pH 2.5 or pH

3.3. ANOVA determined significant differences for both β-damascenone and Nerol with p-values of 0.014 and 0.034 respectively. Furfural, geraniol, and 2-phenylethanol peak areas were not impacted by extraction pH.

3.2 XCMS Detection

The XCMS software produced principal component analysis (PCA) plots for each wine in each vintage (Figure 4.1). Plots A, B, C and D, representing Norton and

Chambourcin wines, were plotted according to principal components 1(PC1) and 2(PC2).

In figure 1A (2013 Chambourcin), the ion exchange and depleted resin chromatograms are distinguished from the control chromatograms based on PC1, which accounts for 17% of the variance. In figure 1B (2014 Chambourcin), the control and depleted resin chromatograms are separated from the ion exchange chromatograms along PC2, which accounts for 24% of the variance. In figure 1C (2013 Norton), the control and depleted resin chromatograms are separated from the ion exchange chromatograms by PC1, which explains 30% of the variance. In figure 1D (2014 Norton), the control and depleted resin 79

chromatograms are separated from the ion exchange chromatograms by PC2, which accounts for 18% of the variance. Figures 1E and 1F, representing Valvin Muscat chromatograms, did not clearly distinguish between control and ion exchange treatment.

XCMS listed the100 most significant differences (Welch’s T-test) in the peak areas of individual ions among the control, depleted resin, and ion exchange chromatograms for the 2013 and 2014 Chambourcin and Norton wines. These features were manually inspected and sorted in the original chromatograms. Features determined to be “false positives” that did not represent chemical compounds were deleted. Ions deemed to be part of an identifiably compound are listed in table 4.2. There were 85 ion features detected as being different between control and ion exchange-treatment chromatograms.

Out of the 85 features, 27 were detected in 2013 and 2014 Chambourcin and Norton wines. For seven ion features, a tentative identification was made based on matching the mass spectral pattern with the NIST library database and published retention indices.

Pure standards were then used for identity confirmation and quantification of these compounds (table 4.3). Benzoic acid, α-phenylethanol, vanillin, and furfuryl alcohol were dramatically reduced by the ion exchange treatment, with the strongest stripping effects observed in the 2013 Chambourcin. Vanillin odor activity values were reduced from 6.3 to 0.3 in 2013 Chambourcin, which should significantly affect the aroma perception of this wine. Equally significant stripping was observed in furfuryl alcohol (from 4 to 0.2 mg/L), α-phenylethanol (248ug/L to >50ug/L) and benzoic acid (from 620 to 220 ug/L).

However, no OAVs for these compounds are readily available, making it unclear whether these compounds are odor-active. Furfuryl alcohol was also reduced dramatically (from

10 to 4mg/L) in the 2014 Chambourcin. 80

3.3 Manual Detection

The chromatograms were also analyzed visually to determine differences among

control and treatments. From visual inspection, 20 compounds were identified as being

potentially different between control and treatment. The peaks were tentatively identified

by their mass spectrum and published retention indices, and their peak areas were

determined manually in each chromatogram (tables 4.4a-4.4c). Reductions in peak area

were determined for 2 unknown monoterpenes, benzyl acetate, eugenol, vitamin A

aldehyde “A” and “B”, isoamyl acetate and several C-6 alcohols in the 2013

Chambourcin. Increases in peak area were determined for 2-hexenyl caproate, ethyl octanoate, and cytolene. In the 2014 Chambourcin, reductions in C-6 alcohol “B”, 2- hexenyl caproate, a linalool derivative, 2 unknown monoterpenes, Vitamin A aldehyde

“A”, benzyl acetate, and eugenol were determined. An increase in vitamin A aldehyde

“B” was also determined. In the 2013 Norton, reductions in isoamyl acetate, C-6 alcohol

“A”, y-nonalactone, ethyl octanoate, unknown monoterpenes “A” and “B”, methyl salicylate, benzyl acetate, eugenol, and vitamin A aldehyde “B” were determined.

Increases in 2-hexenyl caproate and cytolene were determined. In the 2013 Norton, reductions in C-6 alcohol “B”, linalool derivative, unknown monoterpenes “A” and “B”, vitamin A aldehydes “A” and “B”, benzyl acetate, and eugenol were determined. An increase in cytolene was determined. In the 2013 Valvin Muscat, reduction in ethyl octanoate, benzyl acetate, and eugenol were determined, along with an increase in 2- octanone. A reduction in isoamyl acetate was determined for the 2014 Valvin Muscat.

Within each tested wine and each compound measured, the largest peak was 81

assigned a value of 100. The other two peaks were assigned a value based on the

relative proportion of the largest peak. Peaks with relative variation larger than 20 percent

are reported in table 4.5. In total, 36 incidents of variation larger than 20 percent were

determined in 15 compounds. For the other 5 compounds analyzed, variation larger than

20 percent was not observed in any of the wines. In 21 of the 36 incidents, a stripping

effect was observed with the control having the largest peak. This stripping effect was

observed consistently in one or more wines for the following tentatively identified

compounds; ethyl octanoate, 3 unidentified monoterpenes, 2 C6-alcohols, methyl

salicylate, vitamin “A” aldehyde, benzyl acetate, Y-nonalactone, and eugenol. For two

compounds, the largest peak was either in the control or in the ion exchange treatment,

depending on the wine; isoamyl acetate and vitamin A aldehyde “B”. For two compounds

the largest peak was consistently in the ion exchange treatment; cytolene, and 2-octanone.

Monoterpenes were quantified in both Valvin Muscat vintages (table 5.5). In the 2013

vintage, linalool was reduced from 1.7 to 0.9 mg/L in the ion exchange treatments, while

linalool oxide increased from 477 to 543 ug/L, and α-terpineol increased from 0.8 to 1.0

mg/L. In the 2014 Valvin Muscat, α-terpineol also increased (from 0.2 to 0.4mg/L), while

trans-rose oxide and citronellol both decreased from 2.7 to 1.9ug/L and 52 to 38ug/L

respectively.

4. DISCUSSION

4.1 pH Dependency

A 15-20% increase in peak area was determined for compounds extracted in the

higher pH model wine. This finding suggests that neutral species are impacted by

extraction pH during SPE. This is contrary to previous findings (Supelco, 1998). Further 82

experiments will be able to confirm or deny this finding. However, this should put into

suspicion findings of differences in peak areas less than 20 percent, since they may be

due to differences in SPE extraction, not absolute differences among treatments.

4.2 XCMS Detection

The XCMS analysis was able to group the wines according to treatment in the PCA

plots for 2013 and 2014 Norton and Chambourcin (figure 4.1). Since the software does

not list information about the loading plots, it is not possible to determine the principal

components used.

4.3 Manual Detection

All tentatively identified compounds determined to have been stripped during ion

exchange treatment have been identified as aroma-active components of wine. Ethyl

octanoate smells like fruit/fat (Acree, flavornet). The three unknown monoterpenes were

only detected as being stripped in the Norton and Chambourcin wines, where they

probably exist below aroma threshold. C6-alcohols generally contain green/grassy, flower, and bourbon aromas and have been described as characteristic aroma compounds of some white wines (Herbst-Johnstone et al., 2012; Oliviera et al., 2006). Methyl salicylate is a volatile phenol that smells like mint or wintergreen, commonly found in wines made from Vitis riparia grapes (Mansfield, 2008). Vitamin A aldehyde, also known as retinal, is a carotenoid. Carotenoids are odorless precursors which can be converted to C13-norisprenoids during the winemaking process (Kwasniewski et al,

2010). Benzyl acetate is an ester described as having a jasmine aroma, formed from an

esterification reaction between benzyl alcohol and acetic acid during fermentation 83

(Dennis et al., 2012). Y-nonalactone is a sweet, pleasant smelling compound synthesized

by yeast during fermentation (Garbe et al., 2001). Eugenol is a volatile phenol, described

as having a clove or vanilla-like aroma (Kennison et al., 2008). The aroma of 2-

octanone is described as floral or ‘over-ripe fruit’ and subsequently constitutes a major

component of appasamiento-style wines (Lopez de Lerma et al., 2012). Isoamyl acetate

is a fermentation-derived ester with a distinct banana aroma (Plata et al., 2003).

5. CONCLUSION

XCMS software was able to determine significant differences between sets of

chromatograms. It also has the advantage of providing an unbiased assessment of chromatogram differences. However, a significant amount of time was required in order

to get to the point where these differences were confirmed and quantified. More

significant differences were found using manual techniques. Other untargeted

chromatogram software should be investigated. Five phenolic monomers and five

monoterpenes were determined to be stripped from the ion exchange-treated wines.

Therefore, compounds in these classes may be the most likely compounds stripped by ion exchange resin.

REFERENCES

Bazinet, L. & Firdaous, L. (2011). Recent patented applications of ion-exchange membranes in the agrifood sector. Recent patents Chem. Eng. 4 (2), 207-216.

Benitez, P., Castro, R., Barroso, C.G. (2002). Removal of iron, copper, and manganese from white wines through ion exchange techniques: effects on their organoleptic characteristics and susceptibility to browning. Analytica Chimica Acta. 458, 197- 202.

Bonorden, W.R., Nagel, C.W., Powers, J.R. (1986). The adjustment of high pH/high titratable acidity wines by ion exchange. Amer. J. Enol. Vitic. 37 (2), 143-148.

Bousbouras, G.E. & Kunkee, R.E. (1971). Effect of pH on malo-lactic fermentation in wine. Am. J. Enol. Vitic. 22(3), 121-126.

Carreon-Alvarez, A., Herrera-Gonzalez, A., Casillas, N., Prado-Ramirez, R., Estarron- Espinosa, M., Soto, V., de la Cruz, W., Barcena-Soto, M., & Gomez-Salazar, S. (2011). Cu (II) removal from tequila using an ion-exchange resin. Food Chem. 127(4), 1503-9. Cole, J. and Boulton, R. (1989). A study of calcium salt precipitation in solutions of malic acid and tartaric acid. Vtis, 28, 177-190.

Cuvelier, J.M. (1995). Ion exchange resins used in the food industry and in the drinking water preparation in Europe. Reactive Polymers. 24, 109-116. Esau, P. & Amerine, M.A. (1966). Synthetic ion-exchange resins in wine research. Amer. J.Enol. Vitic. 17 (4), 268-276.

Fazzalari, F.A. (1978). Compilation of odor and taste threshold values data. ASTM data series (USA).

Garbe, L.A., Lange, H., Tressl, R. (2001). Biosynthesis of y-nonalactone in yeast. In Aroma Active Compounds in Foods. ACS Symposium Series.

Garcia-Ruiz, J.M., Alcantara, R., & Martin, J. (1991). Evaluation of wine stability to potassium hydrogen tartrate precipitation. Am. J. Enol. Vitic. 42, 336-40. 85

Herbst-Johnstone, M., Araujo, L.D. Allen, T.A., Logan, G., Micolau, L., Kilmartin, P.A. Effects of mechanical harvesting on ‘Sauvignon blanc’ aroma. In I International Workshop on Vineyard Mechanization and Grape and Wine Quality. 978, 179- 186.

Ibeas ,V., Correia, A., Jordao, A. (2015). Wine tartrate stabilization by different levels of cation exchange resin treatments: impact on chemical composition, phenolic profile, and organoleptic properties of red wine. Food Res. Int. 69, 364-372.

Jackson R.S. (2014). Wine science: principles and applications (4th ed.). Amsterdam: Elsevier/Academic Press.

Keith, E.S. & Powers, J.J. (1968). Determination of flavor threshold levels and sub- threshold, additive, and concentration effects. J. Food Sci. 33(2), 213-218.

Kennison, K.R., Gibberd, M.R., Pollnitz, A.P, Wilkinson, K.L. (2008). Smoke-derived taint in wine: the release of smoke-derived volatile phenols during fermentation of Merlot juice following grapevine exposure to smoke. J. Agric. Food Chem. 56, 7379-7383.

King, A.D., Ponting, J.D., Sanshuck, D.W., Jackson, R., Mihara, K. (1981). Factors affecting death of yeast by sulfur dioxide. J. Food Prot. 44, 92-97.

Kwasniewski, M.T., Vanden Heuvel, J.E., Pan, B.S., Sacks, G.L. (2010) Timing of cluster light environment manipulaition during grape development affects C13 norisprenoids and carotenoid concentrations in Riesling. J Agric. Food Chem. 58, 6841-6849.

Lasanta, C. & Gomez, J. (2012). Tartrate stabilization of wines. Trends Food Sci. Technol. 28, 52-59.

Lasanta, C., Caro, I., Perez, L. (2013). The influence of cation exchange treatment on the final characteristics of red wines. Food Chem. 138, 1072-1078.

Lopez de Lerma, N., Bellincontro, A., Mencarelli, F., Morena, J., Peinado, R.A. (2012). Use of electronic nose, validation by GC-MS, to establish the optimum off-vine dehydration time of wine grapes. Food Chem. 130(2), 447-452.

Mansfield, A.K. (2008). Characterization of key volatile compounds in red table wines produced from Frontenac grapes (Vitis spp.) PhD dissertation. University of Minnesota.

Mira, H., Leite, P., Ricardo-Da-Silva, J.M., Curvelo-Garcia, A.S. (2006). Use of ion 86

exchange resins for tartrate wine stabilization. J. Int. Sci. Vigne Vin. 40 (4), 223- 246.

Oliveira, J.M., Faria, M., Sa, F., Barros, F., Araujo, I. (2006). C6-alcohols as varietal markers for assessment of wine origin. Analytica Chimica Acta. 563, 300-319.

Pinero, Z., Palma, M., Barosso, C.G. (2004). Determination of terpenoids in wines by solid phase extraction and gas chromatography. Analytical Chimica Acta. 513(1), 209-214.

Plane, R.A., Mattick, L.R., Weirs, L.D. (1980). An acidity index for the taste of wine. Am. J. Enol. Vitic. 31 (3), 265-268.

Plata, C., Millan, C., Mauricio, J.C., Ortega, J.M. (2003). Formation of ethyl acetate and isoamyl acetate by various species of wine yeasts. Food Micr. 20(2), 217-224.

Qiang, Y. & Yan, Y. (2013). Study on the deacidifying condition of litchi wine by ion- exchange resin 314. Modern Food Sci. & Tech. 29, 380-2. Robinson, A.L., Boss, P.K., Heymann, H., Solomon, P.S., Trengrove, R.D. (2011). Development of a sensitive non-targeted method for characterizing the wine volatile profile using headspace solid-phase microextraction comprehensive two- dimensional gas chromatography time-of-flight mass spectrometry. J. Chromatography A. 1218 (3), 504-517.

Soares, P.A.M.H., Geraldes, V., Fernandes, C., Cameira dos Santos, P., Norberta de Pinho, M. (2009). Wine tartaric stabilization by electrodialysis: Prediction of required deionization degree. Am. J. Enol. Vitic. 60, (2), 183-188. Springer, A.E., Riedl, J., Esslinger, S., Roth, T., Glomb, M.A., Fauhl-Hassek, C. (2014). Validated modeling for German white wine varietal authentication based on headspace solid-phase microextraction online coupled with gas chromatography mass spectrometry fingerprinting. J. Agric. Food Chem. 62 (28), 6844-6851.

Supelco. (1998). Guide to solid phase extraction. Bulletin 910. Supelco, Bellefonte, PA.

Tautenhahn, R., Patti, G.J., Rinehart, D., Siuzdak, G. (2012). XCMS Online: A Web- Based Platform to Process Untargeted Metabolomics Data. Anal. Chem.. 84(11). 5035-5039.

Walker, T., Morris, J., Threlfall, R., Main, G. (2002). pH modification of Cynthiana Wine using cationic exchange. J. Agric. Food Chem. 50 (22), 6346-6352.

Walker, T., Morris, J., Threlfall, R., Main, G. (2004). Quality, sensory and cost comparison for pH reduction of Syrah wine using ion exchange or tartaric acid. J. Food Qual. 27, 483-496

Wenig, P., Odermatt, J. (2010). OpenChrom: a cross-platform open source software for the mass spectrometric analysis of chromatographic data. BMC bioinformatics, 11(1), 405.

Wibowo, D., Eschenbruch, R., Davis, C.R., Fleet, G.H., Lee, T.H. Occurrence and growth of lactic acid bacteria in wine: a review. Am J. Enol. Vitic. 36(4),1985.

Zoecklin, B.W., Fugelsang, K.C., Gump, B.H., Nury, F.S. (1995). Wine Analysis and Production. New York: Chapman & Hall.

TABLES AND FIGURES

APPENDIX