Methoxypyrazines, etc

2/22/212 Methoxypyrazines, etc Gavin L. Sacks *, Justine Vanden Heuvel, Bruce Pan, Justin Scheiner, Imelda Ryona, Sarah Harris * Cornell University, Department of Food Science, NYS Agricultural Experiment Station, Geneva, New York 14456 212 Washington Association of Wine Grape Growers Meeting Acknowledgements Sacks Lab Imelda Ryona (right) Dr. Bruce Pan (left) Sarah Harris Dr. Justine Vanden Heuvel Justin Scheiner Dr. Alan Lakso Dr. Diego Intrigliolo Oren Kaye and John Thorngate (Constellation Brands) Projects were supported by: New York Wine and Grape Foundation Viticulture Consortium East USDA - Federal Formula Funds Kaplan Research Fund Masking and wine Fundamental Theorem of Wine Aroma (FTWA) Vegetal, Herbaceous Fruity, Sweet Unripe aromas and ripe aromas mask each other (and, excessive amounts of either is a problem) 1

2/22/212 Balanced wines lie along this fruity, vegetal continuum Sauvignon Blanc Carmenere Most Dry Reds Muscat Vegetal, Herbaceous Fruity, Sweet Excessive green more likely to be cited as a problem, at least for East coast reds Wine Spectator scores and tasting notes for Long Island red wines 75-85 points (182 wines) 85-9 points (85 wines) Low scoring wines more than 3x likely to have earthy, vegetal, herbal descriptors.9 Fractional usage of term in tasting notes..21.7.7.4 Vegetal Herbal Earthy Adapted from data compiled by Larry Perrine at Channing Daughters Winery Wine Spectator scores: continued 75-85 points (182 wines) 85-9 points (85 wines) High scoring wines nearly 3x more likely to have ripe, fruit descriptors Fractional usage of term in tasting notes.16..58 1.32 Ripe Fruit Adapted from data compiled by Larry Perrine at Channing Daughters Winery 2

2/22/212 The most notorious contributor to green 3-Alkyl-2-methoxypyrazines (MPs) N O C H 3 N R R = alkyl group Detection threshold for in wine IBMP: 5-15 ng/l IPMP:.5-2 ng/l Most closely associated with Bordeaux wine grapes: Cabernet, Merlot, Sauv blanc R Abbr Typical conc. in Cabernet (ng/l) Aroma isobutyl IBMP 5-2 Bell pepper, vegetal Predominant MP in immature grapes sec-butyl sbmp n.d. - 1 Peas, potatoes isopropyl IPMP n.d. 2 Asparagus, peas Predominant MP in Asian lady beetle Overview of talk today Overview of strategies in the winery Previous attempts Something new: juice silicone fining Controlling MPs in the vineyard Empirical studies of accumulation and degradation Moving towards mechanistic explanations (relating MP concentrations to precursors) Future directions MPs and winemaking: extraction during maceration 1 Adapted from Roujou de Boubee 21 thesis IBMP (ng/l) 8 6 4 2 MPs mostly in skins, extraction kinetics comparable to or slightly faster than anthocyanins 2 4 6 Fermentation time (hours) 3

2/22/212 Good correlation between IBMP in whole grapes and skin fermented wines R 2 =.96 Ryona, Pan, Sacks (JAFC, 29) MPs are stable in the bottle IBMP in wine vs. IBMP in same wine 14 months later (Pan and Sacks, unpublished) (n=12 wines) Several ways to remove MPs from wines, but none have selectivity EXAMPLES: Activated charcoal Distillation/thermovinification - MPs are readily extracted during fermentation, stable in the bottle, challenging to remove selectively - End result: attempts to achieve MP remediation on finished wines have been unsuccessful 4

2/22/212 Selectively reducing MPs after harvest: an exercise in futility? Approach Matrix Reduction Caveats Fluorescent & UV lights Wine ns No effect Yeast selection Juice ns No effect Oak sawdust, gallic acid, epicathechin Model Juice ns No effect Synthetic closures Wine 7-8% Lack of selectivity (?) Packaging Tetrapack Wine 26-45% Lack of selectivity (?) Activated charcoal, deodorized oak chips, bentonite Wine ns 34% Lack of selectivity Thermovinification Juice 29-67% Sensory changes Data collected from several publications Other recent ideas for MP removal Flash détente - for grapes Almost certainly removes C6 aldehydes Data for MPs? Odorant Binding Protein (OBP) for wine or juice Pickering et al 21 patent No literature reports yet Silicone fining of grapes Our group Reminder: Classifications of Wine Flavors Primary Derived from grape ( foxy aromas, MPs) Secondary Produced during alcoholic or malolactic fermentations Byproduct of yeast biosynthesis Enzymatic transformation of non-volatile precursors from substrate (e.g. terpenes) Tertiary Develop post-fermentation Oak Abiotic transformations during cellaring 5

IBMP concentration (ng/l) 2/22/212 Could we use a not-so-selective sorbent to remove MPs before fermentation without removing other volatiles? Grape Juice by GC-MS (~1 compounds) Wine by GC-MS (>1 detectable compounds) Idea: Fine juice/must with food grade silicone (PDMS) prior to fermentation to remove MPs Trials 27 Cabernet Franc Rose 28 Cabernet Franc Red 28 Chardonnay White Standard Sequence 1. Add chopped silicone tubing to juice (up to 4 g/l) 2. Inoculate, add nutrients 3. Remove silicone 4. End of Fermentation, <1 g/l RS 21 Riesling White Why it works: polyphenolics & most grape aroma precursors are semi-polar; MPs are non-polar Example results. Ryona, Reinhardt and Sacks, Food Res. Int., 212 in press 27 Cabernet Franc Rosé 9 8 7 6 5 4 3 2 1 5 1 15 2 25 3 35 4 45 Silicone contact (hour) >9% reduction No Silicone (15 L) Silicone (15 L) No Silicone (45 L) Silicone (45 L) Generally, 5-9% reduction in MPs; no significant reduction of other wine volatiles 6

Normalized value (%) 2/22/212 Results from 4 trials over 3 years: Silicone treatment of must reduces MPs, does not reduce most other volatiles 14 12 1 8 6 4 2 A B g/l 1 g/l 2 g/l 4 g/l C D Silicone added to must Most secondary volatiles * ns ns ns ns * (fermentation * * esters, some terpenes) not affected AB B AB A >7 volatiles profiled by SPE- GC-MS A AB A MPs reduced AB AB in Bdose-response AB B fashion Data for 21 Riesling with ladybug taint shown above Results for red winemaking Still works, but more complicated 8 7 6 5 4 g/l 4.4g g/l 13.3 g/l 4. g/l 3 ns ns ns ns 2 4 4 56 * * 6 87 1 * 11 12 With reds, MPs must first be extracted from skins into liquid phase (left) Must held cold, then warmed and inoculated after 48 hours. Silicone removed after Day 5 MP reduction and sensory evaluation 27 Cabernet Franc Rosé (native IBMP) 28 Chardonnay white (Spiked in IPMP) 28 Cabernet Franc Red (Native IBMP) 21 Riesling (IPMP from MALB) > 9% Informal evaluation, less fresh green note on treated wine 93% Informal evaluation, less MALB taint 53% Informal evaluation, less green note 56% 9% of ~1 wine industry workshop participants select silicone treatment as less affected by MALB taint. Formal sensory trial: forced choice test for greenness showed significant effect 7

IBMPconcentration 2/22/212 Issues with silicone fining Not approved for commercial wine production by TTB Currently, using chopped pieces of food grade tubing. Want to identify a higher surface area, more convenient form (but still food-grade!). Not appropriate for wines that rely on primary aromas e.g. rotundone ( black pepper ) Labrusca-type, Muscat Other polymers more effective? Any health concerns? Fermentation kinetics and color are not affected Alternative to control in the winery: Controlling MPs in the vineyard General observations regarding MPs at harvest and growing conditions (from the last two decades) Lower concentrations with higher growing temperatures Lower concentrations with higher cluster light exposure Lower concentrations at less vigorous sites Bogart and Bisson, 26 Mechanistic explanations for these events were scattered or non-existent MP research at Cornell: distinguishing accumulation from degradation Accumulation Degradation Project 1: Role of cluster exposure? ( burn off the MPs? ) Veraison Time 8

IBMP (ppt ) IBMP (ppt) IBMP concentration (ppt) 2/22/212 Variably shoot-thinned vine experiments Block I Block II Block III (27 and 28) Shaded Exposed Exposed Exposed Exposed Shaded Shaded Shaded Unshaded - Compared shaded and unshaded berries within the same vine - Accounts for vine-to-vine variation Two years of study in Finger Lakes (27 and 28) Evidence that cluster sun exposure pre-veraison will inhibit MP accumulation Little evidence that cluster exposure post-veraison impacts MP degradation 25 2 15 1 5 27 Shaded 27 Exposed Cluster exposure to sunlight 5 15 25 35 45 55 65 75 85 95 15 115 125 135 145 Days post-bloom Ryona, et.al. J. Ag Food Chem, 28 Exposed fruits accumulate less IBMP 4 35 3 25 2 15 1 5 28 Shaded 28 Exposed 1 2 3 4 5 6 7 8 9 1 11 12 Days Days From post-bloom Bloom Adapted from Lakso and Sacks, Kliewer Symposium, 29 9

IBMP concentration at harvest (picogram/gram) Harvest IBMP concentration (ppt) 2/22/212 27: Constant difference in MP between exposed and shaded clusters during season **P<.5 *P<.8 veraison The importance of accumulation: Within a region, at same maturity, MP at harvest reflects MP at veraison 25 2 y =.181x -.572 R 2 =.9367 15 1 13 sites on Seneca Lake 5 Ryona, et.al. J. Ag Food Chem, 28 5 1 15 2 25 Pre-veraison IBMP concentration IBMP on Day concentration 47 (picogram/gram) (ppt) Importance of exposure in pre-veraison period: Corroboration from field studies Experiment: Variable timing of leaf removal at two sites - Cabernet Franc (Finger Lakes), 27 and 28 - Merlot (Long Island), 28 Summary of results: Reduction in IBMP at harvest consistently observed only with early season interventions - at berry set - 3 days after berry set J. Scheiner, S. Ennahli, A Wise, L Tarleton, B. Pan, G. Sacks, and J. Vanden Heuvel; AJEV, 21 {Recent corroborating results from UC Davis using shadecloth: A. Koch, et al, Phys. Plantarum, 212} 1

Harvest IBMP (ppt) IBMP concentration at harvest (picogram/gram) 2/22/212 The problem with the Cluster Shading is the Lone Gunman explanation for differences in MPs 25 25 2 13 sites on Seneca Lake y =.181x -.572 R 2 =.9367 Within a site, cluster shading results in in MPs 15 1 5 125 25 5 1 15 2 IBMP concentration Day 47 (picogram/gram) Pre-veraison IBMP (ppt) But, within a region, we see nearly an order of magnitude of range in pre-veraison and harvest IBMP! Multivariate Field Study What variables really matter? 1 sites in NYS 2 Long Island 2 Lake Erie 6 Finger Lakes 1 vines at each site 2 x 5 vine panels Along with: Justine Vanden Heuvel Justin Scheiner Scheiner, et al. AJEV 212 Over 1 viticultural & environmental factors characterized for each vine (e.g. light, water status, cropload, etc.. at multiple time points) MPs measured at harvest and 2x pre-veraison Data analyzed by multivariate statistics (PLSR) 11

IBMP concentration (pg/g) IBMP concentration 2/22/212 Summary of 28-9 Multivariate Studies IBMP accumulation pre-veraison independently associated with higher temperature(!) greater water availability and to a lesser extent, cluster shading IBMP degradation post-veraison correlates with maturity (sugar accumulation, etc) and not much else Scheiner, Vanden Heuvel, Pan, and Sacks. AJEV 212 Surprise: warmer seasons = more IBMP accumulation, although faster degradation Multiple NY sites Single site 21 very warm, wet Peak Harvest 29 cool & wet 6 5 4 3 2 1 Peak Harvest 28 27 warm & wet warm & dry Highest pre-veraison IBMP our lab has ever seen: > 8 ppt from Central Valley (CA) Merlot and Cabernet Sauvignon MP in warm vs. cool climate, assuming adequate water: a proposal IBMP warmer sites IBMP cooler sites Time 12

5-DAA IBMP (pg/g) Harvest IBMP (pg/g) 2/22/212 If conditions that promote vine growth increase MPs, can this inform viticultural strategies? Idea 1: Shoot tipping on Cabernet franc 3 timings of shoot tipping 1 days before anthesis (1DBA) anthesis (AN) 1 days after anthesis (1 DAA) Observations over two years: treatment may result in slight increase pre-veraison, no sig effect at harvest Scheiner, et al unpublished data 3 25 2 15 1 5 v 29 21 CON 1 DBA Anthesis 1 DAA If conditions that promote vine growth increase MPs, can this inform viticultural strategies? Idea 2: Growth inhibitor (chlormequat) 2 application rates (4 and 8 mg/l) 3 application timings 7 days before anthesis (7 DBA) anthesis (AN) 3 days after anthesis (3 DAA) 12 v 21 data from 7 DBA Observations for one year: pre-bloom treatment increases IBMP accumulation Scheiner, et al unpublished data 1 8 6 4 2 a CON b 4 mg/l c 8 mg/l The problem: can we make rational predictions about MP management without understanding MPs at the molecular level? 13

Concentration (ng/l) 2/22/212 Putative MP formation/degradation Pre-veraison Accumulation Degradation Branched amino acid + NH 4 + + α-dicarbonyl HP N R N OH OMT MP N R N OCH 3???? Biosynthesis pathways suggested by Murray, et al J Sci Food Ag 1975 O-methyltransferase (OMT) capable of methylating hydroxyprazines (HP) to MP purified by Hashizume, et al 21, cloned by Dunlevy, et all 21 Proposed MP degradation during ripening Converted back to odorless HP Pre-veraison Accumulation Degradation Branched amino acid + NH 4 + + α-dicarbonyl N N R OH OMT N R N OCH 3 N N R OH Partial glycosylation Ryona, Leclerc, and Sacks, J. Ag. Food Chem., 21 Some VERY recent data Relationship of MP and HP during the growing season IBHP and IBMP (ng/l) v. Date 1 8 IBHP 6 IBMP 4 2 5/2 6/21 8/1 9/29 (left) IBHP and IBMP in California Merlot Harris and Sacks, unpublished IBHP and IBMP are well correlated before veraison. Potentially, MP biosynthesis is dependent on HP concentration If so, what drives HP biosynthesis? 14

2/22/212 One scary thought.... Can IBHP be converted back to IBMP by yeast????? O-methylation 3-isobutyl-2-hydroxypyrazine (IBHP) 3-isobutyl-2-methoxypyrazine (IBMP) 2 pg/ml or more????? Small scale fermentations to evaluate potential of IBHP to be converted to IBMP 24 h at 25 C 24 h at 25 C 48-72 h at 25 C Yeast Yeast Strains Used: AMH CY379 EC1118 Vin 13 Kloeckera apiculata Juice Chemistry: Brix = 2 ph = 3.6 TA = 6.5 g/l 1 ng/ml IBHP Harris, Ryona, and Sacks. Unpublished data The short answer: no conversion of IBHP during fermentation O-methylation 3-isobutyl-2-hydroxypyrazine (IBHP) 3-isobutyl-2-methoxypyrazine (IBMP) No detectable IBMP (limit of detection ~ 1 ng/l) If there s conversion, its <.1%. Phew! Harris and Sacks, unpublished 15

IBMP at 2 weeks preveraison (ng/l) 2/22/212 Future directions for studying MPs Moving beyond the empirical to the mechanistic What explains increased MP accumulation in conditions of high water availability, cluster shading, etc Increased hydroxypyrazine (HP) production? Or increased VvOMT activity? And how do these relate to vine physiology? What controls MP degradation? Enzymatic? What enzyme? Future directions: genomics and grape breeding Several groups (including ours) working on mapping genes associated with high MPs Possibility for marker assisted selection when using non-vinifera species high in MPs (and other undesirable compounds) 8 6 4 2 Individual in population At left, IBMP in 168 individuals from St Pepin X Cabernet franc mapping population Sacks and Owens, unpublished Acc. # IBMP IPMP TSS (ng/kg) (ng/kg) ( Brix) 1 13±1 1±3 2.5 2 11±9 7±2 2.3 3 143±8 17±3 18. 4 17±1 1±.4 15. 5 33±5 8±2 21.2 6 286±12 16±4 23.5 7 251±1 31±5 2.6 8 52±6 n.d. 15.3 9 353±16 n.d. 2.4 1 18±2 n.d. 2.1 Something different: IBMP and IPMP in accessions of V. cinerea (native species, potentially useful to grape breeders) IBMP > 13 ng/kg in all accessions, up to 353 ng/kg. Similar results for V. riparia V. cinerea No correlation between MPs and soluble solids 16

Bell Pepper Intensity Green Pepper Aroma 2/22/212 A final issue: how good a predictor of greenness are MPs? r 2 =.74 Fifty Bordeaux and Loire reds, judged by expert panel Exceptional correlation between IBMP and green pepper IBMP concentration (ng/l) Roujou de Boubee, et.al. (JAFC 2) Correlation of MPs and bell pepper more modest around or below threshold (15 ng/l) 2.5 2 1.5 1.5 Zoom on Roujou data r 2 =.37 5 1 15 2 25 IBMP (ng/l) Weak or non-existent correlations observed in other reports Chapman, et al (JAFC, 24) Falcao, et al (JAFC, 27) Preston, et al (AJEV 28) Scheiner, et.al. (AJEV, 212) Take home messages MPs are an occasional fault (in NYS, ~5% of wines over 2 ppt). Presence or absence of other compounds important to herbaceousness IBMP in NYS wines (and elsewhere) frequently between 5-15 ppt Sacks, Pan; unpublished 17

Intensity of herbaceousness Intensity of herbaceousness 2/22/212 When MPs are < 15 ng/l, they are a lousy predictor of herbaceousness 28 29 IBMP (pg/ml) IBMP (pg/ml) Data from NY State Cabernet franc multi-site study Scheiner, et al AJEV 212 Summary Veggie masks fruity and vice versa. Selective removal of MPs from wine is hard. Removing MPs from juice w/ non-polar sorbents may work MP accumulation is increased in conditions that promote growth (heat, water) and, to a lesser extent, in shaded clusters MP degradation relates to maturity Our molecular level understanding of why is still lousy And, we cannot explain herbaceousness solely by MP concentrations Wine is more than just ethanol, H 2 O, and MPs. Thankfully. 18