YEAST POPULATION DYNAMICS DURING INOCULATED AND SPONTANEOUS FERMENTATIONS AT THREE LOCAL BRITISH COLUMBIA WINERIES JESSICA NICOLE LANGE

YEAST POPULATION DYNAMICS DURING INOCULATED AND SPONTANEOUS FERMENTATIONS AT THREE LOCAL BRITISH COLUMBIA WINERIES by JESSICA NICOLE LANGE B.Sc., The University of British Columbia, 2010 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in The College of Graduate Studies (Biology) THE UNIVERSITY OF BRITISH COLUMBIA (Okanagan) December 2012 Jessica Nicole Lange, 2012

ABSTRACT Little assessment of yeast assemblages has occurred in Canadian wineries, unlike other large wine producing regions (Spain, Italy, Argentina). The aim of this study was to compare yeast assemblages during inoculated and spontaneous fermentations at three Canadian wineries. The wineries (Quails Gate Estate Winery, QGEW; Cedar Creek Estate Winery, CCEW; and Road13 Estate Winery, R13EW) are located in the Okanagan wine region, British Columbia. All three wineries have a history of using commercial yeast. During the 2010 vintage, nine inoculated and three spontaneous Vitis vinifera L. var. Pinot noir fermentations were sampled from four distinct stages of fermentation. Yeast populations from inoculated fermentations were also assessed at QGEW during the following vintage in 2011. Saccharomyces cerevisiae isolates were discriminated at the strain level by microsatellite analysis of hyperviariable trinucelotide loci. Non- Saccharomyces species were identified by sequencing the ITS and the D1/D2 domain regions of the large subunit of rdna. Non-Saccharomyces spp., particularly Henseniaspora uvarum, were the dominant yeasts detected during cold-soak at all three wineries. Spontaneous fermentation appeared to have a greater species/strain diversity/richness than inoculated fermentation at the youngest (R13EW) of the three wineries. Commercial strains were isolated in relatively low frequencies in the spontaneous fermentation at this winery, whereas at the older wineries (QGEW and CCEW) commercial strains dominated fermentation. R13EW was the only winery where the commercial ADY inoculant fully implanted. At QGEW and CCEW, a commercial yeast, Lalvin ICV-D254, was the major non-inoculant strain detected in both inoculated and spontaneous fermentations. Only QGEW and CCEW reported previous use of this ii

strain in other varietals. Nevertheless, the different wineries exhibited unique yeast species/strain assemblages at all stages of fermentation, even cold-soak. During both vintages studied at QGEW, the non-inoculant ADY strain (Lalvin ICV-D254) was dominant or co-dominant in inoculated fermentation. Thus, mixed-strain populations in inoculated tanks were observed in both years. This study emphasizes the need for further research on whether the age of a winery is a major factor in affecting the yeast assemblage of fermenting wine, the source(s) of non-inoculant yeast, and the effects yeasts have on the sensorial attributes of the finished wine product. iii

PREFACE The samples for this study were collected from three local British Columbia wineries, including Quail s Gate, Cedar Creek, and Road13 Estate Wineries. With guidance from my supervisor, Dr. Daniel Durall, I was responsible for developing the experimental design, collecting all experimental data during the 2010 vintage, and writing the thesis. Mansak (Ben) Tantikachornkiat (BSc with Honors candidate, UBCO) was responsible for collecting data from the 2011 vintage at Quails Gate Estate Winery for an Honor s study. His data were utilized for comparing the 2010 and 2011 inoculated populations detected at this winery. The data described in this thesis were presented as an invited oral presentation at the Annual Conference of the American Society of Ecology and Viticulture in Portland, OR, USA and at the Annual Conference of the British Columbia Wine Grape Council in Penticton, BC, Canada in June and July, 2012, respectively. Portions of the abstract were published in the American Journal of Enology and Viticulture. This thesis was reviewed by the following members of my supervisory committee, all faculty at the University of British Columbia (Okanagan): Dr. Dan Durall, Dr. Louise Nelson, and Dr. Cedric Saucier. iv

TABLE OF CONTENTS ABSTRACT....ii PREFACE...iv TABLE OF CONTENTS.....v LIST OF TABLES...viii LIST OF FIGURES.....x LIST OF ABBREVIATIONS.......xii ACKNOWLEDGEMENTS..xiv CHAPTER 1: INTRODUCTION............1 1.1 INTRODUCTION TO WINEMAKING...1 1.2 THE ORIGIN AND NATURAL HABITAT OF WINE YEAST.....2 1.3 SUCCESSION OF WINE YEAST SPECIES DURING FERMENTATION..5 1.4 INOCULATED AND SPONTANEOUS FERMENTATIONS 9 1.5 THE USE OF MICROSATELLITE MARKERS FOR S. CEREVISIAE STRAIN-LEVEL IDENTIFICATION....12 1.6 RESEARCH OBJECTIVES....13 1.7 RESEARCH HYPOTHESES..14 1.8 RATIONALE OF HYPOTHESES..15 CHAPTER 2: AN ASSESSMENT OF YEAST POPULATION DYNAMICS DURING INOCULATED AND SPONTANEOUS FERMENTATIONS AT THREE LOCAL BRITISH COLUMBIA WINERIES.......17 2.1 SYNOPSIS...17 2.2 MATERIALS AND METHODS.....19 2.2.1 Study sites and experimental design... 19 2.2.2 Collection of samples....20 2.2.3 Inoculation....22 v

2.2.4 Yeast isolation... 23 2.2.5 Identification of yeast strains and species..... 25 2.2.6 UBCO ADY commercial S. cerevisiae microsatellite comparative database........ 29 2.2.7 Data analysis..... 31 2.3 RESULTS....34 2.3.1 Isolation and identification of yeasts....34 2.3.2 Succession of non-saccharomyces and S. cerevisiae yeasts....37 2.3.3 Detection of the commercial ADY inoculant...38 2.3.4 Comparison of species/strain diversity and richness between spontaneous and inoculated fermentations...40 2.3.5 Comparison of yeast assemblages between wineries....41 2.3.6 Comparison of inoculated yeast populations between 2010 and 2011 vintages at Quails Gate Estate Winery..42 2.4 DISCUSSION......43 2.4.1 Species/strain accumulation curves and study sample size..43 2.4.2 The effectiveness of microsatellite DNA analysis and comparative commercial ADY database..44 2.4.3 Commercial S. cerevisiae yeast dominate fermentation.....46 2.4.4 Non-Saccharomyces spp. and S. cerevisiae succession....50 2.4.5 Implantation of the commercial ADY inoculant..53 2.4.6 Species/strain diversity and richness of spontaneous fermentation..56 2.4.7 Comparison of cold-soak, inoculated, and spontaneous yeast assemblages between wineries... 58 2.4.8 Comparison of inoculated yeast populations detected between two vintages at Quails Gate Estate Winery........61 2.5 SUMMARY..........63 vi

CHAPTER 3: CONCLUSION..............66 3.1 CONCLUSION SUMMARY.............. 66 3.2 NOVELTY OF THE RESEARCH......67 3.3 MANAGEMENT IMPLICATIONS....... 68 3.4 ASSUMPTIONS AND LIMITATIONS..... 70 3.5 SUGGESTIONS FOR FURTHER RESEARCH.... 71 REFERENCES.............93 APPENDIX A...... 102 vii

LIST OF TABLES Table 1.1: Comparative table of literature on wine yeast ecology studies....74 Table 2.1: Fermentation data of 2010 tanks at (A) QGEW; (B) CCEW; (C) R13EW..76 Table 2.2: Fermentation data of inoculated tanks from the 2011vintage at QGEW.. 79 Table 2.3: Commercial S. cerevisiae strains reported used since the establishment of QGEW, CCEW, and R13EW. x indicates strain used 80 Table 2.4: Total number of commercial and non-commercial unknown S. cerevisiae strains, Saccharomyces spp., and non-saccharomyces yeasts detected at QGEW, CCEW, and R13EW in the 2010 vintage........81 Table 2.5: Simpson s diversity index (D) and abundance-based coverage (ACE) estimators of yeast species/strain richness and diversity of (A) QGEW; (B) CCEW; and (C) R13EW inoculated and spontaneous tanks during cold-soak (CS), early (ER), mid (M), and end (F) stages of fermentation....81 Table 2.6: Similarity indices of inoculated (T1-T3) and spontaneous (W) population assemblages during cold-soak (CS), early (ER), mid (M), and end (F) stages of fermentation at (A) QGEW 1 ; (B) CCEW 1 ; (C) R13EW 1. Grey shading indicates 80% shared similarity yeast species/strain assemblages between tanks and stages of fermentation.....82 Table 2.7: Permutational MANOVAs from QGEW, CCEW, and R13EW s inoculated, spontaneous, cold-soak, and QGEW year-to-year non-metric multidimensional scaling (NMS) ordinations. DF= degrees of freedom; SS= sum of squares...... 83 Table A.1: Commercial S. cerevisiae ADY microsatellite database constructed at UBCO...102 Table A.2: Primer information for ten loci evaluated by microsatellite DNA analysis (bp = basepairs). 103 Table A.3: Yeast species/strain frequency of occurrence from inoculated (T1-T3) tanks during cold-soak (CS), early (ER), mid (M), and end (F) stages of fermentation at QGEW, CCEW, and R13EW in the 2010 vintage..104 Table A.4: Yeast species/strain frequency of occurrence from the spontaneous (W) tank during early (ER), mid (M), and end (F) stages of fermentation at QGEW, CCEW, and R13EW in the 2010 vintage 105 viii

Table A.5: Yeast species/strain frequency of occurrence from inoculated tanks during cold-soak (CS), early (ER), mid (M), and end (F) stages of fermentation at QGEW from the 2011 vintage.....106 Table A.6: Accession number (#) and percent (%) similarity for the BLAST matches of non-saccharomyces isolates detected in inoculated (T1-T3) or spontaneous (W) fermentation tanks for Quails Gate (QG); Cedar Creek (CC); Road13 (R13) Estate Wineries; and Quails Gate 2011 vintage (QG-2011) isolated in cold-soak (CS), early (ER), or mid (M) stages of fermentation. Amplification of ITS and D1/D2 regions of ribosomal DNA using ITS1f/ITS4 and NL1/NL4 universal primer sets.. 107 ix

LIST OF FIGURES Figure 2.1: Yeast species/strain frequency of occurrence at QGEW (2010) during cold-soak (CS), early (ER), mid (M), and end (F) stages of inoculated (A-C) and spontaneous (D) fermentation. (A) QG-T1; (B) QG-T2; (C) QG-T3; (D) QG-W.......84 Figure 2.2: Yeast species/strain frequency of occurrence at CCEW (2010) during cold-soak (CS), early (ER), mid (M), and end (F) stages of inoculated (A-C) and spontaneous (D) fermentation. (A) CC-T1; (B) CC-T2; (C) CC-T3; (D) CC-W.....85 Figure 2.3: Yeast species/strain frequency of occurrence at R13EW (2010) during cold-soak (CS), early (ER), mid (M), and end (F) stages of inoculated (A-C) and spontaneous (D) fermentation. (A) R13-T1; (B) R13-T2; (C) R13-T3; (D) R13-W.....86 Figure 2.4: Species/strain accumulation curves of the 2010 vintage using Mao Tau individual-based rarefaction with replacement during cold-soak (CS), early (ER), mid (M), and end (F) stages of fermentation from inoculated tanks (n=3) at: QGEW (QG- A through D); CCEW (CC- E through H); and R13EW (R13- I through L). Dashed lines indicate Mao Tau 95% upper and lower bound confidence intervals.....87 Figure 2.5: Species/strain accumulation curves using Mao Tau individual-based rarefaction with replacement during early (solid), mid (dashed), and end (double dashed) stages of fermentation from spontaneous tanks (W, n=1) at (A) QGEW; (B) CCEW; and (C) R13EW..88 Figure 2.6: Species/strain accumulation curves using Mao Tau individual-based rarefaction with replacement during cold-soak (CS), early (ER), mid (M), and end (F) stages (A-D) of the 2011 fermentation from inoculated tanks (n=3) at QGEW. Dashed lines indicate Mao Tau 95% upper and lower bound confidence intervals.....88 Figure 2.7: 2D nonmetric multidimensional scaling (NMS) ordination of QGEW (QG - red triangles), CCEW (CC- orange squares), and R13EW (R13- green circles) yeast species/strain assemblages detected in inoculated tanks (T1-T3) for early (ER), mid (M), and end (F) stages of fermentation. The percent of variation is 23.7% for the x-axis and 48.4% for the y-axis. The final stress is 15.90 and the resulting p-value is p = 0.002.89 x

Figure 2.8: 2D nonmetric multidimensional scaling (NMS) ordination of QGEW (QG - red triangles), CCEW (CC- orange squares), and R13EW (R13- green circles) yeast species/strain assemblages detected in spontaneous (W) fermentation tanks for early (ER), mid (M), and end (F) stages of fermentation. The percent of variation is 64.9% for the x-axis and 15.0% for the y-axis. The final stress is 0.096 and the resulting p-value is p = 0.003....90 Figure 2.9: 2D nonmetric multidimensional scaling (NMS) ordination of QGEW (QG - red triangles), CCEW (CC- orange squares), and R13EW (R13- green circles) yeast species/strain assemblages detected in cold-soak (CS) stage of fermentation in tanks (T1-T3). The percent of variation is 46.3% for the x-axis and 34.2% for the y-axis. The final stress is 12.37 and the resulting p-value is p = 0.003..... 91 Figure 2.10: 2D nonmetric multidimensional scaling (NMS) ordination of QGEW inoculated fermentation tanks (T1-T3) yeast species/strain communities between years 2010 (red triangles) and 2011 (blue squares). The percent of variation is 54.4% for the x-axis and 17.8% for the y-axis. The final stress is 9.27 and the resulting p-value is p = 0.110. 92 xi

LIST OF ABBREVIATIONS ACE Abundance-based coverage estimator ADY Active dry yeast AMH Assmanshausen (Enoferm ) AP Arome Plus (Lalvin ) BLAST Basic Alignment Search Tool BSA Bovine Serum Albumin CC Cedar Creek CCEW Cedar Creek Estate Winery CS Cold-soak D Simpson s Index DNA Deoxyribonucleic Acid dntp Deoxynucleoside Triphosphate ER Early EW Estate Winery F End FADSS Fragment Analysis and DNA Sequencing Services FAGE Fast Adaptive Genome Evolution ITS Internal Transcribed Region NCBI National Centre for Biotechnology Information NGS Next Generation Sequencing NMS Nonmetric multidimensional scaling M Mid MCT Microcentrifuge Tube PCR Polymerase Chain Reaction PCuvee Premium Cuvee (Red Star ) permanova Permutational Multivariate Analysis of Variance QG Quails Gate QGEW Quails Gate Estate Winery R13 Road13 R13EW Road13 Estate Winery xii

SD Sobs T UBCO UN Var. W WL YEPD Standard Deviation Species Observed Tank (Inoculated Fermentation) University of British Columbia Okanagan Unknown Varietal Spontaneous Fermentation Wallerstein Laboratory Yeast Extract Peptone Dextrose xiii

ACKNOWLEDGEMENTS A sincere thank-you to my supervisor, Dr. Daniel Durall for all of his direction and invaluable guidance throughout this project. I would also like to thank my supervisory committee members, Dr. Louise Nelson and Dr. Cedric Saucier, for all of their encouragement, interest, constructive criticism, and positive feedback to my thesis. A big thank-you is given to all those who assisted me with lab work and statistical analysis. I was assisted with sampling and molecular work at Quails Gate Estate Winery by Liz Halverson (BSc with Honors, UBCO), plate streaking by Hannah Pawluck (BSc, UBCO), DNA analysis by Sheri Maxwell (UBCO FADSS Technician); and with statistical analysis by Dr. Melanie Jones, post-docs Dr. Matthew Whiteside and Dr. Brian Pickles, and Brian Ohsowski (PhD candidate). Additionally, I would like to thank Mansak (Ben) Tantikachornkiat for processing data collected from Quails Gate Estate Winery during the 2011 vintage. A sincere thanks is given to the winemakers Grant Stanley of Quails Gate Estate Winery, Darryl Brooker of Cedar Creek Estate Winery, and Michael Bartier of Road13 Estate Winery for their patience and support while I collected samples during the 2010 vintage. Also, thank-you to David Ledderhof of Quails Gate Estate Winery for being so informative throughout this project. Funding for this project was graciously provided by Quails Gate Estate Winery and Natural Sciences and Engineering Research Council (NSERC) through a collaborative research development (CRD) grant. xiv

CHAPTER 1: INTRODUCTION 1.1 INTRODUCTION TO WINEMAKING Winemaking is an ancient art, which dates back to the Neolithic period (8500 to 4000 B.C.). One piece of evidence for early winemaking lies in the discovery of Egyptian fermentation jars containing Saccharomyces cerevisiae cell remnants and Vitis vinifera plant material (Cavalieri et al. 2003). By 2000 B.C., the cultivation of Vitis vinifera and production of wine was heavily practiced in Greece and Crete. By 500 B.C., winemaking became well established in Sicily, Italy, France, Spain, Portugal and northern Africa. Subsequently, it spread to Germany and to the greater portion of Europe (Pretorius 2000). The practice of viticulture, storage, and aging of wine has developed tremendously over the years; yet, the primary organisms responsible for alcoholic fermentation during vinification, i.e., Saccharomyces and non-saccharomyces species, remain unchanged. The main yeast involved in alcoholic fermentation is Saccharomyces cerevisiae; however, the transformation of macerated grapes into a finished wine product involves an array of complex interactions between related genera (S. uvarum, S. bayanus, S. paradoxus), non-saccharomyces spp., filamentous fungi (Aspergillus, Botrytis and Penicillium), lactic and acetic acid bacteria (those organisms responsible for malolactic fermentation), mycoviruses, and bacteriophages (Pretorius 2000; Fleet 2007; Garijo et al. 2008). Together, their dynamics, development, succession, and metabolic activity during fermentation generate a variety of secondary metabolites, higher alcohols, esters, aldehydes, ketones, volatile compounds, organic acids, and enzymes, which affect the final organoleptic and aromatic properties of wine (Mateo et al.1992; Romano et al. 1

2003; Bisson and Karpel, 2010). There are two methods of fermentation practiced by winemakers today: spontaneous and inoculated fermentation. Spontaneous fermentation is the traditional approach to winemaking and it relies on the spontaneous entrance of indigenous (native) yeast into the grape must. It is thought to achieve increased yeast diversity and wine complexity (Maro et al. 2007; Tello et al. 2011). Inoculated fermentation is a modern approach to winemaking and involves the addition of a commercial active dry yeast (ADY) to the grape must in order to initiate fermentation. Inoculated fermentation attempts to achieve a yeast monoculture through the dominance of a particular ADY added to the must. Inoculation usually results in an efficient fermentation with a consistent, predictable, and reproducible wine product (Pretorius 2000; Maro et al. 2007). Wine yeast species affect aromatic and sensorial characteristics of wine and have a unique impact on the oenological properties. The frequency, diversity, and dynamics of yeast populations may affect the overall characteristics of the wine; therefore, making the study of yeast population dynamics an important part of the wine industry (Schutz and Gafner, 1993; Mateo et al. 1992; Maro et al. 2007). 1.2 THE ORIGIN AND NATURAL HABITAT OF WINE YEAST The origin and natural habitat of non-saccharomyces and Saccharomyces wine yeast species are not fully understood, although analysis of yeast ecology performed at various wineries has given insight to this topic of study. The standing hypothesis for the initial origin of wine yeast in a newly established winery is as follows: Saccharomyces and non-saccharomyces species are thought to have been deposited naturally onto the surface of the grapes by wind currents and insects including bees, wasps, and Drosophila 2

species (Stevic 1962; Lachance et al. 1994; Mortimer and Polsinelli, 1999). Clusters of grapes analyzed from various areas of the vineyard show the dominance of particular yeast species while some non-saccharomyces species are inconsistent in their occurrence. For example, some grape clusters show a large dominance of the yeast Metschnikowia, while other clusters show no sign of this yeast, indicating that the source of this indigenous yeast may be from an insect that carried the particular yeast species from its origin to the vineyard (Mortimer and Polsinelli, 1999). Samples taken from grape skins in the vineyard report Hanseniaspora uvarum (anamorph Kloeckera apiculata) and Candida stellata as the dominant non-saccharomyces species isolated with minor detection of species belonging to Torulospora, Metschnikowia, Kluyveromyces, Cryptococcus, Oenococcus, Pichia, Issatchenkia, and Rhodotorula genera (Zott et al. 2008; Ocon et al. 2010; Barata et al. 2012). The non-saccharomyces yeast population peaks when the berry is most ripe, as it provides a nutritious medium and a large surface area for growth (Renouf et al. 2005). There are two conceptual views held with respect to the predominance and natural habitat of non-saccharomyces and S. cerevisiae yeasts. The first view argues that non- Saccharomyces spp. are vineyard resident and Saccharomyces spp. are winery resident yeasts. Non-Saccharomyces spp. have been detected in large numbers within the vineyard and as mentioned previously are found to heavily colonize the surface of Vitis vinifera grapes, as opposed to their low detection in the winery establishment (Vaughan- Martini and Martini, 1995; Jolly et al. 2003; Mercado et al. 2007). The view of S. cerevisiae as winery resident yeast is based on widespread sampling of wineries with 3

detection of S. cerevisiae occurring on winery equipment, hoses, walls, stemmercrushers, pumps, (Martini 1993; Ciani et al. 2004; Clavijo et al. 2010) and in the winery air (Garijo et al. 2008). Furthermore, the same S. cerevisiae strains have been isolated over multiple sampling years indicating their ability to reside within the winery from one year to another (Beltran et al. 2002; Santamaria et al. 2005; Clavijo et al. 2010). Although sanitization measures are employed in the winery, some yeasts are able to survive these cleaning measures (Pretorius 2000; Ocon et al. 2010). This may be due to some yeasts ability to produce ascospores, enabling their survival over long periods of time (Vaughan-Martini and Martini, 1995). In contrast, the second view argues that non-saccharomyces and S. cerevisiae yeasts reside in both the vineyard and the winery. Non-Saccharomyces and S. cerevisiae isolates may be detected in greater numbers in the vineyard and within the winery, respectively; however, the detection of both species in the vineyard and winery has been well reported in the literature (Sabate et al. 2002; Le Jeune et al. 2006; Cordero-Bueso et al. 2011; Hall et al. 2011). The transport of Vitis vinifera grapes into the winery creates a large-scale introduction of the non-saccharomyces population into the winery setting, while humans, insects, and the common practice of recycling and distributing lees back into the vineyard as natural fertilizer exposes the vineyard to an S. cerevisiae population making it plausible that S. cerevisiae and non-saccharomyces are both vineyard and winery resident yeasts (Ganga and Martinez, 2004; Clavijo et al. 2010; Hall et al. 2011). 4

The reports of varying composition and transient detection of S. cerevisiae and non-saccharomyces spp. in the vineyard may be due to a number of environmental and conditional factors affecting their survival, such as climatic conditions (rain, temperature, humidity), viticulture practice (leaf thinning), geographical zone, insects, and vintage (Valero et al. 2005; Mercado et al. 2007; Maro et al. 2007). Furthermore, the reporting from some studies of an absence of S. cerevisiae in the vineyard may be due to procedural limitations. Without the use of enrichment methods, some wine yeast species on natural vineyard surfaces, including soil, V. vinifera vines, and grape skins may go undetected (Fleet and Heard, 1993; Beltran et al. 2002; Barata et al. 2012). Future research using next-generation sequencing (NGS) may help to elucidate the seemingly undetectable S. cerevisiae population residing in the vineyard as thousands of samples may be multiplexed using DNA sequence barcodes able to characterize 99.99% of the microbiota (Hamady et al. 2008; Bokulich et al. 2012). A recent study revealed the power of NGS in wine microbial ecology upon analyzing the bacterial diversity of botrytized wine (Bokulich et al. 2012); however, to my current knowledge, NGS has not been employed on wine yeast populations present in the vineyard or during fermentation. This technique appears to analyze populations at more in-depth levels then first-generation profiling technologies can provide (Bokulich et al. 2012), seemingly beneficial for future wine yeast ecology studies. 1.3 SUCCESSION OF WINE YEAST SPECIES DURING FERMENTATION The yeast population dynamics of wine fermentation involve complex interactions between a vast collection of non-saccharomyces and Saccharomyces species and strains. 5

Wine fermentation is generally divided into the following stages: cold-soak, early, mid, and end fermentative stages. The conversion of grape must into wine involves the development and succession of non-saccharomyces and S. cerevisiae populations throughout these stages. Wine yeast fermentation succession includes the development of non-saccharomyces species during the cold-soak stage followed by their replacement with S. cerevisiae yeasts during the early, mid, and end stages of fermentation. The non- Saccharomyces species are generally ethanol-intolerant, heat intolerant, oxygen dependent, and less competitive by nature ultimately leading to their death as fermentation progresses (Hansen et al. 2001; Combina et al. 2005; Le Jeune et al. 2006). Species belonging to Hanseniaspora (anamorph Kloeckera apiculata) and Candida have been isolated in greatest numbers during the cold-soak stage of fermentation along with a minor sub-population composed of Torulospora, Metschnikowia, Cryptococcus, Oenococcus, Pichia, Issatchenkia, and Rhodotorula species (Hierro et al. 2006; Zott et al. 2008; Ocon et al. 2010). The degree of berry ripeness, the cold maceration process, and sulphur dioxide (SO 2 ) addition prior to coldsoak have been shown to greatly affect the predominance and diversity of non- Saccharomyces species during the initial phases of fermentation (Hierro et al. 2006). For example, Candida diversa and Issatchenkia sp. are detected in higher numbers when the must contains a lower sugar concentration due to the maceration of un-ripe grapes. In contrast, the maceration of very ripe grapes generates a must with high sugar concentration promoting the dominance of Candida stellata and H. uvarum populations (Hierro et al. 2006). Furthermore, age of the vineyard, grape variety and, as previously 6

mentioned, climatic conditions, geographical location, insects, and vintage may greatly affect the composition of vineyard yeast and, consequently, affect the yeast population during the initial stages of fermentation (Mercado et al. 2007; Maro et al. 2007; Barata et al. 2012). During the time of cold-soak, the must s nutrient composition and oxygen concentration are high and the tank temperature, ethanol, and competition relatively low. These conditions support the development and growth of the sensitive non- Saccharomyces yeast population (Combina et al. 2004). As fermentation progresses, S. cerevisiae increase in density and non- Saccharomyces species decline (Zott et al. 2008). The resultant increased temperature along with increased ethanol and carbon dioxide create unfavourable living conditions for non-saccharomyces species (Vaughan-Martini and Martini, 1995). The production of ethanol, heat, and carbon dioxide with the ability to ferment in the presence of high glucose and low oxygen concentrations is thought to be an evolved mechanism in S. cerevisiae, known as the Crabtree effect. The Crabtree effect decreases competition with surrounding microbiota and increases the fitness and survival of the S. cerevisiae population (Goddard 2008). Other factors may cause the death of non-saccharomyces species. These include the presence of toxic compounds secreted by killer yeasts (Heard and Fleet, 1987; Zargoc et al. 2001), nutrient depletion (glucose, nitrogen, amino acids), and cell-to-cell contact (Nissen et al. 2003). Cell-to-cell contact may cause the death of non-saccharomyces populations in response to space limitation due to inter-species space competition (Nissen et al. 2003). Furthermore, the non-saccharomyces density may be adversely affected by the addition of sulphur dioxide to the must, which is intended to kill 7

off any bacterial or hyphal fungal microbiota (Ocon et al. 2010). As ethanol concentrations rise during the initial stages of fermentation (cold-soak and early), parabolic death kinetics are expressed by the non-saccharomyces population (Nissen and Arneborg, 2003). The parabolic death curve is expressed when different times of death between various non-saccharomyces species occur. Initially, the least tolerant species to ethanol and depleting sugar concentrations begin to die, followed by the mid-tolerant species, and then lastly, the death of the most tolerant non-saccharomyces species (Nissen and Arneborg, 2003). The occurrence of S. cerevisiae in the must during the early, mid, and end stages of fermentation may occur spontaneously or by the addition of a commercial ADY by the winemaker. On average, a winery is exposed to billions of yeast cells every vintage making the spontaneous implantation of various S. cerevisiae strains into the must very probable (Beltran et al. 2002). Saccharomyces cerevisiae strains are commonly isolated on winery equipment, tanks, hoses, walls, and floors of the winery establishment, are able to survive on winery surfaces for extended periods of time and between annual harvests, and may enter tanks via airborne transfer or by the placement of winery machinery or tools into the must (Vaughan-Martini and Martini, 1995; Constanti et al. 1997; Torija et al. 2000; Valero et al. 2005). Because of these various sources by which S. cerevisiae is introduced, as well as other varying conditions, the S. cerevisiae populations often show considerable diversity in strain dynamics during fermentation within or between wineries (Schutz and Gafner, 1993; Beltran et al. 2000; Mercado et al. 2007). Another likely mode by which S. cerevisiae is introduced into a fermentative tank is through airborne 8

mechanisms. Saccharomyces cerevisiae cells are thought to travel in the form of a single particle, attached to a dust particle, or as a suspended aerosol droplet (Garijo et al. 2008). 1.4 INOCULATED AND SPONTANEOUS FERMENTATIONS Commercial ADY originate from indigenous strains isolated in wine-producing regions (Pretorius 2000; Schuller and Casal, 2007). In 1965, S. cerevisiae strains, Montrachet and Pasteur Champagne, were the first commercial ADY to be manufactured and sold for inoculating purposes as these specific strains appeared to harbour good fermentation parameters (Pretorius 2000). Today, several manufacturing companies exist distributing a wide selection of commercial ADY. Commercial ADY aid in the prevention of stuck fermentations, a risk taken when performing spontaneous fermentations (Bisson 1999). Because of this, the practice of adding commercial ADY to the grape must is becoming more and more popular within the wine industry as it has proven to decrease the risk of wine spoilage and promote a complete and rapid fermentation (Valero et al. 2007). Spontaneous fermentation is the traditional approach to winemaking and is still heavily practiced in Old World wine-producing regions, although the practice of wild ferments in North American wineries is becoming increasingly common. There is a claim that spontaneous fermentation allows for the development of indigenous yeast and a terrior associated with the vineyard and winery, promoting the development of a distinct wine product unique and reflective of its producing region (Pretorius 2000; Vilanova et al. 2005; Fleet 2007). In comparison to inoculated fermentation, spontaneous 9

fermentation arises independently of yeast inoculation and relies on the introduction and completion of fermentation by the S. cerevisiae microflora of the vineyard and winery. Spontaneous fermentation is thought to achieve increased strain diversity as no one particular strain was intentionally added to the tank. This allows for the development of many different wine yeasts (Barrajón et al. 2009; Csoma et al. 2010; Tello et al. 2011). Strains of S. cerevisiae generate unique aromatic, sensorial, and chemical compounds in the grape juice and a fermentation composed of many wine yeasts is thought to yield a wine product composed of many sensorial and organoleptic properties (Mateo et al. 1991). Concerns of competition and dominance of commercial S. cerevisiae over indigenous yeasts have been raised (Santamaria et al. 2005; Maro et al. 2007). Commercial S. cerevisiae strains have been isolated in spontaneous tanks and found to compete with the indigenous strains (Constanti et al. 1997; Beltran et al. 2000; Hall et al. 2011). Commercial ADY are chosen based on specific fermentative characteristics, such as quick fermentation initiation, fermentation at low temperatures, tolerance to high temperatures, low hydrogen sulfide production, and high ethanol intolerance; most of which may aid in their ability to out-compete weaker yeasts (Barrajón et al. 2009). The better adapted commercial ADY for fermentation may decrease the overall diversity of weaker wine yeasts by out-competing them and predominantly performing fermentation (Ganga and Martinez, 2003). It has been suggested that a decrease in strain diversity during fermentation may affect the overall distinctive aromatic qualities of the final wine product (Ganga and Martinez, 2003; Vilanova et al. 2005; Fleet 2007). Furthermore, 10

wineries using the same commercial ADY may begin to generate wine varietals that share similar characteristics since each strain is thought to generate specific organoleptic properties in the wine. Thus, the use of similar commercial ADY between wineries and regions may standardize wine properties and diminish region-specific properties generated by indigenous yeasts (Raineria and Pretorius, 2000). Additionally, inoculated fermentations are also subject to competition with indigenous or other commercial yeast. Studies conducted on inoculated tanks resulted in the detection of S. cerevisiae strains other than the ADY added (Beltran et al. 2002; Clavijo et al. 2011; Tello et al. 2011). A study performed by Barrajón et al. (2009) noted that some ADY were scarce or non-existent in several inoculated fermentation tanks and concluded that ADY implantation is not always guaranteed. Effective implantation of ADY depends on temperature, water hardness, sugar concentration, agitation, and rehydration duration (Soubeyard et al. 2006; Barrajón et al. 2009). Failure to properly rehydrate the starter culture prior to inoculation may adversely modify the viability, physiological, and fermentation behaviour of ADY (Soubeyard et al. 2006). Furthermore, improper rehydration and implantation of the ADY may generate a longer than desired lag phase and allow for the entrance and establishment of other commercial and, or, indigenous yeast (Barrajón et al. 2009; Soubeyard et al. 2006; Lopes et al. 2007). Despite the correct practice and steps involved in ADY rehydration and inoculation, unsuccessful commercial ADY implantation may still be observed, possibly due to competition between indigenous and commercial yeast (Barrajón et al. 2009). An overview of literature from various wine producing regions that have studied various aspects of wine 11

yeast ecology, including yeast population dynamics in inoculated fermentation, spontaneous fermentation, on winery surfaces, and/or in vineyards, and their molecular method of choice, can be viewed in Table 1-1. 1.5 THE USE OF MICROSATELLITE MARKERS FOR S. CEREVISIAE STRAIN-LEVEL IDENTIFICATION The identification of yeast using molecular techniques has gone through ample development over the years. Characterization of S. cerevisiae yeast strains has occurred by various molecular measures including restriction fragment length polymorphism (RFLP), karyotyping, DNA hydridization, PCR-based assays, and mitochondrial restriction digests (Perez et al. 2001). A more modern approach to strain determination of S. cerevisiae wine yeast involves analysis of simple sequence repeats (SSRs) or microsatellites. Microsatellites are short repetitive DNA sequences varying in tandem repeat number, due to DNA replication errors including base pair additions, deletions, and slipped-strand mis-pairing (Strand et al. 1993; Legras et al. 2004). Microsatellites are dispersed throughout the Saccharomyces genome and act as molecular markers used for strain identification. An ideal molecular marker should be highly polymorphic, numerous, and provide reproducible results in order to grant a high degree of discrimination (Field et al. 1996). Various reports in the literature concluded that the use of microsatellites for molecular analysis and identification of S. cerevisiae at the strain-level is highly accurate and preferable over other molecular methods (Techera et al. 2001; Richards et al. 2009; Hall et al. 2011). The molecular techniques for this research utilized microsatellite fingerprint analysis as a means of S. cerevisiae strain identification. 12

1.6 RESEARCH OBJECTIVES Population dynamics of yeasts during fermentation have been well studied in more traditional wine-producing regions such as Spain (Santamaria et al. 2005; Garijo et al. 2008; Clavijo et al. 2011), Argentina (Lopes et al. 2002; Combina et al. 2005; Mercado et al. 2007), and Italy (Ciani et al. 2004; Maro et al. 2007; Tofalo et al. 2011). To my knowledge, with the exception of a previous study conducted at Quails Gate Estate Winery in 2007 (Hall et al. 2011), wine yeast population studies have yet to be reported from Canadian wineries. To better assess wine yeast population dynamics during inoculated and spontaneous fermentations in British Columbia s Okanagan wine region, an intensive sampling was performed during the 2010 and 2011 vintages at three local wineries, including Quails Gate (QG), Cedar Creek (CC), and Road13 (R13) Estate Wineries (EW). The three overarching objectives of this thesis were to: (1) Assess the population dynamics of wine yeast species and strains in both inoculated and spontaneous Vitis vinifera L. varietal (var.) Pinot noir fermentations during the 2010 vintage at three Okanagan wineries: QGEW, CCEW, R13EW. (2) Perform a year-to-year comparison of wine yeast species and strain composition of inoculated Vitis vinifera L. var Pinot noir fermentations between the 2010 and 2011 vintages at QGEW. 13

(3) Construct a comparative commercial S. cerevisiae ADY microsatellite database for CCEW and R13EW and combine it with the already existing QGEW database constructed at UBCO (Hall et al. 2011). 1.7 RESEARCH HYPOTHESES objectives: The following are hypotheses that were tested for two of the three overarching (1a) Commercial S. cerevisiae strains will be more prevalent and will occur in higher frequency than non-commercial S. cerevisiae strains during the early, mid, and end stages of inoculated and spontaneous fermentations at all three wineries. (1b) Non-Saccharomyces spp., particularly Henseniaspora uvarum, will be the dominant yeasts detected in the cold-soak stage at all three wineries. (1c) A succession from non-saccharomyces spp. to S. cerevisiae strains will occur as fermentation progresses from cold-soak to the later fermentation stages in both inoculated and spontaneous fermentations at all three wineries. (1d) The commercial S. cerevisiae ADY inoculant will be the dominant yeast detected during inoculated fermentation at all three wineries. 14

(1e) Yeast species/strain richness and diversity will be greater in spontaneous than inoculated fermentations during the early, mid, and end stages at all three wineries. (1f) At each winery, yeast assemblages will differ between the inoculated and spontaneous fermentations. (1g) Inoculated tanks at QGEW and R13EW will have similar yeast assemblages, whereas at CCEW yeast assemblages will differ between the inoculated tanks of study. (1h) Each winery will express distinct yeast species/strain assemblages during cold-soak, inoculated, and spontaneous fermentations. (2a) The 2010 and 2011 vintages at QGEW will exhibit similar yeast assemblages and strain diversities during the inoculated fermentation. 1.8 RATIONALE OF HYPOTHESES I expected support for hypothesis 1a because commercial S. cerevisiae strains tend to be the most prevalent yeast isolated from fermentation when both inoculated and spontaneous fermentations are conducted at a winery (Kluftinger et al., data unpublished; Constanti et al. 1997; Santamaria et al. 2005; Hall et al. 2011), whereas in traditional wine-producing areas, non-commercial indigenous (native) strains tend to dominate when the sole type of fermentation used is spontaneous (Frezier and Dubourdieu, 1992; Le Jeune et al. 2002; Combina at el. 2005; Le Jeune et al. 2006). I expected support for 15

hypothesis 1b because H. uvarum is typically the most commonly isolated non- Saccharomyces yeast during cold-soak (Maro et al. 2007; Barrajón et al. 2009; Clavijo et al. 2011). Other non-saccharomyces spp. that I expected to observe in lower frequencies include species belonging to Pichia, Wickerhamomyces, Metschnikowia, or Torulaspora genera. I expected support for hypothesis 1c because the succession from non- Saccharomyces yeast to S. cerevisiae yeast during fermentation has been consistently reported in the literature (Combina et al. 2005; Zott et al. 2008; Barrajón et al. 2009; Tello et al. 2011). I expected support for hypothesis 1d because the intention of inoculation is to achieve successful implantation and dominance of the commercial ADY (Barrajón et al. 2009). I expected support for hypothesis 1e because the consensus of the literature suggests that spontaneous fermentation tends not to favour one species/strain over another, which results in a relatively high species/strain diversity (Schutz and Gafner, 1993; Pretorius 2000; Mercado et al. 2007). I expected inoculated and spontaneous yeast assemblages to differ from one another assuming the reasons outlined in hypotheses 1d and 1e are true, thereby supporting hypothesis 1f. I expected support for hypothesis 1g because the same commercial ADY strain was used in each inoculated tank at QGEW and R13EW, and a different commercial ADY strain was added to each of CCEW inoculated tanks. I expected support of hypothesis 1h because the three wineries were located in different regions of the Okanagan Valley and each winery was subjected to different viticulture and vinification practices, which may affect yeast composition (Cordero-Bueso et al. 2011; Tello et al. 2011). I expected to support hypothesis 2a because the two years were identical with respect to ADY inoculants; and viticulture and vinification practices (Cordero-Bueso et al. 2011; Tello et al. 2011). 16

CHAPTER 2: AN ASSESSMENT OF YEAST POPULATION DYNAMICS DURING INOCULATED AND SPONTANEOUS FERMENTATIONS AT THREE LOCAL BRITISH COLUMBIA WINERIES 2.1 SYNOPSIS Inoculated fermentation has been shown to provide a rapid, reliable, and controlled fermentation that facilitates the production of predictable and consistent wine quality while reducing the risk of wine spoilage (Pretorius 2000; Beltran et al. 2002; Santamaria et al. 2005). On the other hand, spontaneous fermentation facilitates the development of a diverse assortment of strains, which may be indigenous to the local region. The presence of various strains in the must is said to generate complexity, sensory attributes unique to the region, and vintage variability in the finished wine product (Maro et al. 2007; Mortimer and Polsinelli, 1999; Pretorius 2000). Interestingly, the assumed inoculated and spontaneous population dynamics described above are not always achieved. In some studies, the ADY inoculant either fail to successfully implant in the must (Barrajón et al. 2009) or appear to behave as competitors able to enter spontaneous and inoculated fermentation tanks (Schutz and Gafner, 1993; Hall et al. 2011; Tello et al. 2011). Commercial S. cerevisiae strains have appeared to: 1) decrease species/strain diversity and richness in spontaneous fermentations (Beltran et al. 2002; Santamaria et al. 2005; Hall et al. 2011); 2) compete against the ADY inoculant (Torija et al. 2001; Lopes et al. 2007; Kluftinger et al., unpublished data; Clavijo et al. 2011); and 3) survive between vintages and re-emerge in the following year s fermentations (Torija et al. 2001; Beltran et al. 2002; Constanti et al. 2007). These findings solidify the complexity and ignorance of understanding towards inoculated and spontaneous population dynamics during fermentation, and highlight the potentially negative or positive impact they may 17

have on sensory attributes of the finished wine product (Mateo et al. 1991; Santamaria et al. 2005; Blanco et al. 2006). Unlike other well-studied wine regions of the world, a large-scale comparative assessment of inoculated and spontaneous fermentation population dynamics among multiple Canadian wineries has not been reported in the literature. In order to evaluate the yeast population dynamics in one of Canada s growing wine regions, a study was conducted to address three main objectives: 1) to assess the population dynamics of wine yeast species and strains in both inoculated and spontaneous Vitis vinifera L. var Pinot noir fermentations of three Okanagan wineries; 2) perform a year-to-year comparison of wine yeast species and strain composition of inoculated var. Pinot noir fermentations between 2010 and 2011 harvests at QGEW; and 3) construct a comparative commercial S. cerevisiae ADY microsatellite database for CCEW and R13EW in order to build onto the already existing QGEW database constructed at UBCO in 2007 (Hall et al. 2011). This third objective allowed for the identification of S. cerevisiae strains from three different wineries. This present work was novel to Canada by specifically answering the question of which specific yeasts were responsible for conducting inoculated and spontaneous var. Pinot noir fermentations at several local wineries in an operational setting. My results provided fuel for future studies that are interested in determining: 1) whether the age of a winery affects fermenting yeast assemblages; 2) the source(s) of non-inoculant yeasts in 18

fermentation; 3) the rationale for unsuccessful ADY implantation; and 4) how a winery s yeast microflora affects the sensory attributes of wine. 2.2 MATERIALS AND METHODS 2.2.1 Study sites and experimental design Samples were collected during the 2010 vintage from QGEW, CCEW, and R13EW located in the Okanagan Valley of British Columbia, Canada. Quails Gate Estate Winery was established in 1989 and produces on average 55,000 cases of wine annually; CCEW was established in 1983 and produces 40, 000 cases of wine annually; and R13EW was established in 1998 and produces on average 15,000 cases of wine annually. Yeasts were isolated and identified from three inoculated and one spontaneous Vitis vinifera L. varietal (var.) Pinot noir fermentation tanks at each of the three wineries. Four distinct stages of fermentation were sampled, including: cold-soak (CS), early (ER), mid (M), and end (F) stages. Sixteen isolates were identified from each stage, which resulted in the identification of 720 yeast species/strains. Similarly, samples were collected from three inoculated V. vinifera L. var Pinot noir fermentation tanks in 2011 at QGEW during the four fermentation stages described previously. Eight isolates were identified from the cold-soak stage and sixteen isolates were identified from the early, mid, and end stages of fermentation. In total, 168 yeasts were identified in 2011 from the inoculated fermentation tanks. All equipment involved in the var. Pinot noir winemaking process, including the receiving area, crusher/de-stemmer, and press was located within the winery 19

establishment. The fields/blocks from which var. Pinot noir grapes were harvested, along with other supplemental data, are listed for QGEW (2010) in Table 2-1a; CCEW in Table 2-1b; R13EW in Table 2-1c; and QGEW (2011) in Table 2-2. Some data were unable to be retrieved due to the succession of winemakers between the 2010 and 2011 year at R13EW, which resulted in missing values. The grapes were located in vineyards within 1 km of the winery and were fermented in 5300 L stainless steel vessels (fermentation tanks) at all three wineries. Each tank of study was treated with sulphur dioxide (SO 2 ) after the grapes were harvested, de-stemmed, crushed, and loaded into the fermentation tanks. Dates of SO 2 addition are shown in Table 2-1a-c for the 2010 vintage (all three wineries) and Table 2-2 for the 2011 vintage (QGEW). All sampling occurred after the completion of these steps. Prior to early stage sampling, the inoculated tanks of study were inoculated with a commercial ADY strain. It should be noted that nutritional supplements, such as Diammonium phosphate (DAP) and Superfood, were not added to any of the ADY at time of rehydration. Furthermore, there was no intentional addition of commercial ADY to any of the spontaneous tanks of study. 2.2.2 Collection of samples Approximately 500 ml of var. Pinot noir must samples were collected from each tank at the following fermentation stages: cold-soak, early, mid, and end stages. The coldsoak stage included freshly crushed and SO 2 treated var. Pinot noir must soaking with skins, seeds, and stems in its chilled 5300 L stainless steel fermentation tank. Cold-soak samples were collected from those fermentation tanks to be inoculated, after SO 2 addition, and prior to any addition of commercial ADY. Cold-soak samples were not collected from 20

the spontaneous tanks; however, the cold-soak samples collected from the fermentation tanks to be inoculated were considered samples representative of the cold-soak stage in the spontaneous fermentation tanks. This is because similar var. Pinot noir blocks (regions of var. Pinot noir growth in the vineyard) used for inoculated fermentations were also used for spontaneous fermentation (Table 2-1a-c). The remainder of the fermentation stages were defined by the Brix concentration (residual sugar) of the must. Residual sugar concentration of the must correlates with yeast fermentation kinetics. For the inoculated tanks, early stage sampling occurred approximately two days after commercial ADY addition and when the must measured between 10-20 Brix. An ADY strain was not added to any of the spontaneous tanks, therefore, an early sample was collected from the spontaneous tanks when the must measured between 10-20 Brix. For both inoculated and spontaneous fermentations, the third sample was taken during the mid stage of fermentation when the must measured between 5-15 Brix, and once the must concentration approached 0 Brix, the end sample was collected. This was when the fermentation was considered finished. All fermentations lasted approximately 10-20 days in duration. Tables 2-1 and 2-2 show the specific dates and fermentation parameters. Samples collected from QGEW were obtained from the center of each tank, approximately 0.5 m below the cap (a thick top-layer of grape skins, stems, and seeds) using a sterilized stainless steel collecting apparatus. The collection apparatus was first washed in caustic soda (sodium hydroxide), neutralized with citric acid, and then sprayed with 90% denatured ethanol immediately prior to every sample collection. The sample was poured into an autoclaved sterile glass or nalgene bottle that was immediately capped 21